• Search Menu
  • Advance articles
  • Author Guidelines
  • Submission Site
  • Open Access
  • About Interacting with Computers
  • About the BCS, The Chartered Institute for IT
  • About Interaction
  • Editorial Board
  • Advertising and Corporate Services
  • Journals Career Network
  • Self-Archiving Policy
  • Dispatch Dates
  • Journals on Oxford Academic
  • Books on Oxford Academic

BCS The Chartered Institute for IT

Article Contents

1 introduction, 2 socio-technical systems design, 3 socio-technical systems design approaches, 4 problems with existing approaches to socio-technical systems design, 5 socio-technical systems engineering, 6 an stse research agenda, acknowledgements.

  • < Previous

Socio-technical systems: From design methods to systems engineering

  • Article contents
  • Figures & tables
  • Supplementary Data

Gordon Baxter, Ian Sommerville, Socio-technical systems: From design methods to systems engineering, Interacting with Computers , Volume 23, Issue 1, January 2011, Pages 4–17, https://doi.org/10.1016/j.intcom.2010.07.003

  • Permissions Icon Permissions

It is widely acknowledged that adopting a socio-technical approach to system development leads to systems that are more acceptable to end users and deliver better value to stakeholders. Despite this, such approaches are not widely practised. We analyse the reasons for this, highlighting some of the problems with the better known socio-technical design methods. Based on this analysis we propose a new pragmatic framework for socio-technical systems engineering (STSE) which builds on the (largely independent) research of groups investigating work design, information systems, computer-supported cooperative work, and cognitive systems engineering. STSE bridges the traditional gap between organisational change and system development using two main types of activity: sensitisation and awareness; and constructive engagement. From the framework, we identify an initial set of interdisciplinary research problems that address how to apply socio-technical approaches in a cost-effective way, and how to facilitate the integration of STSE with existing systems and software engineering approaches.

Socio-technical systems design (STSD) methods are an approach to design that consider human, social and organisational factors, 1 as well as technical factors in the design of organisational systems. They have a long history and are intended to ensure that the technical and organisational aspects of a system are considered together. The outcome of applying these methods is a better understanding of how human, social and organisational factors affect the ways that work is done and technical systems are used. This understanding can contribute to the design of organisational structures, business processes and technical systems. Even though many managers realise that socio-technical issues are important, socio-technical design methods are rarely used. We suspect that the reasons for their lack of use are, primarily, difficulties in using the methods and the disconnect between these methods and both technical engineering issues, and issues of individual interaction with technical systems.

The underlying premise of socio-technical thinking is that systems design should be a process that takes into account both social and technical factors that influence the functionality and usage of computer-based systems. The rationale for adopting socio-technical approaches to systems design is that failure to do so can increase the risks that systems will not make their expected contribution to the goals of the organisation. Systems often meet their technical ‘requirements’ but are considered to be a ‘failure’ because they do not deliver the expected support for the real work in the organisation. The source of the problem is that techno-centric approaches to systems design do not properly consider the complex relationships between the organisation, the people enacting business processes and the system that supports these processes ( Norman, 1993; Goguen, 1999 ).

We argue here that there is a need for a pragmatic approach to the engineering of socio-technical systems based on the gradual introduction of socio-technical considerations into existing software procurement and development processes. We aim to address problems of usability and the incompatibility of socio-technical and technical systems development methods. Our long-term research goal is to develop the field of socio-technical systems engineering (STSE). By this, we mean the systematic and constructive use of socio-technical principles and methods in the procurement, specification, design, testing, evaluation, operation and evolution of complex systems.

We believe that it is not enough to simply analyse a situation from a socio-technical perspective and then explain this analysis to engineers. We also must suggest how socio-technical analyses can be used constructively when developing and evolving systems. Many companies have invested heavily in software design methods and tools, so socio-technical approaches will only be successful if they preserve and are compatible with these methods. We must avoid terminology that is alien to engineers, develop an approach that they can use, and generate value that is proportionate to the time invested.

These are challenging objectives and, to achieve them, we must draw on research from a range of disciplines. There are at least four significant research communities that have explored and addressed socio-technical issues that affect the specification, design and operation of complex computer-based systems:

Researchers interested in work, in general, and the workplace. An interest in the design of work was the original stimulus for proposing socio-technical approaches. Mumford (1983) and Eason’s (1988) research typify the approach of this community. The original objective was to make work more humanistic and the initial focus was on manufacturing systems. As computers have become pervasive in the workplace, however, the community has also examined the relationships between work and its computer-based support noting, for example, that the computer system can shape and constrain work practices ( Eason, 1997 ).

Researchers interested in information systems. Information systems are large-scale systems that support the work of the enterprise and this community recognised at an early stage that socio-technical issues were significant (e.g., Taylor, 1982 ). This community has generally taken a broad perspective on the relationships between information systems and the enterprise rather than focusing on specific aspects of computer-supported work (e.g., Avison et al., 2001 ).

Researchers interested in computer-supported cooperative work (CSCW). This community has focused on the minutiae of work arguing that the details of work, as understood through ethnographic studies, profoundly influence how computer-based systems are used. Suchman’s seminal book (1987) which triggered work in this area, was followed by many ethnographic studies of systems in different settings ( Ackroyd et al., 1992; Bentley et al., 1992a; Heath and Luff, 1992; Heath et al., 1994; Rouncefield, 1998; Clarke et al., 2003 ). Many of these were concerned with co-located work (e.g., in control rooms) and most did not consider wider enterprise issues that affect system requirements and design.

Researchers interested in cognitive systems engineering. This community, exemplified by the work of ( Hollnagel and Woods, 2005; Woods and Hollnagel, 2006 ), has been primarily interested in the relationships between human and organisational issues and systems failure. Their main focus has been on control systems and health care and this community has not been much concerned with broader information systems.

Whilst these communities have had some mutual awareness, we believe that it is fair to say that there has been relatively little cross-fertilisation across communities. For example, in Mumford’s (2006) review article, there are no references to the strands of work in CSCW or cognitive systems engineering, and few references to the information systems literature.

Sitting alongside these communities, with some awareness of socio-technical issues, is the HCI research community. Some areas of HCI have clearly been influenced by socio-technical ideas, including usability (e.g., Nielsen, 1993; Mayhew, 1999; Krug, 2005 ) and human/user centred system design (e.g., Gould and Lewis, 1985; Norman and Draper, 1986; Gulliksen et al., 2003 ). Holistic design, for example, is identified by Gulliksen et al. (2003) as a key principle, and they note the need to explicitly consider the work context and social environment. More generally, much of the focus has been on sensitisation to socio-technical issues (e.g., Dix et al., 2004 has a chapter on this topic). There has been little work on how these socio-technical issues might directly influence the design of an interface to a complex software system (understandably so: we believe this to be a significant research challenge). By the same token, some researchers in the ubiquitous computing community have been influenced by socio-technical thinking ( Crabtree et al., 2006 ), although most research in this general area focuses on the development and evaluation of new technologies.

We believe that we need to integrate the work of these disparate communities under a common heading of socio-technical systems engineering. Our objectives here, therefore, are to summarise the contributions of the different research communities in this area, and to propose a practical vision for further developments. We do not provide a complete survey of socio-technical systems design (that would be impossibly long). Instead we present different perspectives on STSD, which we use as a basis for introducing a pragmatic framework for STSE that is deliberately limited in scope but which leaves room for the application of different STSD approaches. In this paper we have focused our discussions on organisational systems, but we believe that STSE applies to other types of systems based on Commercial Off the Shelf equipment and applications, for example, or domestic systems. After laying out our framework, we go on to propose a research agenda for socio-technical systems engineering where we identify research problems that need to be addressed to make STSE a practical reality.

Section 2 introduces the notion of STSD and Section 3 briefly discusses STSD approaches. Section 4 discusses shortcomings of these existing approaches. Section 5 introduces the notion of socio-technical systems engineering, identifying two main types of STSE activities. We conclude by identifying outstanding research issues that can be used to shape the discipline of socio-technical systems engineering.

The term socio-technical systems was originally coined by Emery and Trist (1960) to describe systems that involve a complex interaction between humans, machines and the environmental aspects of the work system—nowadays, this interaction is true of most enterprise systems. The corollary of this definition is that all of these factors—people, machines and context—need to be considered when developing such systems using STSD methods. In reality, these methods are more akin to philosophies than the sorts of design methods that are usually associated with systems engineering ( Mumford, 2006 ). STSD methods mostly provide advice for sympathetic systems designers rather than detailed notations and a process that should be followed.

The term socio-technical systems is nowadays widely used to describe many complex systems, but there are five key characteristics of open socio-technical systems ( Badham et al., 2000 ):

Systems should have interdependent parts.

Systems should adapt to and pursue goals in external environments.

Systems have an internal environment comprising separate but interdependent technical and social subsystems. 2

Systems have equifinality. In other words, systems goals can be achieved by more than one means. This implies that there are design choices to be made during system development.

System performance relies on the joint optimisation of the technical and social subsystems. Focusing on one of these systems to the exclusion of the other is likely to lead to degraded system performance and utility.

STSD methods were developed to facilitate the design of such systems. We have restricted our scope here to this class of systems, and do not consider deeply embedded systems, for example, where there is usually no social subsystem involved.

From its inception in the period immediately after World War II, by what is now called The Tavistock Institute, until the present day, there have been several attempts at applying the ideas of STSD. Some of these were successful, others less so ( Mumford, 2006 ). The prevailing climate within a particular company (or sometimes within a country) affected attitudes towards the idea of STSD: where attitudes were positive this often led to the successful uptake of the ideas.

Mumford (2006) provides an historical overview of developments in STSD. The general aim was to investigate the organisation of work, with early work in STSD focused mostly on manufacturing and production industries such as coal, textiles, and petrochemicals. The aim was to see whether work in these industries could be made more humanistic. In other words, the intention was to move away from the mechanistic view of work encompassed by Taylor’s (1911) principles of scientific management, which largely relied on the specialisation of work and the division of labour.

The heyday of STSD was, perhaps, the 1970s and the early part of the 1980s. This was a time when there were labour shortages, and companies were keen to use all means available to retain their existing staff. Apart from the usual cultural and social reasons, companies could also see good business reasons for adopting socio-technical ideas. The XSEL (eXpert SELler) system of the Digital Equipment Corporation (DEC), for example, was developed using STSD (see Mumford and MacDonald, 1989 for a retrospective view). It was an expert system designed to help DEC sales staff assist customers in properly configuring their VAX computer installations. This system was a success and at its peak the family of expert systems, including XSEL, that were being used to support configuration and location of DEC-VAX computers was claimed to be saving the company tens of millions of dollars a year ( Barker and O’Connor, 1989 ). Of course, it is impossible to assess the contribution of STSD to this success but the example illustrates that socio-technical approaches can be used effectively in real systems engineering.

By contrast, the latter part of the 1980s and the 1990s were possibly the low point in STSD’s history. The adoption of lean production techniques and business process re-engineering dominated, and STSD was largely sidelined. Dankbaar (1997) , however, suggested that these different methods (STSD, BPR, etc.) can all learn from each other. The late 1980s and early 1990s also saw the emergence of ethnographic studies of work, stimulated by Suchman’s (1987) seminal research at Xerox PARC. These ethnographic approaches (e.g., Heath and Luff, 1991 ) highlighted the significance of socio-technical issues in the design of software-intensive systems (e.g., Blomberg, 1988 ).

The 21st century has seen a revival of interest in socio-technical approaches as industries have discovered the diminishing returns from investment in new software engineering methods. However, socio-technical ideas and approaches may not always be explicitly referred to as such ( Avgerou et al., 2004 ). The ideas appear in areas such as participatory design methods, CSCW and ethnographic approaches to design. Indeed, one of the key tenets of STSD is a focus on participatory methods, where end users are involved during the design process (e.g., Greenbaum and Kyng, 1991 ). However, these methods, all of which have their roots in STSD, differ in important respects. Participatory design, which covers a whole range of methods (e.g., see Muller et al., 1993 ), often involves the users (or user representatives) effectively moving into the territory of the system developers for the duration of the project. By contrast, empathic design ( Leonard and Rayport, 1997 ) and contextual design (e.g., Beyer and Holtzblatt, 1999 ), which reflect STSD ideas, adopt the inverse view and put the developers into the users’ world as part of the development process.

The field of CSCW came about partly in response to a need to discuss the development of group support applications ( Grudin, 1994 ), but it has implicit roots in socio-technical thinking. Bowker et al. (1997) make the link explicit, dealing with the socio-technical system and CSCW, as does the recent special issue of the journal Computer-Supported Cooperative Work which deals with CSCW and dependability in health care systems ( Procter et al., 2006 ). The field of dependability 3 ( Laprie, 1985; Avizienis et al., 2004 ) is also intrinsically concerned with socio-technical systems, although this field sometimes uses the term ‘computer-based systems’ to refer to socio-technical systems.

STSD methods continue to be advocated for systems development and appear to be particularly suited to some application areas. Since the late 1990s, for example, STSD has been frequently advocated within health informatics for the development of health care applications (e.g., Whetton, 2005 ). Many such systems are under-utilised because they introduce ways of working that conflict with other aspects of the user’s job, or they require changes to procedures that affect other people’s responsibilities. One of the keys to developing systems that are acceptable to the users is a detailed understanding of the underlying work structures. In other words, what is required is a socio-technical approach ( Berg, 1999, 2001; Berg and Toussaint, 2003 ).

Most recently, in the UK, the need for STSD has been highlighted by issues surrounding the National Health Service’s ongoing National Programme for Information Technology (NPfIT; see Brennan, 2007 for a commentary on the programme). Even though many of the developments to date within the NPfIT have been imposed in an essentially top–down manner, there are still areas where there is a role for STSD, even if only at a local level ( Eason, 2007 ).

Although the vast majority of applications have been implemented in the workplace, socio-technical ideas are equally applicable in other settings where technology is deployed. In recent years, there has been an increasing uptake of technology in the home, particularly as smart home technologies and assistive technologies. The requirements for home-based systems are somewhat different from those of workplace systems. Sommerville and Dewsbury (2007) , for example, developed a model for the design of dependable domestic systems, which adopts a socio-technical view in which the system comprises the user, the home environment, and the installed technology.

Socio-technical systems design has been manifested in a wide range of different methods. Different traditions developed in different countries at different times have led to different approaches (see Mumford, 2006 for a fairly comprehensive historical review). The individual methods, to some extent, reflect different national cultures and approaches to work and work organisation. The consequence has usually been that each method is tailored to a particular market, which partly explains why there have never been any significant or successful attempts to integrate approaches to create a more general, standardised method of STSD.

There has been limited transferability of the available methods. In general, those who developed a method have had most success in applying it. Mumford’s ETHICS (1983, 1995) , for example, was mostly used in the USA when Mumford worked directly with organisations based there, such as DEC (see Section 2).

As the nature of the different markets has changed, the methods have not always kept pace. In some instances, the methods have been reactively refined—ETHICS, for example has recently been paired with agile methods of software development ( Hickey et al., 2006 ). In most cases, however, there has not been any reconsideration of the role of the earlier fundamental notions of STSD. Whether this is because STSD is not deemed relevant to modern ways of working, or because there is simply ignorance of these approaches is an open question. STSD remains an active field of research and practice, although in many cases it is the ideas, rather than the original methods, that are being applied.

Even though the notion of user participation lies at the heart of STSD, there has been a disappointing uptake of user-centred methods in general. Eason (2001) , for example, found that none of the 10 most widely advocated methods (including socio-technical design) were in common use. Furthermore, even where the methods were being used, user involvement was still largely to assist in the development of a techno-centric system. Users were not seen as participants in an integrated systems development process to produce a system that took appropriate account of social and organisational requirements.

One area where user participation has been taken seriously is in software development using agile methods, such as extreme programming (XP), Dynamic Systems Development Method (DSDM), and Scrum (see Abrahamsson et al., 2002 for a review and analysis of these methods). These methods incorporate at least some face-to-face user involvement—although in practice who plays the role of the user can often depend on who is available to talk to the developers—and use short iterative development cycles to develop evolutionary prototype solutions in a manner that takes account of local contingencies (e.g., see Boehm and Turner, 2004 ). However, agile methods are mostly concerned with end-user requirements, and make the simplistic assumptions that: (a) suitable users are available to interact with the development team and (b) the user requirements are congruent with broader organisational requirements. While there are certainly interesting ideas emerging from agile methods, their focus on interaction with individual users does not address the need for broader socio-technical awareness in systems engineering.

In addition to the approaches covered by Mumford’s (2006) extensive review, we have also identified several other approaches that encompass socio-technical ideas. We believe that these other approaches can also help inform the development of socio-technical systems:

Soft Systems Methodology (SSM; Checkland, 1981; Checkland and Scholes, 1999 ), which builds on ideas from action research, has its roots in systems engineering rather than the social sciences. SSM treats purposeful action as a system: logically linked activities are connected together as a whole, and the emergent property of the whole is its purposefulness. One of SSM’s key features is its focus on developing an understanding of the problem (SSM uses the more generic term problematic situation). This understanding takes into account the roles, responsibilities, and concerns of the stakeholders that are associated with the particular problem. The understanding of the problem provides the basis for the solution, which again takes into account stakeholders’ differing viewpoints. SSM explicitly acknowledges that the final solution is based on attempting to accommodate the views (and needs) of the various stakeholders. We believe that problem understanding is one of SSM’s principal strengths, but it can also be used to develop information models of the more technical aspects of a system. It has been used to evaluate existing information systems too ( Checkland and Poulter, 2006 ).

Cognitive Work Analysis (CWA; Rasmussen et al., 1994b; Vicente, 1999 ) was developed to analyse the work that could be performed by complex socio-technical systems. It is therefore a formative approach based on predicting what a system could do, in contrast to most approaches which are either normative (how work should be done) or descriptive (how work is done).

The socio-technical method for designing work systems ( Waterson et al., 2002 ) focuses on system design. It is used to identify tasks that have to be allocated to machines (and hence implemented using IT) and also considers those tasks that have to be performed by humans (both individually, and as teams). This method is designed for general use in function allocation and socio-technical work systems.

Ethnographic workplace analysis (e.g., Suchman, 1987; Hughes et al., 1997; Viller and Sommerville, 2000; Martin and Sommerville, 2004 ) emphasises the situated nature of action, and has investigated how the results from ethnographic studies can inform the design of socio-technical systems. Ethnographic workplace analysis has largely focused on the operational issues that affect the functionality and use of a system. It has highlighted how workarounds and dynamic process modifications are commonplace and revealed the importance of awareness and the physical workplace in getting work done.

Contextual design ( Beyer and Holtzblatt, 1999 ) is aimed at designing products directly from the designer’s comprehension of how the customer actually performs work. It is founded on the notion that any system inherently embodies a particular way of working, which then largely dictates how the system will be used and how it will be structured. Contextual design gives rise to activities that are focused on the front end of design, and, in particular, on customers and their work.

Cognitive systems engineering ( Hollnagel and Woods, 2005; Woods and Hollnagel, 2006 ) deals with the analysis of organisational issues, and offers some practical support for systems design. CSE uses observation as a tool for analysing work in context, and uses abstraction on the results to identify patterns in the observations that occur across work settings and situations, thereby increasing the understanding of sources of expertise and failure.

Human-centred design ( International Standards Organisation, 2010 ), which follows principles such as basing the design upon an explicit understanding of users, their tasks, and the environments in which those tasks are carried out. It also includes as one of the four main design activities the understanding and specification of the context in which the system will be used, and explicitly refers to consideration of social and cultural factors, including working practices and the structure of the organisation.

STSD methods can be categorised based on the how well they deal with the three broad stages in the systems engineering lifecycle: analysis, design and evaluation. There are also some general sets of principles that provide abstract guidance for developing socio-technical systems, rather than directly supporting detailed aspects of systems development. These include Cherns’ (1976, 1987) and Clegg’s (2000) principles, which cover aspects such as power and authority ( Cherns, 1987 ), and the fact that design should reflect the needs of the stakeholders ( Clegg, 2000 ).

Table 1 indicates how some of the better-known approaches relate to the different phases of the systems engineering life cycle. All of the methods tend to be most strongly related to one particular phase of the life cycle, although they still provide some support for the other phases. Whilst several of the approaches offer support for most phases in the systems engineering lifecycle, our belief is that none of the approaches provide complete coverage for all of the phases.

The development of STSD methods has identified and attempted to address real problems in understanding and developing complex organisational systems which, nowadays, inevitably rely on large-scale software-intensive systems. Despite positive experiences in demonstrator projects, however, these methods have not had any significant impact on industrial software engineering practice. The reasons for this failure to adopt and maintain the use of STSD approaches have been analysed in several places, and from several viewpoints (e.g., Mathews, 1997; Mumford, 2000, 2006 ). We summarise the main problems identified by these authors below, and also discuss other issues that have arisen in our own use of STSD methods.

Relationship between socio-technical systems design approaches and the development phases of the systems engineering life cycle. A double tick (✓✓) indicates that a particular design approach provides strong support for the associated phase of the life cycle; a single tick (✓) indicates some support.

4.1 Inconsistent terminology

There is considerable variation in what people mean by the term socio-technical system and this is inevitably confusing to potential adopters of these approaches. The term has its original roots in organisational and clinical psychology, in work carried out by the Tavistock Institute in the 1950s and 1960s. However, it is also often closely linked with the field of management science in the UK, where the ETHICS method ( Mumford, 1983, 1995 ) was developed at the Manchester Business School.

Nowadays, many different fields have adopted the term, often using their own interpretation—sometimes focusing on the social system, sometimes on the technical, but rarely on both together. This may help to explain the somewhat disparate nature of the literature (e.g., Griffiths and Dougherty, 2001 ).

It is important that people involved in a specific systems development project have an agreed understanding of what is meant by the term socio-technical system. This particularly applies to the development team, in order to make sure that they focus on the appropriate social and technical aspects of the system and how these are interdependent and interact. The critical point is that there needs to be agreement about the social and technical elements of the system that need to be jointly optimised.

4.2 Levels of abstraction

Similar to the problems of terminology are problems in determining the appropriate levels of abstraction to use when analysing and describing socio-technical systems. Rather than using different terms to describe the same thing, though, here we are talking about people describing the same system but using different levels of abstraction, often based on the fact that they draw the system boundaries in different places. There is a tendency by some to decompose the system into separate social and technical systems. The depth of analysis for each of the (sub-)systems is then given different emphasis, with the focus often falling mostly on the technical aspects of the system ( Eason, 2001 ).

Finding the appropriate level of abstraction is critical, but often not easy. Hollnagel (1998) , for example, criticises the work on socio-technical systems for over-emphasising the context, which includes the organisational aspects, at the expense of neglecting the individual. He argues that current approaches cannot satisfactorily explain why humans perform erroneous actions and, hence, cannot be used in human reliability analysis. When this view is taken to the extreme, undesirable events are simplistically seen as the result of organisational failings, which stack the odds against the human operator, who is then portrayed as the innocent victim of these failings. In other words, it overlooks the fact that the context includes individuals, often working as part of a team, who through their own volition could still theoretically perform the correct action.

4.3 Conflicting value systems

In attempting to make sense of the literature, Land (2000) suggested that it can be divided into two basic categories. Each category is based on a set of values that underpins much of the thinking around socio-technical systems.

The first set of values is a fundamental commitment to humanistic principles. In other words, the designer is aiming to improve the quality of working life and job satisfaction of the employee(s). It is argued that increases in productivity will automatically follow, and that these will generate added value for the company. Early approaches to STSD were particularly concerned with ensuring that humanistic principles were considered during the design and deployment of new systems.

The second set is often described as managerial values. In this view, socio-technical principles are regarded as a means of helping to achieve the company’s objectives (particularly economic ones). Humanistic objectives are perceived as having limited inherent value, but if their achievement leads to better employee performance, and the company benefits as a result, then all well and good. Approaches such as contextual design are primarily geared to the use of STSD as a means of building systems that provide more effective organisational support.

Ethnographic analysis can be considered as an intermediate category. Most work in this area has adopted an ethnomethodological approach where, it is claimed, the analysis of the work is not influenced by any particular theoretical framework or intended outcome. The extent to which such analysis is truly value-free is, of course, debatable.

Problems arise when these different sets of values come into conflict. The dichotomy between the first two categories helps to explain why, in some cases, managers and employees (as represented by trades unions, for example) can both be somewhat suspicious of socio-technical ideas, with the former applying managerial values, and the latter, humanistic values.

4.4 Lack of agreed success criteria

There has been significant theorising about the way to design socio-technical systems, but recent published examples of successful use in the design of software-intensive systems are comparatively scarce. Consequently, there has generally been little evaluation of the efficacy of using STSD approaches. Indeed, one of Majchrzak and Borys’ (2001) major criticisms is that existing socio-technical systems theories are not specific enough to allow for empirical testing. Other reasons for the lack of evaluation include the predominant research emphasis on system design rather than evaluation and, in the UK at least, the difficulties of funding long-term, longitudinal research. Large scale complex IT systems often have a lead time that is measured in years, rather than months—in a hospital for example, it may take several years to introduce a new system throughout the organisation.

Another problem of assessing success is the difficulty in establishing evaluation criteria for the social elements of the system. Whilst benchmark tests can be used to determine whether the technical part of the system meets the appropriate criteria (response time, throughput, cost/benefit analysis), it is more difficult to determine if a system is a better fit to organisational needs, or that a system has increased the quality of working life of the staff. The latter often requires examining or measuring derived effects. So, for example, if a system claims to increase job satisfaction (as a first order effect), this might be measured by looking at the change in levels of absenteeism, improvements in health, and increases in productivity ( Land, 2000 ). This evaluation is made harder by the fact that there are other, quite separate, influences on these factors and in many cases it may be impossible to link them directly to some new system.

Furthermore, the success (or otherwise) of the implementation is defined by a range of stakeholders, particularly operators, middle management and top-level management ( Land, 2000 ). Each category of stakeholder is likely to have a different viewpoint on the system and different criteria for success.

Related to the lack of criteria for success is the absence of work that demonstrates the cost-benefits of STSD methods and tools. Similar problems have also affected other (related) fields such as HCI, and more generally, human factors/ergonomics. New methods may be perceived by managers and systems developers as simply adding extra time, effort and cost to what are already long and expensive development projects. Demonstrating the cost effectiveness of STSD methods should be an important goal, as is the need for them to integrate with existing system development processes.

4.5 Analysis without synthesis

Socio-technical design methods have mostly been used to analyse existing systems, but these methods are limited in the support that they provide for the more constructive synthesis where the results of the analyses are systematically used in the software design process. In other words, they have been used to critique existing systems that (may) have failed, but without always suggesting how the problems could be fixed by appropriate re-engineering of the system (e.g., see Kawka and Kirchsteiger, 1999 ). There are not many recorded examples of the successful use of these ideas in a prospective manner, particularly for the first instance of a new type of system. This may be due to the envisioned world problem ( Woods and Dekker, 2000 ) which arises because of the difficulty of imagining or predicting the relation between people, technology and context in a domain that does not yet exist.

There are techniques that can be exploited in the construction of new systems ranging from the general notion of learning from past experience, to utilising existing components (appropriately adapted to the situation at hand). Petroski (1986, 1994, 2006) , for example, has documented how engineering has progressed as a discipline over the centuries by learning from its past failures. At a lower level, the work on patterns of co-operative interaction ( Martin and Sommerville, 2004 ), offers a way of supporting the re-use of insights gained from previous fieldwork in new system design.

4.6 Multidisciplinarity

Some of the failings of STSD can be attributed to the multidisciplinary nature of system development. The need for several disciplines to be involved is widely accepted, but the borders between the disciplines have been largely maintained, despite efforts at creating interdisciplinary teams by involving domain specialists in the design process. The issue is mainly down to failures in understanding and communication, where one discipline does not fully understand what the other disciplines can do ( Bader and Nyce, 1998 ), and hence does not ask them to deliver something that assists the system development processes. Dekker et al. (2003) , for example, have suggested that practitioners of ethnography and contextual design fail to deliver products that can be used by other disciplines. Their argument is that some of the work carried out by ethnographers and those involved in contextual inquiry does not go far enough, because it essentially stops after collecting data, rather than analysing the data to ascribe meaning to it so that it could be more readily used by others. This was reflected in a report on cooperation between software engineers and sociologists, where it was found that differences in both language and culture were major barriers to multidisciplinary work ( Sommerville et al., 1992 ).

In general, the maintenance of boundaries between the various disciplines may be a result of the way that systems development has traditionally been perceived and carried out. Specialised individuals or teams were typically allocated responsibility for a particular stage of development, such as requirements analysis or user interface design, and were rarely involved with other developers. Rather than relying on specialised individuals (or teams), what is required is that an individual (or team) has a working knowledge and appreciation of what the other disciplines have to offer, and can communicate effectively with them.

4.7 Perceived anachronism

Changes in ways of working at organisational, national and global levels were at least partly reflected in changes in attitudes towards STSD. In the late 1980s, for example, companies started to move towards lean production methods and business process re-engineering (BPR), often based on the use of new enterprise systems. The philosophy that underpins these methods ostensibly runs counter to many of the humanistic ideas behind STSD (e.g., Niepce and Molleman, 1998 ), and there were no attempts to try and adapt the STSD methods to the changing business management methods. It is somewhat ironic that it was BPR that made the explicit link to IT innovations, while the socio-technical systems community expended significant energy in the preceding decades on ideological debates ( Mathews, 1997 ) rather than trying to keep pace with technical and organisational developments.

In addition, STSD approaches were largely developed during the 1960s and 1970s, before the advent of the personal computer, and widespread use of interactive computing systems. It was only in the 1980s, however, that HCI achieved widespread recognition as a separate discipline, with its inherent focus on the importance of the interaction between people and technology at the lowest level rather than just the design of the user interface. It explicitly recognised the importance of the roles of the social and technical aspects of work. Many STSD approaches, however, fail to take account of the work in HCI and hence have little to say about interaction design.

The failure to reflect developments in organisational methods and technology can make STSD appear rather anachronistic and unfashionable. This is particularly true when designing new systems that are based on innovative ways of working and novel technology.

4.8 Fieldwork issues

Although STSD methods such as participatory design prescribe the involvement of users, it is comparatively silent on issues such as which users to select, what level of experience in design they need and so on ( Damodoran, 1996; Scacchi, 2004 ). More generally for fieldwork, there are problems with identifying the system stakeholders in the first place, before deciding which groups of stakeholders (and which individuals) should be involved. Traditional approaches involving an embedded ethnographer are expensive and prolonged, although notions such as ‘quick and dirty’ ethnography address this to some extent ( Crabtree, 2003 ).

The key issue, perhaps, is the identification of the focus, extent and level of detail required in the fieldwork. This is not just a problem for STSD. Within HCI, for example, there have often been discussions about the pragmatics of using available methods, which are seen as overly time consuming and unwieldy. Discounted engineering ( Nielsen, 1993 ) and lightweight methods (e.g., Monk, 1998 ) offer possible solutions.

4.9 Summary

The problems that we have identified all need to be solved if socio-technical approaches are to be accepted and effectively used by the systems engineering community. None of them are insurmountable, although the solution to some of the problems, such as the lack of agreed success criteria (Section 4.4) will only emerge as people apply the framework. We have used the problems to inform the requirements for a discipline of socio-technical systems engineering, which we describe next.

In reflecting on the history of socio-technical methods, Mumford (2006) suggested that these methods continue to be relevant, arguing that there is still a role for humanistic, socio-technical ideas in the 21st century. In addition to the humanistic arguments, we believe there is a strong pragmatic case for applying socio-technical approaches to systems engineering. Simply put, the failure of large complex systems to meet their deadlines, costs, and stakeholder expectations are not, by and large, failures of technology. Rather, these projects fail because they do not recognise the social and organisational complexity of the environment in which the systems are deployed. The consequences of this are unstable requirements, poor systems design and user interfaces that are inefficient and ineffective. All of these generate change during development, which leads to delays in the delivery of the system, and to a delivered system that does not reflect the ways that different stakeholders work.

We have noted that the system stakeholders inevitably have different concerns. The main concern of the system developers is usually whether the system meets the specified requirements. The main concern of the users is usually whether the system will help them do their job, without adversely affecting other parts of their work. The main concern of management is whether the system will generate added value to the organisation in a timely manner and whether it is compliant with regulatory requirements. Reconciling these different concerns is not a simple task.

We argue that these concerns can be addressed, at least in part, by evolving current socio-technical methods into a discipline of socio-technical systems engineering (STSE), in which a socio-technical approach pervades the entire systems engineering life-cycle. Our vision is for a discipline that combines the philosophies of the STSD approaches with the complementary methods identified in Section 3. STSE has to be founded on the recognised strengths of socio-technical approaches but must also address the recognised problems in existing approaches (see Section 4). Furthermore, we have to take into account the barriers to introducing any new approach namely:

New methods require upfront investment for an unknown later return.

There is often a high entry cost in terms of tooling and training to use new methods.

The challenge of method usability–experience is required to improve method usability but if initial usability is poor, the methods will not be used.

These constraints mean that, whatever the academic credentials of new techniques and methods, it is hard to get practitioners to adopt them. If STSE is to become a reality, we need to recognise these barriers and develop approaches that minimise the costs of introduction and the associated risks.

In promoting STSE, our intention is to focus on the development of complex IT systems, as well as providing a more effective basis for analysing existing systems. In this way, we hope to overcome the tendency to simply analyse existing systems that has often affected STSD methods (see Section 4.5). Instead, we intend to use the results of the analysis to exploit what we have learned about socio-technical systems (including how they can go wrong, for example) and synthesise the results to help in designing better systems ( Coiera, 2007; Walker et al., 2008 ).

We consider a complex IT system to be a system that includes one or more networked, software-intensive systems that is used to support the work of different types of stakeholder in one or more organisations. In general, we assume that these systems are ‘systems of systems’ involving databases, middleware and personal applications such as MS Excel. We make no assumptions about the technologies used to develop the system, but note that it is increasingly the case that such systems are constructed by configuring off-the-shelf ERP systems such as those provided by SAP ( Pollock and Williams, 2009 ). Nowadays, new systems are rarely completely new, but instead incorporate and inter-operate with a wide range of existing systems. The costs of integration are likely to exceed the costs of developing the new components of the system ( Hopkins and Jenkins, 2008 ).

We fully realise that in order for STSE to be successful we need to bring about something of a change of mind-set among systems engineers. This is no small task, because engineering per se has developed its own culture over a long period of time, and is often slow to change ( Vincenti, 1993 ). It does, however, have a history of changing as a result of learning from failures ( Petroski, 1986, 1994, 2006 ), so we intend to promote STSE by highlighting the socio-technical nature of system failures, and indicating the lessons that need to be learned. In this way we believe that we can help systems engineers become more aware of the usefulness of the social sciences, and hence make them more amenable to socio-technical ideas.

We strongly believe that if we want to make an impact on practical systems engineering, we have to start with existing systems engineering processes. Socio-technical considerations are not just a factor in the systems development process: social-technical factors have to be considered at all stages of the system life-cycle. While systems engineering processes differ considerably between organisations, we have observed four fundamental activities in all complex organisational IT systems development projects ( Fig. 1 ):

 alt=

Systems engineering activities.

Procurement . Decisions are made on what systems to re-use and what new systems to procure from internal or external suppliers. Some analysis will normally precede this, but this is rarely an in-depth analysis of the areas of the organisation where the system will be used.

Analysis . Stakeholders in the system are involved in a process that results in requirements for the new components of the system that is to be introduced.

Construction . The new components of the system are constructed and integrated with existing systems and databases.

Operation . The system is deployed and put into use. Over time, changes to the system are proposed and the development activity continues to create new releases that are deployed and used.

In Fig. 1 , we have deliberately avoided showing these activities as sequential. We believe that they are fundamental to all complex IT systems and that these activities interchange information. The nature and extent of the information interchange varies considerably. For example, a military system may involve an extended analysis phase which culminates in the publication of a detailed requirements document. This is then input to the construction phase with a tightly controlled change management mechanism for feedback to the analysis phase. In contrast, agile development approaches interleave analysis and construction with informal requirements used to drive the construction of the system.

When new business systems (or systems of systems) are introduced, this is often in conjunction with a change process where there is a goal of (usually) implementing significant changes to the business or its processes. Segarra (1999) , for example, highlighted the importance of making sure that IT developments and business change were integrated in the manufacturing of aircraft and cars in Europe. The organisational change process has a structure comparable to the development process, as shown in Fig. 2 . While this change process should (and to some extent does) take into account social and organisational issues, the changes are often deliberately disruptive because the organisation wants to impose process change. There is likely to be a reluctance to invest in understanding existing processes and their fit with the organisation because these processes are seen as obsolete and due for replacement.

This attitude can lead to serious problems because existing processes have been adapted by the people involved to take particular organisational and workplace concerns into account. A failure to understand the details of actual processes may mean that replacement processes are less suited to the work as it is really done and, hence, are considerably less efficient than current processes.

 alt=

The organisational change process.

A major problem in many organisations that we believe is an important contributor to system failure is that there are often only weak connections between change processes and system development processes (although see Segarra, 1999 for one attempt at integrating business processes and IT innovations). There are separate change and systems engineering teams, with the principal communication between them being a requirements document or a set of process workflows. Those involved in the change process may be unaware of technical factors that limit the flexibility of the system that is being developed. Those involved in the development process may have no real understanding of the ways that the proposed workflows will be instantiated in practice, nor of the environment where the system will be deployed.

Proponents of STSD have regularly referred to the process of design as being a socio-technical system itself. However, as noted above, there are actually two distinct processes that often only communicate infrequently. In the worst case, they are linked at the start of the project, when some form of requirements are gathered, and at the end of the project when the system is delivered. In the interim period, both processes are operating simultaneously, usually at different rates, and rarely interchanging information, even though the operation of one often has an impact on the operation of the other. The organisational issues being addressed by the change team are not communicated to the systems engineering team; the technical issues that constrain organisational change are not fed back to the change team.

Our vision of STSE is that it can serve as a means to bridge the system development and change processes as shown in Fig. 3 . The application of this approach should feed information to the development team about socio-technical issues and provide support for using this information constructively in making design decisions in a timely manner. Similarly, STSE should provide the change team with cost-effective approaches to socio-technical analysis and provide information to them about technical factors that constrain the possibilities of change.

To realise our vision we need to improve communications between system stakeholders about socio-technical issues, and provide constructive support for using information about socio-technical factors in both technical systems design and organisational change processes. We therefore envisage two types of STSE activities:

Sensitisation and awareness activities . These are concerned with sensitising stakeholders across the system to the concerns of other stakeholders, and with convincing stakeholders of the value of a socio-technical approach. For example, engineers involved in designing the system database might be made aware of the fact that collecting complete data in some settings may be practically impossible.

Constructive engagement . These activities are concerned with integrating STSD approaches into the practical systems development and change management processes in an organisation. The nature of the constructive engagement varies depending on the development or change activities that are involved.

 alt=

Socio-technical systems engineering.

We discuss below in a little more detail what we mean by sensitisation and constructive engagement. Rather than just noting that we need to take account of the social and technical factors and their interdependencies, we explicitly identify who needs to be made aware of which factors, and provide a focus for the activities that are needed to integrate STSD approaches into the engineering life cycle. We note here, however, that identifying appropriate approaches to sensitisation and constructive engagement and integrating these into development and change processes are the key challenges facing STSE researchers.

As well as bridging the change and system development processes, STSE can inform the change and systems development processes of broader organisational goals and constraints. It therefore acts as an information bridge between the wider organisation and specific projects to develop new complex IT systems.

This notion of STSE as a means of linking and coordinating change processes and systems engineering processes is pragmatic and deliberately limited. Our intention is to provide a framework through which we can use socio-technical approaches in practice and convince practical engineers of their value. While a broader notion encompassing humanistic work practices or organisational re-design could be adopted, we believe that our less ambitious approach has a better chance of adoption. Our approach is less threatening to existing management and can be introduced in an incremental way. If we can succeed in a limited way, we will then be in a better position to extend the scope of STSE.

We cannot and do not claim that this deliberately limited view of socio-technical systems engineering solves all of the problems that we identified in Section 4 of this paper. However, by rooting the approach in the language of business, by explicitly linking to the notion of change management and by proposing close interaction between development and change management teams, we believe that we address some of these problems including inconsistent terminology, lack of agreed success criteria (success is related to the success of the change proposals), analysis without synthesis, multidisciplinarity and perceived anachronism. Other work that we are involved with is concerned with using responsibilities as an abstraction to represent work ( Lock et al., 2009; Sommerville et al., 2009 ). This focuses on appropriate abstractions for STSE and may be incorporated into the approach described here at some later date.

5.1 Sensitisation and awareness

The primary aim of sensitisation activities is to ensure that system stakeholders, including the development engineers, are made aware of the socio-technical issues that may affect the design and use of the system. In short, they have to be convinced that adopting a socio-technical approach is worthwhile and persuaded to actively participate in the process. Based on our experience, we have noted several types of sensitisation activity:

Sensitising system engineers to the notion that socio-technical factors should be considered during system design, and to the cultures of the organisation’s different stakeholder groups. In large organisations, different parts of the organisation may have their own cultures and there is a need for better cross-organisational understanding of these.

Sensitising those involved in procuring a new, complex IT system to the socio-technical considerations that may influence the design and use of the system.

Sensitising system stakeholders to the socio-technical issues that, almost inevitably, are a source of conflict with other stakeholders.

Sensitising system stakeholders to the notion that an analyst will be studying their work with a view to a deeper understanding of it, rather than to assess or audit what they do. Here, concerns such as snooping and reporting to management have to be addressed.

Sensitising stakeholder groups to the different world views of other groups, perhaps from different disciplines, in the organisation. For example, accountants think about financial transactions in one way and are concerned about ensuring accounting regulations are followed; users of financial data may think about these transactions in a totally different way, reflecting their own management responsibilities.

Sensitising management and other system stakeholders to the real technical constraints that limit what is possible with a software system.

The need for sensitisation varies depending on the people in an organisation and the organisation itself. In line with the pragmatic nature of STSE, activities are selectively employed as circumstances dictate. It is clear from our extensive experience in ethnographic studies, however, that sensitisation is essential if the later stages of systems engineering are to succeed. Failure at an early stage will inevitably mean that key system stakeholders will not understand the impact of socio-technical factors on systems and why systems design is not simply a technical process.

A key issue here, of course, is how to we achieve sensitisation in practice. The academic literature is of little help because, naturally, existing socio-technical studies have already crossed this barrier and have convinced companies and other organisations to become involved in these studies. Clearly, practitioners rarely read academic papers and appealing to the canon of work on socio-technical systems is unlikely to be an effective approach. There are three possible approaches that we have previously investigated and that we believe have some potential here:

Taking engineers to the workplace . The idea of bringing users to the software development team is one that is widely accepted (e.g., in agile methods) but we believe that taking software developers into the workplace, even for a short time, can reveal to them the complexity of work and the difficulties faced by system users. This approach is one that we have found to be successful in a number of different situations ( Bentley et al., 1992b; Lock et al., 2008 ).

Workplace vignettes . Of course, the practicalities of achieving this can be daunting, so we have explored the notion of ‘ethnographic vignettes’, textual and video descriptions of situated work, that highlight socio-technical issues for engineers and managers ( Clarke et al., 2003; Martin et al., 2006 ).

War stories . War stories are short illustrative descriptions of problematic situations ( Orr, 2005 ) that have arisen and how these have been addressed. We have catalogued a set of war stories relating to problems that arose in the development and deployment of an electronic patient record system ( Martin et al., 2004; Mackie, 2006 ).

We cannot claim that these are complete solutions to the problems of sensitisation and there are real practical difficulties in presenting both vignettes and war stories. However, the availability of social media such as YouTube, may offer some opportunities to make this information widely and easily accessible.

5.2 Constructive engagement

Constructive engagement activities provide a means of integrating STSD approaches into the systems engineering and the organisational change processes, and synchronising the two processes at appropriate points. The precise nature of the constructive engagement will vary from project to project, largely determined by which particular activities in the development and change processes are involved. Here we discuss three types of constructive engagement.

5.2.1 Problem definition

Software design methods are geared towards developing a solution to ‘the problem’, so if that ‘problem’ is not understood, applying the methods will generate an inappropriate solution. The nature of the identified problem, though, is rarely simple because each group of stakeholders has its own viewpoint about what it really is. Instead of there being one single problem, there is usually a set of overlapping problems with conflicting characteristics. Indeed, some of these ‘problems’ may be no such thing – some stakeholders may be perfectly happy with the status quo and their ‘problem’ is that a new system is being imposed on them because of the requirements of other stakeholders.

STSD approaches have recognised that understanding ‘the problem’ that the system is intended to address is one of the keys to success, which is why many STSD methods are oriented towards analysis and problem understanding. Using an STSD approach will therefore help the stakeholders to focus on the nature of the problems and issues and come to some agreement about what these really are. It will also help systems developers to understand the real problems—rather than what they perceive as being the ‘problem’—their system is supposed to solve.

The alignment of the systems engineering and organisational change processes during problem definition is facilitated by organising, presenting and analysing the process and environmental issues using a coherent framework. The result should be a description of the work context that has been agreed by the stakeholders, accompanied by a set of corresponding requirements based on work performed in that context. These requirements, in principle at least, will define: the purpose of the system within the wider organisational context; the practicalities of its use in its operational environment; and the functionality it provides to system users. Achieving an appropriate balance between these different requirements forms the basis for the construction of a system that will be acceptable to, and used by the end users, as well as delivering the expected benefits to the stakeholders.

In practice, however, expressing what is really required by system stakeholders as a set of requirements means losing some of the richness that is typical of socio-technical analysis. Requirements can state broad functionality, but the way that the functionality is realised and the ways that the system presents information to stakeholders cannot be described using requirements statements. We know that HCI design, for example, depends on prototyping and experimentation; other aspects of STSD such as support for cooperation and collaboration must also be explored and discovered rather than pre-determined.

5.2.2 Constructing the solution

We use the term construction rather than design and implementation because approaches such as agile development and configuration of ERP systems do not distinguish between these activities. The key to success lies in ensuring that the engineers involved in systems construction are aware of socio-technical issues—particularly the interdependence of technical and organisational aspects—and the realities of the environment in which the system will be used. It is also important that there is agreement within the organisation about which methods will be used during development. In this way we can alleviate design and implementation decisions that make it more difficult to incorporate the system into everyday, routine work.

Getting the construction right is not simply a matter of writing better system requirements. In the same way that requirements for a user interface cannot adequately express the richness of the interaction with a particular system, social and organisational complexity cannot be simply distilled into ‘social’ or ‘cooperation’ requirements. System requirements are still needed to provide engineers with a broad understanding of what has to be constructed. The agile approach of involving end-users as ‘owners’ of requirements is a good one but needs to be extended to take into account a broader set of system stakeholders.

An unavoidable constraint on construction is the need to fit with existing procurement and systems engineering processes. For good reasons, organisations are very reluctant to make radical changes to these processes, so STSE has to integrate with them rather than be presented as a new, additional approach. If the procurement process does not consider usability then it should be extended to include it. If it is left to the supplier to decide on the levels of usability, these will be determined by the time and resources available during development, rather than seen as a requirement that has to be met (e.g., see Artman, 2002 ).

The human-centred design methods that have been developed in the field of HCI provide one way of making sure that technical and social aspects are considered together. The use of prototyping, for example, allows users to think about how they would use the system, and offer feedback on the way that the system will look and feel before the final system is delivered. It also provides a way of synchronously linking the systems development and organisational processes.

5.2.3 Evaluation

The evaluation of a socio-technical system involves assessing the deployed system to understand how well it has met the expectations of its stakeholders. In the ideal world, where perfect knowledge of the future was available, it would be possible to lay out all the criteria for evaluation during the analysis of the system, when the system goals are set. In reality, systems are frequently oversold with inflated expectations of how they will perform in a situation that often is unknown during the construction stage— Woods and Dekker’s (2000) envisioned world problem—with the net effect that the final system fails to satisfy those expectations. It is therefore important to recognise that the nature of evaluation changes as the design and the organisation evolve, and that the expectations of the stakeholders will also change accordingly.

Human-centred design approaches advocate evaluation throughout the development process and in the longer term ( International Standards Organisation, 2010 ). Full systematic evaluation of a deployed system is rare, however, partly because organisational issues get marginalised (e.g., Doherty and King, 2001 ). The original system stakeholders may have moved on, and the new stakeholders may have different expectations, based on their experience of the deployed system. Some stakeholders also take a fatalistic approach: they see themselves as being stuck with the system, so there is no point in complaining about it. Other stakeholders who are in a position to complain, simply refuse to use a system that they do not like, and disassociate themselves from it.

Nevertheless, we argue that there is a place for lightweight evaluation as part of the STSE cycle. This should not be seen as a means of criticising the original stakeholders or requirements, but rather as a constructive activity that leads to a more effective operational system. Essentially, the evaluation should be concerned with ‘filling in the gaps’ in the analysis of the system which may arise because of incompleteness or incorrectness, or because of subsequent organisational change. In other words, when new requirements arise, or existing requirements change on the organisational side, or when problems arise with satisfying the original requirements on the systems development side, these need to be assessed in their own right, and in terms of the wider development project. This is because they are likely to change the shape of the delivered system, and hence the nature of the evaluation of whether the system meets its goals.

We see the one of the primary roles of evaluation as being its contribution to the process of ‘domestication’ ( Williams and Edge, 1996 ) where the system gets bedded into the organisation. Domestication is the activity of familiarisation with new software and changing both the software and business processes so that the software becomes an integral part of everyday work. The types of questions asked during evaluation are therefore not ‘does this work?’ but ‘how can we make this work?’ This may, of course, lead to change proposals and further iterations of the analysis and construction activities. However, the changes required may be process changes that people carry out to fit the system into their normal work practice.

In this paper, we have briefly reviewed several methods for developing socio-technical systems and suggested why these methods have not entered the mainstream of system design practice. Based on this and on our own extensive experience—both authors have over 15 year’s experience of working with industry, understanding industrial concerns and transferring research results into practice—we have proposed a pragmatic framework for socio-technical systems engineering. We believe that this framework can be used as a basis for integrating socio-technical analysis and practical, technical systems engineering. We have deliberately designed it as a means of linking organisational change processes and technical systems development and make no claims that our framework provides complete coverage of all socio-technical issues.

The framework is based on almost 20 years of experience of attempting to integrate social and organisational insights from workplace studies into the systems engineering process. The key lesson that we have learned from this work is that there cannot be one simple way to achieve this and that a variety of different techniques, appropriate to the organisations involved should be adopted. We believe that the framework we propose provides a basis for focusing socio-technical analysis around real business concerns and hence increasing the probability of uptake. It establishes a general model that will, inevitably, be instantiated in different ways in different organisations.

The fact that the framework does not exist in isolation from its instantiation and situated use means that an empirical evaluation of the framework is not currently practical. Separating the value of the framework from its instantiation (essential for empirical framework evaluation) is, in our view, impossible. We have qualitatively evaluated our ideas through discussions with industrial collaborators and have received positive feedback from them.

In outlining our framework for STSE we have been particularly influenced by work on ethnographic workplace analysis and on cognitive systems engineering. The STSE framework is also compatible with Resilience Engineering ( Hollnagel et al., 2006 ). In particular, STSE addresses the way that people use everyday workarounds to keep systems running, and how people often intervene to mitigate the effects of failures that could otherwise have serious adverse consequences. Furthermore, the framework is also consonant with human-centred design approaches ( International Standards Organisation, 2010 ), although our framework makes explicit the relationship between system development and organisational change.

We believe that the different socio-technical design methods have much in common and our notions of the basic activities of STSE allow any method of socio-technical analysis to be used. Methods of analysis, in our view, are not the issue. Rather, research in STSE should address the engineering problems of applying socio-technical approaches in a cost-effective way and integrating STSE with existing systems and software engineering processes.

Research in this area requires an interdisciplinary approach and may involve computer scientists, software engineers, HCI designers, psychologists, sociologists and human factors specialists. We believe that all of these areas still have much to learn from each other. We would advocate the use of techniques such as action learning ( Revans, 1982 ) here, so that people can learn to know what things they do not know about, and to ask people in similar positions questions so that they can explore and overcome their ignorance.

Some of the most important areas are:

STSE processes Our model of STSE is based around the notions of sensitisation and constructive engagement. The research issues here relate to the specific activities that might be involved in the STSE process to manifest these notions and how these can be integrated with systems engineering process activities.

How can requirements be made richer to incorporate information about socio-technical processes? In reality, the model of system development where systems are built to a specification of requirements is not going to change for complex systems. Nor, in our view, should it change. However, current requirements documents are usually impoverished descriptions of how work is done and what is really needed. We need to develop guidance for requirements writers that allows them to express a richer picture of the socio-technical systems to the engineers responsible for systems development.

How do we transfer knowledge and experience from one organisation to another? The issue here is discovering how to separate the essential (what applies to all organisations in a sector) from the accidental (the specific ways in which an organisation works). We will then be in a position to transfer process knowledge across organisations.

What tool support is effective in supporting STSE processes? We need to make use of existing tools—both software engineering tools and Web 2.0 tools—that support collaboration and communication (wikis, social networks, and so on). We need to know more about how to deploy existing tools for distributed project support, how to use these tools to support problem solving, how to integrate technical and social tools and so on.

Modelling and abstraction Modelling and abstraction is fundamental to software engineering, with models of different types being used by engineers to communicate. The practical use of socio-technical approaches has to acknowledge this by providing a means of modelling, and by integrating with existing approaches. Examples of research issues in this area are:

What models and abstractions are useful when thinking about systems design and interaction in a distributed multi-organisational system? The abstractions currently used in technical system modelling (e.g., use-cases, objects, etc.) do not seem to us to be sufficient to represent socio-technical considerations.

Can current approaches to system modelling (e.g., the UML) be adapted to reflect socio-technical considerations? What are the benefits and problems of adopting this approach?

Can organisations be meaningfully modelled to provide useful information for socio-technical systems design? This is a longer term issue which involves extending the scope of our framework beyond the change process in organisations to consider broader issues of organisational politics and dynamics.

Integrated human-centred design The importance of effective human-centred design is now generally recognised, if not universally practised ( Woods et al., 2007 ). However, most methods of socio-technical analysis have paid little attention to those areas of design relating to individuals ( Hollnagel, 1998 ). Furthermore, there is a tendency in the engineering community to identify all human, social and organisational issues as problems of the human interacting with the technology (such as “finger trouble”). In doing so, they ignore the relationship between individual interaction and the social organisation of work, and particularly how the latter can influence the former. Research issues here include:

How can we integrate methods of socio-technical analysis with methods that support HCI design and evaluation? Many HCI methods have focused on the individual whereas socio-technical methods focus on the organisation and groups within the organisation. We need to develop practical process guidance that allows organisations to use these methods together and to integrate their results.

How can we use the interface to highlight relevant socio-technical issues, such as awareness of work? The CSCW research community has addressed this issue and there have been a range of proposed techniques to support awareness (e.g., Gross et al., 2005 ). Much of this depended on special purpose systems and has been overtaken by the use of web-based 2systems. This work should be extended and developed to reflect modern interaction and to take organisational rather than situational considerations into account.

How can evaluation methods be extended to take organisational issues into account? Current approaches to evaluating HCI design are often based around the individual using the proposed interface. However, the organisational setting where work is done has a profound influence on the use of systems, and we need to extend evaluation methods to consider how organisational considerations affect the use of an interface. This is particularly relevant when things go wrong and the system has to support coping behaviour.

Organisational learning In many cases, the socio-technical problems that affect a system are not new. They have occurred before but the organisation has no means of learning from these problems or, indeed, from the problems of comparable organisations. We believe that we have to revisit the notion of organisational memory ( Walsh and Ungson, 1991 ) with a view to supporting the organisational learning process and thus reducing the chances of mistakes being repeated. Research issues in this area include:

How can different types of knowledge be captured at low cost and maintained in an accessible way? The problem of low-cost knowledge capture was, we believe, one reason why many attempts to implement organisational memory systems in the 1990s were ineffective. Capturing knowledge for the future distracts people from their everyday work so we need to discover techniques that capture information from normal work activities with minimal intervention from the people involved in these processes.

How can the use of organisational memories and other support for organisational learning be embedded in the STSE process? Organisational memories and learning from experience can only be effective if they are actually used. We need to invent ways of easily accessing such information as part of routine processes and ensuring that the information can be updated with accounts of practical usage experience.

How can we deploy modern tools and technologies (wikis, Google , etc. ) to develop a workable organisational memory system? People are becoming increasingly familiar with Web 2.0 collaboration tools. Using these as a basis for organisational learning means that initial barriers to tool use are lowered. We are convinced that using these web-based systems is the most effective way to reduce the costs of collecting and using organisational information. To do so, however, we need to investigate how to structure these tools to maintain long-term information about an organisation and its processes.

Global systems Existing approaches to STSD are virtually all based on an assumption that systems are located within a coherent organisation where the system stakeholders have similar cultural values and assumptions. However, there is now an increasing trend to create global systems, which may involve several disparate organisations that are located around the world. Similarly, the teams involved in complex systems engineering projects are geographically distributed across timezones and cultures. Research issues in this area of global systems and the globalisation of systems engineering include:

How should socio-technical systems design methods evolve to cover work that is not co-located? The evolution of socio-technical methods to address differences in organisational and social culture that cause problems to be understood and addressed in different ways.

How can fieldwork techniques evolve to collect information about everyday practice at remote sites? Many STSD methods rely on interaction with end-users either through interviews or direct observation of work. This direct interaction is often impractical when users are distributed across the world. Methods of information collection about work practice have to evolve to cope with this situation.

How are electronically mediated computer systems integrated with everyday work? Interaction of distributed teams is normally mediated by electronic systems. While there have been many studies of the use of systems such as email (e.g., Bellotti et al., 2003 ), we need to understand how teams work around the problems that they encounter when using such systems. We also need to understand how social networks and social media can be used effectively in professional situations to support socio-technical systems engineering.

We are under no illusions about the problems of introducing new methods and approaches or the length of time required to introduce them into an organisation. However, we are convinced that the increasing awareness in industry that systems problems are not just technical problems means that there is a real possibility of introducing a cultural change in the practice of systems development.

Whilst this is no easy task, we believe that we can achieve our goal by taking inspiration from Vicente (2008) . His starting point was to understand the failure to date of the human factors/ergonomics field to satisfy one of its main goals of bringing about societal change. So, in particular, we believe that we need to raise the profile of STSE within organisations; to highlight socio-technical failures as a way of promoting a move towards the use of STSE; and to exploit the opportunities presented by failures and service disruptions in a way that will encourage a shift towards the use of STSE. In this way we believe that we can establish a discipline of socio-technical systems engineering that meets the needs of the 21st century.

The authors would like to thank Denis Besnard, John Rooksby and Phil Tetlow for comments on an earlier draft, and the anonymous reviewers whose comments have helped to improve the paper. This work was funded by the EPSRC as part of the Large Scale Complex IT Systems Project.

Google Scholar

Google Preview

Here we use the term organisational to describe factors that are related to the company or business per se, whilst we use the term social to describe factors that are related to the relationships between people who work together within and across organisations.

Here Badham et al. are using the term social subsystem to refer to people, work context and organisations.

See also http://www.dirc.org.uk .

Email alerts

Citing articles via.

  • Recommend to your Library

Affiliations

  • Online ISSN 1873-7951
  • Copyright © 2024 British Computer Society
  • About Oxford Academic
  • Publish journals with us
  • University press partners
  • What we publish
  • New features  
  • Open access
  • Institutional account management
  • Rights and permissions
  • Get help with access
  • Accessibility
  • Advertising
  • Media enquiries
  • Oxford University Press
  • Oxford Languages
  • University of Oxford

Oxford University Press is a department of the University of Oxford. It furthers the University's objective of excellence in research, scholarship, and education by publishing worldwide

  • Copyright © 2024 Oxford University Press
  • Cookie settings
  • Cookie policy
  • Privacy policy
  • Legal notice

This Feature Is Available To Subscribers Only

Sign In or Create an Account

This PDF is available to Subscribers Only

For full access to this pdf, sign in to an existing account, or purchase an annual subscription.

socio technical case study

Gender and climate experts explore and share insights from case studies that utilize socio-technical innovation bundles (STIBs) in India

  • From CGIAR Initiative on Gender Equality
  • Published on 16.02.24
  • Impact Area Climate adaptation & mitigation , Gender equality , Poverty reduction, livelihoods & jobs

Share this to :

socio technical case study

CGIAR GENDER Impact Platform and CGIAR Gender Equality initiative, HER+  began the year with a webinar focused on socio-technical innovation bundling to strengthen women’s resilience to climate change: distilling learnings from case studies. This webinar held on January 25 presented insights on various case studies where STIBs have contributed to women’s empowerment and resilience and learn from them as we are piloting STIBs in multiple learning labs across the globe.

HER+ initiative is geared towards developing guidance on designing and implementing context-specific Socio-Technical Innovation Bundles (STIBs ). With funding from CGIAR GENDER Platform , HER+ work package ‘EMPOWER’ led by International Rice Research Institute (IRRI) documented cases from different parts of India where social, technical, and technological innovations are bundled for enhancing women’s empowerment and resilience.

The knowledge exchange featured speakers from partner organizations who shared key findings on how socio-technical innovation bundling has been applied to empower women and build their climate resilience in different parts of India. Partner organizations including Institute of Social Studies Trust (ISST), Grameen Foundation India , Swayam Sikshan Prayog (SSP), and Satish Satmile Club O Pathagar (SSCOP) in India conducted case studies to analyze their experiences of deploying such bundles and their outcomes.

STIBs help integrate technology development and technology adoption by combining social, technical and technological innovations to improve rural women’s access to agricultural technologies and contribute towards their empowerment and resilience. The webinar featured presentations based on the following four cases highlighting the findings and insights.  –

  • The Institute of Social Studies Trust (ISST) study focused on the work of Uthhan – a grassroot organization working amongst landless and marginalized women farmers in Gujarat. The case study, presented by Pallavi Sobti-Rajpal of Uthhan highlighted how a combination of leadership, training, peer-to-peer learning, and use of sustainable agricultural practices has effectively empowered them and made them aware of climate-smart agricultural practices.
  • Garima Joshi of Grameen Foundation India presented how the uptake of bio-fortified mustard (BFM) seeds increased amongst women farmers in Uttar Pradesh when the new technology was integrated with other social interventions, such as capacity building and knowledge dissemination sessions and collaborating with farmer collectives like Farmer Producer Organizations and Self-Help Groups.
  • SSCOP’s case study, presented by Tapan Chowdhury illustrated the usefulness of bundling innovations and rolling them out through women collectives resulting in increased adoption of technologies like rice transplanters, zero tillage, combined harvester: shade nets, and coco-pits among women farmers in the Eastern Gangetic plains of West Bengal project activities.
  • SSP’s presentation reflected on their experience of implementing the Women-led Climate-Resilient Farming (WCRF) model in the drought-prone region of Maharashtra. The study, presented by Dr. Radheshyam Jadhav, emphasized how the model integrated various technological innovations such as soil management, water harvesting, mixed cropping, vermicomposting and livestock rearing with technical innovations like capacity building training and social innovations like creating awareness around land rights to empower women and enhance their resilience.

This was followed by a presentation of a comprehensive synthesis encapsulating the key insights from the cases. The synthesis study conducted and presented by Dr. Monika Banerjee, an independent consultant working on this project highlighted the necessity to strategically and systematically plan and execute bundled innovations within large-scale, multi-stakeholder projects, guided by a shared vision toward a common goal. Many cases lacked this strategic approach, often combining innovations in an ad-hoc manner, leading to a gap in achieving the envisaged project goals.

The webinar shed light on how bundling approaches are currently used in development interventions. It generated a rich discussion of the suitability of using STIBs as a way of enhancing women’s empowerment and resilience to climate change through access to new technologies.

socio technical case study

                                  The full webinar is available here:  WATCH NOW .

Related news

socio technical case study

NATURE+ launches circular bioeconomy hub in India to inspire, innovate and integrate nature-positive businesses

  • Environmental health & biodiversity
  • Poverty reduction, livelihoods & jobs

CGIAR’s Nature-Positive Solutions Initiative, in collaboration with partner BAIF and other key sta…

socio technical case study

Millions of refugees and displaced people to gain protection, resilience, and long-term climate solutions thanks to new CGIAR-UNHCR partnership

  • Climate adaptation & mitigation

Author: Marianne Gadeberg, Independent Communications Consultant In 2023, CGIAR forged a new partner…

Photo credit: ILRI/Ukama Ustawi/Susan MacMillan

Scaling African innovations for Africa's food systems

CGIAR's partners in a dozen countries of East and Southern Africa are "keeping it real"…

This website uses cookies in order to improve the use experience and provide additional functionality Detail

  • Reviews / Why join our community?
  • For companies
  • Frequently asked questions

Socio-Technical Systems

What are socio-technical systems.

A socio-technical system (STS) is one that considers requirements spanning hardware, software, personal, and community aspects. It applies an understanding of the social structures, roles and rights (the social sciences) to inform the design of systems that involve communities of people and technology. Examples of STSs include emails, blogs, and social media sites such as Facebook and Twitter.

  • Transcript loading...

The basis of STSs is general systems theory, which describes what the disciplines of science have in common—i.e., that they all refer to systems: sociologists see social systems, psychologists cognitive systems, computer scientists information systems, and engineers hardware systems. In general systems theory, no discipline has a monopoly on science— all are valid.

These disciplinary perspectives on computing allow us to view computing through distinct levels and trace its evolution. Computing began at the mechanical level (hardware devices), evolved an information level (devices + software), then acquired a human level (IT + human-computer interaction), and finally a community level (STSs). A community works through people using technology, as people work through software using hardware. Consequently, social requirements are now an important part of computing design.

While sociologists study the social level alone as if it were apart from physicality, and technologists study technology as if it were not part of society, socio-technology is a distinct field of inquiry on how personal and social requirements can be met by IT system design. As such, STSs seek to merge people and technology, viewing the integration of computers into societal systems as the next evolutionary step of humanity. An STS approach to design raises the cost of development but results in complex systems, like social networks, that have far more performance potential. Exploring a design problem by rising to an STS mindset can reveal further dimensions of a design’s use potential and inspire development.

Questions related to Socio-Technical Systems

Social media platforms are prime examples of socio-technical systems. Take Facebook as a case. It combines technology with human interaction. People connect, share, and communicate through this digital platform. Facebook uses complex algorithms and software. It also relies on users to create, share, and engage with content. This interaction between human behavior and technological infrastructure makes it a socio-technical system. It showcases how technology and social dynamics intertwine to form a unified system.

Socio-technical systems theory explores how social and technical elements interact. Organizations work best when their social and technological parts align. Socio-technical systems theory believes people and technology should not be separated in analysis. Instead, you should view them as interconnected parts of a whole system. This approach helps design strategies that consider human needs and technological capabilities. It aims to create a balanced environment where technology supports human roles and social structures. The theory applies to workplace design, software development, and organizational change management.

A socio-technical system (STS) in software engineering is the complex interplay between social aspects (people, organizations, cultures) and technical aspects (machines, software, hardware, etc.) of a system. It underlines the idea that the design and functioning of organizational systems are influenced not only by the technical elements of tools and methods used but also by social factors such as human interactions, values, norms, and expectations.

Watch the video below to learn more about the complexity of understanding socio-technical systems. It emphasizes the challenges people face in grasping these interconnected systems.

Understanding the socio-technical system in software engineering can help teams better design, implement, and maintain software systems. It helps teams acknowledge that a change in the technical parts of a system may impact the social aspects and vice versa. This approach encourages a holistic view of systems, considering all aspects that will play a role in successfully developing, deploying, maintaining, and improving software systems.

Characteristics of socio-technical systems include:

Integration of Social and Technical Elements: These systems blend technology with human social elements.

User-Centric Design : Designers focus on designing systems based on how people interact with the technology.

Adaptability: They are flexible and can adapt to changes in social or technical environments.

Complex Interactions: These systems feature complex interactions between people, technology, and the environment.

Goal-oriented: They aim to achieve specific objectives, balancing technical efficiency with social needs.

Collaborative Nature: They often involve collaboration among stakeholders. 

Evolutionary Development: They evolve through continuous feedback and learning.

Interdisciplinary Approach: Their development and analysis need engineering, sociology, and psychology knowledge.

The five components of a socio-technical system are:

Goals and Values: Define the purpose and guiding principles of the system. They shape how the procedure operates and its ultimate objectives.

Technical: Comprise the tools, technologies, and techniques used in the system. They enable the system to function and achieve its goals.

Structural: Involve the organizational framework and roles within the system. They dictate how you coordinate tasks and distribute responsibilities.

Psychosocial: Relate to the human aspects, including relationships, teamwork, and communication. They impact the well-being and collaboration of individuals within the system.

Managerial: Include leadership and management practices guiding the system. They play a crucial role in decision-making, strategy, and system optimization.

The 6 principles of the socio-technical approach are:

1. Compatibility: The system should be compatible with the organization's objectives, users' skills, and environment. Technology should support and not interfere with organizational activities.

2. Optimization of social and technical elements: The system's social and technical elements must be jointly optimized to ensure system success. This means keeping a balance. Improvements in one aspect should not deteriorate the other.

3. Adaptability: The system should adapt to environmental changes, like market shifts or regulatory adjustments.

4. Human Values: The system should consider the values, comfort, safety, and satisfaction of all stakeholders, leading to a work environment that encourages employee well-being and productivity.

5. Socio-technical Systems are Irreducible: The system viewed as a whole differs from the sum of its parts. Unintended consequences may arise from changing a single component, affecting the entire system.

6. Variety: The systems should be designed to handle the maximum variety of tasks, situations, or problems.

These principles' clear understanding and application lead to more effective and sustainable system designs.

Differences between Social-Systems Theory and Socio-Technical Theory:

Focus: Social-Systems Theory concentrates on social structures, roles, and interactions. The socio-technical theory integrates these social aspects with technical systems.

Components: Social-Systems Theory deals primarily with human relationships and societal norms. Socio-Technical Theory considers both human and technological elements.

Application: Social-Systems Theory applies to understanding and analyzing social groups and organizations. Socio-Technical Theory applies to designing and improving systems involving technology and human interaction.

Goal Orientation: Social-Systems Theory aims to understand social dynamics. Socio-Technical Theory seeks to create balanced and effective systems combining social and technical factors.

To deepen your understanding of socio-technical systems, consider these two resources:

Read an article on Complex Socio-Technical Systems : This article talks in detail about complex socio-technical systems. It covers topics like:

The definition of complex socio-technical systems.

The complexity of the human world.

Strategies for handling complex socio-technical systems.

Watch the ‘21st-century design’ video: In the video, Don Norman discusses the concept of "complex socio-technical systems" and their relation to "wicked problems." He explains why people avoid the term "wicked problem" due to its ambiguity and overuse.

Take the Course with Don Norman: Enroll in " Design for the 21st Century " to learn from Don Norman. It helps you apply your design skills to address global challenges using socio-technical systems.

Socio-technical systems from design methods refer to systems integrating social and technical components. This perspective focuses on creating solutions that consider technology and the human aspects. In this context:

User-Centered Design: Prioritizes the users' needs, experiences, and behaviors. It ensures the system is intuitive, usable, and satisfies user requirements.

Collaborative Design: Involves stakeholders in the design process, including users. This collaboration helps understand the social context and better align with user needs.

Iterative Development : Encourages refining and improving the system through continuous feedback from users and stakeholders.

System Flexibility: Designs systems that can adapt to changes in social or technological environments.

Holistic Approach: Considers the broader impact of the system, including societal, organizational, and ethical implications.

Literature on Socio-Technical Systems

Here’s the entire UX literature on Socio-Technical Systems by the Interaction Design Foundation, collated in one place:

Learn more about Socio-Technical Systems

Take a deep dive into Socio-Technical Systems with our course Design for a Better World with Don Norman .

“Because everyone designs, we are all designers, so it is up to all of us to change the world. However, those of us who are professional designers have an even greater responsibility, for professional designers have the training and the knowledge to have a major impact on the lives of people and therefore on the earth.” — Don Norman, Design for a Better World

Our world is full of complex socio-technical problems:

Unsustainable and wasteful practices that cause extreme climate changes such as floods and droughts.

Wars that worsen hunger and poverty .

Pandemics that disrupt entire economies and cripple healthcare .

Widespread misinformation that undermines education.

All these problems are massive and interconnected. They seem daunting, but as you'll see in this course, we can overcome them.

Design for a Better World with Don Norman is taught by cognitive psychologist and computer scientist Don Norman. Widely regarded as the father (and even the grandfather) of user experience, he is the former VP of the Advanced Technology Group at Apple and co-founder of the Nielsen Norman Group.

Don Norman has constantly advocated the role of design. His book “The Design of Everyday Things” is a masterful introduction to the importance of design in everyday objects. Over the years, his conviction in the larger role of design and designers to solve complex socio-technical problems has only increased.

This course is based on his latest book “Design for a Better World,” released in March 2023. Don Norman urges designers to think about the whole of humanity, not just individual people or small groups.

In lesson 1, you'll learn about the importance of meaningful measurements . Everything around us is artificial, and so are the metrics we use. Don Norman challenges traditional numerical metrics since they do not capture the complexity of human life and the environment. He advocates for alternative measurements alongside traditional ones to truly understand the complete picture.

In lesson 2, you'll learn about and explore multiple examples of sustainability and circular design in practice. In lesson 3, you'll dive into humanity-centered design and learn how to apply incremental modular design to large and complex socio-technical problems.

In lesson 4, you'll discover how designers can facilitate behavior-change , which is crucial to address the world's most significant issues. Finally, in the last lesson, you'll learn how designers can contribute to designing a better world on a practical level and the role of artificial intelligence in the future of design.

Throughout the course, you'll get practical tips to apply in real-life projects. In the " Build Your Case Study" project, you'll step into the field and seek examples of organizations and people who already practice the philosophy and methods you’ll learn in this course.

You'll get step-by-step guidelines to help you identify which organizations and projects genuinely change the world and which are superficial. Most importantly, you'll understand what gaps currently exist and will be able to recommend better ways to implement projects. You will build on your case study in each lesson, so once you have completed the course, you will have an in-depth piece for your portfolio .

All open-source articles on Socio-Technical Systems

Socio-technical system design.

socio technical case study

The Evolution of Computing

socio technical case study

Design Spaces

About this book, the social environment model, socio-technical design, polite computing, online rights, open access - link to us.

We believe in Open Access and the  democratization of knowledge . Unfortunately, world class educational materials such as this page are normally hidden behind paywalls or in expensive textbooks.

If you want this to change , cite this page , link to us, or join us to help us democratize design knowledge !

Cite according to academic standards

Simply copy and paste the text below into your bibliographic reference list, onto your blog, or anywhere else. You can also just hyperlink to this page.

New to UX Design? We’re Giving You a Free ebook!

The Basics of User Experience Design

Download our free ebook The Basics of User Experience Design to learn about core concepts of UX design.

In 9 chapters, we’ll cover: conducting user interviews, design thinking, interaction design, mobile UX design, usability, UX research, and many more!

  • Reference Manager
  • Simple TEXT file

People also looked at

Hypothesis and theory article, how does socio-technical lock-in cause unsustainable consumption in cities a framework and case study on mobility in bangkok.

socio technical case study

  • 1 Graduate School of Environmental Studies, Tohoku University, Sendai, Japan
  • 2 Graduate School of Global Environmental Studies, Kyoto University, Kyoto, Japan

Consumption of raw materials, energy, manufactured goods, and services is increasingly concentrated in cities, as urbanization accelerates globally. Such consumption is influenced by complex interactions arising between the various socio-technical and natural systems that make up cities. To improve understanding of the interlinked factors that can perpetuate—or “lock-in”—unsustainable consumption, we build an explanatory framework that conceptually joins the literature on socio-technical systems and on urban consumption. Two questions guide our study: (1) What are the principal socio-technical systems in cities that influence consumption behavior? (2) How do these systems interact to lock urban dwellers into unsustainable consumption behavior? The resulting framework incorporates theories of socio-technical lock-in with factors relating to both “structure” and “agency” in consumption literature. Specifically, it describes the influence and interactions of physical, non-physical, and human systems on two interlinked scales: macro-scale (structure and collectively shared conditions) and micro-scale (agency and individually shaped conditions). To demonstrate the practical value of this framework, we apply it to a case study on mobility in Bangkok, Thailand. This allows us to systematically identify the interlinked mechanisms contributing to the growing dependence on and lock-in to individually owned passenger vehicles. Our study thus provides a comprehensive understanding of the multiplex drivers of consumption behavior, taking into account both structure and agency. The framework also provides a tool for other scholars to empirically identify lock-in mechanisms that hamper the adoption of more sustainable consumption behavior in other sectors and geographies.

Introduction

The intertwined relationship between cities and consumption is widely recognized ( Hodson and Marvin, 2010 ; McMeekin and Southerton, 2012 ; Mylan et al., 2016 ; Vergragt et al., 2016 ). Urbanization trends are accelerating globally as humans migrate from rural areas, drawn by the rising cultural and economic attractiveness of cities. As a result, economic and social activities in urban areas have become the core drivers of environmental deterioration and climate change around the globe ( Fuchs and Lorek, 2005 ; Schor, 2005 ; EEA, 2012 ; Lorek and Spangenberg, 2014 ).

These impacts arise from two dimensions. Cities, by virtue of their concentrated wealth and population, import and consume vast volumes of raw materials, manufactured goods and services. Studies ( Mckinsey, 2016 ; C40, 2018 ) predicted that from 2015 to 2030, 81% of global consumption will occur in urban areas. The other dimension is that this consumption and these economic activities then discharge tremendous volumes of waste and pollution into the atmosphere, soil, and aquatic environment, with local, regional, and global effects ( Grimm et al., 2008 ). Consumption-based emissions are also strongly linked to global climate change. Indeed, studies show ( C40, 2018 ; C40 Arup, 2019 ) that more than 70% of global greenhouse gases (GHG) are emitted by the consumption activities of cities. Although developed nations are overall responsible for most of these impacts, the contribution of developing nations is rising fast, driven by rapid urbanization over the past half century ( Mckinsey, 2012 , 2016 ).

In trying to explain the drivers of urban consumption, researchers ( Zauberman, 2003 ; Foxon, 2011 ) have predominantly examined them from two perspectives: macro and micro. Macro focuses on collectively provided conditions ( Miles and Paddison, 1998 ; Jackson and Papathanasopoulou, 2008 ; Seto et al., 2014 ) that include rising levels of wealth and of quality of life ( Jayne, 2005 ; Lorek and Spangenberg, 2014 ), government development policies ( Cohen, 2005 ; Schor, 2005 ; UNEP, 2018b ), urban infrastructure, the availability of technologies, and the influence of biophysical conditions ( Orum and Dunleavy, 2019 ). At the micro scale, individually owned or shaped conditions are commonly articulated, including psychological aspects such as individual preferences and values; income; privately owned assets such as housing, vehicles, and household appliances; and demographic features such as marital status and family size.

Urban consumption behaviors also result from complex interactions between all these factors, at both macro and micro scales ( Sudmant et al., 2018 ). The influence of these forces in driving consumption can be constant, long-lasting, and self-reinforcing ( Barnes et al., 2004 ; Jackson and Papathanasopoulou, 2008 ; Cecere et al., 2014 ; Seto et al., 2016 ). Furthermore, urban dwellers can be “locked-in” to unsustainable practices and consumption habits through the cumulative effect of collectively shaped structural conditions and individually shaped circumstances ( Boucher, 2016 ). It is therefore important to consider the intertwined and multi-dimensional nature of various influencing factors if we are to understand what prevents people in cities from adopting more sustainable living patterns and technologies ( Miles and Paddison, 1998 ; Sanne, 2002 ).

In light of this, the existing literature on unsustainable urban consumption behavior has at least two limitations. First, its focus on individual factors from a micro perspective (e.g., psychological factors such as preferences or willingness to adopt new technologies or behaviors) ( Pereira Heath and Chatzidakis, 2012 ; de Koning et al., 2016 ) means that the influence of broader structural conditions has received less attention. Second, its focus on developed countries has meant less focus on developing countries. Although the drivers of rising demand for goods and services are commonly discussed, there have been limited studies on the factors that perpetuate unsustainable consumption behaviors in developing countries. Our knowledge on the conditions necessary to achieve global sustainable consumption—inclusive of the developing world—is thus incomplete ( Gasper et al., 2019 ).

The transitions literature on socio-technical systems offers insights that can help urban-consumption researchers to understand the diverse mechanisms affecting consumer behavior. Cities are well suited to being viewed as socio-technical systems ( Hodson and Marvin, 2010 ). The many systems making up urban areas—including infrastructure, technologies, formal and informal institutions, and people—are highly compatible with conceptions of the basic components that constitute socio-technical systems. Seeing cities as socio-technical systems in this way also reinforces the need to consider the multi-directional and multi-dimensional links that arise between these systems, since interactions across human, technological, and social systems are heavily stressed in transitions scholarship ( Geels, 2004 ; Cohen, 2012 ; Patorniti et al., 2017 , 2018 ; Trencher et al., 2020 ).

This socio-technical perspective of cities also requires that we understand the concept of “lock-in”. This phenomenon, which features heavily in scholarship on socio-technical systems ( Unruh, 2000 ; Foxon, 2002 ; Erickson et al., 2015 ; Seto et al., 2016 ), occurs due to the tendency for policies, technologies, cognitive frames, infrastructure, and social systems to co-evolve and reach a state of stability and self-perpetuation. Lock-in occurs when links or feedbacks between these factors combine to lock a system or set of actors to a particular pathway or behaviors, thereby hampering the adoption of environmentally superior alternatives ( Maréchal, 2010 ; Foxon, 2011 ; Seto et al., 2016 ; Wesseling and Van der Vooren, 2017 ).

In this context, this study aims to conceptually marry the two fields of socio-technical systems and urban consumption, in order to deepen understanding of the drivers of unsustainable consumption behaviors in urban areas. To achieve this, we create an analytical framework guided by two research questions: (1) What are the principal socio-technical systems in cities that influence consumption behavior? (2) How do these systems interact to lock urban dwellers into unsustainable consumption behavior? To show the practical value of our theoretical framework, we apply it to a case study on Bangkok, Thailand, using this to identify the interlinked lock-in mechanisms driving unsustainable mobility behavior.

Reflecting our methodological approach, our paper is structured as follows. The following Section Analytical Framework of Urban Consumption Lock-in consists of two parts. In the first, we draw on scholarship to propose an analytical framework that conceptualizes the key socio-technical systems that affect urban consumption. In the latter half, we propose how lock-in to unsustainable consumption can arise from interactions across these systems. In Section Case Study: Lock-in to Private Passenger Vehicle Usage in Bangkok City, we apply the framework to the case of mobility behavior in Bangkok to identify how lock-in to unsustainable behavior can occur in a developing country context. We identify the relevant lock-in mechanisms through a process of causal loop tracing ( Wesseling and Van der Vooren, 2017 ), which draws on evidence collected from secondary documents.

Our contribution to literature is four-fold. First, we address the need for a comprehensive framework in urban consumption literature that can account for both structure and agency in understanding the drivers of consumption behavior ( Kirchberg, 2007 ; Pekkanen, 2020 ). Second, we increase the relevance and scope of scholarship on socio-technical systems and sustainability transitions, where most studies focus on innovation and technology ( Unruh, 2000 ; Erickson et al., 2015 ; Seto et al., 2016 ; Geels, 2018 ). We do so by extending the application of lock-in theories to a social dimension, and specifically to demand-side behavior. Third, by interpreting the complex socio-technical structures in cities as an interwoven web of collective and individual-level systems, our study explicitly incorporates geographical scale and place-based conditions. These have lacked emphasis in transitions research ( Chandrashekeran, 2016 ). Fourth, and finally, our framework could provide a tool to help other scholars to empirically identify lock-in mechanisms that hamper the adoption of more sustainable consumption patterns in other sectors and geographies.

Analytical Framework of Urban Consumption Lock-In

Systems influencing urban consumption.

In this section, we introduce our novel framework to elucidate the concept of socio-technical lock-in in urban consumption (see Table 1 ). Specifically, a lens of socio-technical systems is applied in order to understand the diverse drivers of urban consumption, due to its suitability for capturing the multiple and complexly interwoven sub-systems that make up cities. The framework is based on insights from literature describing the basic components of socio-technical systems ( Geels, 2004 ; Seto et al., 2016 ; Markolf et al., 2018 ; Trencher et al., 2020 ). It consists of three broad categories: (1) physical systems of infrastructure, technology, and biogeographic conditions; (2) non-physical systems , comprising formal institutions, non-formal institutions, economics, and knowledge and competencies; and (3) human systems , consisting of demographic features. To understand how each system may lock-in unsustainable consumption behaviors, we visualize the framework in two dimensions: horizontal and vertical.

www.frontiersin.org

Table 1 . Framework of lock-in factors affecting individual urban consumption * .

To conceive the horizontal dimension, we examined literature on socio-technical systems and systemic causes of lock-in to extract explanations of the core urban systems that influence consumption behavior ( Unruh, 2000 ; Foxon, 2002 ; Seto et al., 2016 ; Markolf et al., 2018 ; Trencher et al., 2020 ). Previous lock-in literature has focused on interactions and self-reinforcing relationships between institutional and technological factors, often in large-scale socio-technical systems such as energy or transport ( Unruh, 2000 ; Foxon, 2002 ). Recently, Seto et al. (2016) and Ürge-Vorsatz et al. (2018) extended the scope of relevant lock-in causes by integrating behavioral factors. Their framework, however, does not explicitly discuss broader contextual factors such as geographic or economic conditions. Markolf et al. (2018) later address this deficiency, incorporating ecological context in addition to social and technological dimensions. A similar conception of socio-technical lock-in is carried out by Trencher et al. (2020) . Their framework, which also discusses energy systems, integrates the influence of exogenous context, rooting this in conceptions of a “landscape” from sustainability transitions literature ( Geels, 2004 ; Geels and Schot, 2007 ; Edmondson et al., 2019 ), alongside other material, non-material, and human factors. Thus, although these recent studies consider the importance of broader contextual factors such as economic forces, natural conditions, and culture, their discussion on socio-technical lock-in is overwhelmingly placed in the context of understanding the self-perpetuating nature of fossil-fuel-based technologies or infrastructures. However, the field of sustainable consumption requires putting human agents as the focal point, situating these within the structural conditions and barriers posed by urban environments, technologies and social factors ( Bengtsson et al., 2018 ).

To conceive the vertical dimension, we explicitly reflect the influence of both structure and agency on urban consumption behaviors. By depicting the two interlinked and interacting geographical scales of macro and micro, our framework is congruent with descriptions of multi-scalar dimensions in socio-technical systems ( Giddens, 1984 ; Vergragt et al., 2016 ; Elbasha and Wright, 2017 ) and lock-in literature ( Seto et al., 2016 ). The macro scale reflects the collectively provided or shared conditions that may be climatic, economic, technological, infrastructural, institutional and cultural in nature. These conditions are typically long-lasting and difficult for individuals to change, so they exert a powerful force on urban dwellers' consumption behavior ( Jackson and Papathanasopoulou, 2008 ). In contrast, the micro scale depicts smaller-scale conditions, assets, and institutions—typically owned, accessed, or shaped by individual agents. Relevant actors at this scale include individuals, households, social networks and specific organizations, such as workplaces ( Oliveira et al., 2020 ). At the micro scale, these actors often have more power than at the macro scale to decide or control consumption behaviors in relation to their own material possessions, environment, and non-material conditions ( Jørgensen, 2012 ; Tao et al., 2021 ).

In proposing this framework comprised of two scales, macro and micro, we do not explicitly include a third scale—the so-called meso scale—that is sometimes evoked in consumption literature ( Kirchberg, 2007 ; Maréchal, 2010 ; Poças Ribeiro et al., 2019 ). These studies explicitly describe the meso scale as situated between the macro and the micro, arising from interactions between the structural and individual conditions that shape urban environments. Yet conceiving the meso scale can be problematic, since literature concedes that its boundaries with other scales are blurry ( Vergragt et al., 2016 ; Pekkanen, 2020 ). Thus, to avoid conceptual ambiguities and simply but clearly distinguish between structural and individual conditions, we incorporate in our framework only two scales: macro and micro.

The following sections describe each urban system captured by the analytical framework in more detail. We provide examples from a broad range of literature on urban consumption, climate governance, and energy transitions to describe how each can influence consumption behaviors and lock-in unsustainable living patterns. Two points deserve prior emphasis. First, we discuss these systems and scales separately for the sake of conceptual clarity. However, because of their interconnectedness and interdependence, we expect a certain overlap or conceptual blurriness when looking at the messy empirical world of living cities. Second, we do not assume that a particular consumption behavior would experience interactions and influences equally from all systems and scales. Rather, each system's relevance and influence would vary with particular consumption acts by micro-scale actors.

Biogeographic Conditions

Biogeographic conditions are in the nature-in-city realm ( Jim, 2016 ), which is either pre-existing or newly formed to accommodate human activity ( Werner, 2011 ; Frank, 2017 ; Trencher et al., 2020 ). This system includes long-lived natural conditions ( Seto et al., 2016 ), such as topography, vegetation, land and water availability, air quality, and climate. At the macro scale, such conditions can negatively impact consumption behaviors. For example, climate (e.g., heat, cold, humidity, rain, wind), natural lightings, and environmental stresses, such as air and pollen pollution, can force households to adapt. This might entail using heating, cooling, or air-filtration systems to make interior living spaces more comfortable ( Lariviere and Lafrance, 1999 ; Fell et al., 2014 ; Singh et al., 2017 ), or using private passenger vehicles, which offer more comfort than walking, cycling, or public transport ( Burge et al., 2007 ; Tao et al., 2018 ; Böcker et al., 2019 ; Lipson et al., 2019 ; Nissen and Becker, 2020 ). The causal relationship between such energy-intensive consumption behavior and biogeographic conditions may create a vicious circle. Specifically, not only do such behaviors contribute to air pollution and climate change through increased energy use 1 , they can also exacerbate uncomfortable conditions and thus reinforce the need for adaptation measures ( Eom et al., 2020 ). The urban heat island effect 2 is one such example ( Souza et al., 2009 ).

At the micro scale, outdoor natural environments are also shaped by individually owned buildings. These too strongly influence consumption. For example, a lack of space for home gardens can deny households the opportunity to produce their own food. This makes them dependent on buying vegetables and fruits produced by industrial agricultural and food-production systems, typically located far from cities ( Diehl et al., 2020 ). It also creates reliance on municipal waste treatment systems, since organic household waste cannot be treated at the source (i.e., converted to compost) ( Eades et al., 2020 ).

Infrastructure and Built Environment

Infrastructure and built-environment systems include human-made surroundings that support or provide the physical setting for human activity ( Kaklauskas and Gudauskas, 2016 ). At the macro scale, these systems typically consist of long-lived, collectively provided, and shared buildings, structures, and infrastructure such as those supporting transportation (e.g., roads, subways, bike lanes, pedestrian networks), supply chains, telecommunications, energy provision networks, leisure (e.g., parks, public spaces), commerce (e.g., shopping centers), and industry (e.g., technology parks) ( Davis et al., 2019 ; Pekkanen, 2020 ). Accounts of such macro-scale conditions impacting consumption behavior are widespread in literature—mobility behavior being no exception. For example, residents can be initially attracted to using private motor vehicles or taxi alternatives when supportive or optimized urban infrastructures are in place, such as roads, bridges, and expressways ( Burge et al., 2007 ), or when there is a lack of infrastructures that encourage alternative mobility behaviors, such as public transport, cycling, and walking ( McIntosh et al., 2014 ; Hagen et al., 2016 ; Nguyen et al., 2019 ). Urban dwellers may become locked-in to lifestyles dependent on private vehicle use, as cultural norms are established, and as expansionist city planning reinforces the need to own private motor vehicles because of urban sprawl ( Seto et al., 2014 , 2016 ).

At the micro scale, infrastructure and built environment may include individually owned or rented buildings and external connections to telecommunications or energy infrastructure ( Sanne, 2002 ; Seto et al., 2016 ). Literature also describes how unsustainable behaviors can be perpetuated at the micro scale. For example, the fixed architecture and limited roofing space of apartment or office buildings can pose a barrier to installing solar panels, prolonging dependence on grid-based electricity ( Mah et al., 2018 ; Roberts et al., 2019 ; Reindl and Palm, 2021 ). Large dwelling sizes or poorly insulated buildings also tend to lock-in households to high energy consumption ( Guan et al., 2014 ; Jones et al., 2015 ; Sakah et al., 2019 ). Meanwhile, physical limitations in high-rise apartment buildings may prevent urban dwellers from installing charging infrastructure for electric vehicles, perpetuating dependence on gasoline vehicles ( Ou et al., 2018 ).

Availability of Technologies, Goods, and Services

The availability and characteristics of technologies, goods, and services is a core consumption enabler ( Foxon, 2002 ; Sanne, 2002 ; Lorek and Fuchs, 2013 ). At the macro scale, the sustainability of consumption behavior is structurally influenced by the general availability of technological artifacts (e.g., passenger vehicles, smartphones, household appliances), goods (e.g., foods, clothes, eco-bags), and services (e.g., sharing bike, Internet, and cellphone networks; repair services; restaurants) ( McKinsey, 2020 ), and by their properties: size, energy efficiency, embodied emissions, environmental footprint, and so on. The influence of availability in the consumption patterns of people who migrate from rural to urban areas has been observed in some countries undergoing rapid urbanization, since consumption tends to increase because of more opportunities to buy and use technologies, goods, and services ( Kurniawan et al., 2020 ). Availability can also structurally restrain consumers' choices, when green goods such as electric vehicles or organic foods are lacking ( De Rubens et al., 2018 ; Vermeir et al., 2020 ) or when platforms for sharing bicycles or vehicles, for example, are underdeveloped ( Sun et al., 2018 ; Meelen et al., 2019 ; Merfeld et al., 2019 ).

At the micro scale, the availability of technologies, goods, and services under possession or control will also dictate consumption behavior. For example, the distance and frequency of motorized travel tends to increase when individuals or households acquire their own vehicle ( Yang et al., 2017 ; Haque et al., 2019 ). Likewise, the availability of items such as foods, drinks, services, and appliances in households or offices (e.g., free bottled water, dishwashers, central heating) will also stimulate consumption and usage ( Cohen and Murphy, 2001 ; Sanne, 2002 ; Lorek and Fuchs, 2013 ) and, in some cases, lock-in individuals to unsustainable practices.

Formal Institutions

As part of non-physical systems, formal institutions are the formalized and shared rules, commitments and principles that govern behavior at the level of society or individual organizations ( Foxon, 2011 ; Spangenberg, 2014 ; Trencher et al., 2020 ). At the macro scale, national or city-level institutions may include policies, rules, laws, roadmaps, master plans, and regulations ( Clammer, 1997 ; Sanne, 2002 ; Jayne, 2005 ; Dawkins et al., 2019 ). These can heavily influence and lock-in consumption patterns ( Hult and Larsson, 2016 ; Seto et al., 2016 ), in their absence as much as their presence. For example, motorists can be discouraged from driving in downtown areas if regulations such as road tolls or car-free zones are in place ( Sareen et al., 2021 ; Trencher et al., 2021 ). In the absence of such progressive policies, motorists can be encouraged to use private vehicles instead of public transport. The absence of strict energy-efficiency regulations for housing or electronic goods can increase energy consumption among consumers ( Zhu and Li, 2015 ). The absence of policies to reduce plastic packaging around the world is exacerbating the problem of consumer waste ( UNEP, 2018a ).

At the micro scale, relevant formal institutions may include contracts, organizational rules, and procedures that influence the behavior of individuals or groups ( Schulte, 2015 ). For example, building rental or leasing contracts often prevent energy-saving measures. Some contracts can restrict or economically discourage renters from modifying a building's design or material to enhance its energy efficiency ( Weatherall et al., 2020 ). In the workplace, dress codes such as those mandating a suit and tie ( Niinimäki, 2010 ) have been associated with necessitating greater use of air-conditioning and thus higher energy consumption.

Informal Institutions

Informal institutions are the social regularities or standards, informally generated and maintained, that regulate human consumption behaviors ( Coşgel, 1997 ). Widely described as a cause of socio-technical lock-in ( Geels, 2004 ; Trencher et al., 2020 ; Duygan et al., 2021 ), at the macro scale this system includes collectively produced and widely shared political or societal views, beliefs, values, culture, trends, norms, and religions ( Fuchs and Lorek, 2005 ; Biswas and Roy, 2015 ; Pekkanen, 2020 ). For example, the highly contagious fashion culture and emphasis on appearance in developed cities induces high levels of purchasing goods such as clothes, shoes, accessories, and makeup ( Rudd, 1997 ; Rajagopal, 2011 ). Shared preferences and culture affect dietary behavior, such as fast food or meat-eating ( Wirtz, 2019 ), which has varying impacts on sustainability ( Riordan and Stoll-kleemann, 2015 ). Societal norms such as the use of hot water for bathing, or cultural identification with automobiles, has increased the consumption of fossil fuels ( Seto et al., 2016 ; Sovacool and Martiskainen, 2020 ).

Viewed from the micro scale, informal institutions such as the worldviews, preferences, routines, or experiences of individuals or small groups can also perpetuate socio-technical lock-in ( Sanne, 2002 ; Shove and Walker, 2010 ; Schulte, 2015 ). For example, people's fashion, dietary, transport, or lifestyle routines are strongly influenced by interpersonal relations with family, peers, and workplace or social networks ( Algesheimer et al., 2005 ; Niu, 2013 ).

Economic Forces

Collectively produced economic conditions are primary drivers of the production, marketing, supply, and consumption of technologies, goods, and services ( Goodwin et al., 2018 ). At the macro scale, these may include economic vitality and growth, employment opportunities, salary conditions, and access to finance, which are the collective result of individual economic agents and organizations ( Sanne, 2002 ; Jayne, 2005 ; Mckinsey, 2012 ; Sudmant et al., 2018 ). Economic conditions can also arise from business models, marketing strategies, and the costs or prices of goods and services ( Foxon, 2011 ; Biswas and Roy, 2015 ; UNEP, 2019 ). Studies show how declines in the cost of goods such as air-conditioners ( Sovacool, 2016 ) and cars ( Tang, 2009 ), driven by mass production and technological learning, have sharply increased their consumption. Another powerful economic mechanism that drives consumption behavior is the ideology embedded into marketing campaigns about constantly “upgrading” to the latest model—of smartphones and computers, for example ( Seto et al., 2016 ).

At the micro scale, economic conditions shaped by individuals or organizations also influence consumption practices. These conditions include disposable household income and financial resources ( Sudmant et al., 2018 ; Trencher et al., 2020 ). An extreme example is the tremendous individual and organizational wealth possessed by the so-called “1%”, which enables them to enjoy lavish but unsustainable lifestyles ( Robeyns, 2019 ; Wiedmann et al., 2020 ). This may encompass the ownership and use of multiple residences, vehicles, boats, and jets. As a more mundane but widespread example, increased wealth at the household level has been shown to stimulate more consumption of meat and processed food in some developing countries ( Delgado, 2003 ; Bereznicka and Pawlonka, 2018 ).

Knowledge and Competencies

The nature of knowledge, skills, and intellectual capacities, whether collective or individual, also affects the sustainability of consumption behavior ( Dawkins et al., 2019 ; Trencher et al., 2020 ). At the macro scale, the managerial and technical capacity of local governments, industry, and professional organizations is often reported to hamper the introduction of more sustainable policies and urban-planning practices ( Cohen and Murphy, 2001 ; Lorek and Spangenberg, 2014 ; Hult and Larsson, 2016 ; Sudmant et al., 2018 ). For example, without the knowledge or skills to tackle environmental pollution and waste or to promote sustainable energy and infrastructure, city planners in both developing and developed countries may find themselves locked-in to maintaining and reproducing unsustainable policies and urban infrastructures ( Zhu and Li, 2015 ; Verma et al., 2016 ; Trinh et al., 2021 ). Thus, escaping the shackling effect of existing cognitive capacities may require experts and best practices to be imported from elsewhere ( Alizadeh, 2017 ).

At the micro scale, intellectual resources such as knowledge, awareness, and skills also influence the consumption behavior of individual agents and organizations ( Ueno et al., 2005 ; Zhu and Li, 2015 ; Lazzarini et al., 2018 ). For example, even when green products are available in markets, a lack of environmental awareness or knowledge can hamper individuals' sustainable-consumption choices, because they may not understand, notice, or appreciate the explanatory labeling or value propositions ( Biswas and Roy, 2015 ). Similarly, lack of awareness of the environmental consequences of unsustainable energy systems may hamper consumers from choosing to purchase renewable electricity ( Chapman and Itaoka, 2018 ) or electric vehicles ( Krishna, 2021 ).

Demographic Features

The demographic characteristics of human populations in urban areas are another important consumption driver that can contribute to the lock-in of unsustainable behavior. Viewed from the macro scale, relevant features include population size, growth, age, density, and geographic distribution ( Mckinsey, 2012 , 2016 ). With many cities attracting increasing numbers of residents, particularly from rural areas, urban locations are frequently locked-in to increasing needs for energy, manufactured goods, and natural resources ( Meng et al., 2018 ). Conversely, however, low population density can encourage larger flooring space in offices and households, increasing per capita energy consumption ( Sudmant et al., 2018 ; UNEP, 2018b ).

At the micro scale, factors such as the size and attributes of households and organizations (e.g., age, education, gender, health, marital status) influence choices around energy and material consumption ( Soltani et al., 2019 ). For example, households with young dependents tend to consume more electricity, because of more time spent at home and more sensitivity toward the indoor environment ( Su, 2019 ). This then affects transport behavior, since cars are often used for children's schooling ( Yang et al., 2018 ) or for household tasks such as shopping and running errands ( Yao and Wang, 2018 ). Older people have been found to be less likely than younger people to upgrade household appliances such as refrigerators and air-conditioners ( Yagita and Iwafune, 2021 ). This perpetuates energy-intensive living patterns, because of the lower energy efficiency of older appliances.

Conceptual Interaction on Urban Consumption Lock-In

In studying urban consumption behavior, we identified that structural lock-in can wed agents to current and unsustainable practices as well as prevent agents from adopting sustainable alternatives. This draws on three fundamental principles from previous literature. First, lock-in arises from the persistent and multidirectional interactions that occur in multiple and interconnected urban systems, at differing scales ( Seto et al., 2016 ). Second, feedback loops between these systems may emerge—by design or consequence. This occurs when one system influences another, which then reciprocally influences the first, reinforcing the collective impact ( Wesseling and Van der Vooren, 2017 ; Markolf et al., 2018 ). Third, one factor might be more important than another, and thus may become a precondition for others or exert cascading effects on others.

Thus, behavioral lock-in, like technological lock-in, is created by persistent interactions and feedbacks occurring across the various physical, non-physical, and human systems that make up cities and urban areas. These systems are collectively provided or shaped at the macro scale or individually owned or shaped at the micro scale. The ability of individual agents to break free from these hampering conditions can be determined by the extent of their dependence on the collectively provided structural conditions at the macro scale, and by their ability to overcome various restraining factors originating from the micro scale.

To systematically capture these dynamic forces affecting urban consumption behavior, this study follows the distinction between structure and agency in structuration theory ( Whittington, 2015 ). As shown in Figure 1 , we apply two perspectives to visualize the origin and direction of these shapers of consumption behavior: (1) top-down, describing forces originating from the macro scale (i.e., from structural conditions), and (2) bottom-up, describing forces originating at the micro scale (i.e., from conditions under the control of individual agents and organizations).

www.frontiersin.org

Figure 1 . Mechanisms contributing to the lock-in of urban consumption (Source: Authors).

In conceiving the top-down perspective, we make two proposals about the nature of consumption drivers and lock-in forces that originate from the macro scale. The first is that various systems at the macro scale will interact and self-reinforce, magnifying the cumulative influence of collectively provided structural conditions in urban areas ( Bettencourt, 2013 ). Road transport, if viewed as a socio-technical system comprising several interlinked, self-reinforcing components ( Geels, 2004 ), is an illustrative example. Here, a nexus of co-evolving, interlocked, and self-perpetuating factors at the macro scale may arise from policies that support passenger vehicle use through the provision of infrastructure like roads and bridges, the rise of a middle-income class, increasing production or availability of passenger vehicles in the market, and cultural identification with the car as a symbol of desire or a transportation norm ( Cohen, 2012 ; Seto et al., 2016 ; Sovacool, 2017 ; Canitez, 2019 ).

The second proposal is that individual conditions at the micro scale are directly influenced and perpetrated by the collectively provided structural conditions at the macro scale ( Janssen and Jager, 1999 ). For instance, increasing GDP at the macro scale often manifests as higher incomes for individuals and households, providing more opportunities for the consumption of resources, goods, and services ( Spangenberg, 2014 ; Zaharia et al., 2019 ). This relationship has also been noted when the urban population size directly influences the density of the workplace or dwelling sizes due to less land availability ( Aranda-Mena et al., 2018 ). Also commonly discussed, shared social norms or values can pose constraints or pressures on an agent's beliefs and preferences ( Foxon, 2011 ; Biswas and Roy, 2015 ; Pekkanen, 2020 ). These can influence the practices of individuals, households, and organizations in diverse living activities such as bathing habits ( Sovacool and Griffiths, 2020 ), dietary preferences ( Einhorn, 2020 ), and choice of transport mode ( Belgiawan et al., 2017 ).

The bottom-up perspective is inspired by the idea of dynamic agency ( Elbasha and Wright, 2017 ). This encompasses two views. Conditions under the control or influence of individual agents may interact and self-reinforce to perpetuate unsustainable consumption behavior at the micro scale. Individual agents, organizations, or communities may also exercise their agency to contribute to continuing or reproducing larger structural factors originating from the macro scale ( Phipps, 2001 ; Whittington, 2015 ). In addition, though it is beyond the scope of this paper, certain individual agents may succeed in transforming structural conditions ( Corman, 2008 ).

Two postulations emerge from here. In the first, multiple systems at the micro scale can interact, reinforcing each other's influence to perpetuate the conditions that drive unsustainable consumption. For example, higher levels of disposable income or access to credit can incite the purchase of larger households or business premises, resulting in greater demand for energy, furniture, and household appliances due to the increased flooring area ( Li et al., 2018 ; Sakah et al., 2019 ).

In the second postulation, individual factors at the micro scale can exert an upward influence on collective conditions at the macro scale, even stimulating change. Some have argued that agents have little control over structural forces ( Jackson and Papathanasopoulou, 2008 ). But examples do exist where individual behavior has succeeded in influencing structural conditions at the macro scale ( Spangenberg, 2014 ; Pekkanen, 2020 ). For example, the intellectual resources of individuals agents or organizations, when widely shared, can enrich the collective knowledge base of society ( Kanger and Schot, 2016 ). And, though it is an atypical example, the individual beliefs and travel behavior of Greta Thunberg have circulated globally, arguably influencing collective norms and knowledge on climate change in many countries ( Martiskainen et al., 2020 ).

Our attempt to conceptually integrate concepts of socio-technical lock-in with discussions of unsustainable consumption behavior draws inspiration from others. Notably, Kim and Kim (2020) highlighted that lock-in mechanisms can explain the loyalty of some consumers to certain consumption behaviors. Meanwhile Jackson and Papathanasopoulou (2008) explain that lock-in to unsustainable consumption patterns is exacerbated by the general inability of consumers to control the broader structural conditions shaping their surrounding urban environment.

Yet by adopting theories of lock-in and socio-technical systems in the domain of urban consumption, we acknowledge several challenges that arise. For example, though human agents can play a crucial role in triggering or inhibiting change in socio-technical systems ( Borrás and Edler, 2014 ), the sustainability transitions literature on socio-technical systems and lock-in has been criticized for overemphasizing the influence of structural conditions and paying inadequate attention to individuals and agency ( Duygan et al., 2021 ). However, by applying insights about socio-technical systems and lock-in while heeding the limitations of scholarship on urban consumption and sustainability transitions, an opportunity arises for a novel research perspective with a more comprehensive view of urban consumption. Importantly, this perspective can account for structural and individual factors, as well as interactions and lock-in that arise from the socio-technical systems that comprise cities.

Case Study: Lock-In to Private Passenger Vehicle Usage in Bangkok City

This section applies the above framework to a case study on Bangkok, to show how the various systems comprising urban environments can structurally lock-in consumption behavior to unsustainable practices and restrain the adoption of sustainable alternatives. We focus on urban mobility because individual transportation choices have a significant impact on the sustainability of cities ( C40, 2018 ). Bangkok is an ideal case study to illustrate this. Private passenger vehicles have surpassed 6 million units to become the dominant means of urban transport ( OTP, 2018 ). Applying the framework of lock-in factors affecting individual urban consumption shows that the deepening dependence on passenger vehicles can be largely explained by a suite of interactions and feedbacks occurring across multiple systems.

The case study first provides background information on car production and usage in Thailand. We then apply the lock-in framework to explore the interactions among various urban systems at macro and micro scales, showing how these structurally lock-in individual mobility behavior to relying on privately owned passenger vehicles. We focus on two consumption stages: vehicle purchasing and vehicle use. The core aim of this case study is to show how the interactions theorized in this paper can arise in the real world and perpetuate unsustainable mobility behavior. A full-fledged empirical analysis is beyond the scope of this study, but evidence is sourced from a wide range of documents, including policy documents, gray literature, and academic studies. To ensure a comprehensive understanding of the transport situation in Bangkok, literature was selected to ensure coverage of factors affecting a wide range of transport choices, such as passenger cars, walking, and public transport (including trains, buses, and bus rapid transit).

Background of Bangkok's Mobility

This case examines mobility behavior in Bangkok City (henceforth referred to as Bangkok). Surrounded by five vicinities, Bangkok lies at the heart of the broader Bangkok Metropolitan Region. The whole region comprises a diversity of density and land uses, ranging from dense downtown areas, semi-urban and industrial areas, over a vast area of 7,762 km 2 . National level policies are also examined when they affect urban development in Bangkok.

Triggered by its accelerating development and population growth, Bangkok's traffic situation is worsening, provoking the environmental and social consequences of a steep surge in privately owned passenger vehicles. The city's population increased from 6.4 million in 2000 to 10.7 million in 2021, a rate of 2.6% per year. It currently accommodates more than 13% of Thailand's total population and is home to some 6 million on-road passenger vehicles. The ratio of 0.6 vehicles per resident is twice as high as Tokyo's and three times higher than Beijing's. 3

This explosion in the use of private passenger vehicles (see Figure 2 ) has occurred in a series of waves, one from 2011 to 2013 and another from 2018 to 2019. Despite the availability of diverse transport modes, such as buses and rapid rail transit, privately owned passenger vehicles have become the dominant mode of travel in Bangkok, making up around of 40% of journeys, followed by motorcycles with nearly 24% ( OTP, 2018 ). The rapidly expanding fleet of motor vehicles has worsened the city's traffic conditions and air quality and has raised the urban temperature by contributing to the heat island effect ( Attavanich, 2017 ; Khamchiangta and Dhakal, 2019 ). Motor vehicles thus damage public health, human welfare, and the economy ( Uttamang et al., 2017 ). Indeed, the impacts of air pollution on health and life satisfaction have been estimated to cost around $6.17 billion annually during the first wave of car purchasing ( Attavanich, 2017 ). Bangkok is often also alerted with days of harmful air pollution, its annual average for PM2.5 levels being more than four times the World Health Organization's target ( IQAir, 2019 ). The dominant sources coming from traffic activities, PM2.5, PM10, O3, and SO3, are observed to be heavily contributing to non-accidental mortality, cardiopulmonary diseases, and lung cancer in the city ( Chuersuwan et al., 2008 ; Guo et al., 2014 ; Fold et al., 2020 ).

www.frontiersin.org

Figure 2 . Thailand total vehicle sales 1980–2020 data ( Trading Economics, 2021 ).

As we now demonstrate, the lock-in of unsustainable mobility behavior in Bangkok is the result of long-term interactions among multiple urban systems, at both macro and micro scales, which promote the purchase and use of privately owned passenger vehicles. The lock-in mechanism is summarized in Figure 3 .

www.frontiersin.org

Figure 3 . Visualization of mechanisms contributing to the lock-in of mobility based on individual passenger vehicles in Bangkok (Source: Authors).

Influence of Lock-In Factors on Vehicle Purchasing

We begin with vehicle purchases, both new and used. We focus first on two clusters of lock-in mechanisms, shown as [1] and [2] in Figure 3 . The first cluster shows how informal institutions, formal institutions, economic forces, and technology availability at the macro scale intertwine to exert a downward impact on informal institutions at the micro scale. The second cluster starts with informal institutions at the micro scale, showing how these exert an upward influence on economic forces at the macro scale. To illustrate the persistence of the lock-in situation over time, while also addressing the changing relevance of various factors, we focus explicitly on different time periods when describing each cluster. When referring to the different systems of the lock-in framework, we use codes in the text and in Figure 3 . For example, [ECO-MA] in upper case refers to economic forces at the macro scale, while [eco-mi] in lower case refers to economic forces at the micro scale.

The description of cluster [1] is strongly based on the first wave of car buying, between 2011 and 2013. Beginning with economic forces at the macro scale [ECO1-MA], rising GDP has boosted households' income and expenditure levels [eco2-mi].

In 2010, with the economy rebounding, Thailand's GDP per capita regained positive growth of 7.2% compared to 2009 ( Pasuk and Pornthep, 2012 ). Bangkok's average monthly income per household continued to be ranked the highest in Thailand, hitting a peak of 49,000 THB (Thai baht) in 2011–2013 ( NSO, 2019 ). The driving effect of this rising income on car ownership is stressed in multiple studies ( Pongthanaisawan and Sorapipatana, 2010 ). For example, Taylor and Yue (2011) predicted that car ownership surpasses motorcycle ownership when household income exceeds around 50,000 THB per month. Thus, the increasing income triggered by economic growth gave many Bangkok residents the opportunity to purchase their first passenger vehicle.

Thailand has been by far the largest automobile producer [ECO3-MA] among ASEAN countries for the last 20 years ( Figure 4 ). The high availability of automobiles for the domestic market [TECH1-MA] provides Thais with wide opportunities for purchasing. Indeed, Thailand has been called the “Detroit of Asia” ( Wad, 2009 ; Wonglimpiyarat, 2016 ) owing to its success at attracting foreign automobile brands, mainly Japanese, to build factories on the outskirts of Bangkok ( Warr and Kohpaiboon, 2017 ). Thus, compared to the populations of other ASEAN countries, Thai people benefit from low purchase costs for new vehicles, since locally produced vehicles are exempted from import tariffs.

www.frontiersin.org

Figure 4 . Yearly automobile production in ASEAN countries (Retrieved from AAF and OICA on July 15, 2021).

Informal institutions at the macro scale [INFORIN1-MA] also contributed to increased purchasing of private automobiles. National government policy propagated a political ideology around the automotive industry's important position in ensuring economic growth, stressing its creation of employment ( Thailand Board of Investment, 2017 ) and its contribution of 12% to average annual GDP since 2012 ( Pollio, 2012 ). Policies favorable to the automobile market were implemented as a result, aimed at shortening the gap between customers' income and purchase prices. For example, since the early development of the automotive industry in the 1990s, low import duties [FORIN1-MA] have persistently provided attractive market conditions, by ensuring affordable price choices of new vehicles [ECO4-MA] ( Wu and Pojani, 2016 ). The used-vehicle market was also targeted by policies at the macro scale: the rate of annual registration taxes [FORIN2-MA] was lowered to make automobiles more affordable for lower-income households ( Techakanont and Leelahanon, 2015 ).

Other informal institutions have also propelled vehicle purchases. After experiencing severe consequences from the Asian economic crisis of 1997 ( Kittiprapas, 2000 ) and the recession during post-Thaksin's regime (2001–2006), the new government of Yingluck Shinawatra in 2011 continued the populist development strategy from their predecessor. This notably included establishing preferential policies to support the outputs of the domestic vehicle production sector ( Chambers, 2013 ; Pollio and Rubini, 2021 ). In the second half of 2011, the government's first-time car-buyer tax rebates scheme [FORIN3-MA] was introduced to stimulate automobile purchasing over motorcycles, which were historically Thais' preferred choice for urban mobility. The scheme offered a tax rebate up to 100,000 THB ( Muthitacharoen et al., 2019 ), equivalent to the average yearly income of low-income households. At the same time, banks' lending policies [ECO5-MA] acted as a supporting economic force, providing more opportunities for the rising urban middle class to obtain personal credit [eco6-mi]. The additional car purchases were even boosted in this period by competition across the banking industry, as many introduced subprime loans ( Muthitacharoen et al., 2019 ). With loans authorized for many low-income applicants, and the above-mentioned subsidy accepted as a down payment, this interconnected web of economic forces and formal institutions at the macro scale shrank the gap between income and affordability. It thus shifted the purchase preferences of many first-time buyers [inforin2-mi] from motorcycles to automobiles. As a consequence of these driving forces originating from the macro scale, national sales reached 1.44 million in 2012, up 80% from 2011. This momentum was sustained, with 1.33 million purchases in 2013 ( Krungsri Research, 2020 ). Demonstrating the shift, in Bangkok the number of registered automobiles exceeded, for the first time ever, the number of registered motorcycles in 2012 and 2013 ( Figure 5 ).

www.frontiersin.org

Figure 5 . Number of newly registered vehicles in Bangkok (1995–2018) ( Department of Land Transport, 2021 ).

Cluster [2] is linked to the second wave of car buying, from 2018 to 2019. Although car sales dropped dramatically after the first wave, the accumulated car ownership at the micro scale [tech2-mi] triggered a series of economic repercussions at the macro scale. The vast number of cars purchased during the first car buyer scheme (2011–2013) [FORIN3-MA] led to demand for replacement [inforin3-mi] as households became accustomed to the convenience of private automobile use. This occurred after around five years of ownership, largely during 2018 and 2019, as owners finished loan repayments and opted for replacement after considering the increasing repair costs for vehicles exiting the warranty period [eco7-mi] ( Krungsri Research, 2020 ).

Macro-scale conditions reacted to this new market opportunity. Car dealers adopted aggressive marketing activities [ECO8-MA], such as cash-back deals for owners trading in vehicles early. Vehicle upgrading then reinforced many Bangkok residents' preference for private automobiles [inforin2-mi]. This created a stock of affordable used vehicles, lowering the financial barrier for lower-income households to acquire their first automobile. At the same time, Thailand's per capita GDP [ECO1-MA] recovered after a slight drop due to the influence of political uncertainty on the economy ( BOT, 2014 ), reaching ~10% growth in 2017 ( KNOEMA, 2021 ). The stimulated economic activity at the macro scale raised household incomes further [eco2-mi] ( NSO, 2019 ).

Though the second wave was less pronounced than the first, dependence on private automobiles in Bangkok deepened once again, as a result of these interactions across multiple sub-systems, pushing annual car sales to pass 1 million in the following 2 years in 2018 and 2019 ( Figure 2 ).

Influence of Lock-In Factors on Vehicle Use

We now examine how various urban systems have collectively fostered and locked-in urban mobility behavior that depends on private car use. We focus on two clusters of intertwined factors, marked as [3] and [4] in Figure 3 . Cluster [3] illustrates an upward trajectory: how technology availability, informal institutions, and knowledge at the micro scale collectively influence biogeographic conditions at the macro scale. Cluster [4] illustrates a downward trajectory: how formal institutions along with knowledge and competencies at the macro scale collectively influence informal institutions at the micro scale. In contrast with the previous section, where we focused on two temporal events of vehicle purchasing, here we draw on different time periods and pieces of evidence to show how a temporally accumulated array of factors is perpetuating dependence on automobile use.

In cluster [3], mobility behavior is closely correlated with the availability of privately owned automobiles in households [tech2-mi]. Such ownership leads to more driving, and decreases the likelihood of owners walking, cycling, or using public transport ( Pongprasert and Kubota, 2017 ; Witchayaphong et al., 2020 ). This causal trend is also stimulated by economic forces at the micro scale because, as more households enjoy higher incomes [eco2-mi], they become less sensitive to travel costs ( Witchayaphong et al., 2020 ).

Moreover, for high-income Thais, the persistence of informal institutions at the micro scale, such as personal preference for car usage [inforins4-mi], can be explained not only by the relative levels of convenience, comfort, and safety—just as importantly, vehicle ownership gives them a desirable symbol of economic success and social status ( Dissanayake and Morikawa, 2010 ; Pongthanaisawan and Sorapipatana, 2010 ; Matsuyuki et al., 2020 ). The correlation between private automobile ownership and social status has been described as a way for Bangkok's wealthy to save face and avoid the embarrassment of using public transport. These individual perceptions are influenced by widely circulated norms and societal values at the macro scale [INFORIN5-MA], with Thai society rooted in traditional class divisions separating the wealthy from the working class [DEM1-MA].

The literature also describes a self-reinforcing loop of exhaust fumes from the growing fleet of on-road vehicles and higher vehicle usage rates. Among all road transportation, passenger cars are the major contributor to worsening air quality in Bangkok [BIO1-MA] ( Chuersuwan et al., 2008 ; Pongthanaisawan and Sorapipatana, 2010 ; Cheewaphongphan et al., 2017 ). Interestingly, the deterioration of air quality is exacerbated by low levels of environmental awareness [knocom1-mi], because many of Bangkok's citizens accept environmental problems such as air pollution as part of urban living. Furthermore, the extreme climate conditions [BIO2-MA] posed by high temperatures are worsened by the combined influence of climate change and the heat island effect. Bangkok's average surface temperatures have increased by more than 6°C from 1991 to 2016 ( Khamchiangta and Dhakal, 2020 ); the use of automobiles, though it is not the major cause, has contributed to this anthropogenic heat flux ( Khamchiangta and Dhakal, 2019 ). Alongside concerns about residents' physical and mental health on account of heat stress ( Arifwidodo and Chandrasiri, 2020 ), the high relative humidity all year in Thailand ( Taweekun and Tantiwichien, 2013 ) causes perspiration and body odor when walking to public transport stations. The air-conditioned comfort of private automobiles thus offers an attractive mobility option for many Bangkok residents [inforin4-mi].

In cluster [4], we highlight a further web of interactions and self-reinforcing loops that collectively contribute to locking-in reliance on private automobiles. In particular, we draw attention to the aggregated impacts of formal and informal institutions, infrastructure development, and the deficiency of knowledge and competences around integrated, sustainable transport planning. The first set of relations occurs between formal institutions and the expansion of road infrastructure. Thailand's road policies [FORIN4-MA] aim to meet the increasing demand for road space that has ensued from the surge in private vehicle purchases and the increase of freight traffic serving the automotive industry ( Charoentrakulpeeti et al., 2006 ).

Specifically, the construction of vehicle-production factories on the outskirts of Bangkok has necessitated the investment of road infrastructure [INFRA1-MA] to attract foreign development investment and enable supply-chain development ( Lecler, 2002 ; Chalermpong and Ratanawaraha, 2015 ). Japanese investments in particular have powered this trend. Indeed, 41% of foreign direct investment flows to Thailand for automobile and other industries between 2010 and 2017 have come from Japanese companies and government agencies ( OECD, 2021 ). In parallel, after the first and second waves of car buying, successive infrastructure-development policies have sought to increase the convenience of car ownership by easing traffic congestion ( Peungnumsai et al., 2020 ; Witchayaphong et al., 2020 ). This has resulted in a disproportionate allocation of public spending toward increasing Bangkok's road networks over other transport infrastructure, such as rail. Government data shows that road development has accounted for about 80% of the total annual investment budget for infrastructure development since 2009 ( Ministry of Transport, 2020 ).

In particular, from 2016 to 2020, national highway development was the only category of transport infrastructure to receive increasing investment from government budgets and loans, rising around 10% compared to 2011–2015 ( Ongkittikul, 2014 ). This directed public spending was underpinned by the government's goal to add 6,612 km of roadways across the country as part of its Intercity Motorway Development Master Plan. Formal commitments to expanding Thailand's road network continue. Setting a 20-year development plan for 2017–2036, the government is currently working to add more than 390,000 km of roads. These public investment policies have widened the gap between the development of road infrastructure and of alternative transport modes ( OECD, 2020 ). And the resulting superior road network has increased people's preferences for using cars [inforin4-mi] ( Wu and Pojani, 2016 ).

Use of private automobiles is also stimulated by other macro-scale formal institutions, such as government policies on buildings, fuel taxes, and tollway fees. The Thai Building Control Regulation [FORIN4-MA] has been criticized for failing to limit how much parking space can be installed around residential and office buildings ( Chalermpong and Wibowo, 2007 ; Pongprasert and Kubota, 2018 ). This has led to a marked increase of parking spaces [infra2-mi] in private-sector-led real estate development, boosting the convenience of park-and-ride behavior for car users. Government policies—e.g., subsidies and tax breaks for consumers and the oil and gas industry—also incite vehicle use by reducing the burden of fuel costs [FORIN6-MA] ( ADB, 2015 ). Meanwhile, tollway prices are subsidized to encourage road use ( Charoentrakulpeeti et al., 2006 ).

With regard to the underdevelopment of public transportation, this is largely explained by the prioritization of road development in formal institutions [FORIN4-MA]. This has manifested as frequent political debates on the need for the state to invest in public transport. The development of public bus and rail networks has received inconsistent and less support [FORIN7-MA] than road development in terms of public funds, private investments, and development plans ( Charoentrakulpeeti et al., 2006 ; Hossain, 2006 ) despite the its overloaded capacity ( Peungnumsai et al., 2020 ). For example, a lack of political support has delayed the development and expansion of Bangkok's bus rapid transit system (BRT). The BRT—originally planned for 12 routes covering 185 km, as proposed in 2004—was delayed several times due to change in government and complicated institutional structures and bidding procedures. It finally opened in 2010, with coverage confined to a single 15 km route with low ridership capacity ( Wu and Pojani, 2016 ). The BRT's limited coverage and capacity [INFRA3-MA] ( Munira et al., 2013 ; Sanko et al., 2014 ; Pongprasert and Kubota, 2017 ) pose physical barriers to Bangkok residents able to access or ride public transport. These structural conditions at the macro scale thus reinforce preferences at the micro scale for using privately owned automobiles [inforins4-mi] ( Witchayaphong et al., 2020 ).

Development of the public transportation network is also hampered by the impact of knowledge and competencies on infrastructure development. This occurs in two sequences of lock-in causes. The first starts from the low degree of coordination between the different government departments and agencies involved in Bangkok's transportation system [KNOCOM2-MA]. Though several national and city departments oversee urban transport planning, operation, regulation, monitoring, and so on, there is no agency that leads this effort or oversees collaboration among the relevant bodies. This has led to fragmented transport planning and inter-agency conflict ( Hossain, 2006 ; Marks, 2020 ); it has also ultimately prevented the formation of an authoritative multi-modal transport master plan ( Wu and Pojani, 2016 ).

The lack of coordinated competencies and planning to guide the sustainable development of Bangkok's transportation system has also hampered the emergence of efforts to integrate the different modes of transport (BRT, Metropolitan Rapid Transit, skyrail, etc.) [INFRA4-MA] ( Ongkittikul, 2014 ). For example, Bangkok transport planners have missed an important opportunity to link and streamline fare collection for Metropolitan Rapid Transit and skyrail, which currently use separate systems ( Chalermpong and Ratanawaraha, 2015 ). Most of BRT, meanwhile, lacks any integration with rail stations ( Wu and Pojani, 2016 ). Inconveniences posed by the absence of an integrated ridership experience reduce accessibility to public transport, again increasing the probability of Bangkok residents using privately owned automobiles to fulfill their mobility needs.

The second sequence of lock-in causes relates to government agencies' limited competences in the design of safe and user-friendly infrastructures to promote the use of public transport [KNOCOM3-MA]. Understaffing and a lack of qualified technical consultants in public transportation agencies has resulted in poorly designed interconnectivity infrastructure [INFRA5-MA] around stations ( Wu and Pojani, 2016 ). This occurs with the design of walkways in stations themselves ( Prasertsubpakij and Nitivattananon, 2012 ) as much as the footpaths, pedestrian bridges, and shading devices (e.g., roofing or trees) around stations in the urban environment. The few Bangkok residents who choose to walk or cycle around thus feel uncomfortable and unsafe [inforin5-mi] ( Prasertsubpakij and Nitivattananon, 2012 ; Munira et al., 2013 ). These structural conditions discourage residents from accessing stations on foot and push them further away from public transport, perpetuating demand for motorized travel, even for trips of just a few hundred meters.

Conclusion and Contributions to Literature

Unraveling the diverse factors that influence urban consumption is crucial for tracing the causes of unsustainable behaviors, and for understanding the interlinked processes that contribute to their persistence, reproduction, and self-reinforcement. Much literature has explored the drivers and barriers of urban consumption. Until now, however, limited attention has been paid to the interdependences and interactions among them. This is especially so for factors that cross the traditional division between “structure” and “agency” in consumption literature. Furthermore, the influence of broader structural conditions on consumption behavior has received less attention, because many studies have focused on individually shaped circumstances, such as psychological factors and willingness to adopt new technologies, services, or practices.

To fill this gap in scholarship, this study married theoretical insights from the field of urban consumption with conceptions of socio-technical lock-in from sustainability and energy transition studies. Combining these into a comprehensive framework, we developed a theoretical tool for understanding the interlinked and self-reinforcing drivers of consumption behavior from two perspectives: macro scale (shared structural factors) and micro scale (individually shaped factors). This framework accounts for the power of both structure and agency, and allows scholars to understand the complex and intertwined causal relations that can lock urban dwellers into unsustainable consumption patterns, stopping them adopting new and more sustainable behavior.

As well as contributing to consumption literature, our study helps to overcome some of the limitations of scholarship on socio-technical lock-in to date. By interpreting cities as a socio-technical system with biophysical and human features, our framework contains an explicitly geographical dimension that accounts for place-specific characteristics. This connection to place has largely lacked in the existent literature. Our study also focused on social aspects and the consumption behavior of urban dwellers. This helps to overcome the tendency for literature on socio-technical transitions and lock-in to overemphasize structural dimensions and to underemphasize the importance of agency and micro-scale factors.

To illustrate the explanatory power of this framework, we applied it to a case study of Bangkok to identify the numerous and varied causes of lock-in to automobile-based mobility. Our analysis revealed that a complex web of influencing factors and bidirectional relations has contributed to this lock-in. At the macro scale, important factors included informal institutions such as political ideologies about the economic importance of local automotive industry. These drive the formation of formal institutions that encourage the purchase of vehicles and the expansion of supporting infrastructure, such as roads. At the micro scale, important factors included the intertwined effect of economic forces (e.g., rising household income) and informal institutions (e.g., perceptions of increased social status from owning vehicles) on increasing tendencies to purchase and use vehicles. We noted several feedbacks occurring across these factors. One involved the increase in individually owned vehicles, where the need to replace these spurred vigorous marketing strategies from car dealers. New car purchases led to more vehicles in the second-hand market being affordable for low-income households. Another self-reinforcing loop occurs as vehicle numbers grow, causing a demand for greater road capacity, which is subsequently met through government policy. The expanded road network increases the convenience of driving, which reinforces preferences for and lock-in to mobility based on individually owned automobiles.

The case study provides several important insights for policies to rupture the dynamics that perpetuate lock-in to unsustainable passenger mobility in the vast urban areas within and around Bangkok. Overall, development policies should avoid creating additional structural conditions that lock residents further into environmentally polluting mobility systems based on conventional automobiles. First and foremost, urban planning policies and greater public investments are needed to provide a diversity of alternatives to car-based travel and to realize the full potential of Bangkok's inefficient and poorly linked public transport network. Second, this study determined multiple situations where economic policies are fundamental drivers of individual consumption—in this case, the purchase of new vehicles. Therefore, policy changes should aim to shift consumer preferences toward sustainable vehicle choices, balancing development with environmental protection goals. For example, while gradually abolishing incentives for the purchase of gasoline engines along with fuel subsidies, new policies should be introduced to spur the purchase of compact-sized electric or hybrid vehicles. This would also necessitate public investments in charging infrastructure. Furthermore, government authorities should consider taking early actions on the eventual need to curb automaker investments in the production of vehicles with internal combustion engines. Governmental policies in many nations are increasingly restricting the production of this technology while setting target years for a complete phase-out. Authorities in Bangkok and Thailand may therefore be in the process of supporting the development of an industry with limited prospects for export. The state could look to other Asian neighbors such as China for lessons on how to steer the automobile industry toward a sustainable growth trajectory, focusing on electric, automated, and connected vehicles.

Future research could address the limitations of this study. Scholars could apply the framework empirically to other types of consumption behavior beyond mobility, such as food, fashion, housing, and energy. This study was unable to systematically determine the relative importance of influencing factors. Future studies could tackle this, perhaps by including quantitative methods such as cross-impact analyses that use survey data from consumers or experts, for example. By inviting respondents to rank the importance of particular factors, such approaches could sketch causal relationships and hierarchical structures more systematically and objectively. This could boost the practical value of the framework, by identifying both known and unknown sources. The results could then be used to identify strategies for disrupting the most important drivers of lock-in to unsustainable consumption behavior in specific cities.

Data Availability Statement

The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.

Author Contributions

NT, GT, and KM designed the study and edited the final version. NT collected and analyzed data. NT and GT wrote the original draft. All authors provided input into the draft and final manuscript.

The research was supported by Kaken Funds (Grant Numbers 19K20501 and 21H03666) from the Japan Society for the Promotion of Science.

Conflict of Interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher's Note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

1. ^ Globally, the 2.3% increase in energy consumption is responsible for an energy-related CO 2 emissions rise to 33.1 Gt CO 2 , up 1.7% in 2018 ( IEA, 2019 ).

2. ^ The urban heat island phenomenon occurs when urban areas experience a higher temperature at day or night than outlying areas. This typically results from land transformation after the reduction of forest or vegetation cover, the increase of artificial surfaces such as concrete, bitumen and buildings, and exhaust heat from vehicles, air-conditioners, and industrial activities.

3. ^ The total number of automobiles registered in Tokyo in 2019 was nearly 4 million, equivalent to around 0.3 per resident ( Automobile Inspection and Registration Information Association, 2019 ). In Beijing in 2017 the ratio was around 0.2 per resident ( Gao et al., 2020 ).

ADB. (2015). Fossil Fuel Subsidies in Thailand: Trends, Impacts, and Reforms . Available online at: www.adb.org (accessed August 15, 2020).

Google Scholar

Algesheimer, R., Dholakia, U. M., and Herrmann, A. (2005). The social influence of brand community: evidence from European car clubs. J. Mark. 69, 19–34. doi: 10.1509/jmkg.69.3.19.66363

CrossRef Full Text | Google Scholar

Alizadeh, T (2017). An investigation of IBM's smarter cites challenge: what do participating cities want? Cities 63, 70–80. doi: 10.1016/j.cities.2016.12.009

Aranda-Mena, G., Dahl Lund, J., and Dahl, C. (2018). “Learning from dense cities: Hong Kong spatial constructs as narratives,” in 52nd International Conference of the Architectural Science Association, December , 753–759. Available online at: https://www.researchgate.net/publication/330823624 (accessed August 15, 2020).

Arifwidodo, S. D., and Chandrasiri, O. (2020). Urban heat stress and human health in Bangkok, Thailand. Environ. Res. 185:109398. doi: 10.1016/j.envres.2020.109398

PubMed Abstract | CrossRef Full Text | Google Scholar

Attavanich, W (2017). “Impact of the first-time car buyer program on the environmental cost of air pollution in Bangkok,” in Kuala Lumpur Seminar Series , 83170.

Automobile Inspection Registration Information Association. (2019). Number of Vehicles Owned by Prefecture/Vehicle Type . Available online at: https://www.airia.or.jp/publish/statistics/number.html (accessed November 10, 2020).

Barnes, W., Gartland, M., and Stack, M. (2004). Old habits die hard: path dependency and behavioral lock-in. J. Econ. Issues 38, 371–377. doi: 10.1080/00213624.2004.11506696

Belgiawan, P. F., Schmöcker, J. D., Abou-Zeid, M., Walker, J., and Fujii, S. (2017). Modelling social norms: Case study of students' car purchase intentions. Travel Behav. Soc. 7, 12–25. doi: 10.1016/j.tbs.2016.11.003

Bengtsson, M., Alfredsson, E., Cohen, M., Lorek, S., and Schroeder, P. (2018). Transforming systems of consumption and production for achieving the sustainable development goals: moving beyond efficiency. Sustain. Sci. 13, 1533–1547. doi: 10.1007/s11625-018-0582-1

Bereznicka, J., and Pawlonka, T. (2018). Meat consumption as an indicator of economic well-being—case study of a developed and developing economy. Acta Sci. Polon. Oecon. 17, 17–26. doi: 10.22630/aspe.2018.17.2.17

Besagni, G., and Borgarello, M. (2018). The determinants of residential energy expenditure in Italy. Energy. 165, 369–386. doi: 10.1016/j.energy.2018.09.108

Bettencourt, L. M. A (2013). The origins of scaling in cities. Science. 340, 1438–1441. doi: 10.1126/science.1235823

Biswas, A., and Roy, M. (2015). Leveraging factors for sustained green consumption behavior based on consumption value perceptions: testing the structural model. J. Clean. Prod. 95, 332–340. doi: 10.1016/j.jclepro.2015.02.042

Böcker, L., Priya, T., Liu, C., and Dijst, M. (2019). Weather and daily mobility in international perspective : a cross-comparison of Dutch, Norwegian and Swedish city regions. Transport. Res. Part D 77, 491–505. doi: 10.1016/j.trd.2019.07.012

Borrás, S., and Edler, J. (2014). The governance of change in socio-technical and innovation systems: three pillars for a conceptual framework. Govern. Socio-Tech. Syst. Exp. Change 2018, 23–48. doi: 10.4337/9781784710194.00011

BOT. (2014). Thailand's Economic Conditions in 2014 . Available online at: https://www.bot.or.th/English/MonetaryPolicy/EconomicConditions/AnnualReport/AnnualReport/Annual_eng_2014.pdf (accessed on February 5, 2021).

Boucher, J. L (2016). Culture, carbon, and climate change: a class analysis of climate change belief, lifestyle lock-in, and personal carbon footprint. Soc. Ecol. J. Environ. Thought Sociol. Res. 25, 53–80. doi: 10.17234/SocEkol.25.1.3

Burge, P., Fox, J., Kouwenhoven, M., Rohr, C., and Wigan, M. R. (2007). Modeling of motorcycle ownership and commuter usage: a UK study. Transp. Res. Rec. 2031, 59–68. doi: 10.3141/2031-08

C40 Arup. (2019). The Future of Urban Consumption in a 1.5°C World . Available online at: https://www.c40.org/consumption (accessed January 8, 2020).

C40. (2018). Consumption-based GHG emissions of C40 cities (Issue March) . Available online at: https://c40-production-images.s3.amazonaws.com/researches/images/68_C40_CBE_Report_FINAL.original.pdf?1520350162 (accessed January 8, 2020).

Cachinho, H (2014). Consumerscapes and the resilience assessment of urban retail systems. Cities 36, 131–144. doi: 10.1016/j.cities.2012.10.005

Canitez, F (2019). Pathways to sustainable urban mobility in developing megacities: a socio-technical transition perspective. Technol. Forecast. Social Change 141, 319–329. doi: 10.1016/j.techfore.2019.01.008

Cecere, G., Corrocher, N., Gossart, C., and Ozman, M. (2014). Lock-in and path dependence: an evolutionary approach to eco-innovations. J. Evol. Econ. 24, 1037–1065. doi: 10.1007/s00191-014-0381-5

Chalermpong, S., and Ratanawaraha, A. (2015). How land use affects station access behaviors of bus rapid transit passengers in Bangkok, Thailand. Transp. Res. Rec. 2533, 50–59. doi: 10.3141/2533-06

Chalermpong, S., and Wibowo, S. S. (2007). Transit station access trips and factors affecting propensity to walk to transit stations in Bangkok, Thailand. J. Eastern Asia Soc. Transp. Stud. 7, 1806–1819. doi: 10.11175/eastpro.200

Chambers, P (2013). Economic guidance and contestation: an analysis of Thailand's evolving trajectory of development. J. Curr. Southeast Asian Aff. 32, 81–109. doi: 10.1177/186810341303200104

Chandrashekeran, S (2016). Multidimensionality and the multilevel perspective: territory, scale, and networks in a failed demand-side energy transition in Australia. Environ. Plan. A 48, 1636–1656. doi: 10.1177/0308518X16643728

Chapman, A., and Itaoka, K. (2018). Curiosity, economic and environmental reasoning: public perceptions of liberalization and renewable energy transition in Japan. Energy Res. Social Sci. 37, 102–110. doi: 10.1016/j.erss.2017.09.026

Charoentrakulpeeti, W., Sajor, E., and Zimmermann, W. (2006). Middle-class yravel patterns, predispositions and attitudes, and present-day transport policy in Bangkok, Thailand. Transp. Rev. 26, 693–712. doi: 10.1080/01441640600746927

Cheewaphongphan, P., Junpen, A., Garivait, S., and Chatani, S. (2017). Emission inventory of on-road transport in bangkok metropolitan region (BMR) development during 2007 to 2015 using the GAINS model. Atmosphere. 8:167. doi: 10.3390/atmos8090167

Chuersuwan, N., Nimrat, S., Lekphet, S., and Kerdkumrai, T. (2008). Levels and major sources of PM2.5 and PM10 in Bangkok Metropolitan Region. Environ. Int. 34, 671–677. doi: 10.1016/j.envint.2007.12.018

Clammer, J (1997). Contemporary Urban Japan—A Sociology of Consumption . Hoboken, NJ: Blackwell.

Cohen, M (2012). The future of automobile society: a socio-technical transitions perspective. Technol. Anal. Strat. Manag. 24, 377–390. doi: 10.1080/09537325.2012.663962

Cohen, M., and Murphy, J. (2001). Exploring Sustainable Consumption: Environmental Policy and the Social Sciences . Newark, NJ: New Jersey Institute of Technology.

Cohen, M. J (2005). Sustainable consumption in national context: an introduction to the special issue. Sustain. Sci. Pract. Policy 1, 22–28. doi: 10.1080/15487733.2005.11907962

Corman, S. R (2008). Structuration Theory. International Encyclopedia of Communication . Oxford, UK: Oxford University Press. doi: 10.1002/9781405186407.wbiecs111

Coşgel, M. M (1997). Consumption institutions. Rev. Soc. Econ. 55, 153–171. doi: 10.1080/00346769700000031

Davis, D. R., Dingel, J. I., Monras, J., and Morales, E. (2019). How segregated is urban consumption? Joan Monras. J. Polit. Econ. 127, 1684–1738. doi: 10.1086/701680

Dawkins, E., Andr,é, K., Axelsson, K., Benoist, L., Swartling, Å. G., and Persson, Å. (2019). Advancing sustainable consumption at the local government level: a literature review. J. Clean. Prod. 231, 1450–1462. doi: 10.1016/j.jclepro.2019.05.176

de Koning, J. I. J. C., Ta, T. H., Crul, M. R. M., Wever, R., and Brezet, J. C. (2016). GetGreen Vietnam: towards more sustainable behaviour among the urban middle class. J. Clean. Prod . doi: 10.1016/j.jclepro.2016.01.063

De Rubens, G. Z., Noel, L., and Sovacool, B. K. (2018). Dismissive and deceptive car dealerships create barriers to electric vehicle adoption at the point of sale. Nat. Energy 3, 501–507. doi: 10.1038/s41560-018-0152-x

Delgado, C. L (2003). Rising consumption of meat and milk in developing countries has created a new food revolution. J. Nutr. 133, 3907–3910. doi: 10.1093/jn/133.11.3907s

Department of Land Transport. (2021). Statistics on the Number of Newly Registered Vehicles in Bangkok, From 1995 to 2018 . Available online at: http://apps.dlt.go.th/statistics_web/statistics.html (accessed May 7, 2021).

Diehl, J. A., Sweeney, E., Wong, B., Sia, C. S., Yao, H., and Prabhudesai, M. (2020). Feeding cities: Singapore's approach to land use planning for urban agriculture. Glob. Food Sec . 26, 100377. doi: 10.1016/j.gfs.2020.100377

Ding, C., Cao, X., and Liu, C. (2019). How does the station-area built environment influence Metrorail ridership? Using gradient boosting decision trees to identify non-linear thresholds. J. Transp. Geogr. 77, 70–78. doi: 10.1016/j.jtrangeo.2019.04.011

Dissanayake, D., and Morikawa, T. (2010). Investigating household vehicle ownership, mode choice and trip sharing decisions using a combined revealed preference/stated preference Nested Logit model: case study in Bangkok Metropolitan Region. J. Transp. Geogr. 18, 402–410. doi: 10.1016/j.jtrangeo.2009.07.003

Duygan, M., Kachi, A., Oliveira, T. D., and Rinscheid, A. (2021). Introducing the Endowment-Practice-Institutions (EPI) framework for studying agency in the institutional contestation of socio-technical regimes. J. Clean. Prod. 296:126396. doi: 10.1016/j.jclepro.2021.126396

Eades, P., Kusch-Brandt, S., Heaven, S., and Banks, C. J. (2020). Estimating the generation of gardenwaste in England and the differences between rural and urban areas. Resources 9, 1–23. doi: 10.3390/resources9010008

Edmondson, D. L., Kern, F., and Rogge, K. S. (2019). The co-evolution of policy mixes and socio-technical systems: towards a conceptual framework of policy mix feedback in sustainability transitions. Res. Policy 48:103555. doi: 10.1016/j.respol.2018.03.010

EEA. (2012). Consumption and the Environment–2012 Update: The European Environment—State and Outlook 2010. Brussels: EEA.

Einhorn, L (2020). Normative Social Influence on Meat Consumption Laura Einhorn MPIfG Discussion Paper. MPIfG Discussion Paper . Available online at: https://pure.mpg.de/rest/items/item_3196378/component/file_3204446/content (accessed March 5, 2021).

Elbasha, T., and Wright, A. (2017). Reconciling structure and agency in strategy -as-practice research: Towards a strong- structuration theory approach. Management 20, 107–128. doi: 10.3917/mana.202.0107

Eom, J., Hyun, M., Lee, J., and Lee, H. (2020). Increase in household energy consumption due to ambient air pollution. Nat. Energy 5, 976–984. doi: 10.1038/s41560-020-00698-1

Erickson, P., Lazarus, M., and Tempest, K. (2015). Carbon Lock-in from Fossil Fuel Supply Infrastructure. SEI Discussion Brief, 2015 . Available online at: https://mediamanager.sei.org/documents/Publications/Climate/SEI-DB-2015-Carbon-lock-in-supply-side.pdf

Fell, H., Li, S., and Paul, A. (2014). A new look at residential electricity demand using household expenditure data. Int. J. Indus. Organ. 33, 37–47. doi: 10.1016/j.ijindorg.2014.02.001

Fold, N. R., Allison, M. R., Wood, B. C., Thao, P. T. B., Bonnet, S., Garivait, S., et al. (2020). An assessment of annual mortality attributable to ambient PM2.5 in Bangkok, Thailand. Int. J. Environ. Res. Public Health 17, 1–14. doi: 10.3390/ijerph17197298

Foxon, T. J (2002). Technological and institutional ‘ lock-in ' as a barrier to sustainable innovation. Change , 1–9. Available online at: http://www.iccept.ic.ac.uk (accessed January 8, 2021).

Foxon, T. J (2011). A coevolutionary framework for analysing a transition to a sustainable low carbon economy. Ecol. Econ. 70, 2258–2267. doi: 10.1016/j.ecolecon.2011.07.014

Frank, B (2017). Urban systems: a socio-ecological system perspective. Sociol. Int. J. 1, 1–8. doi: 10.15406/sij.2017.01.00001

Fuchs, D. A., and Lorek, S. (2005). Sustainable consumption governance: a history of promises and failures. J. Consumer Policy 28, 261–288. doi: 10.1007/s10603-005-8490-z

Gao, Y., Newman, P., and Kenworthy, J. (2020). The rise and decline of car use in Beijing and Shanghai. Sustain. Urban Plan. Des . 279–290. doi: 10.5772/intechopen.90130

Gasper, D., Shah, A., and Tankha, S. (2019). The framing of sustainable consumption and production in SDG 12. Global Policy 10, 83–95. doi: 10.1111/1758-5899.12592

Geels, F. W (2004). From sectoral systems of innovation to socio-technical systems: insights about dynamics and change from sociology and institutional. Res. Policy . 33, 897–920. doi: 10.1016/j.respol.2004.01.015

Geels, F. W (2018). Disruption and low-carbon system transformation: progress and new challenges in socio-technical transitions research and the Multi-Level Perspective. Energy Res. Social Sci 37, 224–231. doi: 10.1016/j.erss.2017.10.010

Geels, F. W., and Schot, J. (2007). Typology of sociotechnical transition pathways. Res. Policy 36, 399–417. doi: 10.1016/j.respol.2007.01.003

Giddens, A (1984). The Constitution of Society: Outline of the Theory of Structuration . California: University of California.

Goodwin, N., Development, G., Library, S. S., Thinking, F., Development, S., Well-being, H., et al. (2018). Macroeconomics in Context . London: Routledge.

Grimm, N. B., Faeth, S. H., Golubiewski, N. E., Redman, C. L., Wu, J., Bai, X., et al. (2008). Global change and the ecology of cities. Science. 319, 756–760. doi: 10.1126/science.1150195

Guan, H., Soebarto, V., Bennett, J., Clay, R., Andrew, R., Guo, Y., et al. (2014). Response of office building electricity consumption to urban weather in Adelaide, South Australia. Urban Climate 10, 42–55.

Guo, Y., Li, S., Tawatsupa, B., Punnasiri, K., Jaakkola, J. J. K., and Williams, G. (2014). The association between air pollution and mortality in Thailand. Sci. Rep. 4, 1–8. doi: 10.1038/srep05509

Hagen, J. X., Pardo, C. F., and Valente, J. B. (2016). Motivations for motorcycle use for Urban travel in Latin America: a qualitative study. Transport Policy 49, 93–104. doi: 10.1016/j.tranpol.2016.04.010

Haque, M. B., Choudhury, C., Hess, S., and Sourd, R. C (2019). dit Modelling residential mobility decision and its impact on car ownership and travel mode. Travel Behav. Soc. 17, 104–119. doi: 10.1016/j.tbs.2019.07.005

Hodson, M., and Marvin, S. (2010). Can cities shape socio-technical transitions and how would we know if they were? Res. Policy 39, 477–485. doi: 10.1016/j.respol.2010.01.020

Hossain, M (2006). The issues and realities of BRT planning initiatives in developing Asian cities. J. Public Transp. 9, 69–87. doi: 10.5038/2375-0901.9.3.4

Hult, A., and Larsson, J. (2016). Possibilities and problems with applying a consumption perspective in local climate strategies—the case of Gothenburg, Sweden. J. Clean. Prod. 134, 434–442. doi: 10.1016/j.jclepro.2015.10.033

IEA. (2019). Global Energy and CO 2 Status Report: the Lastest Trends in Energy and Emissions in 2018 . Paris: IEA

IQAir. (2019). World Air Quality Report. 2019 World Air Quality Report, August , 1–35. Available online at: https://www.iqair.com/world-most-polluted-cities/world-air-quality-report-2019-en.pdf

Jackson, T., and Papathanasopoulou, E. (2008). Luxury or “lock-in”? An exploration of unsustainable consumption in the UK: 1968 to 2000. Ecol. Econ. 68, 80–95. doi: 10.1016/j.ecolecon.2008.01.026

Janssen, M., and Jager, W. (1999). An integrated approach to simulating behavioural processes: a case study of the lock-in of consumption patterns. J. Artif. Soc. Social Simul. 2, 21–35.

Jayne, M (2005). Cities and Consumption . London: Routledge.

Jim, C. Y (2016). Urban biogeography. International Encyclopedia of Geography: People, the Earth, Environment and Technology: People, the Earth, Environment and Technology. New York, NY: Wiley, 1–4.

Jones, R. V., Fuertes, A., and Lomas, K. J. (2015). The socio-economic, dwelling and appliance related factors affecting electricity consumption in domestic buildings. Renew. Sustain. Energy Rev. 43, 901–917. doi: 10.1016/j.rser.2014.11.084

Jørgensen, U (2012). Mapping and navigating transitions - The multi-level perspective compared with arenas of development. Res. Policy 41, 996–1010. doi: 10.1016/j.respol.2012.03.001

Kaklauskas, A., and Gudauskas, R. (2016). “Intelligent decision-support systems and the Internet of Things for the smart built environment,” in: Start-Up Creation: The Smart Eco-Efficient Built Environment . Amsterdam: Elsevier.

Kanger, L., and Schot, J. (2016). User-made immobilities: a transitions perspective. Mobilities 11, 598–613. doi: 10.1080/17450101.2016.1211827

Khamchiangta, D., and Dhakal, S. (2019). Physical and non-physical factors driving urban heat island: Case of Bangkok Metropolitan Administration, Thailand. J. Environ. Manag. 248:109285. doi: 10.1016/j.jenvman.2019.109285

Khamchiangta, D., and Dhakal, S. (2020). Time series analysis of land use and land cover changes related to urban heat island intensity: case of Bangkok Metropolitan Area in Thailand. J. Urban Manag. 9, 383–395. doi: 10.1016/j.jum.2020.09.001

Kim, B., and Kim, D. (2020). Attracted to or locked in? Explaining consumer loyalty toward airbnb. Sustainability 12, 1–19. doi: 10.3390/su12072814

Kirchberg, V (2007). Cultural consumption analysis: Beyond structure and agency. Cult. Sociol. 1, 115–135. doi: 10.1177/1749975507073926

Kittiprapas, S (2000). “Thailand: The Asian financial crisis and social changes,” in The Social Impact of the Asia Crisis . London: Palgrave Macmillan, 35–56.

PubMed Abstract | Google Scholar

KNOEMA. (2021). Thailand - Gross Domestic Product Per Capita in Current Prices . Available online at: https://knoema.com/atlas/Thailand/topics/Economy/National-Accounts-Gross-Domestic-Product/GDP-per-capita?action=exportandgadget=indicator-preview-host (accessed February 5, 2021).

Krishna, G (2021). Understanding and identifying barriers to electric vehicle adoption through thematic analysis. Transp. Res. Interdiscip. Perspect. 10:100364. doi: 10.1016/j.trip.2021.100364

Krungsri Research (2020). Thailand Industry OUtlook 2020–2022: Automobile Industry . Krungsri Research

Kurniawan, R., Trencher, G. P., Edianto, A. S., Setiawan, I. E., and Matsubae, K. (2020). Understanding the multi-faceted drivers of increasing coal consumption in Indonesia. Energies 13, 1–22. doi: 10.3390/en13143660

Lariviere, I., and Lafrance, G. (1999). Modelling the electricity consumption of cities: effect of urban density. Energy Econ. 21, 53–66.

Lazzarini, G. A., Visschers, V. H. M., and Siegrist, M. (2018). How to improve consumers' environmental sustainability judgements of foods. J. Clean. Prod. 198, 564–574. doi: 10.1016/j.jclepro.2018.07.033

Lecler, Y (2002). The cluster role in the development of the Thai car industry. Int. J. Urban Reg. Res. 26, 799–814. doi: 10.1111/1468-2427.00419

Li, C., Song, Y., and Kaza, N. (2018). Urban form and household electricity consumption: a multilevel study. Energy Build. 158, 181–193. doi: 10.1016/j.enbuild.2017.10.007

Lipson, M. J., Thatcher, M., and Hart, M. A. (2019). Climate change impact on energy demand in building-urban-atmosphere simulations through the 21st century. Environ. Res. Lett. 14:125014. doi: 10.1088/1748-9326/ab5aa5

Lo, K., Mah, D. N., Wang, G., Leung, M. K. H., Lo, A. Y., and Hills, P. (2018). Barriers to adopting solar photovoltaic systems in Hong Kong. Energy Environ. 29, 649–663. doi: 10.1177/0958305X18757402

Lorek, S., and Fuchs, D. (2013). Strong sustainable consumption governance—precondition for a degrowth path? J. Clean. Prod. 38, 36–43. doi: 10.1016/j.jclepro.2011.08.008

Lorek, S., and Spangenberg, J. H. (2014). Sustainable consumption within a sustainable economy—beyond green growth and green economies. J. Clean. Prod. 63, 33–44. doi: 10.1016/j.jclepro.2013.08.045

Mah, D. N., yin Wang, G., Lo, K., Leung, M. K. H., Hills, P., and Lo, A. Y. (2018). Barriers and policy enablers for solar photovoltaics (PV) in cities: perspectives of potential adopters in Hong Kong. Renew. Sustain. Energy Rev. 92, 921–936. doi: 10.1016/j.rser.2018.04.041

Maréchal, K (2010). Not irrational but habitual: the importance of “behavioural lock-in” in energy consumption. Ecol. Econ. 69, 1104–1114. doi: 10.1016/j.ecolecon.2009.12.004

Markolf, S. A., Chester, M. V., Eisenberg, D. A., Iwaniec, D. M., Davidson, C. I., Zimmerman, R., et al. (2018). Interdependent Infrastructure as Linked Social, Ecological, and Technological Systems (SETSs) to Address Lock-in and Enhance Resilience. Earth's Future 6, 1638–1659. doi: 10.1029/2018EF000926

Marks, D (2020). An urban political ecology of Bangkok's awful traffic congestion. Journal of Political Ecology 27, 732–758. doi: 10.2458/V27I1.23604

Martiskainen, M., Axon, S., Sovacool, B. K., Sareen, S., Furszyfer Del Rio, D., and Axon, K. (2020). Contextualizing climate justice activism: knowledge, emotions, motivations, and actions among climate strikers in six cities. Global Environ. Change 65, 102180. doi: 10.1016/j.gloenvcha.2020.102180

Matsuyuki, M., Aizu, N., Nakamura, F., and Leeruttanawisut, K. (2020). Impact of gentrification on travel behavior in transit-oriented development areas in Bangkok, Thailand. Case Stud. Transp. Policy 8, 1341–1351. doi: 10.1016/j.cstp.2020.09.005

McIntosh, J., Trubka, R., Kenworthy, J., and Newman, P. (2014). The role of urban form and transit in city car dependence: analysis of 26 global cities from 1960 to 2000. Transp. Res. Part D: Transp. Environ. 33, 95–110. doi: 10.1016/j.trd.2014.08.013

Mckinsey. (2012). Urban World: Cities and the Rise of the Consuming Class. Washington, DC: McKinsey Global Institute.

Mckinsey. (2016). Urban World—the Global Consumers To Watch. McKinsey Global Insitute (Issue April) . Available online at: https://www.mckinsey.com/~/media/McKinsey/FeaturedInsights/Urbanization/UrbanworldTheglobalconsumerstowatch/Urban-World-Global-Consumers-Full-Report.ashx (accessed July 8, 2020).

McKinsey. (2020). The Great Consumer Shift : Ten Charts That Show How US Shopping Behavior is Changing . Washington, DC: McKinsey Global Institute.

McMeekin, A., and Southerton, D. (2012). Sustainability transitions and final consumption: Practices and socio-technical systems. Technol. Anal. Strat. Manag. 24, 345–361. doi: 10.1080/09537325.2012.663960

Meelen, T., Frenken, K., and Hobrink, S. (2019). Energy Research and Social Science Weak spots for car-sharing in The Netherlands? The geography of socio- technical regimes and the adoption of niche innovations. Energy Res. Social Sci. 52, 132–143. doi: 10.1016/j.erss.2019.01.023

Meng, F., Liu, G., Hu, Y., Su, M., and Yang, Z. (2018). From production to consumption: a multi-city comparative study of cross-regional carbon emissions. Energy Proc. 152, 744–749. doi: 10.1016/j.egypro.2018.09.239

Merfeld, K., Wilhelms, M., Henkel, S., and Kreutzer, K. (2019). Technological Forecasting and Social Change Carsharing with shared autonomous vehicles : uncovering drivers, barriers and future developments—a four-stage Delphi study. Technol. Forecast. Soc. Change 144, 66–81. doi: 10.1016/j.techfore.2019.03.012

Miles, S., and Paddison, R. (1998). Urban consumption: an historiographical note. Urban Stud. 35, 815–823. doi: 10.1080/0042098984565

Ministry of Transport. (2020). Transport Statistics 2018 . Bangkok: Ministry of Transport.

Munira, S., Sarm, S., and Santoso, D. S. (2013). Perception of public van users in Bangkok. J. Eastern Asia Soc. Transp. Stud. 10, 1501–1515. doi: 10.11175/easts.12.2215

Muthitacharoen, A., Samphantharak, K., and Chantarat, S. (2019). Fiscal stimulus and debt burden: evidence from Thailand's first-car-buyer tax rebate program. Int. Tax Public Fin. 26, 1383–1415. doi: 10.1007/s10797-019-09549-6

Mylan, J., Holmes, H., and Paddock, J. (2016). Re-introducing consumption to the “circular economy”: A sociotechnical analysis of domestic food provisioning. Sustainability 8:8080794. doi: 10.3390/su8080794

Nguyen, M. Q., Thanh, T., Pham, X., Thuy, T., and Phan, H. (2019). Traffic Congestion—a prominent problem in vietnam current situation and solutions. Eur. J. Eng. Res. Sci. 4, 112–116. doi: 10.24018/ejers.2019.4.9.1524

Niinimäki, K (2010). Eco-clothing, consumer identity and ideology. Sustain. Dev. 18, 150–162. doi: 10.1002/sd.455

Nissen, K. M., and Becker, N. (2020). How does weather affect the use of public transport in Berlin? OPEN ACCESS How does weather affect the use of public transport in Berlin?

Niu, H. J (2013). Cyber peers' influence for adolescent consumer in decision-making styles and online purchasing behavior. J. Appl. Soc. Psychol. 43, 1228–1237. doi: 10.1111/jasp.12085

OECD. (2020). The Overview of Economic Assessment of Thailand. Paris: OECD Publishing.

OECD. (2021). OECD Investment Policy Reviews: Thailand (OECD Inves). Paris: OECD Publishing.

Oliveira, M. G., Vidal, D. G., and Ferraz, M. P. (2020). Urban lifestyles and consumption patterns. Sustain. Cities Commun. 2020, 851–860. doi: 10.1007/978-3-319-95717-3_54

Ongkittikul, S (2014). Chapter 9 Thailand Country Report , vol 15. Jakarta: ERIA. 351–396.

Orum, A. M., and Dunleavy, P. (2019). Collective consumption. Wiley Blackwell Encycloped Urban Reg. Stud. 2019, 1–8. doi: 10.1002/9781118568446.eurs0052

OTP (2018). Travel Demand Freight Movement Survey For National Transport Planning. Beaverton, OR: OTP.

Ou, S., Lin, Z., He, X., and Przesmitzki, S. (2018). Estimation of vehicle home parking availability in China and quantification of its potential impacts on plug-in electric vehicle ownership cost. Transp. Policy 68, 107–117. doi: 10.1016/j.tranpol.2018.04.014

Pasuk, P., and Pornthep, B. (2012). Locked in the middle-income trap: Thailand ' s economy between resilience and future challenges. Friedrich Ebert Stiftung, March . Available online at: http://library.fes.de/pdf-files/bueros/thailand/09208.pdf (accessed February 5, 2020).

Patorniti, N. P., Stevens, N. J., and Salmon, P. M. (2017). A systems approach to city design: exploring the compatibility of sociotechnical systems. Habitat Int. 66, 42–48. doi: 10.1016/j.habitatint.2017.05.008

Patorniti, N. P., Stevens, N. J., and Salmon, P. M. (2018). A sociotechnical systems approach to understand complex urban systems: a global transdisciplinary perspective. Hum. Factors Ergon. Manuf. 28, 281–296. doi: 10.1002/hfm.20742

Pekkanen, T. L (2020). Institutions and agency in the sustainability of day-to-day consumption practices: an institutional ethnographic study. J. Bus. Ethics . 168, 241–260. doi: 10.1007/s10551-019-04419-x

Pereira Heath, M. T., and Chatzidakis, A. (2012). “Blame it on marketing”: consumers' views on unsustainable consumption. Int. J. Consum. Stud. 36, 656–667. doi: 10.1111/j.1470-6431.2011.01043.x

Peungnumsai, A., Miyazaki, H., Witayangkurn, A., and Kim, S. M. (2020). A grid-based spatial analysis for detecting supply—demand gaps of public transports: a case study of the bangkok metropolitan region. Sustainability 12, 1–27. doi: 10.3390/su122410382

Phipps, A. G (2001). Empirical applications of structuration theory. Geogr. Ann. Ser. B Human Geogr. 83, 189–204. doi: 10.1111/j.0435-3684.2001.00106.x

Poças Ribeiro, A., Harmsen, R., Rosales Carreón, J., and Worrell, E. (2019). What influences consumption? Consumers and beyond: purposes, contexts, agents and history. J. Clean. Prod. 209, 200–215. doi: 10.1016/j.jclepro.2018.10.103

Pollio, C (2012). The Auto Industry in Thailand : Value Transfer, Technological Dependence and Relations Between Local and Foreign Capital (C. MET Working Paper, Issue Febuary 2012).

Pollio, C., and Rubini, L. (2021). Who drives the automotive sector? Thailand selective policies. Int. J. Emerg. Markets 16, 370–390. doi: 10.1108/IJOEM-02-2018-0084

Pongprasert, P., and Kubota, H. (2017). Why do TOD residents still use cars? Understanding the factors affecting the automobile ownership and use of residents living near transit stations of Bangkok. Urban Reg. Plan. Rev. 4, 231–250. doi: 10.1061/9780784481547.018

Pongprasert, P., and Kubota, H. (2018). “The influences of built environment factors on mode switching of TOD residents from car use to transit dependence: case study of Bangkok, Thailand,” in International Conference on Transportation and Development 2018 , 186–196. Available online at: http://www.asce-ictd.org/ (accessed April 20, 2020).

Pongthanaisawan, J., and Sorapipatana, C. (2010). Relationship between level of economic development and motorcycle and car ownerships and their impacts on fuel consumption and greenhouse gas emission in Thailand. Renew. Sustain. Energy Rev. 14, 2966–2975. doi: 10.1016/j.rser.2010.07.034

Prasertsubpakij, D., and Nitivattananon, V. (2012). Evaluating accessibility to Bangkok Metro Systems using multi-dimensional criteria across user groups. IATSS Research 36, 56–65. doi: 10.1016/j.iatssr.2012.02.003

Rajagopal, R (2011). Consumer culture and purchase intentions toward fashion apparel in Mexico. Datab. Market. Customer Strat. Manag. 18, 286–307. doi: 10.1057/dbm.2011.33

Reindl, K., and Palm, J. (2021). Installing PV: barriers and enablers experienced by non-residential property owners. Renew. Sustain. Energy Rev. 141:110829. doi: 10.1016/j.rser.2021.110829

Riordan, T. O., and Stoll-kleemann, S. (2015). The challenges of changing dietary behavior toward more sustainable consumption. Environ. Sci. Policy Sustain. Dev. 57, 4–13. doi: 10.1080/00139157.2015.1069093

Roberts, M. B., Bruce, A., and MacGill, I. (2019). Opportunities and barriers for photovoltaics on multi-unit residential buildings: reviewing the Australian experience. Renew. Sustain. Energy Rev. 104, 95–110. doi: 10.1016/j.rser.2018.12.013

Robeyns, I (2019). What, if anything, is wrong with extreme wealth? J. Human Dev. Capabil. 20, 251–266. doi: 10.1080/19452829.2019.1633734

Rudd, N. A (1997). Cosmetics consumption and use among women: ritualized activities that construct and transform the self. J. Ritual Stud. 1997, 59–77.

Sakah, M., de la Rue du Can, S., Diawuo, F. A., Sedzro, M. D., and Kuhn, C. (2019). A study of appliance ownership and electricity consumption determinants in urban Ghanaian households. Sustain. Cities Soc. 44, 559–581. doi: 10.1016/j.scs.2018.10.019

Sanko, N., Dissanayake, D., Kurauchi, S., Maesoba, H., Yamamoto, T., and Morikawa, T. (2014). Household car and motorcycle ownership in Bangkok and Kuala Lumpur in comparison with Nagoya. Transp. A: Transp. Sci. 10, 187–213. doi: 10.1080/18128602.2012.726285

Sanne, C (2002). Willing consumers—or locked-in? Policies for a sustainable consumption. Ecol. Econ. 42, 273–287. doi: 10.1016/S0921-8009(02)00086-1

Sareen, S., Remme, D., Wågsæther, K., and Haarstad, H. (2021). Energy Research and Social Science A matter of time: explicating temporality in science and technology studies and Bergen ' s car-free zone development. Energy Res. Social Sci. 78:102128. doi: 10.1016/j.erss.2021.102128

Schor. (2005). Prices and quantities: Unsustainable consumption and the global economy. Ecol. Econ. 55, 309–320. doi: 10.1016/j.ecolecon.2005.07.030

Schulte, B. K (2015). Staying the Consumption Course: Exploring the Individual Lock-In Process in Service Relationships. Berlin: Springer Gabler.

Seto, K. C., Davis, S. J., Mitchell, R. B., Stokes, E. C., Unruh, G., and Ürge-Vorsatz, D. (2016). Carbon lock-in: types, causes, and policy implications. Annu. Rev. Environ. Resour. 41, 425–452. doi: 10.1146/annurev-environ-110615-085934

Seto, K. C., Dhakal, S., Bigio, A., Blanco, H., Delgado, G. C., Dewar, D., et al (2014). Human Settlements, Infrastructure, and Spatial Planning . Available online at: https://www.researchgate.net/publication/275035277_Human_Settlements_Infrastructure_and_Spatial_Planning (accessed May 20, 2020).

Shove, E., and Walker, G. (2010). Governing transitions in the sustainability of everyday life. Res. Policy 39, 471–476. doi: 10.1016/j.respol.2010.01.019

Singh, J., Mantha, S. S., and Phalle, V. M. (2017). Characterizing domestic electricity consumption in the Indian urban household sector. Energy Build. J. 170, 74–82. doi: 10.1016/j.enbuild.2018.04.002

Soltani, A., Pojani, D., Askari, S., and Masoumi, H. E. (2018). Socio-demographic and built environment determinants of car use among older adults in Iran. J. Transp. Geogr. 68, 109–117. doi: 10.1016/j.jtrangeo.2018.03.001

Soltani, M., Rahmani, O., Pour, A. B., Ghaderpour, Y., Ngah, I., and Misnan, S. H. (2019). Determinants of variation in household energy choice and consumption: case from Mahabad City, Iran. Sustainability 11:11174775. doi: 10.3390/su11174775

Souza, L. C. L., Postigo, C. P., Oliveira, A. P., and Nakata, C. M. (2009). Urban heat islands and electrical energy consumption. Int. J. Sustain. Energy 28, 113–121. doi: 10.1080/14786450802453249

Sovacool, B. K (2016). How long will it take? Conceptualizing the temporal dynamics of energy transitions. Energy Res. Social Sci. 13, 202–215. doi: 10.1016/j.erss.2015.12.020

Sovacool, B. K (2017). Experts, theories, and electric mobility transitions: Toward an integrated conceptual framework for the adoption of electric vehicles. Energy Res. Social Sci. 27, 78–95. doi: 10.1016/j.erss.2017.02.014

Sovacool, B. K., and Griffiths, S. (2020). The cultural barriers to a low-carbon future: a review of six mobility and energy transitions across 28 countries. Renew. Sustain. Energy Rev. 119:109569. doi: 10.1016/j.rser.2019.109569

Sovacool, B. K., and Martiskainen, M. (2020). Hot transformations: governing rapid and deep household heating transitions in China, Denmark, Finland and the United Kingdom. Energy Policy 139:111330. doi: 10.1016/j.enpol.2020.111330

Spangenberg, J (2014). Institutional change for strong sustainable consumption: sustainable consumption and the degrowth economy. Sustain. Sci. Pract. Policy 10, 62–77. doi: 10.1080/15487733.2014.11908125

Su, Y. W (2019). Residential electricity demand in Taiwan: consumption behavior and rebound effect. Energy Policy 124, 36–45. doi: 10.1016/j.enpol.2018.09.009

Sudmant, A., Gouldson, A., Millward-Hopkins, J., Scott, K., and Barrett, J. (2018). Producer cities and consumer cities: using production- and consumption-based carbon accounts to guide climate action in China, the UK, and the US. J. Clean. Prod. 176, 654–662. doi: 10.1016/j.jclepro.2017.12.139

Sun, F., Chen, P., and Jiao, J. (2018). Promoting public bike-sharing: a lesson from the unsuccessful Pronto system. Transp. Res. Part D 63, 533–547. doi: 10.1016/j.trd.2018.06.021

Tang, R (2009). The Rise of China ' s Auto Industry and Its Impact on the U.S. Motor Vehicle Industry. Ithaca, NY: Cornell University

Tao, S., Corcoran, J., Rowe, F., and Hickman, M. (2018). To travel or not to travel : ‘Weather' is the question. Modelling the effect of local weather conditions on bus ridership. 86, 147–167. doi: 10.1016/j.trc.2017.11.005

Tao, Y., Wang, Q., and Zou, Y. (2021). Simulation and analysis of urban production–living–ecological space evolution based on a macro–micro joint decision model. Int. J. Environ. Res. Public Health 18:9832. doi: 10.3390/ijerph18189832

Taweekun, J., and Tantiwichien, A.-U.-W. (2013). Thermal comfort zone for Thai people. Engineering 05, 525–529. doi: 10.4236/eng.2013.55062

Taylor, M. A. P., and Yue, W. L. (2011). An investigation of vehicle ownership and the effects of income and vehicle expenses in mid-sized city of Thailand. J. Eastern Asia Soc. Transp. Stud. 9, 437–451. doi: 10.11175/easts.9.437

Techakanont, K., and Leelahanon, S. (2015). “The impact of economic crises on the Thai automobile industry,” in Global Automobile Demand (pp. 159–184). Palgrave Macmillan, London.

Thailand Board of Investment (2017). Thailand's Automotive Industry The Next-Generation . Malaysia: TBI.

Trading Economics (2021). Thailand Total Vehicle Sales 1980–2020 Data . Available online at: https://tradingeconomics.com/thailand/total-vehicle-sales (accessed August 25, 2021).

Trencher, G., Rinscheid, A., Duygan, M., Truong, N., and Asuka, J. (2020). Revisiting carbon lock-in in energy systems: Explaining the perpetuation of coal power in Japan. Energy Res. Social Sci. 69:101770. doi: 10.1016/j.erss.2020.101770

Trencher, G., Truong, N., Temocin, P., and Duygan, M. (2021). Energy Research and Social Science Top-down sustainability transitions in action: how do incumbent actors drive electric mobility diffusion in China, Japan, and California? Energy Res. Social Sci. 79:102184. doi: 10.1016/j.erss.2021.102184

Trinh, L. T. K., Hu, A. H., and Pham Phu, S. T. (2021). Situation, challenges, and solutions of policy implementation on municipal waste management in Vietnam toward sustainability. Sustainability 13. doi: 10.3390/su13063517

Ueno, T., Inada, R., Saeki, O., and Tsuji, K. (2005). “Effectiveness of displaying energy consumption data in residential houses-Analysis on how the residents respond,” in Proceedings of European Council for an Energy-efficient Economy, Panel 6 . (ECEEE), 1289-1299. Available online at: https://www.eceee.org/static/media/uploads/site-2/library/conference_proceedings/eceee_Summer_Studies/2005c/Panel_6/6100ueno/paper.pdf

UNEP. (2018a). Single-Use Plastics: A Roadmap for Sustainability . Available online at: http://hdl.handle.net/20.500.11822/25496 (accessed July 20, 2020).

UNEP. (2018b). The Weight of Cities Resource Requirements . Nairobi: UNEP.

UNEP. (2019). Emissions Gap Report 2019 Executive Summary. In Emissions Gap Report 2019 . Available online at: https://wedocs.unep.org/bitstream/handle/20.500.11822/30798/EGR19ESEN.pdf?sequence=13 (accessed July 20, 2020).

Unruh, G. C (2000). Understanding carbon lock-in. Energy Policy 28, 817–830. doi: 10.1016/S0301-4215(00)00070-7

Ürge-Vorsatz, Di., Rosenzweig, C., Dawson, R. J., Sanchez Rodriguez, R., Bai, X., Barau, A. S., et al. (2018). Locking in positive climate responses in cities. Nat. Clim. Chang. 8, 174–177. doi: 10.1038/s41558-018-0100-6

Uttamang, P., Aneja, V., and Hanna, A. (2017). Assessment of Air Pollution in Bangkok Metropolitan Region, Thailand. Atmos. Chem. Phys. Discuss. x , 1–28. doi: 10.5194/acp-2017-1063

Vergragt, P. J., Dendler, L., de Jong, M., and Matus, K. (2016). Transitions to sustainable consumption and production in cities. J. Clean. Prod. 134, 1–12. doi: 10.1016/j.jclepro.2016.05.050

Verma, R. L., Borongan, G., and Memon, M. (2016). Municipal solid waste management in Ho Chi Minh City, Viet Nam, current practices and future recommendation. Proc. Environ. Sci. 35, 127–139. doi: 10.1016/j.proenv.2016.07.059

Vermeir, I., Weijters, B., Houwer, J., De Geuens, M., Slabbinck, H., Spruyt, A., et al. (2020). Environmentally sustainable food consumption : a review and research agenda from a goal-directed perspective. Front. Psychol. 11(July). doi: 10.3389/fpsyg.2020.01603

Wad, P (2009). The automobile industry of Southeast Asia: Malaysia and Thailand. J. Asia Pac. Econ. 14, 172–193. doi: 10.1080/13547860902786029

Warr, P., and Kohpaiboon, A. (2017). Thailand' s Automotive Manufacturing Corridor, ADB Economics Working Paper Series (Issue 519) . Mandaluyong: Asian Development Bank.

Weatherall, D., Mccarthy, F., and Bright, S. (2020). Property law as a barrier to energy upgrades in multi-owned properties: insights from a study of England and Scotland. May 2010, 1641–1655. doi: 10.1007/s12053-017-9540-5

Werner, P (2011). The ecology of urban areas and their functions for species diversity. Landsc. Ecol. Eng. 7, 231–240. doi: 10.1007/s11355-011-0153-4

Wesseling, J. H., and Van der Vooren, A. (2017). Lock-in of mature innovation systems: the transformation toward clean concrete in the Netherlands. J. Clean. Prod. 155, 114–124. doi: 10.1016/j.jclepro.2016.08.115

Whittington, R (2015). “Giddens, structuration theory and strategy as practice,” in Cambridge Handbook of Strategy as Practice, Second Edition, January (Cambridge: Cambridge University Press), 145–164.

Wiedmann, T., Lenzen, M., Keyßer, L. T., and Steinberger, J. K. (2020). Scientists' warning on affluence. Nat. Commun. 11, 1–10. doi: 10.1038/s41467-020-16941-y

Wirtz, J (2019). Jollibee Foods Corporation . Pasig: Jollibee Food Corporation.

Witchayaphong, P., Pravinvongvuth, S., Kanitpong, K., Sano, K., and Horpibulsuk, S. (2020). Influential factors affecting travelers' mode choice behavior on mass transit in Bangkok, Thailand. Sustainability 12, 1–18. doi: 10.3390/su12229522

Wonglimpiyarat, J (2016). Towards the Detroit of Asia: empirical research insights of Thailand's OEM strategy. J. High Technol. Manag. Res. 27, 78–87. doi: 10.1016/j.hitech.2016.04.007

Wu, I., and Pojani, D. (2016). Obstacles to the creation of successful bus rapid transit systems: the case of Bangkok. Res. Transp. Econ. 60, 44–53. doi: 10.1016/j.retrec.2016.05.001

Yagita, Y., and Iwafune, Y. (2021). Residential energy use and energy-saving of older adults: a case from Japan, the fastest-aging country. Energy Res. Social Sci. 75:102022. doi: 10.1016/j.erss.2021.102022

Yang, M., Wu, J., Rasouli, S., Cirillo, C., and Li, D. (2017). Exploring the impact of residential relocation on modal shift in commute trips: Evidence from a quasi-longitudinal analysis. Transp. Policy 59, 142–152. doi: 10.1016/j.tranpol.2017.07.005

Yang, Y., Wang, C., Liu, W., and Zhou, P. (2018). Understanding the determinants of travel mode choice of residents and its carbon mitigation potential. Energy Policy 115, 486–493. doi: 10.1016/j.enpol.2018.01.033

Yao, M., and Wang, D. (2018). Mobility and travel behavior in urban China: The role of institutional factors. Transp. Policy 69, 122–131. doi: 10.1016/j.tranpol.2018.05.012

Zaharia, A., Diaconeasa, M. C., Brad, L., Lădaru, G. R., and Ioanăs, C. (2019). Factors influencing energy consumption in the context of sustainable development. Sustainability 11, 1–28. doi: 10.3390/su11154147

Zauberman, G (2003). The intertemporal dynamics of consumer lock-in. J. Consum. Res. 30, 405–419. doi: 10.1086/378617

Zhu, J., and Li, D. (2015). Current situation of energy consumption and energy saving analysis of large public building. Proc. Eng. 121, 1208–1214. doi: 10.1016/j.proeng.2015.09.140

Keywords: consumption, lock-in, mobility, socio-technical system, Thailand, urban

Citation: Truong N, Trencher G and Matsubae K (2022) How Does Socio-Technical Lock-In Cause Unsustainable Consumption in Cities? A Framework and Case Study on Mobility in Bangkok. Front. Sustain. Cities 4:770984. doi: 10.3389/frsc.2022.770984

Received: 05 September 2021; Accepted: 04 January 2022; Published: 08 February 2022.

Reviewed by:

Copyright © 2022 Truong, Trencher and Matsubae. This is an open-access article distributed under the terms of the Creative Commons Attribution License (CC BY) . The use, distribution or reproduction in other forums is permitted, provided the original author(s) and the copyright owner(s) are credited and that the original publication in this journal is cited, in accordance with accepted academic practice. No use, distribution or reproduction is permitted which does not comply with these terms.

*Correspondence: Nhi Truong, truong.thi.ai.nhi.r2@dc.tohoku.ac.jp

This article is part of the Research Topic

Environmental Data, Governance and the Sustainable City

A socio-technical analysis of work with ideas in NPD: an industrial case study

  • Original Paper
  • Published: 24 July 2013
  • Volume 24 , pages 411–427, ( 2013 )

Cite this article

  • Liv Gish 1 &
  • Claus Thorp Hansen 2  

677 Accesses

14 Citations

Explore all metrics

The work with ideas is perceived by both academics and industry to be an important element in new product development and has thus been the subject of investigation in many different research fields. In this paper, we investigate a specific case showing how a successful product was developed based on piecing together a number of ideas that were developed and disseminated in a large industrial company. We do this through an in-depth case study of the development of the energy-labeled circulation pump Alpha Pro, developed by one of the world’s leading pump manufacturers, Grundfos. Using a socio-technical approach, we focus especially on the actors involved and the contextual factors, and less on the detailed development of technical ideas. In our study, we observe that (1) ideas are pieced together from previous ideas and results; (2) ideas are implemented through continuous mobilization of support and development of legitimate arguments; and (3) idea work is also a socio-technical process, because contextual factors matter. We observe that idea work is an ongoing process undertaken across different projects, actors, departments, strategies, and visions within Grundfos, while also involving external actors outside Grundfos. Based on our socio-technical approach and observations, we begin to develop a broader understanding of idea work in new product development and sketch an analytical framework that can be used to make socio-technical analyses of product development projects in future research.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA) Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Rent this article via DeepDyve

Institutional subscriptions

In the EU, a company is considered large when it exceeds 250 employees and has a net turnover of at least €50 million.

Altshuller G, Shulyak L (1998) 40 principles: TRIZ keys to technical innovation. Technical Innovation Center, Worcester, MA

Google Scholar  

Bakker H, Boersma K, Oreel S (2006) Creativity (ideas) management in industrial R&D organizations: a crea-political process model and an empirical illustration of corus RD&T. J Creat Innov Manage 15:296–309

Article   Google Scholar  

Björk J, Magnusson M (2009) Where do good innovation ideas come from? Exploring the influence of network connectivity on innovation idea quality. J Prod Innovation Manage 26:662–670

Boeddrich H (2004) Ideas in the workplace: a new approach towards organizing the fuzzy front end of the innovation process. Creat Innov Manage 13:274–285

Bryman A (2001) Social research methods. Oxford University Press, Oxford

Bylinsky (1985) Smart-power chips. Fortune Int 111:48–50

Chesbrough H (2007) Open innovation: the new imperative for creating and profiting from technology. Harvard Business School, Boston, MA

Chulvi V, González-Cruz MC, Mulet E, Aguilar-Zambrano J (2013) Influence of the type of idea-generation method on the creativity of solutions. Res Eng Design 24:33–41

Cooper RG (2001) Winning at new products—accelerating the process form idea to launch. Basic Books, New York

Detterfelt J, Lovén E, Lakemond N (2009) Suggestion systems for engineering designers—a case study. In: 17th international conference on engineering design, ICED’09, Stanford University, Stanford, CA, USA, The Design Society

Dorst K, Cross N (2001) Creativity in the design process: co-evolution of problem-solution. Des Stud 22:425–437

Eisenhardt KM (1989) Building theories from case study research. Acad Manage Rev 14:532–550

Flyvbjerg B (2011) Case study. In: Denzin NK, Lincoln YS (eds) The Sage handbook of qualitative research, 4th edn. Sage, Thousand Oaks, CA, pp 301–316

Fry A (1987) The post-it note: an intrapreneurial success. SAM Adv Manage J 52:4–9

Gable (1994) Integrating case study and survey research methods: an example in information systems. Eur J Inf Syst (UK) 3:112–126

Garud R, Karnoe P (2001) Path creation as a process of mindful deviation. In: Garud R, Karnoe P (eds) Path dependence and creation. Lawrence Erlbaum Associates, Mahwah, NJ, pp 1–38

Garud R, Gehman J, Kumaraswamy A (2011) Complexity arrangements for sustained innovation: lessons from 3 M corporation. org stud 32:737–767

Geschka H (1983) Creativity techniques in product-planning and development—a view from West-Germany. R D Manage 13:169–183

Gish, Clausen C (forthcoming) The framing of product ideas in the making—a case study of the development of an energy saving pump. Technol Anal Strateg

Hansen CT, Andreasen MM (2005) On the content of a product idea. In: 15th international conference on engineering design, ICED’05, Melbourne, Engineers Australia

Hathuel A, Weil B (2009) C-K design theory: an advanced formulation. Res Eng Design 19:181–192

Howard TJ, Dekoninck EA, Culley SJ (2010) The use of creative stimuli at early stages of industrial product innovation. Res Eng Design 21:263–274

Howard TJ, Culley S, Dekoninck EA (2011) Reuse of ideas and concepts for creative stimuli in engineering design. J Eng Des 22:565–581

Jensen TE (2001) Performing social work: competence, orderings, spaces, and objects. PhD Dissertation, Department of Psychology, University of Copenhagen

Khurana A, Rosenthal SR (1998) Towards holistic “front ends” in new product development. J Prod Innov Manage 15:57–74

Kijkuit B, van den Ende J (2007) The organizational life of an idea: integrating social network, creativity and decision-making perspectives. J Manage Stud 44:863–882

Killander AJ (2001) Why design methodologies are difficult to implement. Int J Technol Manage 21:271–276

Koen PA, Kohli P (1998) Idea Generation: who has the most profitable ideas. Eng Manage J 10:35–40

Koen PA, Ajamian GM, Boyce S et al (2002) Fuzzy front end: effective methods, tools, and techniques. In: Belliveau P et al (eds) The PDMA toolbook for new product development. Wiley, New York, pp 5–35

Kvale S (2009) InterViews: learning the craft of qualitative research interviewing. Sage, Thousand Oaks

Legardeur J, Boujut JF, Tiger H (2010) Lessons learned from an empirical study of the early design phases of an unfulfilled innovation. Res Eng Design 21:249–262

Liikkanen LA, Perttula M (2010) Inspiring design idea generation: insights from a memory-search perspective. J Eng Des 21:545–560

Lopez-Mesa B, Bylund N (2011) A study of the use of concept selection methods from inside a company. Res Eng Design 22:7–27

Mak TW, Shu LH (2008) Using descriptions of biological phenomena for idea generation. Res Eng Design 19:21–28

Markham SK, Ward SJ, Aiman-Smith L, Kingon AI (2010) The valley of death as context for role theory in product innovation. J Prod Innov Manage 27:402–417

Montoya-Weiss M, Calantone R (1994) Determinants of new product performance: a review and meta-analysis. J Prod Innov Manage 11:397–417

Montoya-Weiss M, O’Driscoll TM (2000) From experience: applying performance support technology in the fuzzy front end. J Prod Innov Manage 17:143–161

Mulet E, Vidal R (2008) Heuristic guidelines to support conceptual design. Res Eng Design 19:101–112

Osborn AF (1965) Applied imagination: principles and procedures of creative problem-solving. Scribner, NY

Perttula M, Sipila P (2007) The idea exposure paradigm in design idea generation. J Eng Des 18:93–102

Rubenstein AH (1994) At the front end of the R&D/innovation process: idea development and entrepreneurship. Int J Technol Manage 9:652–677

Ryhammar L, Brolin C (1999) Creativity research: historical considerations and main lines of development. Scand J Educ Res 43:259–273

Shah JJ, Finger S, Lu S et al (2004) ED2030: strategic plan for engineering design. Final report. NSF workshop on engineering design in year 2030. March 26–29th, 2004, Gold Canyon, Arizona, Hosted by DEMAP/Design Automation Lab, Arizona State University

Smith PG, Reinertsen DG (1998) Developing products in half the time: new rules, new tools. Wiley, New York

Stasch SF, Lonsdale RT, LaVenka NM (1992) Developing a framework for sources of new-product ideas. J Consum Market 9:5–15

Strauss A, Corbin J (1990) Basics of qualitative research: Grounded theory procedures and techniques. Sage, Newbury Parl, CA

Tidd J, Bessant J (2009) Managing innovation: integrating technological, market and organizational change. Wiley, Hoboken, NJ

Ulrich KT, Eppinger SD (2003) Product design and development. Irwin McGraw-Hill, Boston, MA

van de Ven A (1986) Central problems in the management of innovation. Manage Sci 32:590–607

van Dijk C, van den Ende J (2002) Suggestion systems: transferring employee creativity into practicable ideas. R D Manage 32:387–395

von Hippel E (2005) Democratizing innovation: the evolving phenomenon of user innovation. J Betriebswirtschaft 55:63–78

Wheelwright SC, Clark KB (1992) Revolutionizing product development: quantum leaps in speed, efficiency, and quality. Free Press, New York

Wodehouse A, Ion W (2012) Augmenting the 6-3-5 method with design information. Res Eng Design 23:5–15

Yang M (2009) Observations on concept generation and sketching in engineering design. Res Eng Design 20:1–11

Yin RK (2009) Case study research: design and methods. Sage, London

Zwicky F (1969) Discovery, invention, research through the morphological approach. MacMillan, Canada

Download references

Acknowledgments

This research is based on the first author’s PhD project, which is funded by Grundfos Management A/S, the Technical University of Denmark, and the Danish Ministry of Science, Technology and Development, with one-third each. We would like to give special thanks to Grundfos for allowing us unrestricted access to information about the case in question. We would also like to send special thanks to all the interviewees and workshop participants who kindly shared their points of view and information about the case. Finally, we acknowledge the reviewers for constructive comments on our manuscript.

Author information

Authors and affiliations.

Department of Management Engineering, Technical University of Denmark, Produktionstorvet, Building 424, Kgs. Lyngby, 2800, Denmark

Department of Mechanical Engineering,, Technical University of Denmark, Produktionstorvet 426, Kgs. Lyngby, 2800, Denmark

Claus Thorp Hansen

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to Liv Gish .

Rights and permissions

Reprints and permissions

About this article

Gish, L., Hansen, C.T. A socio-technical analysis of work with ideas in NPD: an industrial case study. Res Eng Design 24 , 411–427 (2013). https://doi.org/10.1007/s00163-013-0159-z

Download citation

Received : 24 November 2011

Revised : 03 July 2013

Accepted : 08 July 2013

Published : 24 July 2013

Issue Date : October 2013

DOI : https://doi.org/10.1007/s00163-013-0159-z

Share this article

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

  • New product development
  • Early phases
  • Socio-technical analysis
  • Find a journal
  • Publish with us
  • Track your research

Issue Cover

  • Previous Article
  • Next Article

1 Introduction

2 literature review, 3 model world design case study, 4 evaluating representativeness of the engine design model world, 5 conclusions, acknowledgment, conflict of interest, designing representative model worlds to study socio-technical phenomena: a case study of communication patterns in engineering systems design.

Contributed by the Design Theory and Methodology Committee of ASME for publication in the J ournal of M echanical Design .

  • Split-Screen
  • Article contents
  • Figures & tables
  • Supplementary Data
  • Peer Review
  • Open the PDF for in another window
  • Cite Icon Cite
  • Permissions
  • Search Site

Chaudhari, A. M., Gralla, E. L., Szajnfarber, Z., Grogan, P. T., and Panchal, J. H. (September 25, 2020). "Designing Representative Model Worlds to Study Socio-Technical Phenomena: A Case Study of Communication Patterns in Engineering Systems Design." ASME. J. Mech. Des . December 2020; 142(12): 121403. https://doi.org/10.1115/1.4048295

Download citation file:

  • Ris (Zotero)
  • Reference Manager

The engineering of complex systems, such as aircraft and spacecraft, involves large number of individuals within multiple organizations spanning multiple years. Since it is challenging to perform empirical studies directly on real organizations at scale, some researchers in systems engineering and design have begun relying on abstracted model worlds that aim to be representative of the reference socio-technical system, but only preserve some aspects of it. However, there is a lack of corresponding knowledge on how to design representative model worlds for socio-technical research. Our objective is to create such knowledge through a reflective case study of the development of a model world. This “inner” study examines how two factors influence interdisciplinary communication during a concurrent design process. The reference real world system is a mission design laboratory (MDL) at NASA, and the model world is a simplified engine design problem in an undergraduate classroom environment. Our analysis focuses on the thought process followed, the key model world design decisions made, and a critical assessment of the extent to which communication phenomena in the model world (engine experiment) are representative of the real world (NASA’s MDL). We find that the engine experiment preserves some but not all of the communication patterns of interest, and we present case-specific lessons learned for achieving and increasing representativeness in this type of study. More generally, we find that representativeness depends not on matching subjects, tasks, and context separately, but rather on the behavior that emerges from the interplay of these three dimensions.

Design and development of large-scale complex systems, such as aircraft and spacecraft, takes multiple years and involves large numbers of individuals who may be spread across different organizations, each geographically and culturally distinct. Scientific understanding of such design and development projects requires an integrated view of not only the technical aspects of the system being designed but also the social phenomena within and across organizations and the interdependencies between social and technical aspects. However, studying these phenomena in real organizations is limited by both empirical and practical challenges such as difficulties in gaining access to observe or collect data, presence of large numbers of geographically distributed individuals, and projects spanning multiple years or decades.

Therefore, researchers in systems engineering and design have started to rely on abstracted model worlds that aim to be representative of the reference socio-technical system being studied, but only preserve some aspects of it [ 1 ]. Such model worlds enable more tractable observation and data collection as well as, potentially, experimental manipulation. Model worlds may involve theoretical models [ 2 ] or simulated environments in laboratory settings [ 3 – 5 ] or field settings [ 6 – 8 ]. The realism, fidelity, and level of abstraction differ significantly across these model worlds, but they all aim to preserve particular aspects of the reference systems they represent—the aspects relevant to answering the research question of interest. In other words, they aim to be representative of the reference system with respect to the research question.

While there is increasing acceptance of the importance of carefully selecting or designing representative model worlds, so far, there is a dearth of specific guidance on rigorous implementation of the concept. Creating a model world involves a series of research design choices about which features of the real world to include, and which to abstract. For example, a popular airplane manufacturing simulation uses Lego blocks instead of airplane subsystems to drastically simplify the process while maintaining a tactile interaction that is known to ground behavior [ 9 ]. However, while each of these choices can have a strong impact on one’s ability to generalize outcomes, there is no wide agreement on standards for assessing whether an abstracted model world is sufficiently representative of the reference real-world setting to answer a given research question.

To that end, our motivating question for this paper is: How should we design representative model worlds for socio-technical research in systems engineering and design? The specific objective of this paper is to begin to answer this question by illustrating relevant issues through a “case study” of model world design: we design a representative model world for a research purpose and compare it to a real-world reference system, in order to (1) elaborate on and illustrate the large number of research design choices that go into generating a representative model, and (2) synthesize lessons learned about the consequences of these choices. In our case study, the model world was designed to answer a separate, “inner” research question: to study how two factors influence interdisciplinary communication in the design process. This “inner” research question is described further in Sec. 3.1 , since it is only a means to the end of exploring how to design representative model worlds. We see this paper as a step on the path toward more formal approaches to adopting and/or evaluating representativeness, such as dimensionless scaling relationships that relate real-world problems to representative model worlds.

This paper is organized as follows. Section 2 discusses the concepts of model worlds and representativeness, then reviews related work in the engineering design literature. Next, Sec. 3 describes the case study in which we design a model world to answer a research question about a real-world reference system. We introduce the research question, describe the real-world reference setting and the features we wish to preserve in the model world, then discuss the thought process in designing the model world experimental setting. Section 4 utilizes data collected in both the model world and real-world settings to compare the phenomena we attempted to preserve, discusses the generalizability of model world results to the real-world setting, and collects insights and “lessons learned” for designing this and future model worlds. Finally, Sec. 5 concludes the paper.

This section first introduces the concept of a model world in engineering research. Next, a discussion of model world representativeness compares similar concepts such as generalizability and validity in other fields. Finally, a review of related work in the engineering design literature focuses on design task similarity and subject expertise to assess model world representativeness.

2.1 Model Worlds in Engineering Research.

Engineering research traditionally relies on physical models as a simplified setting with reduced spatial or temporal scales where engineers and researchers perform experiments. For example, architectural scale models visualize aesthetics of new building concepts, model engines help understand thermodynamic cycles in a classroom, and wind tunnels analyze aerodynamic lift and drag at reduced scales based on equivalence in Reynold’s number. Advances in information technology enable computational model worlds where mathematical expressions of material properties, system state, and dynamics permit numerical simulation and analysis of system attributes and behaviors. Unbounded by the physical domain, modeling and simulation applications seek to represent real-world behavior as computable expressions.

Research in engineering design and systems engineering goes beyond engineering artifacts to also consider the social and organizational elements influencing their design. In these settings, more details about the human participants, their responsibilities, and the surrounding environment are critical elements. Broadening the concept of a model to include these contextual factors leads to our characterization of a “model world.” The term groups all research settings that aim to proxy a real world setting, including both traditional models and lab experiments.

the subject is the individual, organization, or artifact under study,

the task is the research-relevant function or behavior performed by the subject, and

the context consists of the environment in which subjects interact with tasks; it may include access to information, incentive structure, environmental conditions, etc.

While these elements are often considered separately, we demonstrate in this paper how the interplay among them is more critical in generating useful (representative) model worlds. This interdependency can describe any number of physical relationships (between product components), social interactions, or causal relationships in a model. Later, we use communication patterns to compare two settings in systems design.

While this definition is most closely tied to design research where the subjects are human designers, the task is design work, and the context is the design environment, it can also be applied to other topics such as aerodynamics (subject: airfoil, task: provide lift, context: fluid flow) and optimization (subject: algorithm, task: minimize mass, context: input parameters).

2.2 Model World Representativeness.

Statistical representativeness describes whether inferences made about characteristics of a sample can be transferred to the population from which it was drawn [ 10 ].

Analytical generalizability describes whether theory developed by studying an empirical instance can be applied to other similar instances [ 11 ].

Ecological validity describes whether a subject’s response to stimuli in an artificial setting such as an experimental laboratory corresponds to their behavior in the associated real world setting (i.e., their ecology) [ 12 ].

These examples all consider the relationship between observations in a limiting environment (e.g., statistical sample, case study, or experiment) and some other, target environment (e.g., real world).

Enabled by the emergence of computational models, the field of modeling and simulation formalizes representativeness as model validity, defined as “substantiation that a model within its domain of applicability possesses a satisfactory range of accuracy consistent with the intended application of the model” [ 13 ]. Establishing relationships between different system representations (e.g., between a source system and a model world) relies on a morphism mathematical relationship [ 14 ]. The simplest form of replicative validity assesses the input/output correspondence between source system and generative model. Stronger forms of predictive and structural validity add additional requirements of state and component state correspondence to predict and explain previously unseen system behavior. Model worlds with human actors (described as games or gaming simulations) demand a different type of validation compared to computational model worlds. Raser distills game validity into four criteria: psychological realism for players, structural isomorphism, process isomorphism, and outcome correspondence, with weightings dependent on each application [ 15 ].

Whether the model world is adopted or designed, it is defined by the interplay between subject and task in a context where at least one aspect is abstracted from the real world. Evaluation of model world representativeness must consider how selection of the subject(s), task(s), and context influence outcomes with respect to the research objective. Some research objectives may permit substantial abstraction (e.g., using computational model worlds) while others may not.

We define model world representativeness as a measure of whether, and the extent to which, observations made in a model world can be used to answer a specific research question about the real world . This definition encompasses many of the other perspectives on representativeness but focuses only on the model world itself and not on other elements of research design such as sampling, replication, and randomized controls. Additionally, this definition does not seek absolute correspondence between the model world and the real world—selected abstractions specific to the research question may allow comparisons only at a theoretical level.

2.3 Representativeness in Engineering Design Literature.

Assessing the representativeness of research settings is fundamental to produce generalizable insights and a necessary precursor to build on each other’s research. Within the engineering design literature, a number of efforts evaluate how well the research studies real design scenarios. Specifically, the studies can be classified into two types: those that focus on the representativeness and similarity of design problems/tasks, and those that evaluate the differences between experts and novices. We discuss the design literature in these two areas in the following subsections.

2.3.1 Studies Focused on Design Problem/Task.

It is well accepted in the design community that the nature of the problem plays a significant role in various design phenomena. One way to characterize the space of design problems is to identify dimensions of similarity. Anandan et al. [ 16 ] argue that the similarity of design problems depends on their (i) product design specifications, (ii) constraints (cost, time, weight, safety), (iii) function structures, (iv) concepts, (v) CAD models, (vi) shapes, and (vii) manufacturing process plans. Durand et al. [ 17 ] identify the following structural features of design problems that should be considered while designing problems for design research: problem size, functional coupling, participants’ familiarity with the design problem/solutions and underlying principles, nature of solution space: size and constraints, effort required to solve the problem, domain of design problem and degree to which analogous solutions can be retrieved.

Kumar et al. [ 18 ] compare 55 design problems from the literature based on their representation using protocol analysis and latent semantic analysis. They characterize the problems based on five structural elements: goals of the problem, functional requirements, non-functional requirements, information about end user, and reference to an existing product. Sosa [ 19 ] characterize design tasks for experimental creativity research (idea generation and conceptual design) using the metrics of semantic score, lexical ambiguity, precedent analysis, and readability. Some studies have used experiments to compare different design problems. Levy et al. [ 20 ] experimentally compare four design problems (peanut shelling, corn husking, coconut harvesting, and personal alarm clock) for equivalence in design ideation tasks. Problem comparison uses the ideation metrics of quantity, quality, novelty, variety, and completeness.

2.3.2 Studies Focused on Expertise.

One of the key aspects of representativeness of model worlds relates to ensuring that the behaviors of the individuals (engineers) in the real world match the behaviors of the subjects (typically, students) in the model world. It is commonly argued that students behave differently than experts. There is extensive literature on expertise in general, and design expertise in particular. Cross [ 21 ] presents a detailed review of the literature on the differences between experts and novices. Experts and novices differ in their problem solving strategies. Novices adopt a “depth-first” approach to problem solving, whereas experts adopt a top-down and breadth-first approach. Experts are able to mentally to stand back from the specifics of the problem and draw upon abstract knowledge in their domain of expertise. Experts can recognize underlying principles, rather than focusing on the surface features of problems. Based on the study of two expert designers, Cross and Cross [ 22 ] conclude that design experts take a systemic view of the design situation, choose to frame their view of the problem in a challenging way, and draw upon first principles to guide both their overall concept and detailed design.

Through a comparison of expert and novice designers, Ahmed et al. [ 23 ] conclude that novice designers tend to reason backwards and to use a deductive approach, whereas experienced designers tend to reason forward and, when solving more complex problems, alternate between forward and backward reasoning. Novice designers rely on trial and error while experienced designers use particular design strategies. The differences between experts and novices are also studied in the literature on design education that compares students with experts. Crismond and Adams [ 24 ] highlight that students differ from informed designers (experts) in how they understand the design task, build knowledge around the problem, generate ideas, represent ideas, weigh options and make decisions, conduct experiments, troubleshoot problems, revise/iterate, and reflect on the design process. Atman et al. [ 25 ] concluded that experts spend more time than novices (students) in product realization activities, which encompasses decision-making and communication.

Cash et al. [ 26 ] compare the behaviors of subjects in laboratory experiments with design practitioners for three core design activities: information seeking, ideation, and design review. They find that many of the observed behaviors are common across design contexts (experiments versus practice) and populations. They found differences in the frequency and time spent on debating and communication. As in the study by Atman et al. [ 25 ], the practitioners had a greater frequency of communication and spent more time per instance than in the laboratory settings. They attributed the differences to the embedded nature of practice within a pre-existing design process and the associated importance of communication.

While much of the existing literature is focused on independent analysis of the design problem or the expertise, few studies analyze the differences between model worlds and the real design settings that they are intended to represent. There is a lack of studies that systematically analyze the process of designing model worlds for specific phenomena under investigation. There is a need to consolidate insights from these previous studies and to investigate additional (potential) dimensions of representativeness and the interplay among these dimensions, through empirical comparisons between model worlds and real worlds. This paper takes a step on that path.

To explore the various possible dimensions of representativeness, this paper uses a case study in which a model world was designed to represent a real world setting to advance a study objective. Data were collected in both settings to explore model world representativeness. Section 3.1 describes the “inner” research question that motivates the design of the model world. Section 3.2 discusses the reference system—a NASA concurrent design facility. A model world created to represent the reference system is presented in Sec. 3.3 , along with the thought process followed in creating the model world.

3.1 The “Inner” Research Question.

Recall from Sec. 1 that the objective of this paper is to learn from a “case study” in which a model world was designed and compared to a real-world reference system, in order to (1) elaborate on the research design choices involved in generating a representative model world and (2) to synthesize lessons learned. Within this case study, the model world was designed to answer a separate, “inner” research question: How do two factors influence interdisciplinary communication during the design process: (i) whether designers can search for solutions using catalogs with a limited selection of options (a discrete design space) or using a simulation tool (a continuous design space) and (ii) whether designers have access to global information about the status of the design, such as the design variables determined by other designers, through a shared parameter database. Thus, this paper is a “study-within-a-study”: a study of communication in design is used to explore how to design representative model worlds.

The motivation for this “inner” research question is to support design organizations (such as NASA, in this case) to determine the appropriate information technology infrastructure and communication pathways to use in their design processes, e.g., whether to develop catalogs or continuous search tools for designers, and how frequently designers should receive design variable updates. Communication is an important aspect of systems engineering and design [ 27 ]. Good communication in an organization is essential to address technical dependencies, ensure good design outcomes and promote efficient design and systems engineering processes. Poor communication is a major cause of failure in design processes. Communication is particularly important in large scale complex systems design such as automobiles, aircraft, and spacecraft.

We would ideally perform studies within the real organization. However, there are many barriers to conducting such studies, including the size of the organization, the long duration of design projects, and the costs involved. Collecting data during real projects are also restricted. Since the real-world setting often does not allow us to control the environment and collect fine-grained data, we need to rely on other settings such as model worlds. Establishing representativeness of the model world with respect to certain communication aspects would help advance research because the model world affords more granular data collection over time including repeated studies in quasi-experimental research designs [ 28 , 29 ].

3.2 Real World Reference System: NASA’s Concurrent Design Facility.

The real-world reference setting is a conceptual spacecraft design process at the NASA Goddard Space Flight Center’s concurrent design facility, the Mission Design Lab (MDL). The MDL develops high-level conceptual designs for spacecraft, relying on a team of disciplinary experts co-located in a single room who focus on this task for one entire week. Figure 1 shows the layout of the room and where each disciplinary expert sits. Using the “classic” categories of subject, task, and context, the setting is described in detail in the following paragraphs. These descriptions focus on those aspects of the subjects, task, and context whose interplay may influence communication during the design process, such as the subjects’ expertise, the nature of the task, the formal and informal organizational processes and structure, the available communication channels, the incentives, and the tools that facilitate communication and design.

The seating chart of disciplines in NASA’s mission design lab

The seating chart of disciplines in NASA’s mission design lab

3.2.1 Subjects.

The MDL is staffed by NASA engineers who are experts in their disciplines. They occasionally participate in MDL studies in addition to their regular jobs at Goddard. MDL engineers also have experience in the MDL concurrent engineering process. Typically, when an engineer starts working with the MDL, there is an “apprenticeship” period in which they train with an experienced MDL team member from their discipline. As a result, they understand how the MDL team works together, what is expected to happen when, and how to use the tools provided in MDL.

3.2.2 Task.

The task is to produce a conceptual design for a spacecraft, captured in a report consisting of detailed slides, an oral presentation, and other documentation. A conceptual design consists of a feasible “point design” for a spacecraft with most of the key engineering parameters determined and costed (Concept Maturity Level 4, according to Ref. [ 30 ]). The objective, not explicitly stated mathematically, typically consists of finding a feasible design that meets the engineering and science requirements and a cost cap; implicitly, mass is minimized as a proxy for cost.

The design task is accomplished by a team of 12–20 experts in typical aerospace disciplines such as mechanical systems, thermal systems, power, propulsion, avionics, attitude control, command and data handling, and systems engineering. A disciplinary expert is responsible for the design of a corresponding subsystem on the spacecraft (and some disciplinary experts have integrative roles). The design of each subsystem is dependent on that of others through shared design variables. Some subsystems are tightly coupled and others are loosely or not at all coupled. The MDL completes the task in four work days of approximately eight hours per day. Some decisions must be made relatively quickly to complete a feasible design in available time. Some engineers experience time pressure to solve problems rapidly, but there is typically enough time to finish a design.

3.2.3 Context.

There is limited organizational structure in the MDL. A team leader runs the full-team meetings and sets expectations. Two systems engineers help to ensure needed side meetings happen, important problems and trades are surfaced and solved, and the design closes on time. There is, however, a clear process. Days begin with a briefing on progress and key problems. A similar briefing is held just after lunch. There are clear expectations about what kind of work will happen when during the study and how communication will occur.

Communication is typically face-to-face. All disciplinary experts may easily communicate in person, since they are in the same room and typically seated near others with whom they usually need to speak (see Fig. 1 ). Because they are co-located, the barrier to communication is low. In addition to face-to-face communication, a spreadsheet maintains key design parameters, but it is not typically used to communicate design parameters. Experts may also phone or e-mail others in their disciplines outside the room for assistance. Engineers have shared expectations of many goals and constraints that are not explicitly described, such as acceptable amounts of risk in the design, what constitutes a reasonable mass and how it translates into cost.

The incentives for each individual are varied, and have not been specifically measured. Success in design closure is indirectly relevant to career advancement. Engineers are motivated by the work itself and by learning from each other. They are also, of course, paid a fixed salary for their jobs, of which the MDL is a part. Engineers use a combination of catalogs and simulations to inform design decisions. Some disciplines have specific tools that search and optimize over the solution space, but many also pick from a set of existing and proven designs.

3.2.4 Data Collection.

The collected data include time stamped observational records of interdisciplinary communication over the course of the week-long study: who communicated with whom and for how long. The NASA engineers sometimes talk in groups of three or more, but, for comparison with the model world data, we consider that a group discussion consists of one-on-one interactions among those involved. Such data are collected from seven different studies, all of which took place in the NASA MDL.

Part of the motivation for designing a model world to represent this real-world context is that there are many kinds of data that cannot be collected in this setting. Because MDL studies are proprietary, there are restrictions on sharing the final designs. It is not feasible to capture all the potential design solutions explored by designers (such as what they thought about, did back-of-envelope calculations for, or tested in models) nor the exact content of all of their communications. Finally, it is not possible to “experiment” with alternative tools and processes. Model worlds, in which such data can be collected and experiments run, can potentially support deeper analyses of certain types of questions, if they can be made representative of the real-world setting.

3.2.5 Real World Features to Preserve.

It is important to reflect on which aspects of the real world setting need to be preserved in a representative model world to answer the research question on interdisciplinary communication in design. There are many factors that affect communication in design teams, including technical dependencies among subsystems (which drives what design information needs to be passed among team members), the formal organizational structure (which creates the explicit channels and processes for communicating among individual roles), and the informal cultural norms (which drive expected behavior, such as preferences among which sources of information to trust or what constitutes a good enough solution) [ 31 , 32 ]. In reflecting on the way these concepts are embodied at the MDL, the following key areas emerged as crucial to preserve in the model world because they are expected to drive communication.

1. Technical interdependencies in the design problem: Greater technical dependence among subsystems requires greater coordination between the teams working on those subsystems [ 33 ]. Therefore, strongly coupled systems are more likely to have greater communication links, compared to weakly coupled subsystems. In the MDL, the patterns of communication appear to be strongly driven by the technical dependencies in the physical system (i.e., the spacecraft), and we expect limited communication among disciplines that do not share any technical parameters. Moreover, some interdependencies are particularly challenging to resolve; these drive a significant portion of the communication. Therefore, the first important feature of the model world is a design task for a technical system with complex interdependencies, where some are more important than others.

2. Organization structure, process, and knowledge and expertise: Generally, the organization structure and processes, both formal and informal, drive communication across disciplines. In the MDL, however, since all subject matter experts (SMEs) are collocated in the same physical space and explicitly encouraged to work collaboratively and quickly, there is limited formal organizational structure. Instead, the process is driven by (1) clearly defined disciplinary roles and (2) an informal but well understood expectation of the design process. Regarding (1), for example, there is a defined facilitator role (similar to a project manager) who ensures that the values of key design parameters are communicated to the other disciplines who need them by walking around the room and stimulating pairwise exchanges. Regarding (2), there is a shared expectation for how the design process should go, based on thousands of past MDL studies. While each new spacecraft has unique features, experts report that there is a “rhythm to the room” that is common across studies. By this they mean that most of the design decisions are consistent, even though the specific values are study-specific. For example, the communication engineer always needs to make a link budget which depends on data needs of the science instrument and opportunities for downlink coverage with ground stations. Both of these are driven by orbit choices and by mission operations timelines. All of this necessitates design trades and associated discussion, but those discussions usually happen around the same time in the study and among the same people. While some studies involve unique features (e.g., multiple components flying in formation), these new discussions represent a small fraction of the overall interdisciplinary communication in the study.

In short, rather than formal organizational structures and processes, communication in the MDL is driven more by the knowledge and expertise of its members. Teams of experts coordinate their behavior through shared mental models [ 34 , 35 ]. As a team, they are aware of each others’ role on the team, knowledge, skills, and goals. Experts have a better understanding about which other subsystems they should communicate with to improve performance or to resolve conflicts, as compared to novices. In the MDL in particular, since all the MDL participants are SMEs, we would not necessarily expect to see equal communication among all disciplines that share design parameters. The experts are expected to know which technical dependencies are more important, and hence, focus their communication on the corresponding dependencies. Part of why the MDL is efficient is that there is a strong expectation for which design trades are most salient and will drive future decisions; therefore, we expect the SMEs to spend most of their communicating time on the subset of technical dependencies that are most important to the final system. To embody these characteristics, the subjects in the model world should have insight into their discipline-specific design spaces, and should know which trades with other disciplines are likely to be most important.

In addition, the subjects must be incentivized toward a collaborative goal, so that they are motivated to identify and analyze these tradeoffs with one another .

3. Nature of the design process: A third aspect of the MDL that should be preserved is the design search behavior of the SMEs. Given the short timeline, the MDL culture tends to focus on “good enough” solutions. Their primary goal is to assess whether the design constraints are feasible, not to come up with a final design for the spacecraft. As a result, most of the disciplines work from catalogues from trusted manufacturers, rather than doing de novo design at the component level. For example, if the propulsion engineer is asked by a trajectory analyst whether a particular burn profile is feasible, (s)he will not start designing from first principles; instead, (s)he might answer with the operating limits from a known manufacturer’s design.

Although the MDL has a long history of efficiently delivering spacecraft designs in a week, its success has led to requests for more varied kinds of spacecraft, and several have questioned the scope of applicability of the process. It is efficient precisely because it does not search the whole design space, but how much efficiency is good? What if valuable designs are missed because the experts in the room are not talking as they should. Part of the motivation for a representative model world is to be able to probe these what-ifs, outside of the professional setting. The model world should enable us to explore the “counterfactual” of continuous/optimizing rather than discrete/good-enough search of the design space, which we cannot observe nor experiment with directly in the professional setting.

In summary, the model world for studying communication in concurrent design settings should consist of a reasonably complex task with dependencies among disciplines, subjects with relevant knowledge of respective disciplines and shared mental models, a collaborative context where subjects can interact and exchange information, and incentives to collaborate and make design tradeoffs. Finally, it should enable exploration of “counterfactual” scenarios that cannot be observed nor experimented with directly in the professional setting, such as continuous design search rather than choosing among off-the-shelf solutions.

3.3 Designing the Model World: An Abstract Engine Design Setting

3.3.1 key decisions in designing the model world..

In designing this model world, our first decision was selecting the subject pool. This decision was driven in part by the opportunity to recruit students from a homogeneous subject pool from a large-enrollment design course. One of the authors was teaching a junior-level machine design course with an enrollment of over 200 students across multiple sections. Recruiting the students from the course allowed a reasonable sample size across different treatments.

Once the subject pool was fixed, the next step was to design the task. The factors that were considered included the subjects’ knowledge of the domain, the need for multiple subsystems/components and technical dependence among them, and a reasonable completion time. We also aimed to replicate key features of the real world reference system in the design problem, including multiple objectives, time pressure, discipline interdependence, and the need for collaboration.

Considering desired factors and the context of the machine design course, we designed an abstract engine design problem with five components—connecting rod, crankshaft, piston, piston pin, and flywheel. The students were familiar with these components through the course and the scientific knowledge (e.g., stress analysis, fatigue analysis) required to design these components. The task was framed as a parametric design problem of the five engine components with the objective to maximize the overall factor of safety and to minimize the overall mass of the engine. At a higher level, they had a shared understanding of the role of the components in the engine, and some understanding of how their design decisions affect, and are affected by, other components.

Real engine design is a complex design activity. However, we decided to keep the engine design problem significantly simpler than the spacecraft design problem in MDL, with a smaller number of subsystems and only a few variables in each subsystem. This decision was mainly driven by the need to complete the activity in a shorter time frame of around 30 min instead of multiple days. (The short timeframe fit the requirements of the design course in which the study was run, and also offered us the opportunity to evaluate whether a much-shorter task could still be representative of the real world.) The component-level simulations were also simpler than the sub-system design activities in the MDL study, which allowed us to conduct multiple replications of the experiment. Despite the simplicity, the engine design preserved the characteristics of dependencies between subsystems which required the subjects to communicate with each other to complete the task.

In contrast to domain experts at MDL, however, the students do not have experience with the specific design problem. Therefore, the shared mental model of the design team is expected to be different from the shared mental model of the experts at MDL, resulting in differences in communication patterns between MDL and the model world. Studies on team effectiveness show a strong reciprocal linkage between team cognition and behaviors such as communication. According to Kozlowski and Ilgen [ 36 ], “repeated interactions among individuals that constitute processes tend to regularize, such that shared structures and emergent states crystallize and then serve to guide subsequent process interactions.” Hence, the lack of shared mental model is expected to diminish as the students start interacting with each other within the experiment. For example, at the start of the experiment, the students know which design variables from other components influence their design, but do not know the extent of influence and relative importance of the variables. This knowledge is gained over time as they perform the design task. Therefore, we expect students’ behavior to better represent experts in the latter part of the experiment.

To represent discipline-specific knowledge, each student was seated at a different workstation and given different information as seen from the room layout in Fig. 2 . The extent of their knowledge about their own subsystem and that of the others depended on the information and tools provided, how they used those tools to explore the design space, and how they used the available communication channels to explore the design space collaboratively. (The details of these communication channels and search tools are provided later.) Finally, there was an incentive for the team to perform well on this task (also detailed below). This setup was intended to give the students discipline-specific expertise and to motivate them to communicate with teammates to discover and make important trade-offs.

The disciplines in the engine experiment were randomly spread across the experiment room

The disciplines in the engine experiment were randomly spread across the experiment room

Not all aspects of the experiment were designed to represent behavior in the MDL. We randomly assigned individuals to components and teams, instead of having individuals self-select components and form their own teams. The rationale for this decision was to control for any individual-specific effects, any effects of prior collaborations, personality effects, and acquaintances on their collaboration. While communication in real world design organizations (including the MDL) is affected by these factors, our intent was neither to study these factors, nor was it feasible to identify and replicate such influences within the experiment.

Although MDL engineers can communicate in smaller and larger groups, we decided to limit the communications in the model world to pairwise (one-to-one) and a limited “broadcast” where subjects can push design information and other disciplines can pull the current information. This was partially because of the small team size in the model world (5 in engine design versus 18–20 in MDL) and partially because allowing multi-person chat conversations may lead teams to create one single team-wide chat, which would not be able to distinguish who was really talking to whom, as in the MDL data set. On a related note, although in the MDL the SMEs can communicate verbally and use various visual aids, the students were setup at different screens and could only communicate through the user interface (either one-on-one chat or the broadcast function).

The experiment did not include a formal coordinating role. The MDL has a facilitator and one or two systems engineers, whose roles involve identifying key trades among disciplines and ensuring that the respective SMEs communicate and resolve those trades. Given the small size of the engine design task and teams, we believed such a role would be unnecessary to enable coordination.

Finally, to address the research objective presented in Sec. 3.1 , we designed a factorial experiment with two variables: (i) the nature of the design solution space and (ii) the availability of global information. We also designed the user interface to implement different experimental treatments. Since the experiment was intended to enable exploration of the “counterfactual” cases of continuous/optimizing rather than discrete catalog-based search of the design space, we allowed the first factor to vary along two levels: local search of design space using just simulation (label S) and local search of design space using just catalogs (label C). The experiment was also intended to explore the importance of a central data-sharing repository such as the parameter database. Hence, we allowed the second factor to vary along two levels: where the parameter database is absent (label P 0 ) and where it is present (label P 1 )

Further details of the subjects, design task, and the organizational context are discussed in the following sections. These details are also summarized in Table 1 .

A comparison of the model world and the reference world

3.3.2 Subjects.

The subjects in the model world are undergraduate mechanical engineering students from a required junior-level machine design class. The student subjects have the knowledge of relevant fundamentals of mechanisms, machine design, and failure analysis. They do not have experience in working on real engine design problems, but there is familiarity with component-level parametric design problems similar to the abstract engine design problem used in the model world.

3.3.3 Task.

The engine design task is a parametric design problem that requires specification of the geometry parameters of an internal combustion (IC) engine to optimize the mass and factor of safety. At the discipline level (i.e., the component level), the objective is to minimize the subsystem component’s mass and maximize its factor of safety. The system-level objectives are to maximize the overall factor-of-safety (which is the minimum of the subsystem-level factors of safety) and minimize the total mass of all engine components. The students remain seated during the entire duration to avoid disturbing other teams in the same room. Each team consists of five disciplines corresponding to the five engine components: piston, crankshaft, flywheel, connecting rod, and piston pin. There is no central discipline (e.g., system engineering) to manage global requirements and team interactions. Instead, a broadcast screen displays current values of the system objectives (total mass and factor-of-safety). In the P 1 conditions, it also displays the key design variables.

Different disciplines are coupled through the shared variables similar to the reference setting. Therefore, we define the design interdependence between a discipline pair as the number of variables shared. There are discipline pairs with strong coupling (e.g., flywheel-crankshaft) as well as discipline pairs with weak coupling (e.g., piston pin-connecting rod). Each discipline’s subsystem outputs, such as subsystem mass and subsystem factor-of-safety, may depend on the design parameters of other subsystems. For example, the crankshaft discipline requires the piston bore diameter, which is controlled by the piston discipline, to determine the maximum gas piston force on the crankshaft pin and estimate its factor-of-safety. The shared variables between any discipline and “broadcast,” the broadcast screen that displays system objectives and key design variables, include the system-level mass, system factor-of-safety, discipline-specific design parameters, and design parameters shared by the discipline with other disciplines. All pairwise design interdependencies are summarized in the design structure matrix (DSM) shown in Fig. 3(b) , and compared with that of the real world reference system (Fig. 3(a) ). The “broadcast” element, a counterpart of “systems engineering” and “customer” from the NASA MDL, is supposed to be an integrative agent that facilitate information exchange [ 37 ], and therefore, it is included in the engine DSM. While the two matrices are not identical, they both show variability in the strength of coupling across various disciplines pairs.

Design structure matrix for (a) the spacecraft design problem and (b) the engine design problem. Pairwise design interdependencies, i.e., the number of shared design variables, are shown in the matrix cells.

Design structure matrix for ( a ) the spacecraft design problem and ( b ) the engine design problem. Pairwise design interdependencies, i.e., the number of shared design variables, are shown in the matrix cells.

An exact match between the problems is not critical for representativeness; rather, what matters is whether the interplay among this design task, the students’ expertise, and the design problem context generates representative patterns of communication. As in the real world reference system, subjects need to search the design space and work in a team to learn the technical dependencies in the engine design problem. Design collaboration is essential because the subjects need to work together to maximize the global objectives. Whether this achieves representativeness is evaluated later, in Sec. 4 .

3.3.4 Context.

With the task and subject pool determined, the context was designed to balance the constraints of a classroom setting with the aim of replicating important contextual features of the real world reference system. Unlike in the real world reference system, the organizational structure is entirely decentralized. There is no systems engineer—only a broadcast that displays current system objective values and, in condition P 1 , design variables. The broadcast feature enabled the team to keep in mind the global goals and infer tradeoffs among disciplines, as a systems engineer might do.

The following are specifics of how we implemented simulation modules and catalogs in the experiment. When using simulations for evaluations of the mass and factor of safety, the subjects feed design parameters to the discipline-specific models. Simulations do not have an associated cost, which implies the subjects can run as many simulations as they like. Evaluations from past simulations are available on the screen for reference, as shown in Fig. 4(a) . On the other hand, when using catalogs for design exploration, the pre-evaluated performance is known at discrete points in the design space. The disciplines also have access to a testbed where they can view different catalogs by changing values of the design variables (shared variables as well as those under their own control). Figure 4(b) provides a screenshot of the user interface available while using catalogs. Further, Fig. 5 is a screenshot of broadcast showing design parameter values (only in the P 1 conditions) and global objectives information.

The user interface for two types of design exploration processes: (a) simulation-based and (b) catalog-based

The user interface for two types of design exploration processes: ( a ) simulation-based and ( b ) catalog-based

An example of the global information shown on broadcast, which displays key design variables for condition P1 only

An example of the global information shown on broadcast, which displays key design variables for condition P 1 only

To motivate the students to achieve the system-level objectives, the engine design task included a monetary incentive proportional to the system performance, i.e., total mass and the overall factor-of-safety. Based on the preference over their values, the system outputs were categorized into five quality levels: poor , fair , good , very good , and excellent . In general, total engine assembly mass and factor-of-safety exhibit negative correlation. The system level quality was the minimum of the two quality levels. The poor quality level payed $10 per discipline, whereas the excellent quality level payed $20 per discipline. The payments per discipline for fair, good, and very good levels, respectively, were $12, $15, and $17. We chose equal reward for all subjects on a team because group-level incentives are important for initiating effective collaboration [ 38 ], and, also, this ensures incentive-compatibility by aligning subject behaviors with the task objective.

3.3.5 Data Collection.

In the model world, the collect data include timestamped logs of (i) pairwise text messages and (ii) discipline-wise updates to and from the “broadcast.” Such pairwise interactions are used to represent the amount of communication.

In this section, we evaluate the representativeness of the engine design setting with respect to the NASA MDL by comparing observed team communication patterns. We begin by specifying the quantities used for measuring team communication patterns. Since these quantities depend on the task and the context, it is necessary to state some basic parameters of the two research settings. Such parameters may be different between the settings, but note that the representativeness should be judged based on multidimensional subject-task-context interrelationships and not based on singular dimensions along subject, task, or context. First, the instances of communication between different disciplines are counted as pairwise interactions. A pairwise interaction in the NASA MDL setting represents one instance of face-to-face talk, whereas a pairwise interaction in the engine design setting may represent either one text message, one update or one check of the broadcast element. These measures are not exactly the same, since an MDL conversation likely involves the equivalent of several text messages. However, we do not aim to compare these numbers directly but rather to compare trends and relative changes in the number of interactions, such as changes over time or relative differences in communication across subsystems.

Second, discipline pairs are dichotomized into tightly-coupled and loosely coupled pairs using a threshold number of shared variables. The threshold divides the discipline pairs into two similar-sized groups in either setting. The discipline pairs with the number of shared variables greater than the threshold are tightly-coupled pairs, and the rest of pairs are loosely-paired. The threshold for the engine design task is 0.5 and it is 15.5 for the spacecraft design task. Third, since the time duration of the tasks in the two settings is different, we define unit interval time for each setting so that we can compare communication patterns on the same timeline. The unit interval for the engine design setting is one minute and it is one hour for the NASA MDL setting. These particular unit intervals are selected because (i) they partition the respective task duration into a similar number of time-steps (≈35) in two settings and (ii) the average number of interactions by each discipline (average network degree) within a unit interval is approximately same (between 3 and 5) in two settings.

Technical-Communication Mirroring : This is defined as the correlation between the number of shared variables and the number of interactions between discipline pairs.

Amount of Communication : This is the total number of interactions among all disciplines.

Technical-Communication Mirroring over Time : To observe changes in mirroring over time, we define fractional mirrored interactions , which is the ratio of number of interactions between tightly-coupled discipline pairs and the total number of interactions among all disciplines.

Discipline-Wise Centrality Indices: Two indices are used to quantify network centrality of disciplines with respect to incoming and outgoing interactions. Hub index estimates a discipline’s centrality based on outgoing interactions, whereas authority index that estimates a discipline’s centrality based on incoming links. The estimation procedure for the two indices is given in Ref. [ 39 ] and its implementation is available in python library networkx [ 40 ].

Section 4.1 describes the specific differences and similarities along the aforementioned quantities.

4.1 Comparison of Communication Patterns.

In this subsection, we present the results from quantitative comparison of communication patterns in the reference setting and the model world. Based on the results, we make a qualitative evaluation of which key features of the reference setting are preserved in the model world in Table 2 .

The evaluation of whether key communication-related features are preserved in the model world

4.1.1 Communication Among Subsystem Designers.

A very basic expectation was that the design process would create a need for communication among the subsystem designers. Due to the smaller size of the engine design problem and design team, we expected a smaller amount of communication in the engine case. However, surprisingly, the results suggest that the total number of interactions is similar in the NASA MDL setting and engine experiment. This can be seen from Fig. 6 which plots the empirical estimates for the probability distribution of total pairwise interactions per team. The surprising similarity in the number of interactions is due to the different ways in which interactions are counted in the two settings: at NASA, a whole conversation counts as one interaction, whereas in the engine experiment, several interactions (one-way messages) might make up a conversation. For our purposes, the total number is not important, because we only compare relative trends in the number of interactions. At this point, we simply note that, indeed, the design problem did drive communication among designers.

Estimated distributions of total pairwise interactions per team. Kernel density estimates are found using a statistical data visualization library, seaborn [41]

Estimated distributions of total pairwise interactions per team. Kernel density estimates are found using a statistical data visualization library, seaborn [ 41 ]

4.1.2 Communication Mirroring Technical Dependencies.

We expected to see greater communication among strongly coupled disciplines than among weakly coupled disciplines; in other words, that the strength of coupling would drive the amount of communication between pairs of disciplines. We therefore examine the correlation between the number of shared variables and the number of interactions between discipline pairs, in the engine and NASA MDL settings.

Figure 7 shows the results. As expected, the correlation between the number of shared variables and interactions is large and significant for both the NASA MDL and the engine case, particularly when a catalog is present. Specifically, for the NASA MDL teams, the estimated slope of the linear fit between the number of interactions and the number of shared variables is 1.39 ( R 2 = 0.30) with a two-sided p -value is less than 0.0001, using Wald’s test whose null hypothesis is that the slope is 0 and the degrees-of-freedom (DoF) is 124. For the engine setting when catalogs are present, the C-P 0 and C-P 1 conditions, the slope coefficients are 3.97 ( R 2 = 0.30, DoF = 52, p -value < 0.0001) and 4.14 ( R 2 = 0.33, DoF = 46, p -value < 0.0001), respectively. When a catalog is not present, the S-P 0 and S-P 1 conditions, the correlation is lower (the slope coefficient is smaller). The slope coefficients are 1.91 ( p -value < 0.001, DoF = 82) and 0.42 ( p -value = 0.1, DoF = 52) for the S-P 0 and S-P 1 conditions, respectively. The implications of these results, especially for condition S-P 1 , are explored in Sec. 4.2 .

Mirroring between technical dependence and the amount of communication (a) in the engine design experiment and (b) in the NASA concurrent design facility

Mirroring between technical dependence and the amount of communication ( a ) in the engine design experiment and ( b ) in the NASA concurrent design facility

The implication is that, indeed, the engine design task resulted in some degree of the mirroring seen in the reference NASA setting: communication patterns were correlated with technical dependencies. When catalogs were not present, this effect was diminished, which suggests that catalogs were important in enabling representativeness. This is intuitive because catalogs were intended to help engine designers spend less time searching for subsystem-optimal solutions, perhaps enabling a larger focus on finding good solutions for the entire system by communicating with the other subsystem designers.

4.1.3 Expertise in Resolving Interdependencies.

We expected that the students’ communication patterns should become more like those of the experts in the NASA setting over time, after they have learned about the problem—and in particular, learned which of the technical dependencies are most important in driving the performance of the final system. The NASA experts should know from the start which trades are most important and spend their communication effort on resolving those trades, while students would need to learn this first before their communication patterns would exhibit this characteristic.

To explore this in the data, we examine the “fractional mirrored interactions” (defined earlier). Intuitively, this metric indicates what percentage of the interactions were between tightly coupled disciplines. A value near 1 indicates that nearly all interactions were between tightly coupled pairs, and almost no interactions were between less-coupled pairs.

Figure 8 shows a moving average of this metric over time for ( a ) the engine study and ( b ) the NASA reference setting. From Fig. 8(a) , we observe that the fractional mirrored interactions in the engine setting remains close to 1 for first 4 to 5 time-steps, but it reduces to about 0.7 or 0.8 after 15 time-steps and then remains in this range. On the other hand, the NASA MDL teams start the spacecraft design task in the range of 0.7–0.8, with some slight increases up to about 0.9 toward the end of the time horizon.

Fractional mirrored interactions over time with fifth and 95th percentiles calculated for all discipline pairs

Fractional mirrored interactions over time with fifth and 95th percentiles calculated for all discipline pairs

The possible reason for this behavior is that the information about the design interdependencies that the students receive through the problem statement may drive their communication at an early stage. Indeed, they seem to communicate only about the most important interdependencies, to the exclusion of all else. However, partway through the task, this fraction reduces as low interdependence pairs communicate more, perhaps to explore trades that may help improve the system objectives. On the other hand, the NASA engineers know important interdependencies before the start of the task, but also focus throughout the task on surfacing issues that may require reconvening at the system level or coordinating with less-coupled subsystems. Toward the end, their focus may have narrowed to resolving one or two key trades among tightly coupled subsystems.

The results suggest that, indeed, the students’ communication patterns became more like those of the NASA experts after an initial learning period (the first 10–15 time-steps). After this period, the fractional mirrored interactions were in the same range of 0.7 to 0.8 for most of the study. (The students did not exhibit the same slight rise in fractional mirrored interaction toward the end of the study.)

However, a closer look at the data suggests that a large portion of the students’ mirrored communication was through the broadcast function—which shows the state of the global objectives and, in the P 1 experimental setting, also shows shared design parameters. Figure 9 shows the fraction of mirrored interactions over time excluding integrative disciplines such as broadcast (in the engine experiment) and the systems engineer and customer (in the NASA MDL). While the students’ proportion of interdependency-driven interactions went down significantly to around 0.5, the NASA experts’ proportion remained roughly equal to its value when integrators were included (Fig. 8 ), around 0.7–0.8.

Fractional mirrored interactions over time with fifth and 95th percentiles calculated after excluding integrative disciplines such as systems engineer, customer, and broadcast screen

Fractional mirrored interactions over time with fifth and 95th percentiles calculated after excluding integrative disciplines such as systems engineer, customer, and broadcast screen

These results suggest that the students came to rely upon the broadcast function for managing many of their trades, whereas the experts utilized more pairwise communication. It is possible that this difference in behavior is driven by the smaller size of the problem and the smaller number of pairwise interdependencies in the engine setting and/or by the lack of a designer assigned to an integrative role in the engine setting. Better representativeness might require a more similar problem size and interdependency structure and/or a specific integrative discipline such as systems engineering. Further study is necessary for testing this hypothesis.

4.1.4 Discipline-Specific Communication Patterns.

To further understand how the lack of a designated systems engineer might influence communication patterns in the engine experiment, we examine each discipline’s hub and authority indices [ 39 ]. Hub and authority indices quantify how “central” a discipline is in the communication network based on the number outgoing interactions and the number of incoming interactions, respectively. These indices use pairwise interactions from the entire duration of the tasks.

For NASA, Fig. 10 suggests that the systems engineer is central to driving communication. In the engine experiment, on the other hand, there is no systems engineer. The broadcast function was intended to substitute for some functions of the systems engineer – computing the global objective value and, in the P 1 condition, sharing key design parameters. The results in Fig. 11 show that the broadcast function is particularly important for outgoing links (see hub indices), which represent disciplines pulling updates about shared global objectives and design parameters. However, the broadcast has the lowest importance in the communication network as a receiver of information (indicated by small authority indices). Because our NASA data did not provide the direction of communication, it is not possible to distinguish the systems engineer’s role in outgoing versus incoming communication. However, based on our qualitative understanding from extensive conversations with MDL systems engineers, their role is not the same as that apparently performed by the broadcast function – simply to update designers on objective and parameter values. Therefore, the results suggest that the broadcast function did not fulfill the same integrator role as the systems engineer and that a more representative model world might require a designated systems engineer.

Network centrality for the NASA MDL disciplines. Hub and authority indices [39] are the same for NASA MDL teams because pairwise communication is undirected

Network centrality for the NASA MDL disciplines. Hub and authority indices [ 39 ] are the same for NASA MDL teams because pairwise communication is undirected

Hub indices and authority indices [39] representing the disciplines’ network centrality based on outgoing and incoming links, respectively. These network indices are computed using a network analysis library, networkx [40].

Hub indices and authority indices [ 39 ] representing the disciplines’ network centrality based on outgoing and incoming links, respectively. These network indices are computed using a network analysis library, networkx [ 40 ].

Figures 10 and 11 also provide insights into the communication patterns of the core disciplines excluding systems engineering and broadcast. The plots suggest that the differences in discipline-wise authority and hub indices are statistically insignificant in the engine experiment because of large variance in their values. However at NASA MDL, these differences, albeit small, are statistically significant. The results suggest that the disciplines in the engine experiment contribute equally to the communication over the entire time horizon, and the NASA MDL disciplines are selective about with whom and how much they interact, likely due to their greater knowledge and experience in solving similar problems. Additionally, we do not observe significant differences in the centrality indices because of conditions C, S, P 0 , and P 1 . This suggests that the core disciplines’ relative contributions to the overall communication likely remain unchanged despite changing how they search the design space and whether they can access global status of the design.

4.2 Generalizability of Engine Experimental Results.

Next, we examine the results of the engine experimental treatments—the counterfactuals that could not be tested in the NASA setting. The results provide insights about the effects of the design search method (catalogs versus simulations) and global information availability (presence or absence of a parameter database) on the communication metrics. Subsequently, we consider the extent to which these results may be generalized to the NASA setting based on what was learned in Sec. 4.1 about which aspects of the engine experiment communication patterns matched those of the NASA reference setting—i.e., the engine setting’s representativeness.

4.2.1 Effects of Using Catalogs Versus Simulations.

When catalogs are used, interdisciplinary communication is more frequent (both in total across the task duration and at any given moment) as seen from Fig. 6 . The total number of interactions in the catalog treatments is greater on average than the total number of interactions based on an equal-variance t -test ( t (38) = 6.4, p = 0.001). Also, the correlation between the number of interactions and the number of shared variables is also larger in the catalog treatments than the simulation treatments (see Fig. 7 ). The implications of these results are presented in Sec. 4.2.3 .

4.2.2 Effects of Using Shared Parameter Databases.

Within the engine design study, the global availability of design parameters does not appear to affect the total number of interactions (Figure 6 ). On the other hand, the mirroring results in Fig. 7 show that, among all four experimental conditions, only the S-P 1 fails to exhibit mirroring (correlation between the number of interactions and the number of shared variables is insignificant). It is not entirely clear why this is the case. A possible reason is that more than half of the interactions, there occur through text-based channels, as seen from Fig. 12 . And, as we observed before in Figs. 8 and 9 , the students’ text-based interactions are in general less aligned with technical dependencies, while their interactions through broadcast are more aligned with technical dependencies (which also include many high-interdependence pairings). The condition S-P 1 means that students have easy access to their colleagues’ design parameters but do not have easy access to a set of pre-computed solutions throughout their solution space. Perhaps the latter makes them less able to iterate on their designs and the former makes them less likely to communicate because they already have access to the information required. Further investigation is needed to confirm and to understand this result.

The fraction of total interactions between the core disciplines, equivalently text-based interactions, for the experimental conditions

The fraction of total interactions between the core disciplines, equivalently text-based interactions, for the experimental conditions

4.2.3 Discussion and Implications.

Across both experimental treatments, there is more and better-aligned communication with a catalog than with search within a continuous design space. This is a consequence of having ex-ante evaluations of subsystem outputs available in a catalog. Quicker design exploration when using catalogs may facilitate a better understanding of input-output relationships, of how disciplines behave in response to the changes from shared variables, and of which designers need to coordinate with one another. Also, a catalog provides flexibility in broadcasting any design point from a large design space, thus encouraging more interactions with the broadcast and with other designers. If we believe these findings are representative of NASA, they imply that design organizations such as NASA should consider maintaining or enabling catalog-based search, since it appears to lead to more and better-aligned communication. It remains to be seen, however, whether this translates into superior performance. On the other hand, global availability of design parameters does not, surprisingly, appear to improve communication patterns, suggesting that investment in IT backbones may need to focus on information exchange beyond just sharing design parameters, particularly in concurrent design settings with low barriers to interpersonal communication.

The key question, then, is whether we believe these findings are representative of NASA. The major differences found in the analysis in Sec. 4.1 were around the key role of the systems engineer and the extent to which mirroring occurred through the broadcast function in the engine setting. It is possible that a skilled systems engineer could, in a continuous-search environment, compensate for the lower and less-aligned communication by prompting the right people to talk to one another, but the apparent advantages of catalog-based search would still be relevant, and therefore we consider it likely that catalogs would still prompt better communication patterns in the NASA setting. On the other hand, it is less clear why a shared parameter database made little difference in communication, so we recommend further investigation before applying these latter findings in practice.

The results presented in this paper are limited by the measurement choices made in the two settings. Specifically, we chose a 1-minute time-step in the engine design experiment and an hour in the NASA MDL setting. The unit of communication in the engine study is a chat message or a broadcast message, whereas in the NASA MDL setting, the unit of communication is a face-to-face conversation. While these measures allow us to compare the trends in communication patterns, they do not capture similar amount of information exchanged across the two studies. The content of information exchanged in one chat message is clearly much lower than the content of information exchanged in one face-to-face discussion. Similarly, the information exchanged in one minute in the engine experiment may not be equivalent to the information exchanged in one hour at NASA MDL. Further investigation is necessary to establish equivalent measures of information content across the two studies, which are likely to depend on the complexity of the problem, the mode of communication, and the prior knowledge and expertise of the subjects.

4.3 Lessons Learned for Designing Representative Model Worlds.

This section presents suggestions for designing representative model worlds with respect to the key features of system design process described in Sec. 3.2.5 , based on the “lessons learned” from our analysis and on knowledge from extant literature.

4.3.1 Align Design Expertise and Task Complexity for Subject-Task Interactions.

The design expertise of the subjects and the task complexity should be aligned to ensure that the subjects can achieve the given task objectives and that it prompts the desired types of behavior. For example, expert designers deploy different search strategies (depth-first more than breadth-first) depending on whether the design requirements are complex [ 42 ]. Engineering students in controlled studies take longer to complete coupled system design tasks as the number of design variables increases [ 3 , 43 ]. Our study has shown that, in the context of studying interdisciplinary communication for parametric design, the engine design task possesses appropriate complexity for engineering students to successfully achieve the given objectives, as discussed in Sec. 3.3.3 . Moreover, this interplay between the subject and task shows strong technical-communication mirroring because the students can understand the interdependencies of the problem and decide whom to interact with (see Fig. 7 ). One particularly important result of this subject-task interaction is the design search behavior it prompts. This appeared crucial to driving representative communication patterns, since the teams with access to a catalog had more representative mirroring patterns (see Fig. 7 ).

Thus, our results suggest that when designer expertise and task complexity are appropriately calibrated, novices can behave much more like experts. The engine problem is simpler than the spacecraft problem, but it was designed to be approximately as hard for the students as the spacecraft problem is for NASA engineers, and Sec. 4.1 suggests at least partial success in meeting this goal. Therefore, in designing a model world, it is critical to focus on matching subjects' behavior, which results from the interplay among the task complexity, time, information, & resources available, and subjects’ expertise, rather than on matching task complexity or “absolute” expertise level. To that end, Dorst’s expertise framework [ 44 ] is useful for thinking about the representativeness of subject-task interactions. Dorst categorizes such interactions into seven levels, for example, (i) a naive designer makes a one-off choice from available options, (ii) a novice designer follows strict rules or a formal process to meet fixed requirements given by exerts, and (iii) an advanced designer adapts a formal process for considering situational aspects, etc. If the subject-task interactions in different settings fall under the same level of Dorst’s framework, then we can be more confident that behavior will be representative across the settings.

4.3.2 Allow a Burn-in Period Before Observations.

Compare behaviors after subjects are attuned to making decisions in the given controlled experimental setting. There are several reasons for this. First, subject behaviors from a transient period before becoming fully aware of the setting can be a result of framing effects or human biases such as anchoring bias and availability heuristic [ 45 ]. For example, in Fig. 8 , we observe a high mirroring at the beginning of the engine design task because its problem framing indicated possible dependencies with other subsystems. Cash et al. [ 26 ] observed similar phenomenon where their student teams spent a lot of time initially finding information within a source such as a website for a product, however the equivalent time spent by advanced designers on information seeking initially was much lower (see Fig. 6 in Ref. [ 26 ]). Yu et al. [ 46 ] compared student behaviors and practitioner behaviors for the parametric design of desalination systems. They observed that the initial jump from a given design point towards the desired design space was significantly bigger and faster for practitioners with high knowledge levels than for students with no knowledge.

The second reason why it is important to wait for burn-in is because it might be necessary in order to build subject expertise in the problem, in order to appropriately match the expertise to task complexity as advised in Sec. 4.3.1 . If the problem cannot be simplified to match subjects’ expertise (in our case because it was necessary to include a number of coupled design variables—see Sec. 4.3.4 ), then it appears possible to build task-specific expertise within the model world, by allowing this burn-in period. In the engine experiment, while novices did not behave like experts initially, after a relatively short period of exploration, they learned which technical dependencies drove performance and settled into a pattern that better replicated expert behavior. This is akin to a burn-in period in reliability testing.

Other reasons why burning the first few decision steps may be necessary are as follows: (i) when subjects need to get familiar with the user interface [ 47 ]; (ii) if two consecutive conditions require conflicting skills (order effects), e.g., using two different coding languages; initial decisions in the second condition may be tainted, because using one coding language may inhibit the skills required for using the other coding language; (iii) if subjects receive benefits during the experiment, then the endowment effect will influence decisions of the remaining experiment [ 48 ]. For tips on building an environment for economic experiments, the reader may refer to Refs. [ 49 , 50 ].

Therefore, our results and the literature suggest that students behave more like experts (but not exactly like experts) after they have had time to learn about the problem setting.

4.3.3 Consider Organizational Structure and Incentives.

The organizational structure and incentives must be considered for their roles in constraining and motivating behavior. There are several aspects to this issue that arose in our model world.

First, imposed roles or organizational structure can influence behavior. One particularly important aspect of organizational structure in this case was the formal integrator role—the “systems engineer” in the NASA MDL setting. Having an integrative discipline or designer assigned in the engine setting could have improved representativeness with respect to the nature of system design process and communication among disciplines. For instance, Figs. 7 and 9 suggest that incorporating the broadcast discipline into the engine design teams enabled stronger technical-communication mirroring, comparable to the NASA MDL setting. However, the lack of a designated integrator role also led to differences in some communication patterns, which might have been better matched if such a role had been included in the engine setting. It is well known that engineers spend 40 % – 60 % of their time communicating with peers working on the same projects [ 51 ]. Among such peers, there are individuals, called “gatekeepers” (equivalently integrative disciplines), whom others heavily depend upon for internal as well as external sources of information. The “gatekeepers” of engineering teams provide efficient means to disseminate outside and internal information [ 52 ]. Computer-based tools can also be supportive of systems engineering for dispersed teams if appropriate sufficient means of data exchange and communication are incorporated [ 53 ]. Designers of future model worlds should consider the potential importance of an integrator role or tool in prompting representative behavior.

Second, more generally, other aspects of our organizational structure and incentives were more successful in constraining and motivating representative behavior. The similarities in amount and mirroring of communication between NASA and the engine experiment suggest that our incentives and organizational structure successfully motivated students to work together to find a good solution and constrained their communication pathways to represent those of the experts. Had we enabled the small five-person teams to discuss things all together, rather than restricting them to pairwise communication, the patterns would likely have been very different and less representative of the MDL patterns. The literature bears out the importance of appropriate incentives and their interplay with design communication. Success in cross-functional teams requires setting appropriate project goals for effective communication [ 54 ]. Teams can address design inconsistencies through design dialogue if common goals are identified early in the process [ 32 ]. Moreover, formal and informal organizational channels for communication are known to influence how communication happens [ 52 , 53 ], so it is important to set those up in a representative manner.

It is important to realize that “representative” does not always mean “same.” In the MDL, the incentives for good performance revolve around career goals and employment. In the engine setting, performance was incentivized with a small monetary reward for better-performing designs. These incentives are a relatively poor match to the real world reference system, but it is very difficult to replicate long-term career incentives in a classroom setting. Our results suggest that classroom incentives motivated designers to design good systems and that was sufficient for representativeness.

4.3.4 Select Appropriate Degree of Coupling for Tasks.

For studying system design processes, the disciplines in a controlled setting should be selected so that they possess between-disciplines coupling that is comparable to the reference setting. It is well understood that a larger degree of coupling increases individual effort and total completion time while decreasing solution quality [ 55 , 56 ]. In this study, to reduce effects of overly coupled problems, we used a mix of tightly coupled and loosely coupled disciplines in the controlled experiment to match the NASA MDL settings—see the distribution of shared design variables in Fig. 3 —and the results suggest that it was sufficient to generate relatively similar communication patterns. Based on our results, it is not clear what degree of match in coupling among design variables is “enough” to generate similar communication patterns, but roughly matching the coupling worked well in this case.

Broadly, we can compare the overall degree of coupling between two settings using eigenvalue analysis of their design structure matrices (DSMs). Since a DSM quantifies degrees of coupling between discipline pairs, eigenvectors represent directions to different discipline groups in a latent space where nearby disciplines have similar coupling. The eigenvalues represent the degree of separation between different disciplines in the latent space [ 57 ]. The larger the eigenvalue the larger is the separation of its discipline from other disciplines. If we normalize a DSM with the maximum possible coupling, the eigenvalues greater than 1 represent dominant disciplines that are highly coupled. For instance, the largest eigenvalue of the spacecraft DSM in Fig. 13 corresponds to systems engineering which is the most coupled discipline. For the engine DSM, two largest eigenvalues correspond to broadcast and crankshaft. Interestingly, when broadcast is omitted from the engine DSM, the relative differences in eigenvalues reduce. This implies that the disciplines are less separable in the latent space without broadcast and that broadcast enhances the degree of coupling for the crankshaft discipline.

Eigenvalues of the normalized design structure matrices

Eigenvalues of the normalized design structure matrices

4.3.5 Select Team Size Appropriate for Technical-Communication Mirroring.

The size of a design team may be secondary to choosing an appropriate coupling, because successful collaboration is possible in small teams as well as in teams with large numbers of people [ 58 , 59 ]. One can use various statistical analyses to determine the team size that provides the same strength of technical-communication mirroring between a model world and a reference setting. Suppose we are interested in understanding the correlation between two given properties of discipline pairs, say X and Y . The question is how many discipline pairs we need for such comparison. This question can be answered based on some rules-of-thumb or classical statistical tables [ 60 ]. If tables are inaccessible, alternative methods for approximating the sample size are available [ 61 ]. For multiple regression with m predictors and the estimated correlation of R 2 (equivalently effect size f 2 = R 2 /1 − R 2 ), the suggested number of samples is N ≥ (6.4 + 1.65 m − 0.05 m 2 )/ f 2 . If the goal is to find the partial correlation of each individual predictor, then the number of samples should be N ≥ 8/ f 2 + ( m − 1). For example, consider the correlation between the number of shared variables ( X ) and the number of interactions ( Y ) for which m = 1. If the observed correlation from the NASA MDL teams is R 2 = 0.30 ( f 2 = 0.43), then the number of samples (i.e., the number of discipline pairs) for observing similarly significant correlation from the controlled setting is N ≥ 18.6. Further, since the number of discipline pairs N = ( n 2 − n )/2 is a function of the number of teams n , we can estimate the number of disciplines as n ≥ 6.6. Then, in hindsight, the choice of number of disciplines (recall, total number of components including broadcast is 6) in the engine experiment is roughly correct.

4.3.6 Gather Independent Samples and Compare Aggregate Behaviors.

Large number of observations of team behaviors are always preferable because repetition is helpful for identification of noise and accurate comparison of behaviors at the aggregate level. For instance, the results from Figs. 6 and 8 show noisy communication patterns for the student teams as well as the NASA MDL teams which are different when mean statistics are compared. However, large sample may not be always necessary. It is possible to get estimates of the sample size using classical statistical techniques. Suppose we are comparing team property X between a reference and a model world and interested in accessing the validity of null hypothesis that X has the same average in both the settings. The number of teams required per setting then can be determined from tables or using the normal cumulative distribution function if we assume that observations of X have normal distribution [ 62 ]. For example, if we assume that both X ’s have the same known variance σ but different means, then, to reject the null hypothesis against an alternative hypothesis that the mean difference is Δ u , we require N ≥ ( z α + Φ − 1 ( 1 − β ) Δ u / σ ) number of teams [ 62 ], where Φ −1 is the inverse cumulative distribution function and z α = Φ −1 ( α ) is the quantile of probability α . This method requires a priori specification of the probability of Type I error α and the probability of Type II error β .

This paper set out to illustrate the importance and nuances of designing representative model worlds. By collecting data on both the reference setting and the model world, we were able to compare the results and evaluate the extent to which our model world design succeeded in generating representative behavior. This enabled us to contribute lessons in what to worry about when designing model worlds more generally.

Specifically, our analysis generated several lessons learned about how to design model worlds for a purpose similar to that in our case study (detailed in Sec. 4.3 ): (1) choose the subjects and task such that the interplay between subject expertise and task complexity prompts the desired types of behavior; (2) if expertise in the particular task and context are important, allow a burn-in period so that subjects can learn the setting; (3) consider two aspects of context, organizational structure and incentives, because they constrain and motivate behavior; (4) in studying system design processes, ensure there is appropriate coupling between disciplines; (5) select an appropriate team size to support this need for coupling; (6) gather independent samples and compare aggregate behaviors.

It is worth highlighting that all of the above insights relate to the interplay among subject, task and context, a key feature of our model world definition. This is a critical takeaway because, to the extent that past studies have considered representativeness, it has been in terms of single dimensions (e.g., student subjects versus experienced engineers). While these considerations are important too, our study reveals that it is not enough to consider these dimensions on their own. Moreover, it appears that it might be feasible in some cases to obtain overall representativeness of the model world, without high fidelity on any one dimension (e.g., lesson learned (1)).

The overarching lesson is that model world designers should assess whether they have achieved a representative model world in terms of whether the behaviors match, rather than matching the dimensions of subject, task, and context individually . Achieving matching of behavior usually involves representative interplay among subjects, tasks, and context. For example, the combination of the students’ (limited) expertise and the simpler engine design task can generate communication behaviors observed at NASA MDL which consists of expert engineers working on a complex spacecraft design task. Similarly, what matters is not whether the same incentives are used for students and professionals but choosing (potentially different) incentives that are motivating for each group.

In other words, representativeness is different from similarity . Increasing similarity in individual dimensions may not necessarily increase representativeness—in fact it may even reduce representativeness if the effect is not balanced across the model world. What can be done if a model world does not preserve the intended behaviors and features needed to achieve representativeness? Our case study points to two ways forward. First, assess whether these non-preserved behaviors are crucial to answering the research question. Model worlds need not match all aspects of a real world in order to be representative for a particular research question. Second, if these behaviors are crucial, the model world must be modified to increase representativeness. These modifications will be specific to the setting and the research question; we discussed several such modifications for our particular setting in Sec. 4.3 .

Future work should focus on refining and expanding the lessons learned from the use of the model world (Section 4.3 ). For instance, future work may investigate whether larger problem size, better-aligned interdependency structure, and the availability of an integrative discipline improves the model world representativeness in the collaborative design context. Future work should also explore model worlds for studying phenomena other than communication—such as creativity or search—to determine what is critical for representativeness in those cases. To the extent possible, using theories to identify the impact of choices in the model world on behaviors can help in reducing the number of iterations for developing a good model world.

As highlighted earlier, designing good model worlds of socio-technical systems is challenging, and there is a lack of shared understanding and standards for developing model worlds to support engineering systems design research. There is a need for further research in this direction. In the long term, we hope the community can move towards a set of “scaling relationships” for capturing key dynamics of real-world problems in representative model worlds, but this will be a challenging task since we are dealing with socio-technical systems.

In the meantime, in an effort towards achieving shared understanding within the engineering systems design community, we recommend that publications that use model worlds should clearly specify the objective of the model world, the target real world and the phenomena they are trying to represent. The publications should also clearly specify the thought process used in developing model worlds.

The authors gratefully acknowledge financial support from the National Science Foundation through NSF CMMI EDSE collaborative Grant Nos. 1841192 [ZS, EG], 1841062 [JP], and 1841109 [PG] and also NSF CMMI 1563408 and 1662230. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation. The authors also wish to acknowledge Nikolai Joseph, Connor Forsythe, and Victoria Nilsen for the NASA MDL data collection, and Vikranth Reddy for design of the engine design problem.

There are no conflicts of interest.

Get Email Alerts

Related articles, related proceedings papers, related chapters, affiliations.

  • Accepted Manuscripts
  • ASME Journal Media
  • About the Journal
  • Editorial Board
  • Information for Authors
  • Call for Papers
  • Rights and Permission
  • Online ISSN 1528-9001
  • Print ISSN 1050-0472

ASME Journals

  • About ASME Journals
  • Submit a Paper
  • Title History

ASME Conference Proceedings

  • About ASME Conference Publications and Proceedings
  • Conference Proceedings Author Guidelines

ASME eBooks

  • About ASME eBooks
  • ASME Press Advisory & Oversight Committee
  • Book Proposal Guidelines
  • Frequently Asked Questions
  • Publication Permissions & Reprints
  • ASME Membership

Opportunities

  • Faculty Positions
  • ASME Community

American Society of Mechanical Engineers Logo

  • Accessibility
  • Privacy Statement
  • Terms of Use
  • Get Adobe Acrobat Reader

This Feature Is Available To Subscribers Only

Sign In or Create an Account

Methodology for the Description of Socio-Technical Systems: A Case Study Approach

Affiliation.

  • 1 Institut of Medical Informatics, Medical Faculty of RWTH University Aachen, Aachen, Germany.
  • PMID: 37203772
  • DOI: 10.3233/SHTI230230

The ethical implications and regulatory requirements of AI applications and decision support systems are generally the subjects of interdisciplinary research. Case studies are a suitable means to prepare AI applications and clinical decision support systems for research. This paper proposes an approach that describes a procedure model and a categorization of the contents of cases for socio-technical systems. The developed methodology was applied to three cases and serve the researchers in the DESIREE research project as a basis for qualitative research and for ethical, social, and regulatory analyses.

Keywords: bioethical issues; clinical decision support systems; health technology assessment; privacy; socio-technical system; telemedicine.

To read this content please select one of the options below:

Please note you do not have access to teaching notes, socio‐technical systems theory: an intervention strategy for organizational development.

Management Decision

ISSN : 0025-1747

Article publication date: 1 August 1997

Integrating organizational development (OD) and technological intervention into a total system is one of the more difficult tasks for an executive or consultant to execute. Organizations are profoundly affected by technological advancements and require a flexible customized change model to fit the social network of the specific organization into which technology is being introduced. Examines sociotechnical systems (STS) theory and presents classical organization theories of Burns and Stalker, Woodward, Perrow, Thompson and Trist to develop a contemporary OD intervention in terms of self‐regulating work groups (self‐leading or self‐managing teams) performing interrelated technological tasks. Finally, presents some pointers for executives and consultants in assessing STS interventions via 31 diagnostic questions intended to identify interactions among elements of the system.

  • Group working
  • Organizational development
  • Self‐managing teams
  • Socio‐technical theory
  • Technological innovation

Appelbaum, S.H. (1997), "Socio‐technical systems theory: an intervention strategy for organizational development", Management Decision , Vol. 35 No. 6, pp. 452-463. https://doi.org/10.1108/00251749710173823

Copyright © 1997, MCB UP Limited

Related articles

We’re listening — tell us what you think, something didn’t work….

Report bugs here

All feedback is valuable

Please share your general feedback

Join us on our journey

Platform update page.

Visit emeraldpublishing.com/platformupdate to discover the latest news and updates

Questions & More Information

Answers to the most commonly asked questions here

  • Share full article

Advertisement

The Morning

What’s going on with the u.s. economy.

The economic picture is simpler than the swirl of economic indicators sometimes suggests.

People carrying bags crossing a busy street.

By David Leonhardt

The economy is probably headed toward a recession. No, it’s actually booming again.

Inflation is plummeting. No, it has started rising again.

If you find the cacophony of economic indicators to be confusing, don’t feel bad. It is confusing. Some numbers point in one direction, while others point in the opposite. Partisans from both political parties have an interest in promoting one type of news and downplaying the other.

Yesterday’s inflation report added to the muddle. Inflation — a major concern for many families and already a 2024 campaign issue — has been falling for the past year and half. Americans’ economic mood has improved slightly as a result . But the numbers released yesterday showed inflation to be higher than forecasters had expected. In response, the S&P 500 index fell 1.4 percent , its second-biggest daily decline this year.

In today’s newsletter, I’ll give you a framework for thinking about the state of the U.S. economy by describing the four main phases of the past several years. I can’t tell you what will happen next, but I do think the picture is simpler than the swirl of economic indicators sometimes suggests.

1. The pre-Covid boom

The American economy has been disappointing for much of the past half-century. Income and wealth growth has been slow for most families, and inequality has soared. Perhaps the starkest sign of the problems: Life expectancy in the U.S. is now lower than in any other high-income country, and it isn’t especially close .

Still, there have been a few brief periods when the economy has boomed. One of them began late in Barack Obama’s presidency and continued during Donald Trump’s term. The country finally emerged from its hangover after the housing crash, as businesses expanded and consumers spent more.

The unemployment rate fell below 5 percent in 2016 and below 4 percent in 2018. The tight labor market — combined with increases in the minimum wage in many states — raised income for all income groups.

Change in median weekly income

Compared with two years earlier, adjusted for inflation

socio technical case study

By early 2020, the U.S. economy’s short-term performance was as healthy as it had been since the dot-com boom two decades earlier.

2. The Covid crash

Then the pandemic arrived.

People stayed home and cut back on spending. Businesses laid off workers, and the unemployment rate exceeded 10 percent. Experts understandably worried that the economy could fall into a vicious cycle, in which companies went out of business, families couldn’t make loan payments and banks went under.

Many economists believed that the federal government had been too timid with stimulus after the early-2000s housing crash. During the pandemic, members of Congress vowed not to repeat the mistake and passed huge stimulus bills, which both Trump and President Biden signed.

3. The too-hot recovery

Those stimulus programs worked — a bit too well.

For all the misery that Covid caused, it ended up being less economically destructive than the housing crash. The unemployment surge was temporary. And thanks to the stimulus, the typical American family’s finances improved during the pandemic — a very different situation than the aftermath of the housing crash.

In effect, Washington overlearned the lesson of the previous crisis. It pumped so much money into the economy during the pandemic that families were able to go on a spending spree. Businesses, in response, raised prices.

The pandemic’s supply-chain disruptions played a big role, too. This combination — high demand and low supply — led to sharp price increases.

Year-over-year change in Consumer Price Index

socio technical case study

excluding food

in Jan. 2024

4. The healthier recovery

Once inflation rises, it can remain high for years (as it did from the late 1960s to the early 1980s). Workers demand bigger raises, and businesses, facing higher costs, raise prices even more. The dynamic is reinforcing.

When inflation soared in 2021, some economists thought the pattern was repeating. They predicted that only a deep recession would bring a return to normalcy. These dark predictions turned out to be wrong , though. Instead, inflation started falling rapidly in 2022. The end of the stimulus programs, combined with the end of most supply-chain problems, was enough to reduce price increases.

For the past several months, the economy has again looked healthy, with both employment and wages growing nicely. Yesterday’s report doesn’t change that: Annual inflation fell to 3.1 percent in January, from 3.4 percent in December. It’s just that forecasters had expected it to fall more than it did.

Why didn’t it? Maybe last month’s number was just a statistical blip. But maybe it is a sign that the economy has again become too strong for inflation to keep falling.

“Hiring picked up in January, wage growth was solid, and consumers continue to spend,” as my colleague Jeanna Smialek explains. “Some analysts have suggested that in an economy this hot, wrestling inflation the rest of the way to normal will prove more difficult than the progress so far.” Economists generally consider the ideal inflation rate to be around 2 percent.

By almost any measure, the economy is in better shape today than most forecasters predicted even a year ago. Still, it hasn’t fully recovered from the pandemic.

Related: The American economy has seen a burst of productivity recently. Read more about economists’ views on whether it will continue.

THE LATEST NEWS

Alejandro mayorkas.

House Republicans impeached Alejandro Mayorkas , Biden’s homeland security secretary, despite no evidence that he had committed impeachable offenses.

He is the first sitting cabinet member to be impeached. Read what could happen next .

The number of migrants illegally crossing into the U.S. from Mexico has dropped by half in the past month.

More on Politics

Tom Suozzi, a Democratic former congressman, won a special House election to replace George Santos in a Queens and Long Island district. Suozzi’s strategy offers a playbook for other Democrats .

Biden criticized Trump for suggesting that he would let Russia attack NATO countries that don’t spend enough on defense. “It’s dumb, it’s shameful, it’s dangerous, it’s un-American,” Biden said.

Biden urged Speaker Mike Johnson to let the House vote on a bill to aid Ukraine and Israel. If Johnson refuses, House members may use a procedural trick — a discharge petition — to go around him.

Defense Secretary Lloyd Austin has been released from hospital after his admission for a bladder issue.

A diamond ring and a “James Bond” phone: Court filings revealed new details in Senator Robert Menendez’s bribery case .

Top Senate Republicans endorsed Kari Lake for Senate in Arizona. Lake refused to admit her loss in the state’s 2022 governor’s race.

Israel-Hamas War

Arab governments are privately discussing a new vision for governing Gaza, according to a senior adviser . Their idea includes an independent Palestinian leader and an Arab peacekeeping force.

Negotiations in Cairo over a pause in fighting in Gaza have been extended by three days, according to Egyptian officials.

Hezbollah fired missiles into northern Israel , injuring at least two people.

Two Michigan students have opposing views on the war. Read about their civil, if awkward, conversation on campus .

After a crackdown on drunken driving, motorbikes are filling impound lots in Vietnam.

A South Korean court found three former police officers guilty of destroying evidence tied to a Halloween crowd crush that killed nearly 160 people in Seoul in 2022.

Two major parties in Pakistan agreed to form a coalition government. They withheld power from candidates aligned with former Prime Minister Imran Khan, even though his allies recently won the most seats.

Myanmar’s junta said it would impose a military draft to replenish forces that have been depleted fighting pro-democracy rebels.

China, faced with declining foreign investment, is trying to soften its image abroad. One Communist Party official has played a prominent role.

Other Big Stories

The East Coast is sinking , research has found. A major culprit: overpumping of groundwater.

Better memory: OpenAI is releasing a new version of ChatGPT . It will store what users write and apply the information to future chats.

A vehicle crashed into a hospital emergency room in Austin, Texas, killing at least one person and injuring five others.

As artificial intelligence takes over technical tasks, the human side of work — collaboration, empathy, communication — will become more valuable , Aneesh Raman and Maria Flynn argue.

Don’t buy flowers for your valentine . There are more environmentally friendly ways to show love, Margaret Renkl says.

Here are columns by Ross Douthat on inflation and Thomas Edsall on isolationism .

MORNING READS

Anti-Valentine’s Day: Single, anti-consumerist or just not a romantic? There’s a market for you, too .

Small acts: Times readers shared the little things they do to show affection all year long.

Your brain on love: Roses are red, violets are blue. This is how romance can mess with you .

Modern Love podcast: Hear how a couple navigated getting “un-married” while staying together .

New gig: Grover, the furry blue Muppet from “Sesame Street,” is becoming a reporter. Journalists aren’t optimistic .

Lives Lived: Bob Moore leveraged his grandfatherly image to turn the artisanal grain company Bob’s Red Mill into a $100 million-a-year business. He died at 94 .

N.F.L. draft: Read about the top 100 prospects in this year’s class.

N.B.A.: The Knicks are said to be filing a protest to dispute their 105-103 loss to the Rockets .

No. 32: The Orlando Magic retired Shaquille O’Neal’s jersey , making him the second player in N.B.A. history to have a jersey retired by three different franchises.

ARTS AND IDEAS

The closer: When the comedian Taylor Tomlinson began her tour, she was happy with everything in her set except for her closing joke. So, over months of live shows, she refined it, adding punchlines and tweaking details. A new story by the Times comedy critic Jason Zinoman follows her through the process, showing how small changes led to big laughs .

More on culture

Meghan Markle has found a new podcast partner . Markle and Prince Harry’s deal with Spotify ended after they delivered just one show, Variety reports.

The creators of “Six,” about the six wives of Henry VIII, announced that their second musical , “Why Am I So Single?,” will premiere in London in August.

Paramount, which owns Nickelodeon and MTV, is laying off hundreds of employees as it moves away from traditional television.

Late-night hosts discussed Trump recommending Lara Trump for co-chair of the R.N.C.

THE MORNING RECOMMENDS …

Combine just a few ingredients to make a simple, romantic spaghetti carbonara .

Listen to love songs about crushes .

Shop more sustainably .

Clean your blankets .

Here is today’s Spelling Bee . Yesterday’s pangrams were applicable and clippable .

Celebrate Valentine’s Day with this month’s bonus crossword puzzle , which has a rom-com theme.

And here are today’s Mini Crossword , Wordle , Sudoku and Connections .

Thanks for spending part of your morning with The Times. See you tomorrow. — David

Sign up here to get this newsletter in your inbox . Reach our team at [email protected] .

David Leonhardt runs The Morning , The Times’s flagship daily newsletter. Since joining The Times in 1999, he has been an economics columnist, opinion columnist, head of the Washington bureau and founding editor of the Upshot section, among other roles. More about David Leonhardt

IMAGES

  1. (PDF) Socio-technical case study method in building performance

    socio technical case study

  2. (PDF) An Explorative Study of Socio-Technical Information Systems

    socio technical case study

  3. (PDF) The effects of socio-technical enablers on knowledge sharing: An

    socio technical case study

  4. (PDF) Socio-technical systems theory: An intervention strategy for

    socio technical case study

  5. A Socio -Technical Analysis of the Standards Based Apprenticeship in

    socio technical case study

  6. The socio-technical framework (based on [17])

    socio technical case study

VIDEO

  1. Case Study: Unleash your productivity

  2. HSSC CET Socio-Economic Court Case Update

  3. Socio-Technical Integration Research Project Workshop 4 pt4

  4. Socio-Technical Integration Research Project Workshop 4 pt3

  5. Case Studies

  6. CASE STUDY PHY210 VIDEO PRESENTATION

COMMENTS

  1. Assessing transitions through socio-technical configuration analysis

    Socio-technical configuration analysis Geography of transitions Socio-technical alignments Discourse Modular water technologies 1. Introduction Understanding fundamental sector transformations has become a major field of research in innovation studies and related social science disciplines ( Smith et al., 2010 ).

  2. Socio-technical systems: From design methods to systems engineering

    1 Introduction Socio-technical systems design (STSD) methods are an approach to design that consider human, social and organisational factors, 1 as well as technical factors in the design of organisational systems. They have a long history and are intended to ensure that the technical and organisational aspects of a system are considered together.

  3. A Socio-Technical Perspective on Urban Analytics: The Case of City

    This paper demonstrates that a shift from a purely technical to a more socio-technical perspective has significant implications for the conceptualization, design, and implementation of smart city technologies. Such implications are discussed and illustrated through the case of an emerging urban analytics tool, the City-scale Digital Twin.

  4. Socio-technical case study method in building performance evaluation

    Socio-technical case study method in building performance evaluation. Raymond J. Cole's body of work, spanning sustainable design, system complexity and human agency, has encouraged researchers to reconceptualize the notions of comfort and building performance. However, methods for predicting energy use and assessing environmental performance ...

  5. Gender and climate partners explore and share insights from case

    CGIAR GENDER Impact Platform and CGIAR Gender Equality initiative, HER+ began the year with a webinar focused on socio-technical innovation bundling to strengthen women's resilience to climate change: distilling learnings from case studies. This webinar held on January 25 presented insights on various case studies where STIBs have contributed to women's empowerment and resilience and […]

  6. Designing Socio-technical Systems

    First Online: 28 July 2021 2692 Accesses Abstract Technical system design processes are typically based on systems engineering vee models where designers move between functional and physical domains as they develop detailed designs of the overall system and its sub-systems and component parts.

  7. What are Socio-Technical Systems?

    A socio-technical system (STS) is one that considers requirements spanning hardware, software, personal, and community aspects. It applies an understanding of the social structures, roles and rights (the social sciences) to inform the design of systems that involve communities of people and technology.

  8. Internationally Management of Socio-technical Projects

    They hold regular socio-technical studies, in the earlier stages of realization of the production ways to: a) approach all socio-technical areas, including the possible problems that require training, standardize the activities and information exchange, where several factories are involved; b) verify the convergence of actions to achieve the ...

  9. Frontiers

    Consumption of raw materials, energy, manufactured goods, and services is increasingly concentrated in cities, as urbanization accelerates globally. Such consumption is influenced by complex interactions arising between the various socio-technical and natural systems that make up cities. To improve understanding of the interlinked factors that can perpetuate—or "lock-in"—unsustainable ...

  10. Completely online group formation and development: Small groups as

    Design/methodology/approach: The authors conduct a mixed methods study that integrates interviews, grounded theory analysis, case study methods and social network analysis to build a multi-layered view of completely online group and community development. Findings: Completely online group formation is explicated as a socio-technical system.

  11. Fairness in Socio-Technical Systems: A Case Study of Wikipedia

    Wikipedia content is produced by a complex socio-technical systems (STS), and exhibits numerous biases, such as gender and cultural biases. We investigate how these biases relate to the concepts of algorithmic bias and fairness defined in the context of algorithmic...

  12. A socio-technical analysis of work with ideas in NPD: an industrial

    We do this through an in-depth case study of the development of the energy-labeled circulation pump Alpha Pro, developed by one of the world's leading pump manufacturers, Grundfos. Using a socio-technical approach, we focus especially on the actors involved and the contextual factors, and less on the detailed development of technical ideas.

  13. (PDF) Socio-technical case study method in building performance

    Socio-technical case study method in building performance evaluation Robert Lowe , Lai Fong Chiu and Tadj Oreszczyn UCL Energy Institute, University College London, London, UK

  14. Methodology for the Description of Socio-Technical Systems: A Case

    Abstract The ethical implications and regulatory requirements of AI applications and decision support systems are generally the subjects of interdisciplinary research. Case studies are a...

  15. Innovation niches and socio-technical transition: A case study of bio

    Innovation niches and socio-technical transition: A case study of bio-refinery production - ScienceDirect Futures Volume 43, Issue 1, February 2011, Pages 27-38 Innovation niches and socio-technical transition: A case study of bio-refinery production Antonio Lopolito a , Piergiuseppe Morone b , Roberta Sisto b Add to Mendeley

  16. Socio-technical case study method in building performance evaluation

    socio-technical methods:. philosophical assumptions and theoretical stances. investigative logics - when, where and why. guidelines for practice. contributions to system perspective Applying the socio-technical case study method Philosophical assumptions and theoretical stances The Probe project (Bordass et al., 2001; Cohen et al.,

  17. Designing Representative Model Worlds to Study Socio-Technical

    Abstract. The engineering of complex systems, such as aircraft and spacecraft, involves large number of individuals within multiple organizations spanning multiple years. Since it is challenging to perform empirical studies directly on real organizations at scale, some researchers in systems engineering and design have begun relying on abstracted model worlds that aim to be representative of ...

  18. Towards a socio-technical understanding of discretion: a case study of

    In order to conduct this search I draw on Science and Technology Studies (STS) to develop a case study of the 'less lethal' electric-shock weapon the Taser in England and Wales. While it may seem unusual to blend criminology, policing and STS, the latter is concerned with questions about the role of, and distinctions between, 'technology ...

  19. Socio-technical systems theory: An intervention strategy for

    Examines sociotechnical systems (STS) theory and presents classical organization theories of Burns and Stalker, Woodward, Perrow, Thompson and Trist to develop a contemporary OD intervention in...

  20. Methodology for the Description of Socio-Technical Systems: A Case

    1 PMID: 37203772 DOI: 10.3233/SHTI230230 Abstract The ethical implications and regulatory requirements of AI applications and decision support systems are generally the subjects of interdisciplinary research. Case studies are a suitable means to prepare AI applications and clinical decision support systems for research.

  21. Socio‐technical systems theory: an intervention strategy for

    Books and journals Case studies Expert Briefings Open Access. Publish with us Advanced search. To read this content please select one of the options below: Access and purchase options ... Socio‐technical theory; Technological innovation; Citation. Appelbaum, S.H. (1997), "Socio‐technical systems theory: ...

  22. Implementation of an intelligent supply chain control tower: a socio

    This study applies Socio-Technical systems (STS) theory using the technology typology (long-linked, mediating, intensive) in the context of industry 4.0. ... The case study protocol included the ...

  23. Implementation of an intelligent supply chain control tower: a socio

    This study applies Socio-Technical systems (STS) theory using the technology typology (long-linked, mediating, intensive) in the context of industry 4.0. The company followed a three-phase implementation plan (I: Initiation, II: Live, III: Continuous improvement) to transit from outsourcing supply chain control to developing the SCCT control ...

  24. A case study in a socio-technical approach in ontology engineering

    A socio-technical ontology engineering approach was applied within all of the phases in the life-cycle of the ontology and involved about ten domain experts from research centers and universities. This paper describes results from the study case in developing an ontology for the Indonesian medicinal plants. The approach to the new way of developing ontology is taken from the point view of a ...

  25. What's Going on with the U.S. Economy?

    The economic picture is simpler than the swirl of economic indicators sometimes suggests. By David Leonhardt The economy is probably headed toward a recession. No, it's actually booming again ...

  26. Towards a Framework for Improving Quality of User ...

    Towards a Framework for Improving Quality of User-Centered Services in Socio-technical Systems: A Case Study of Airport System February 2024 DOI: 10.1007/978-981-99-7569-3_30