Analytical sociology for sociotechnical transition research: Bridging case study and system
This paper is motivated by the development in analytical sociology and sociotechnical transition
research. This paper explores the potential for cross fertilization between the two that can improve
current, transition research methodological practice, and provide a response to a number of
criticisms on the Multi-Level Perspective on transition research. The paper proposes a retroductive
transition research methodology to identify and test social mechanisms for their explanatory power
in transition case studies. The methodology consists in the joint use of transition case study and
system dynamics modeling and simulation. The paper discusses how the two methods are used
iteratively, and how each one complements the strengths and counters the weaknesses of the other.
The methodology has particular strengths and implications for the agenda of issues that research on
future transitions to sustainability faces currently. It highlights the central role system dynamics can
potentially play in sociotechnical transition research and community.
Keywords: sustainability, retroduction, transition, mechanisms, simulation
Sociotechnical transitions research is a relatively new field that places a particular focus on
sustainability transitions (Elzen and Wieczorek, 2005; Foxon, 2011; van den Bergh et al., 2011;
Markard et al., 2012; Coenen et al., 2012; McMeekin and Southerton, 2012). One of the most
widely adopted transition research frameworks is the Multi-Level Perspective (MLP). MLP
originates in the work of Rip and Kemp (1998) and has been developed subsequently in Geels
(2002;2004;2005;2010), Geels and Schot (2007), Markard and Truffer (2008), Papachristos et al.
(2013), Papachristos (2018) and Geels et al. (2016a). The MLP has received a number of critical
remarks some of which emphasize issues of agency and the need to develop its methodology (Smith
et al., 2005; Shove and Walker, 2007; Genus and Coles, 2008; Vasileiadou and Safarzynska, 2010;
Svensson and Nikoleris, 2018). Such a methodology should also improve the outline and
measurement of niches, regimes and landscapes, the comparability of cases, and enable theory
development in a cumulative way (Smith et al., 2010; Holtz et al., 2015).
A number of these critical remarks motivate this paper: methodological consistency in
transition studies, transparency of underlying theoretical research assumptions, researcher choices,
system boundary definition, case study data use, and the parsimony and tractability of the analysis
(Genus and Coles, 2008; Smith et al., 2010). The second motivation for the paper is to address the
call for reflexive transition governance (Voss et al., 2006), particularly of future, multi-system
transitions to sustainability perspective that requires an endogenous perspective and iterative
process of problem definition, intervention and response.
The call to identify and theorise transition mechanisms was identified early on (Geels, 2002),
but very little research followed up on this (Geels and Schot, 2010; Papachristos, 2014b;
Papachristos and Adamides, 2016; Papachristos, 2018). Other analytical approaches, such as system
dynamics, must be applied to address future, multi-system sustainability transitions because MLP
cases are primarily retrospective (Turnheim et al., 2015). In order to address this call and part of the
MLP critique, this paper presents a detailed conceptualization of transition mechanisms to: address
the call to study transition patterns and mechanisms (Geels, 2002), keep the MLP relevant for
contemporary and future transitions by explicitly incorporating mechanisms in transition research, a
requirement for reflexive governance, address a number of critical remarks the MLP received and
increase confidence in the use of the framework, and develop a mechanism based approach that will
bridge case studies and modelling and simulation for the study of future sustainability transitions.
The concept of mechanisms is discussed on several literature streams including analytical
sociology and critical realism (Archet, 1995; Hedström and Swedberg, 1998; Archer et al., 1998;
Archer, 2000; Adamides et al., 2012; Hedstrom and Bearman, 2011; Mingers, 2014; Mingers and
Standing, 2017; Keuschnigg et al., 2018; Sorrell, 2018). Identifying causal mechanisms and
“process tracing”, in transitions is an appropriate approach when explanatory and independent
variables are temporally separated (Mahoney, 2000; Bennett and Checkel, 2014). The paper
proposes retroduction as a mechanism oriented, explanatory approach based on analytical sociology
(Keuschnigg et al., 2018) to address the call to identify particular patterns and mechanisms in
transition processes (Geels, 2002). Retroduction is a metaprocess through which an explanation for
a particular phenomenon is developed by identifying and/or postulating mechanisms that can
generate it (Sayer, 1992; Collier, 1994; Sayer, 2000; Bhaskar, 2008).
The application of retroduction to sociotechnical transitions faces challenges due to their
complexity, and the difficulty to identify mechanisms in operation, and deduce their effects
(Hedström and Swedberg, 1998; Hedstrom and Bearman, 2011). In this respect, modelling and
simulation has been proposed as a complementary methodological tool to transition case studies
(Papachristos, 2012; Papachristos, 2014a). The paper argues that the application of the retroductive
methodology in sociotechnical transitions research will: (i) help to identify common mechanisms
across cases, context and time, (ii) equip the MLP methodologically to address contemporary and
future sustainability transitions of multi-energy systems, where changes in one system may induce
major ones in others, (iii) provide a response to the critique of MLP, and increase confidence in
understanding transitions, and (iv) bridge case study and modelling and simulation, and increase the
transparency of research design choices, for example with regard to system boundary definition.
The contribution of the paper, along with the retroductive methodology proposed, is the
definition and extensive discussion on mechanisms, which in turn enables the provision of
guidelines for how to conceptualise, search, and when to stop the search for transition mechanisms
in a case. Such guidelines are missing in prior transition research despite the explicit, early call to
identify transition patterns and mechanisms (Geels, 2002). The paper outlines the implications,
synergies, strengths and weaknesses of bridging case study work and modelling and simulation,
induction and deduction (Van de Ven, 2007), and provides a process overview of how the two can
The rest of the paper is structured as follows. Section 2 offers an overview of the MLP, of the
criticism it has received, and singles out the particular points that the proposed methodology aims to
address. Then section 3 sets the scene for the methodology and provides an overview of
mechanisms and retroduction. Section 4 discusses retroduction and why modelling and simulation
is needed, and 5 discuss the methodology. Section 6 concludes the paper.
2. Sociotechnical Transitions: the MLP Framework
The MLP is a framework for the study of sociotechnical system change, with a focus on system
interconnections and the dynamics of social groups that influence technological change and inertia.
The central analytical MLP concept is the sociotechnical regime, which facilitates analysis of what
underlies the activities of actors who reproduce system elements. The actors are embedded in
interdependent social groups, each with its own regime (set of rules). The MLP distinguishes
between technological, culture, science, markets, industry and policy regimes (Geels and Schot,
2007). The sociotechnical regime refers to the inter-regime alignment and coordination of
intergroup activities that stabilize sociotechnical trajectories. The MLP has two additional analytical
concepts (Geels, 2004): (i) the landscape at the macro level provides gradients for sociotechnical
regime trajectories, and (ii) the niche level where radical innovations incubate and proliferate
protected from external influences.
Transitions in the MLP framework come about when the sociotechnical regime is destabilised
through reinforcing and disrupting interactions that develop between these three levels by (Geels
and Schot, 2007): (i) innovations that may develop in niches through learning processes,
price/performance improvements and support from powerful groups, (ii) pressures that events may
generate or trends at the landscape level that act on the regime (economic, cultural, demographic
and other), (iii) internal regime tensions that can accumulate and create windows of opportunity for
innovations in niches and, (iv) external influence from other systems, regimes or niches
(Papachristos et al., 2013). The transition is finally completed when the social and technical aspects
of novel innovations become embedded in the new sociotechnical system.
Transition process research, in its simplest form, focuses on how the nature, timing and
intensity of interactions between landscape pressures, the build-up of niches innovations and
internal regime tensions may unfold over time, enable or constrain a transition process (Geels and
Schot, 2007; Papachristos, 2014a). MLP case studies follow a process rather than a variance
explanatory style and they don’t attribute transitions to single causes or interactions but to to
ensembles of multiple interlocking causal influences that reinforce or disrupt each other (Geels,
2011). Different interaction configurations can result in different single-system transition pathways
(Geels and Schot, 2007). The range of interaction configurations has been extended to include
multi-system interactions (Papachristos et al., 2013).
2.1 Critique to the MLP and Responses
This section focuses on five points of critique for the MLP to which this article responds to1. The
first, perhaps most severe critique concerns the epistemological status of the MLP: “the potential
contribution of the MLP/transitions framework could be limited to offering a heuristic device that
can be used to organise sets of data” (Genus and Coles, 2008, p1442). This critique is consistent
with organisational research work about what is a theoretical contribution, where a list of variables
or constructs (like niche, regime and landscape) does not constitute theory without an underlying
causal logic (Whetten, 1989; Sutton and Staw, 1995; DiMaggio, 1995; Weick, 1995). The
distinction between a temporal antecedent and a causal one is obscured in narrative and thus
additional information and insight is required (Griffin, 1993).
Geels (2011) responded that an open, heuristic framework for transition research may be more
suitable in guiding the researcher to think on relevant questions in order to flesh out relevant
patterns. In doing so the researcher can utilize “auxiliary theories” to develop middle range theory
(Geels, 2007; Geels, 2011) and avoid the development of either grand theories of social life or
smaller scope theories (Merton, 1968). For example, a vast literature exists on strategic
management (Helfat et al., 2007), organizational change (Poole and Van de Ven, 2004), and
analytical sociology (Hedstrom and Swedberg, 1998; Hedstrom and Bearman, 2011). The aim of
middle range theory is to provide a satisficing trade-off between the criteria of good theory:
accuracy of representation, generality, and parsimony (Weick, 1979).
The second criticism concerns the lack of methodological consistency and accuracy in the
operationalization of MLP concepts like mechanisms and related methodological requirements. The
1 For a detailed overview the interested reader can refer to Coenen et al. (2012), Genus and Coles (2008), Markard and
Truffer (2008), Shove and Walker (2007), and Smith et al. (2005; 2010), and for a response to Geels and Schot (2007),
Geels (2010) and Geels (2011).
large uptake of MLP was also driven by the early emphasis placed on technological niches and the
availability of many historical cases that were analysed retrospectively as bottom up, transition
cases (Smith et al., 2005; Shove and Walker, 2007; Genus and Coles, 2008). The advantage of this
bottom up character of these studies is that the system boundary is better outlined. However,
historical cases are different than contemporary ones as system interactions and therefore
boundaries might change during a transition (Papachristos et al., 2013). Thus, there is a definite
need to use methods to map transitions, their boundaries, and maintain a tractable, parsimonious
analysis (Smith et al., 2010). Geels (2011) responded that the analyst should ﬁrst demarcate the
object of analysis and then operationalise them. Nevertheless, no prescription is offered on MLP
boundaries of analysis.
The third criticism concerns how to identify early signs of systems trajectories, anticipate and
reorient those that are unsustainable (Shove and Walker, 2007). They point to the need to bridge
understanding transitions through retrospective studies with developing the capacity to steer
contemporary and future transitions through prospective studies. While Geels (2011) offers no
explicit response, this issue is clearly linked to critique on agency, operationalization of regimes and
specification of regimes, heuristics, epistemology and explanatory style and methodology.
This critique is important, as a wide variety of future outcomes is possible in path dependent
systems (Pedriana, 2005). Knowledge of causal mechanisms can indicate why certain phenomena,
and not other became possible in the first place, and allow system interventions to produce a useful
outcome (Epstein, 2006; Mingers and Standing, 2017). While considerable work has identified
transitions patterns (Geels and Schot, 2007; de Haan and Rotmans, 2011; Naber et al., 2017), less
empirical work has addressed mechanisms with the exception of Geels (2005), Papachristos
(2014b), Papachristos and Adamides (2016), and Papachristos (2018).
The fourth criticism concerns the evolving/contested nature of the issue of agency (Gibson
and Earley, 2007; Wilson, 2007; Kaisesoja, 2013). Smith et al. (2005), Shove and Walker (2007),
and Genus and Coles (2008) raise the issue of relative neglect of agency in the MLP. Geels provides
no clear argument for why agency can be traced at the individual or collective level. In fact, He
admits that the point of agency is not well developed and calls for: “…understanding of the
mechanisms…” (Geels 2011, p30) through which various levels interact.
Fifth, future sociotechnical transitions require system reorientation towards low carbon
intensity trajectories unlike those witnessed in the past (Papachristos, 2014a). Low carbon
transitions are systemic changes that encompass energy, the economy, and the environment (Rogelj
et al., 2015). Their scope requires a search for alternatives to economic growth (van den Bergh,
2011), and a shift in business models towards product service systems (Tukker, 2015). Such
changes require more than a sociological, retrospective angle to society and technology coevolution.
Nevertheless, little has been done so far to integrate systems thinking approaches, multiple methods
and tools in transition research.
2.2 Conclusions Regarding the MLP Critique
Part of the critique on the MLP being nothing more than a “heuristic device” (Genus and Coles,
2008), comes from openness of the MLP framework which may be more suited to investigate
transitions than precise models and to transition case studies that aim to illustrate and explore rather
than do systematic research (Geels, 2011). Openness, flexibility and even vagueness (Bacharach,
1989) are unavoidable in the initial stages of any new research program such as transition research,
and they facilitate creativity and links to other theories and ontologies (Geels, 2010).
Sociological explanation requires that events and their contexts be theorized openly, factual
material abstracted and generalized, and that causal connections in narrative sequences are
established in a way that can be replicated and criticized. Properties such as closure, precision and
accuracy in operationalising concepts, having pragmatic usefulness and achieving theoretical
cumulation through systematic research are also necessary (Scharpf, 1997). Progress from a stage of
openness to a stage of closure implies that theory is seen as a continuum (Weick, 1995) rather than a
There are two potential responses to the critique of Genus and Coles (2008). The first is that it
comes from a particular perspective on organization science theory and it doesn’t stand if a different
perspective is adopted (Hall, 1999; Kagan, 2009; Goertz and Mahoney, 2012). The second is to
adopt the organization science perspective on theory and develop MLP into a more closed
theoretical proposal. A substantive response to the Genus and Coles criticism, if this perspective on
what constitutes a theoretical contribution is adopted, requires that the MLP theorizing process and
its foundations are reconsidered and improved (Weick, 1989).
The discussion in sections 3 and 4 provides the theoretical background for the methodology
proposed in section 5. Most importantly its attributes constitute an improvement for the five points
of MLP critique discussed in section 2.1. Subsequent sections try to address the foundations and
requirements of mechanism related theorizing in transition research.
3. Foundations of the Proposed Methodology
Several definitions of theory have been proposed (Harre, 1970; Bacharach, 1989; Gioia and Pitre,
1990; Mintzberg, 2005). Theory development is the outcome of a long-winded process in which
theories at work approximate fully fledged theories (Corley and Gioia, 2011). During this process,
all the researcher has to work with usually are appropriate references, data, definitions, variables,
narratives, diagrams and hypotheses, which do not constitute theory on their own, but are part of the
2 There are several views on theory (Blaikie, 2007; Hall, 1999). The author thanks Frank Geels for this point.
building blocks for it (Griffin, 1993; Sutton and Staw, 1995). The implication is that theory can be
viewed as a dimension (Weick, 1995). Researchers progress along this continuum through an
“interim struggle” until they arrive (if ever) at a fully-fledged theory at the end (Mohr, 1982;
Theory development has been likened to a process of evolutionary “disciplined imagination”
(Weick, 1989), where researchers propose, test and select among theory variants. A significant
bottleneck in this is that researchers are both the source of variations and the ones that apply
selection criteria. In effect, they run a mini evolutionary system in: “a head that suffers from
bounded rationality” (Weick, 1989, p529). As the system description is progressively enriched, it
becomes difficult to maintain a clear picture of how mechanisms may be linked to behaviour in a
narrative (Griffin, 1993), and trace their interactions through to their impact because of complexity,
thus selection becomes more difficult.
Narrative descriptions and mental experiments are essential but insufficient as an explicit and
replicable causal framework (Griffin, 1993). They face limiting returns that make the identification
and collection of relevant case data progressively more difficult. The only way to complete the
selection step is to test these variations through real world, or computer simulation experiments
(Weick, 1989). The nature of sociotechnical transitions makes the use of real world experiments
difficult. So far, narratives have been used primarily although simulation has gained ground
(Papachristos, 2011; Papachristos, 2014b; Holtz et al., 2015; Papachristos, 2017; Köhler et al.,
The use of an established narrative approach leads to theoretical accuracy (Weick, 1989).
Modelling and simulation can cope with complexity and leads to parsimonious theoretical
conceptualizations and simplicity (Davis et al., 2007; Harrison et al., 2007). The combination of the
two methods offers a better trade-off between accuracy, simplicity and generality as they sit on
opposite sides of the research strategy spectrum (Figure 1), and has some implications when
research is viewed as a decision-making process (Morgan, 1983).
Figure 1 Research strategies process theory development (adapted from Langley, 1999).
Selection among alternative explanations with an open-ended case research approach may delay
potentially the decision on when a satisfactory explanation has been reached, because confidence in
a case study is built through analysis and corroboration of data. The use of modelling and
simulation can work in synergy with case study research in the proposed methodology as a decision
support tool for the entire data collection and research process rather than its final step (Oreskes et
al., 1994). The iterative search for an explanation and relevant data can proceed in parallel with
modelling and simulation and case study research as both require rich, detailed qualitative
description (Homer, 1996; Sterman, 2000; Yin, 2003). A milestone is reached when the model can
reproduce the phenomenon of interest and thus be considered as an explanation. The following
sections 3.1-3 provides the foundation of the methodology and section 4 shows how it is appllied.
3.1 Social Mechanisms
Several mechanism definitions exist in the literature (Glennan, 1996; Machamer et al., 2000;
Hedström, 2008; Hedström and Bearman, 2009; Hedström and Ylikoski, 2010). The definition
adopted here is that (Torres, 2009, p247): mechanisms are complex systems composed of entities
organized in time such that (i) through engaging in activities they produce a phenomenon, and (ii)
the activities in which the mechanism’s entities engage are characterizable as invariant
generalizations. An invariant generalization implies stability under some related changes and
interventions (Tsoukas, 1989; Woodward, 2002), and requires that mechanism components must be
independent in order to allow one to trace how changes in them impact the overall system
behaviour. Invariance implies that mechanisms should allow for counterfactual inferences that are
associated with control (Machamer et al., 2000; Woodward, 2002).
A mechanism has further characteristics (Hedström and Ylikoski, 2010): (i) it is linked to a
particular effect it produces, (ii) it is an irreducibly causal notion, (iii) it forms a hierarchy with
other mechanisms, (iv) it has a structure which can be linear or non linear (Bunge, 2004), (v) it may
refer to an easy or difficult to observe causal factor (Gerring, 2010), and (vi) it may refer to an
explanation that presumes probabilistic or highly contingent relations (Gerring, 2010). A final
characteristic of mechanisms is their temporal nature.
3.2 The Search for Mechanisms
There is no established method for mechanism search because processes do not come as discernible
given units (Bunge, 2004). The search for mechanisms involves imagination (Weick, 1989), use of
available data and making conjectures (Bunge, 2004). The search for mechanisms presupposes
epistemologically that generalizable properties can be abstracted from historical processes. The
researcher picks a sequence of the ongoing process and justifies why it is considered as the
explanandum (Mayntz, 2004). Then through process tracing the researcher can identify mechanisms
and initial conditions that link explanatory with independent variables and generate jointly the
The focus on mechanisms breaks up the original, explanation seeking, “why” question into a
series of smaller questions about the causal process: What are the participating entities and their
relevant properties? How the interactions of these entities are organized temporally? What factors
could prevent or modify their outcome? (Hedström and Ylikoski, 2010). A mechanism may not be
observed directly (Tsoukas, 1989), but its operation can be observed in a specific instance provided
that some observable, real event sequences have similar properties, in the same way that analytical
constructs are observed via the indicators that operationalize them (Mayntz, 2004).
Further insight into searching for mechanisms is provided by taxonomies of mechanism types
(Coleman, 1986; Tilly, 2001; McAdam et al., 2001; Opp, 2005). The first type covers the macro-to-
micro link (Figure 2). Hedström and Swedberg (1998) call it a situational mechanism, a systematic
link from a social structure, event or state, to the beliefs, desires, and opportunities of individual
actor(s). Examples are the belief-formation and preference-formation mechanisms such as those
expressed in the idea of reference groups (Merton and Rossi 1968; Boudon, 1988). The second type,
at the micro level, is called action-formation mechanism and links specific combinations of
individual desires, beliefs, and action opportunities to a specific action. The mechanism based
approach to social science does not subscribe to an axiomatic position according to which a specific
action theory should be used for all purposes (Hedström and Ylikoski, 2010). The requirement that
mechanism based explanations have to cite the actual causes of the phenomenon to be explained
often makes rational choice explanations unacceptable, as they are built upon implausible
psychological and sociological assumptions.
Finally, transformation mechanisms cover the micro-to-macro link of how interactions of
groups or individuals produce some collective outcome. Examples include Schelling's (1978)
segregation model and game-theoretic models such as the tragedy of the commons, and neoclassical
market models (Scharf, 1997). From this taxonomy it follows that a mechanism based explanation
is not identical with a bottom up explanation (Opp, 2005).
Figure 2. Macro-Micro-Macro model adopted from Hedström and Swedberg (1998)
Processes identified in the causal reconstruction of a particular transition case can then be
formulated as mechanism statements. They must: (i) make mechanisms explicit and detailed, and
(ii) support its assumptions by relevant empirical evidence (Hedström and Ylikoski, 2010). The
statements should provide (Mayntz, 2004): (i) the level of reality mechanisms refer to, (ii) their
degree of conceptual abstraction, (iii) the scope of their claimed applicability at a given level of
abstraction, and (iv) their sequence and timing (Sutton and Staw, 1995).
3.3 Concluding the Search for Mechanisms
The question of when does the search for mechanisms end can be answered from an
epistemological and ontological perspective. From an epistemological perspective, mechanisms
traced to lower levels implies a hierarchy (Hedström and Ylikoski, 2010), and the assumption that
within every covariation relation (Macro 1 Macro 2) there is a mechanism (Micro 1) and within
that covariation relation even lower level mechanisms (Micro 1a,b) which can be known (Figure 3).
A first condition is that this process should end when the researcher is confident about the
underlying assumptions of mechanisms (Hedström and Ylikoski, 2010). A second condition is when
the boundaries of the existing literature have been exceeded and new knowledge has been created
(Anderson et al., 2006). A third condition suggests the investigation of contiguous levels of analysis,
one level above and one level below their focal phenomena (bracketing rule) (Hackman, 2003). For
example, if a researcher is interested in understanding individual behaviour, the key mechanisms
may reside in group structures and processes (one level above) and in individual cognition and
emotions (one level below).
Figure 3. Causal regression to lower level mechanisms
From an ontological perspective, methodological individualism assumes that all causal powers
ascribed to social groups are reducible to the aggregates of causal powers of their individual
members. Nevertheless, organized social groups and larger social structures are among the core
explanatory factors of large scale long term social changes in historical and comparative macro-
sociology (Kaidesoja, 2013), political processes (Scharpf, 1997), and corporate competition
(Stinchcombe, 1998). The search for mechanisms can stop at macro social mechanisms that form
hierarchies of entities above the level of individuals, if they are shown to be relatively durable
collective agents with characteristic emergent capacities and activities (Kaidesoja, 2013). Thus,
agency can be attributed to collective social entities as long as they endure (Weick and Roberts,
4. A Retroductive Methodology for Sociotechnical Research
According to the critical realist perspective retroduction is a metaprocess through which given a
description of an empirical phenomenon: “…events are explained by postulating (and identifying)
mechanisms which are capable of producing them…” (Sayer, 1992, p.107), and the conditions for
the occurrence of the phenomenon are reconstructed (Bhaskar, 2008). Retroduction is suggestive of
what may be part of the real world and through this a novel understanding about the world is
introduced (Wuisman, 2005).
Retroduction starts with a description of the significant features of the Unexplained
Phenomenon X for which a Theoretical Gap exists in the knowledge (Figure 4). Then X is
recontextualised as a more general phenomenon by abduction, and a range of Dynamic Hypotheses
H are put forward, each consisting of an ensemble of generative mechanisms. Subsequently, their
validity is assessed by Deduction of their logical consequences. Finally, if the patterns of the
deduced consequences do not match those derived through case analysis, then a new cycle of
abduction, deduction and induction takes place.
Figure 4. The cycle of discovery (based on Wuisman, 2005)
4.1 The need to test transition mechanisms
The need to test mechanisms arises out of the methodological requirement to evaluate their
explanatory power for some Unexplained Phenomenon X (Figure 4). The evaluation requires the
demonstration of mechanism consequences, observable or not (Sayer, 1992). There are three
reasons to do so. First, the use of inductive methodology can lead to a correlative logic, where “if
Condition then Pattern” statements are made based on the regularity with which transition patterns
Certain events placed on a timeline may suggest causation but their correlation with transition
patterns does not imply necessarily a causal relation (Sayer, 1992). Furthermore, such conditional
MLP logic was derived from observed regularities in historical transitions and they may not repeat
(Smith et al., 2010).
Second, observing a transition pattern and inductively stipulating conditions for it, can lead to
fallacious inferences even in a stable social context. For example, mechanisms that reinforce and
oppose a transition may operate with equal intensity and result in stagnation (Tsoukas, 1989). A
researcher may infer that these conditions will always result in a stagnant situation, but any change
in their nature, intensity or timing may alter the result. The converse is also possible. Thus, while it
is necessary to identify empirical correlations, an analytical shift requires tests of mechanism
nature, timing and intensity in the deductive step of retroduction to provide a measure of their
Third, transition research is concerned principally with structural causes, where multiple
factors lie between cause and effect (Mahoney, 2000), due to system feedback, delays and processes
of accumulation (Geels, 2005; Raven, 2007; Naber et al., 2017). For each covariational relation
there is a plethora of plausible causal mechanisms available in the literature (Gerring, 2010). The
abductive step from effects to transition mechanisms must consider the fact that innovation at a
fundamental level is not linear (Abernathy and Townsend, 1973), and that innovation mechanisms
may not be linearly additive (Manicas, 2006). The resulting mechanisms must be tested and there
are some difficulties related to this.
4.2 Overcoming Difficulties in testing MLP explanations
Learning about complex social systems while being embedded in them is difficult because of a
number of cognitive limitations that apply to organization theorists too (Astley and Van de Ven,
1983; Sterman, 1994; Kahneman, 2011). Humans adopt a simplistic cause and effect view of events,
they do not appreciate correctly system delays, feedback processes (Sterman, 1989a,b; 2008; 2010)
and the possibility that the intensity of feedback loops in the system may change (Diehl and
Sterman, 1995). Research provides empirical evidence that people understand poorly accumulation
or depletion processes (Sterman and Sweeney, 2007; Cronin and Gonzalez, 2007; Cronin et al.,
2009; Aramburo et al., 2012). Even scientists can fall into traps if their results look reasonable
(Nuzzo, 2015). The implications of this are ubiquitous, they are relevant for climate change
research (Sterman, 2008), and link inevitably with sustainability transitions research.
Hence, the deductive step in Figure 4 cannot be performed mentally with confidence because
transition phenomena involve multiple feedbacks and accumulation processes (Geels, 2005; Geels
and Schot, 2007; Raven, 2007; Naber et al., 2017). Furthermore, empirical learning through
feedback and experimentation is also not practical as transitions unfold over decades and real-world
experiments are time consuming and costly. Thus, it is not possible to calibrate human intuition
about how such systems might work in the future by virtue of human experience alone. The fact that
the development of MLP transition typology has been based primarily on single system case studies
is an indication of human cognitive limitations when it comes to multi-system analysis. A multi-
regime transition typology has not been developed despite the role of outsiders in historical and
contemporary transitions (Papachristos et al., 2013; Arranz, 2017).
The combination of modelling and simulation and case studies is a way to overcome the
cognitive limitations discussed and address transitions to sustainability. The combination retains the
context and detail of MLP transition narratives while increasing confidence in understanding
transition dynamics and drawing conclusions. The effectiveness of the use of narrative and
simulation is based on the assumption that they do not share the same weaknesses or potential for
bias, and that the strengths of each method counter the weaknesses of the other (Jick, 1979; Johnson
et al. 2017).
5. Methodology Outline
The proposed retroductive methodology integrates case study and modelling and simulation and this
section discusses how they are combined together in three steps. In the first step, mechanisms are
identified, temporal assumptions are made explicit, and their representations are made more
accurate and integrated in a model. In the second step, these mechanisms are varied to generate
alternative explanations that are tested systematically through a diverse range of tests. Finally, one
is selected by applying consistently more diverse criteria to them at the selection step. The two
methods work in synergy as a case study provides a rich description of a transition or input for
models, and the necessary context for model results (Winsberg, 2006), and evaluate their
implications because the resultant type of knowledge is itself complex and is a statement of research
choices and constraints (Pidd, 2004). Hence, while the methodology steps are outlined in a linear
fashion in the following sections, the researcher will engage iteratively in retroduction.
5.1 Step 1: Dynamic Hypothesis Development
A necessary first step to go from data to a theoretical understanding, is to postulate explanatory
mechanisms for the process (Van de Ven, 1992; Langley, 1999). The MLP employs ‘process theory’
as explanatory style (Geels and Schot, 2010), and process theories explain outcomes in terms of
event sequences and the timing and conjunctures of event-chains (Pettigrew, 1997; Abbott, 2001;
Langley, 2007; Langley et al., 2013). The development of a Transition Narrative is necessary to
capture accurately the complex interactions between agency and changing contexts, time, event
sequences, and actions (Griffin, 1993; Pentland, 1999; Abell, 2004).
Such narratives are developed commonly in transition case studies (e.g. Turnheim and Geels,
2012; 2013). The context and accuracy of narratives is consolidated in order to match the transition
case to one, or more of the proposed transition pathways (Geels and Schot, 2007). A causal
narrative that conveys a System Understanding of Transition can then be produced using the MLP
as a heuristic device (Pedriana, 2005; Geels, 2011). The aim in this first step is to identify and
categorize the key processes and interactions that drive system behaviour in niche(s), regime(s) and
landscape, and identify mechanisms in line with the discussion in section 3.2-3.3.
The question of why an event sequence happened is broken down into a series of questions
about actors and entities, their properties, and their interactions in time. For example, an aggregated
diffusion process of a disruptive technology is the outcome of a series of adoption events and
interactions between agents. This involves situational, action formation and transformation
mechanisms that can be broken down into the study of interactions of early adopters, early majority,
mainstream users and laggards (Rogers, 2003) and further disaggregated to the individual level.
However, the search for mechanisms often stops at the level of certain user groups of market
segments that can be characterised reasonably from data in terms of their preferences and thus their
propensity to exert agency under a range of options.
Causal loop diagrams (CLD) can then be developed to produce a Dynamic Hypothesis of how
interlocking, reinforcing or disrupting feedback loops operate for each user group and other relevant
agents e.g. competing technology firms. The use of CLDs to represent multiple interlocking
mechanisms is a standard part of system dynamics methodology (Sterman, 2000; Lane, 2008). The
development of a CLD requires additional factual and counterfactual questions about the observed
events in order to focus on those that are deemed to be causally significant, a process termed
“disciplined imagination” (Weick, 1989; Griffin, 1993). CLDs can also provide the means to
address the need to map transitions in a more stylized way (Smith et al., 2010). For example,
transition processes of de-alignment, re-alignment (Geels and Schot, 2007) can be cast in terms of
shifting dominance of disrupting and reinforcing mechanisms that unravel a regime and constitute
something new in its position.
The Dynamic Hypothesis is a provisional explanation about what generates system behaviour
in theory (Sterman, 2000). The observed system behaviour can be represented through temporal
patterns by Abstracting and Simplifying from the narrative, and provide a description that contains
less than the total detail of their underlying transition process while they still reflect some of its
fundamental aspects (Richardson, 2005; Janssen et al., 2009). It is crucial that such temporal
patterns reflect all kinds of available data, quantitative or qualitative. Qualitative data are used in
transition case studies and can serve as a main source of information in system dynamics
methodology (Forrester, 1961; Wolstenholme, 1990; Sterman, 2000; Luna-Reyes and Andersen,
2003), that can be elicited through interviews (Ford and Sterman, 1998). Subsequently, a Dynamic
Simulation is constructed.
The benefit of modelling and simulation is the rigour it introduces, as it compels researchers
to face issues and assumptions they may have acknowledged but addressed or stated vaguely
(Harrison et al., 2007; Davis et al., 2007). Nevertheless, model simulation has limitations too (Geels
et al., 2016b). The use of a model along a case study can enhance the reliability and validity of
research, and help the researcher explore system behaviour further. The aim to develop better
explanations through the combined use of methods requires step 2 and 3.
5.2 Step 2: Alternative Hypothesis Development
The development of alternative Dynamic Hypotheses to select one through a diverse range of
criteria is preferable to persistence onto a single one, that is stretched to fit the data (Mitroff and
Emshoff, 1979). The researcher varies the assumptions and mechanisms that underlie a Dynamic
Hypothesis through disciplined imagination (Weick, 1989), to generate a range of corresponding,
alternatives and simulation models3. Such boundary adequacy tests are an integral part of system
dynamics methodology (Sterman, 2000; Burton and Obel, 2011). They require rigorous search and
consideration of available and additional data to determine the concepts, and variables that are
significant and treated endogenously, those treated exogenously and those omitted (Richardson and
Pugh, 1981; Sterman, 2000).
The search for alternative hypotheses can serve different purposes. If the study is
retrospective, the aim is to explain a transition and the search stops when a satisfactory explanation
is reached. If the study is prospective the aim is to achieve some targets in time through robust
policies and enhance system trajectory adaptation to exogenous effects. The search stops if a
threshold of performance is achieved, or it is demonstrated that it is not possible to achieve it. In
either case the model sensitivity to variables has to be explored. If the model is not sensitive to a
variable then it could be set constant or be omitted altogether to create a more parsimonious
transition explanation. If the model is sensitive then the search for what influences the variable, or
ways to influence the variable can generate more elaborate hypotheses.
For example, the model boundary may need to be expanded to include new variables and
feedbacks that can overcome policy resistance. Expanding the boundary of a model can have effects
much greater than those introduced by disaggregating existing model structure (Rahmandad and
3 The generation of many explanations does not imply the development of many models from scratch. Differences in
models can range from one equation to whole modules, or the architecture of the model. The inclusion, or removal, of
those from a model suffices to treat it as a new model.
Sterman, 2008). Boundary exploration is particularly relevant for future oriented studies as the
greater the time horizon the greater the system boundary under consideration. Furthermore, a range
of different representations of the system that can draw on different ontologies or scientific
paradigms may be produced (Gioia and Pitre, 1990).
Boundary tests are difficult because they encompass articulation of all, theoretically plausible
mechanisms for a given covariational relation. Unlike statistical correlation tests, this is an open
ended task as mechanisms may contain underlying ones, or other mechanisms may offer better
alternatives (Gerring, 2010). The syntax of a modelling language facilitates the systematic
generation and test of candidate explanatory mechanisms (Harrison et al., 2007) to substantiate their
theoretical validity and increase the confidence in selecting one (Goldspink, 2002).
5.3 Step 3: Selection & Retention
In this step one candidate Dynamic Hypothesis is selected based on the criteria of parsimony,
testability, logical coherence and explanatory power i.e. how well it reproduces the real
phenomenon (Pfeffer, 1982). Gerring (2005) provides a more nuanced list of criteria. Explanatory
power increases when evidence in support of its assumptions are presented and if it is demonstrated
why alternative hypotheses and their assumptions are unlikely to hold (Siggelkow, 2007; Hedström
and Ylikoski, 2010).
A process of Pattern and Timing Testing underlies these three steps where a comparison of
observed patterns with those derived theoretically from the dynamic hypothesis for the process
(Trochim, 1989; Bitektine, 2008). The theoretical pattern is derived by deduction from the
postulated mechanisms for the transition process. If it matches the empirical pattern then this
increases confidence that the proposed theory is an explanation of the observed transition. Pattern
matching techniques have been explored in system dynamics where the behaviour of the model is
compared to real world patterns (Sterman, 2000), and elsewhere, for example in ecological
modelling (Railsback and Johnson, 2011).
As transition research different events can be found at different ontological levels, any pattern
matching attempt must consider their sequence and evolution in time, and the level they are found
(Mohr, 1982; Lerner and Kaufman, 1985; Abbott, 1990). Pattern matching requires the compilation
of a temporal sequence of observed events to attempt to determine causality. Then a comparison
with theoretically deduced events can be made to match their pattern and timing by observing the
following conditions (Yin, 2003): (i) some events precede others while the reverse is impossible, (ii)
some events occur contingent on others, (iii) some events can follow other events only with a
certain delay, and (iv) certain types of events occur only during specific time periods. Pattern
matching does not remain opaque to the reader if model documentation and reporting guidelines are
followed (Sterman, 2000; Rahmandad and Sterman, 2012; Martinez-Moyano and Richardson,
Nevertheless, a gap exists in transition literature and a challenge for the proposed
methodology. Descriptive MLP studies abstract usually their empirical findings into stylised
patterns, but the development of cross-case quantitative indicators for transitions remains a major
theoretical gap. Indicators are necessary to compare these empirical patterns to a simulation derived
theoretical pattern. A challenge in bridging cases and models is the development of a broadly shared
conceptualisation and framework that can act as a channel for dialogue between case and model
based work. The consistent application of the proposed methodology to transition cases is a step in
The application of the methodology steps proceeds from left to right in Figure 5, towards
greater simplicity. The process starts from a Transition Narrative development which provides the
necessary context and details of the transition “. A System Understanding of Transition is developed
which is relatively simplified but conveys the essent of transition dynamics. Then a Dynamic
Hypothesis is produced by Abstracting & Simplifying the key elements and patterns of the
Transition Narrative. Disciplined Imagination yields the underlying mechanisms drawing on the
researcher’s System Understanding of Transition which is turned into a simulation model to deduce
its consequences, and corroborated with the Dynamic Hypothesis through Pattern and Timing
Testing using data documented in the case analysis (Figure 5).
Figure 5. Progressing from accuracy to generality
The proposed methodology does not offer a decision heuristic for “what” will be the subject of
research: a complete historical transition, or part of an ongoing transition. The “how” involves a
combination of case study and modelling work. The application of the methodology might reveal
the need to better account for behavioural aspects of transitions, or political processes, in the
narrative and in the model. The “how” might result in “what” the study aimed to describe being
different than that it describes at the end.
In summary, section 5 discussed the proposed methodology and how in each step transition
narratives, and modelling and simulation complement each other in a pragmatic way. The
methodology provides a response to the 5 critique points discussed in section 2.1 (Table 1). First, it
provides a way to identify mechanisms, build greater confidence in the nature of causality, and take
MLP beyond its current status as a heuristic. Second, mechanisms can be identified and traced to
the collective and/or individual level, depending on the needs of the case at hand. Thus, they are an
improvement on the neglect of agency that MLP has been criticized for. Third, the introduction of
an established modelling methodology would also address the lack of methodological consistency,
and increase rigour, transparency, and tractability as system dynamics has established rules for
communicating models and results. Fourth, the identification of mechanisms provides actionable
knowledge for how to change the behavior of system elements and thus its trajectory. Fifth, the
integration of system dynamics, equips the MLP for the study of complex, multi-system transitions
to sustainability and thus, keeps it relevant for the future. Moreover, there is an element of
convergence between the foundation of system dynamics as a method and MLP as a framework as
both deal with long term issues and large-scale system change.
Table 1 Attributes of the methodology most relevant for response to MLP critique
Critique Attributes of Methodology
1. MLP is a “heuristic device”. Provide a way for identifying causal mechanisms.
2. Neglect of agency in the MLP Trace agency to multi-level collective agents and mechanisms
3. Lack of methodological consistency and accuracy.
Need to maintain a tractable parsimonious analysis
and methods to map transitions.
Increase methodological consistency and accuracy through
modeling. Provide tractability and parsimony of transitions
4. Need to identify the trajectories that systems take
and reorient them.
Identify the signs of change in system trajectories and provide
actionable knowledge in order to reorient system trajectories.
5. Little integration of systems approaches, methods
and tools in transition research.
Integrate causal loop diagrams and modelling in transition
research. Equip the MLP for multi system transitions to
The contribution of this paper is its proposed a retroductive methodology for transitions research
that combines case study and modelling and simulation. Two issues motivated this research: (i) the
critique the MLP received (Table 1) for which the methodology is proposed to address it and
identify transition mechanisms and patterns, and (ii) the call to equip MLP for multi-regime and
multi-system future sustainability transitions. The proposed methodology has retroduction at its
core, and requires that generative mechanisms for a transition are postulated. Modelling and
simulation is used to test them. Simulation is necessary to study the nature, intensity and timing of
interactions during a transition and identify the signs of system change. The use of mechanisms in
an MLP context goes beyond the provision of a way to categorise events hence, the methodology
overcomes the criticism of Genus and Coles (2008) about the theoretical contribution of the MLP.
The integration of case studies and co-development with modelling and simulation increases
considerably methodological consistency while it maintains the tractability of transition analysis.
Causal loop diagrams and the simulation model provide a record of the assumptions that underlie
transition explanations and enable greater comparability across transition cases. If mechanism based
explanations are identified that are generic across several case studies then this will result in further
integration of theory between different transition pathways, various mechanisms for niche regime
interactions, and geography.
Retroduction along with modelling and simulation provides a means for a shift from historical
research of single system/technology transitions to multi-system/technology sustainability
transitions, which are central to the future research agenda of the MLP. The methodology can be the
basis for the development of principles to bound and measure niches, regimes and landscapes
because it guides the classification of variables that can help operationalize the theoretical concepts
on a case by case basis.
The challenge of modeling and simulation integration in transition research lies in
characteristics that are pertinent to transitions (Köhler et al., 2018): changes in culture, societal
values and political struggles. The combination of case study and simulation requires that
researchers are versed in both strands of research, and then undertake an analysis that is going to be
considerably more time consuming. This effort would be catalysed by the development of a broadly
shared conceptualisation and framework that can act as a channel for dialogue between transition
case and model based work. The consistent application of the proposed methodology to transition
cases is a step in this direction. Moreover, the growing availability of computational social science
related tools is likely to help eradicate the rather unproductive quantitative-qualitative divide
(Keuschnigg et al., 2018).
Finally, the application of proposed methodology should be sensitive to power and political
processes. Researchers engaging in this kind of endeavour should be acutely aware of the politics
and power struggles of transitions at two levels: regarding actor interests that are stake in the
transition and actor interests that are supported through the particular case study conceptualization
and modelling results. Since transition research is a relatively new area, it is hoped that the
proposed methodology will contribute to development of better theory and a more consistent,
rigorous and hence powerful methodological approach in transition studies.
The author would like to thank Frank Geels for valuable feedback on the paper.
Abbott, A. 1990. A primer on sequence methods. Organizational Science 1(4), 375-392.
Abbott, A. 2001. Time Matters: On theory and method. University of Chicago press, Chicago, IL.
Abell, P. 2004. Narrative explanation: an alternative to variance-centered explanation? Annual Review of Sociology 30,
Abernathy, W.J., Townsend, P.L. 1973. Technology, productivity and process change. Technological Forecasting and
Social Change 7, 379-396.
Adamides, E., Papachristos, G., Pomonis, N., 2012. Critical realism in supply chain research: Understanding the
dynamics of a seasonal goods supply chain. International Journal of Physical Distribution & Logistics
Management 42(10), 906 - 930.
Anderson, O.J.J., Blatt, R., Christianson, M.K., Grant, A.M., Marquis, C., Neuman, E.J., Sonenshein, S., Sutcliffe, K.M.
2006. Understanding mechanisms in organizational research: Reflections from a collective journey. Journal of
Management Inquiry, 15, 102-113.
Aramburo, S.A., Castaneda Acevedo, J.A., Morales, Y.O. 2012. Laboratory experiments in the system dynamics field.
System Dynamics Review, 28(1), 94-106.
Archer, M.S., 1995. Realist social theory: The morphogenetic approach. Cambridge University Press, Cambridge.
Archer, M., Bhaskar, R., Collier, C., Lawson, T., Norrie, A. 1998. Critical Realism: Essential Readings. Routledge,
Archer, M.S. 2000. Being human: The Problem of Agency. Cambridge University Press, Cambridge.
Arranz, A.M., 2017. Lessons from the past for sustainability transitions? A meta-analysis of socio-tehnical studies.
Global Environmental Change 44, 125-143.
Astley, W.G., Van de Ven, A.H. 1983. Central perspectives and debates in organizational theory. Administrative Science
Quarterly 28(2), 245-273.
Bacharach, S.B. 1989. Organizational theories: Some criteria for evaluation. Academy of Management Review 14(4),
Bennett, A., Checkel, J.T. (Eds.). 2014. Process Tracing. Cambridge University Press, Cambridge, England.
Bhaskar, R. 2008. A Realist Theory of Science. Verso, London, UK.
Bitektine, A. 2008. Prospective case study design qualitative method for deductive theory testing. Organizational
Research Methods 11(1), 160–180.
Boudon, R. 1998. Social mechanisms without black boxes. In P., Hedström, R., Swedberg (Eds.) Social mechanisms.
An analytical approach to social theory., Cambridge University Press, p 172-203.
Bunge, M. 2004. How does it work? The search for explanatory mechanisms. Philosophy of the Social Sciences 34,
Burton, R.M., Obel B. 2011. Computational modelling for what-is, what-might-be, and what-should-be studies-and
triangulation. Organisation Science 22(5), 1195-1202.
Coenen, L., Benneworth, P., Truffer, B. 2012. Toward a spatial perspective on sustainability transitions. Research Policy
Coleman, J.S., 1986. Social theory, social research, and a theory of action. American Journal of Sociology 91(6), 1309-
Collier, A. 1994. Critical realism: An introduction to the philosophy of Roy Bhaskar. Verson, London.
Corley, K.G., Gioia, D.A. 2011. Building theory about theory building: What constitutes a theoretical constitutes a
theoretical contribution? Academy of Management Review 36(1), 12-32.
Cronin, M., Gonzalez, C., Sterman, J.D. 2009. Why don’t well-educated adults understand accumulation? A challenge
to researchers, educators, and citizens. Organizational Behavior and Human Decision Processes 108(1), 116–
Cronin, M.A., Gonzalez, C. 2007. Understanding the building blocks of dynamic systems. System Dynamics Review
Davis, J.P., Eisenhardt, K.M., Bingham, C.B. 2007. Developing theory through simulation methods. Academy of
Management Review 32(2), 480-499.
De Haan, J., Rotmans, J. 2011. Patterns in transitions: Understanding complex chains of change. Technological
Forecasting and Social Change 78(1), 90-102.
Diehl, E., Sterman, J.D. 1995. Effects of feedback complexity on dynamic decision making. Organizational Behavior
and Human Decision Processes 62(2), 198–215.
DiMaggio, P.J. 1995. Comments on "What theory is not". Administrative Science Quarterly 40(3), 391-397.
Elzen, B., Wieczorek, A. 2005. Transitions towards sustainability through system innovation. Technological Forecasting
and Social Change 72, 651-661.
Epstein, J. 2006. Generative Social Science. Studies in Agent-Based Computational Modeling. Princeton University
Press, Princeton, NJ.
Ford, D.N., Sterman, J.D., 1998. Expert knowledge elicitation to imrpove formal and mental models. System Dynamics
Review 14, 309-340.
Forrester, J.W., 1961. Industrial Dynamics. Cambridge, MA, MIT Press.
Fouquet, R., Pearson, P.J.G. 2012. Past and prospective energy transitions: Insights from history. Energy Policy 50, 1-7.
Foxon, T.J. 2011. A coevolutionary framework for analysing a transition to a sustainable low carbon economy.
Ecological Economics 70, 2258-2267.
Geels, F.W. 2002. Technological transitions as evolutionary reconfiguration processes: a multi-level perspective and a
case-study. Research Policy 31, 1257-1274.
Geels, F.W. 2004. From sectoral systems of innovation to socio-technical systems: Insights about dynamics and change
from sociology and institutional theory. Research Policy 33, 897-920.
Geels, F.W. 2005. Processes and patterns in transitions and system innovations: refining the co-evolutionary multi-level
perspective. Technological Forecasting and Social Change 72, 681-696.
Geels, F.W. 2007. Feelings of discontent and the promise of middle range theory for STS: Examples from technology
dynamics. Science Technology and Human Values 32(6), 627-651.
Geels, F.W. 2010. Ontologies, sociotechnical transitions (to sustainability) and the multi-level perspective. Research
Policy 39, 495 - 510.
Geels, F.W. 2011. The multi-level perspective on sustainability transitions: Responses to seven criticisms.
Environmental Innovation and Societal Transitions 1(1), 24-40.
Geels, F.W., Schot, J. 2007. Typology of sociotechnical transition pathways. Research Policy 36, 399 - 417.
Geels, F.W., Schot, J.W. 2010. The dynamics of transitions: A socio-technical perspective. in: Grin, J., Rotmans, J.,
Schot, J. in collaboration with Geels, F.W. and Loorbach, D., 2010, Transitions to Sustainable Development:
New Directions in the Study of Long Term Transformative Change, Routledge, pp. 9-87.
Geels, F.W., Kern, F., Fuchs, G., Hinderer, N., Kungl, G., Mylan, J., Neukirch, M., Wassermann, S., 2016a). The
enactment of socio-technical transition pathways: A reformulated typology and a comparative multi-level
analysis of the German and UK low-carbon electricity transitions (1990-2014). Research Policy 45, 996-913.
Geels, F.W., Berkhout, F., van Vuuren, D.P. 2016b. Bridging analytical approaches for low-carbon transitions. Nature
Climate Change 6(6), 576–583.
Genus, A., Coles, A.M. 2008. Rethinking the multi-level perspective of technological transitions. Research Policy 37,
Gerring, J. 2005. Causation: A unified framework for the social sciences. Journal of Theoretical Politics 17, 163-198.
Gerring, J. 2010. Causal mechanisms: Yes, but... Comparative Political Studies 43(11), 1499-1526.
Gibson, C.B., Earley, P.C. 2007. Collective cognition in action: Accumulation, interaction, examination and
accommodation in the development and operation of group efficacy beliefs in the workplace. Academy of
Management Review 32(2), 438-458.
Giddens, A., 1984. The Constitution of Society: Outline of the Theory of Structuration. University of California Press,
Gioia, D.A., Pitre, E. 1990. Multiparadigm perspectives on theory building. Academy of Management Review 15(4),
Glennan, S.S. 1996. Mechanisms and the nature of causation. Erkenntnis 44, 49-71.
Goertz, G., Mahoney, J. 2012. A tale of two cultures: Qualitative and quantitative research in the social sciences.
Princeton University Press, Princeton, New Jersey.
Goldspink, C. 2002. Methodological implications of complex systems approaches to sociality: simulation as a
foundation for knowledge. Journal of Artificial Societies and Social Simulation 5(1)
Griffin, L.T. 1993. Narrative, event structure analysis and causal interpretation in historical sociology. American Journal
of Sociology 98(5), 1094-1133.
Hackman, J.R. 2003. Learning more by crossing levels: Evidence from airplanes, orchestras and hospitals. Journal of
Organizational Behavior 24, 1-18.
Hall, J.R. 1999. Cultures of Inquiry: From Epistemology to Discourse in Sociohistorical Research. Cambridge
University Press, Cambridge.
Harre, R. 1970. The principles of scientific thinking. University of Chicago Press, Chicago.
Harrison, J.R., Zhiang, L., Carroll, G.R., Carley, K.M. 2007. Simulation modelling in organizational and management
research. Academy of Management Review 32(4), 1229-1245.
Hedström, P. 2008. Studying mechanisms to strengthen causal inferences in quantitative research, in: Box-
Hedström, P., Bearman, P. 2009. What is analytical sociology all about? An introductory essay. in: Hedström, P.,
Bearman, P., (Eds), The Oxford handbook of analytical sociology. Oxford University Press, pp 3-24.
Hedström, P., Swedberg, R. 1998. Social Mechanisms: An Analytical Approach to Social Theory. Cambridge University
Hedström, P., Ylikovksi, P. 2010. Causal mechanisms in the social sciences. Annual Review of Sociology 36, 49-67.
Hedstrom, P., Bearman, P. 2011. The Oxford handbook of analytical sociology. Oxford University Press, NY.
Helfat, C.E., Finkelstein, S., Mitchell, W., Peteraf, M.A., Singh, H., Teece, D.J., Winter, S.G. 2007. Dynamic
capabilities: Understanding strategic change in organizations. Malden, MA, Blackwell, Publishing.
Holtz, G., Alkemade, F., de Haan, F., Köhler, J., Trutnevyte, E., Luthe, T., Halbe, J., Papachristos, G., Chappin, E.,
Kwakkel, J., Ruutu, S. 2015. Prospects of transition modelling: Position paper of the transition modelling
community. Environmental Innovation and Societal Transitions, 17, 41-58.
Homer, J. 1996. Why We Iterate: Scientific Modeling in Theory and Practice. System Dynamics Review 12(1), 1-19.
Janssen M.A., Radtke, N.P., Lee A. 2009. Pattern-oriented modeling of commons dilemma experiments. Adaptive
Behavior 17(6), 508-523.
Johnson, B.R., Russo, F., Shoonenboom, J. 2017. Causation in mixed methods research: The meeting of philosophy,
science and practice. Journal of Mixed Methods Research, 1-20.
Jick, T. 1979. Mixing qualitative and quantitative methods: Triangulation in action. Administrative Science Quarterly
Kagan, J. 2009. The Three Cultures: Natural Sciences, Social Sciences and the Humanities in the 21st Century,
Cambridge University Press, Cambridge, UK.
Kahneman, D. 2011. Thinking fast and slow. Allen Lane, US.
Kaidesoja, T. 2013. Overcoming the biases of microfoundationalism: Social mechanisms and collective agents.
Philosophy of the Social Sciences 43(3), 301-322.
Keuschnigg, M., Lovsjo, N., Hedstrom, P., 2018. Analytical sociology and computational social science. Journal of
Computational Social Science 1(1, 3-14.
Köhler, J., de Haan, F., Holtz, G., Kubeczko, K., Moallemi, E.A., Papachristos, G., Chappin, E., 2017. Modelling
Sustainability Transitions: An assessment of approaches and challenges. Journal of Artificial Societies and Social
Simulation 21(1), 8.
Lane, D.C., 1999. Social theory and system dynamics practice. European Journal of Operational Research 113, 501-
Lane, D.C., 2008. The emergence and use of diagramming in system dynamics: a critical account. Systems Research
and Behavioral Science 25, 3-23.
Langley, A. 1999. Strategies for theorizing from process data. Academy of Management Review 24(4), 691 – 710.
Langley, A. 2007. Process thinking in strategic organization. Strategic Organization 5,271-282.
Langley, A., Smallman, C., Tsoukas, H., Van de Ven, A.H. 2013. Process studies of change in organization and
management: Unveiling temporality, activity, and flow. Academy of Management Journal 56(1), 1-13.
Laurisden, E.H., Jørgensen, U. 2010. Sustainable transition of electronic products through waste policy. Research Policy
Lerner, R.M., Kauffman, M.B. 1985. The concept of development in contextualism. Developmental Review 5, 309 –
Luna-Reyes, L.F., Andersen, D.L. 2003. Collecting and analysing qualitative data for system dynamics: Methods and
models. System Dynamics Review 19(4), 271–296.
Machamer, P., Darden, L., Craver, C.F. 2000. Thinking about mechanisms. Philosophy of Science, 67(1), 1-25
Mahoney, J. 2000. Strategies of causal inference in small-N analysis. Sociological Methods and Research, 28, 387-424.
Manicas, P.T. 2006. A Realist Philosophy of Social Science: Explanation and Understanding. Cambridge University
Press, Cambridge, U.K.
Markard, J., Raven, R., Truffer, B. 2012. Sustainability transitions: An emerging field of research and its prospects.
Research Policy, 41, 955-967.
Markard, J., Truffer, B. 2008. Technological Innovation systems and the multi-level perspective: Towards an integrated
framework. Research Policy, 37, 596-615.
Martinez-Moyano, I.J., Richardson, G.P., 2013. Best practices in system dynamics modeling. System Dynamics Review
Mayntz, R. 2004. Mechanisms in the analysis of social macro-phenomena. Philosophy of the Social Sciences, 34, 237-
McAdam, D., Tarrow, S., Tilly, C. 2001. Dynamics of Contention. Cambridge University Press.
McMeekin, A., Southerton, D. 2012. Sustainability transitions and final consumption: practices and socio-technical
systems. Technology Analysis and Strategic Management, 24(4), 345-361.
Merton, R.K. 1968. Social Theory and Social Structure. The Free Press, New York.
Merton, R.K., Rossi, A.S. 1968. Contributions to the Theory of Reference Group Behavior. p. 279-334 in Social Theory
and Social Structure. New York: The Free Press.
Mingers, J. 2014. Systems Thinking, Critical Realism and Philosophy: A Confluence of Ideas. Routledge, Oxon, UK.
Mingers, J., Standing, C., 2017. Why things happen – Developing the critical realist view of causal mechanisms.
Information and Organization 27(3), 171-189.
Mintzberg, H. 2005. Developing theory about the development of theory. in: K.G., Smith, M.A., Hitt (Eds.), Great
minds in management: The process of theory development, 355-372, Oxford University Press.
Mitroff, I., Emshoff J. 1979. On strategic assumption making: a dialectical approach to policy and planning. Academy
of Management Review 4(1), 1 – 12.
Mohr, L.B. 1982. Explaining organizational behaviour. Jossey-Bass, San Fransisco.
Morgan, G. 1983. Beyond method. Sage, Newbury Park, CA.
Naber, R., Raven, R., Kouw, M., Dassen, T., 2017. Scaling up sustainable energy innovations. Energy Policy 110, 342-
Nuzzo, R., 2015. How scientists fool themselves - and how they can stop. Nature 526(7572), 182-185.
Opp, K.D. 2005. Explanations by mechanisms in the social sciences. Problems, advantages and alternatives. Mind and
Society 4, 163-178.
Oreskes, N., Shrader-Frechette, K., Belitz, K. 1994. Verification, validation and confirmation of numerical models in
the earth sciences. Science 263(5147), 641-646.
Papachristos, G. 2011. A System Dynamics Model of Socio-technical Regime Transitions. Environmental Innovation
and Societal Transitions 1(2), 202 – 233.
Papachristos, G., 2012. Case study and system dynamics research: Complementarities, pluralism and evolutionary
theory development. 30th International Conference of the System Dynamics Society, St Gallen, Switzerland
Papachristos, G. 2014a. Towards multi-system sociotechnical transitions: why simulate. Technology Analysis and
Strategic Management 26(9), 1037-1055.
Papachristos, G., 2014b. Transition inertia due to competition in supply chains with remanufacturing and recycling: A
system dynamics model. Environmental Innovation and Societal Transitions 12, 47-65.
Papachristos, G., Sofianos, A., Adamides, E. 2013. System interactions in socio-technical transitions: Extending the
multi-level perspective. Environmental Innovation and Societal Transitions 7, 53 - 69.
Papachristos, G., Adamides, E. 2016. A retroductive systems-based methodology for socio-technical transitions
research. Technology Forecasting and Social Change 108, 1-14.
Papachristos, G., 2017. Diversity in technology competition. The link between platforms and sociotechnical transitions.
Renewable and Sustainable Energy Reviews 73, 291-306.
Papachristos, G. 2018. A mechanism based transition research methodology: Bridigng analytical approaches. Futures
article in press. https://doi.org/10.1016/j.futures.2018.02.006
Pentland, B.T. 1999. Building process theory with narrative: from description to explanation. Academy of Management
Review 24, 711-724
Pedriana, N. 2005. Rational choice, structural context and increasing returns: A strategy for analytic narrative in
historical sociology. Sociological Methods and Research 33, 349-382.
Pettigrew, A.M. 1997. What is a processual analysis. Scandinavian Journal of Management 13, 37-348.
Pfeffer, J. 1982. Organizations and organization theory. Pitman, Boston.
Pidd, M. 2004. Bringing it all together, in: Pidd, M. (Ed.), Systems modelling: Theory and practice. Wiley, Chichester,
Poole, M.S., Van de Ven, A.H. 2004. Handbook of organizational change and innovation. Oxford University Press, New
Rahmandad, H., Sterman, J. 2008. Heterogeneity and network structure in the dynamics of diffusion: Comparing agent-
based and differential equation models. Management Science 54(5), 998-1014.
Rahmandad, H., Sterman, J.D. 2012. Reporting guidelines for simulation based research in social sciences. System
Dynamics Review 28(4), 396-411.
Railsback, S.F., Johnson, M.D. 2011. Pattern oriented modelling of bird foraging and pest control in coffee farms,
Ecological Modelling 222(18), 3305-3319.
Raven, R. 2007. Niche accumulation and hybridisation strategies in transition processes towards a sustainable energy
system: An assessment of differences and pitfalls. Energy Policy 35, 2390-2400.
Richardson, G.P., Pugh III, A.L. 1981 Introduction to System Dynamics Modeling with Dynamo. MIT Press,
Richardson, K. 2005. The hegemony of the physical sciences: an exploration in complexity thinking. Futures 37, 615-
Rip, A., Kemp, R. 1998. Technological change. In (eds) S. Rayner, E.L. Malone, Human Choice and Climate Change,
vol. 2. Battelle Press, Columbus, OH, 327-399.
Rogers, E. 2003. The diffusion of innovations. 5th edition, Free Press.
Rogelj, J., Luderer, G., Pietzcker, R.C., Kriegler, E., Schaeffer, M., Krey, V., Riahi, K. 2015. Energy system
transformations for limiting end-of-century warming to below 1.5C. Nature Climate Change 5, 519-527.
Sayer, A. 1992. Method in social science: A realist approach. Routledge, UK.
Sayer, A. 2000. Realism and social science. Sage Publications, UK.
Scharpf, F.W. 1997. Games real actors play: Actor-centered institutionalism in policy research. West Press, Boulder, US.
Schelling, T.C. 1978. Micromotives and Macrobehavior. New York: W. W. Norton.
Shove, E., Walker, G. 2007. Caution! Transitions ahead: politics, practice, and sustainable transition management.
Environment and Planning A 39, 763-770.
Siggelkow, N. 2007. Persuasion with case studies. Academy of Management Journal 50(1), 20-24.
Smith, A., Stirling, A., Berkhout, F. 2005. The governance of sustainable sociotechnical transitions. Research Policy 34,
Smith, A., Voß, J.P., Grin, J. 2010. Innovation studies and sustainability transitions: The allure of the multi-level
perspective and its challenges. Research Policy 39, 435-448.
Sterman, J.D, Sweeney, L.B. 2007. Understanding public complacency about climate change: adults’ mental models of
climate change violate conservation of matter. Climatic Change 80(3–4), 213-238.
Sterman, J.D. 1989a. Misperceptions of feedback in dynamic decision making. Organizational Behaviour and Human
Decision Processes 43(3), 301-335.
Sterman, J.D. 1989b. Modelling managerial behaviour: Misperceptions of feedback in a dynamic decision making
experiment, Management Science 35(3), 321-339.
Sterman, J.D. 1994. Learning in and about complex systems. System Dynamics Review 10, 291-330.
Sterman, J.D. 2000. Business Dynamics: systems thinking and modelling for a complex world. McGraw Hill, New
Sterman, J.D. 2008. Risk Communication on climate: mental models and mass balance. Science 322, 532-533.
Sterman, J.D. 2010. Does formal system dynamics straining improve people's understanding of accumulation?. System
Dynamics Review, 26(4), 316-334.
Stinchcombe, A.L. 1998. Monopolist Competition as a Mechanism: Corporations, Universities, and Nation-States in
Competitive Fields. In (eds) P. Hedström, R. Swedberg, Social Mechanisms: An Analytical Approach to Social
Theory, Cambridge University Press, 267-305.
Sorrell, S., 2018. Explaining sociotechnical transitions: A critical realist perspective. in press Research Policy
Sutton, R.I., Staw, B.M. 1995. What theory is not. Administrative Science Quarterly, 40(3), 371-384.
Svensson, O., Nikoleris, A., 2018. Structure reconsidered: Towards new foundations of explanatory transitions.
Research Policy 47, 462-473.
Tashakkori, A., Teddlie, C. 1998. Mixed Methodology. Combining Qualitative and Quantitative Approaches. Sage,
Tilly, C. 2001. Mechanisms in political processes. Annual Review of Political Science, 3, 21-41.
Torres, P.J. 2009. A modified conception of mechanisms. Erkenntnis, 71, 233-251.
Trochim, W.M.K. 1989. Outcome pattern matching and program theory. Evaluation and Program Planning 12, 355-366.
Tsoukas, H. 1989. The validity of idiographic research explanations. Academy of Management Review 14(4), 551-561.
Tukker, A. 2015. Product services for a resource-efficient and circular economy - a review. Journal of Cleaner
Production 97, 76-91.
Turnheim, B., Geels, F.W. 2012. Regime destabilisation as the ﬂipside of energy transitions: lessons from the history of
the British coal industry (1913–1997). Energy Policy 50, 35–49.
Turnheim, B., Geels, F.W., 2013. The destabilisation of existing regimes: Confronting a multi-dimensional framework
with a case study of the British coal industry (1913-1967). Research Policy 42, 1749-1767.
Turnheim, B., Berkhout, F., Geels, F., Hof, A., McMeekin, A., Nykvist, B., van Vuuren, D. 2015. Evaluating
sustainability transitions pathways: Bridging analytical approaches to address governance challenges. Global
Environmental Change 35, 239-253.
Van de Ven, A.H. 1992). Suggestions for studying strategy process: a research note. Strategic Management Journal 13,
Van de Ven, A.H. 2007. Engaged scholarship: A guide for organizational and social research. Oxford, Oxford University
Van den Bergh, J.C.J.M., 2011. Environment versus growth - A growth of 'degrowth' and a plea for 'a-growth'.
Ecological Economics 70,881-890.
Van den Bergh, J.C.J.M., Truffer, B., Kallis, G. 2011. Environmental innovation and societal transitions. Environmental
Innovation and Societal Transitions 1, 1-23.
Vasileiadou, E., Safarzynska, K. 2010. Transitions: Taking complexity seriously. Futures 42, 1176-1186.
Voss, J., Bauknecht, D., Kemp, R., (Eds.), 2006. Reflexive Governance for Sustainable Development. Cheltenham,
Weick, K. 1979. The social psychology of organizing. Reading, MA, Addison-Wesley.
Weick, K.E. 1989. Theory construction as disciplined imagination. Academy of Management Review 14(4), 516-531.
Weick, K.E., Roberts, K.H. 1993. Collective mind in organizations Heedful interrelating on flight decks. Administrative
Science Quarterly 38(3), 357-381.
Weick, K.E. 1995. What theory is not theorizing is. Administrative Science Quarterly 40(3), 385-390.
Whetten, D.A. 1989. What constitutes a theoretical contribution. Academy of Management Review 14(4), 490-495.
Wilson, J.M., Goodman, P.S., Cronin, M.A. 2007. Group learning. Academy of Management Review 32(4), 1041-1059.
Winsberg, E. 2006. Models of success versus the success of models: Reliability without truth. Synthese 152, 1-19.
Wolstenholme, E. 1990. System Enquiry: A System Dynamics Approach. Wiley, Chichester.
Woodward, J. 2002. What is a mechanism? A counterfactual account. Philosophy of Science 69, 366-377.
Wuisman, J.J.J.M. 2005. The logic of scientific discovery in critical realist social scientific research. Journal of Critical
Realism 4(2), 366-394.
Yin, R.K. 2003. Case study research: Design and methods, 3rd edition, Sage Publications, CA, US.