Bringing Together Linked Coding and Realist Thematic Analysis: A Worked Example for Initial [Programme] Theory Development

Introduction

Eliciting a [programme] theory is a foundational step in realist inquiry (Smeets et al., 2021). As a theory-driven approach, realist inquiries (research, synthesis, and evaluation) begin and end with a theory that can be tested and refined through empirical data. Social programmes embody assumptions about how change is supposed to happen, for whom, and under what conditions. Realist approaches are increasingly adopted in health and social policy research due to their potential to unpack complex phenomena and interventions in open systems (Marchal et al., 2018; Mukumbang et al., 2024). As the field grows, questions persist regarding how best to develop [programme] theories in ways that are methodologically robust, transparent, and practically applicable (McEwan et al., 2023). There is growing recognition for the need to meaningfully engage with interest group holders to take up the findings of realist theorising. This means consistently engaging in dialogue to validate or refute working theories. Interest group holders are individuals with professional or lived experience relevant to the topic of study. It is a term derived from the MuSE consortium network, and advocates for inclusive and decolonial language and reflexivity around groups that are affected by colonization, thus we deliberatively do not use the term ‘stakeholder’ (Akl et al., 2024). This paper reports on the theory-elicitation from Phase 1 of a five-year participatory realist project focused on integrated mental health care for refugees, drawing on data generated through a realist workshop, six sector-specific deliberative dialogues (n = 24), and an in-person World Café event (n > 30) conducted in British Columbia between 2022 and 2024.

While there is broad agreement on the importance of [programme] theories in realist inquiries, there is no single agreed-upon method for constructing them. This lack of prescription is not a weakness. Indeed, it reflects a realist concern with adhering to ontological and epistemological principles rather than enforcing a fixed approach (Salter & Kothari, 2014). The RAMESES II reporting standards offer valuable guidance on good practice, such as making [programme] theories explicit, realist in form, and grounded in evidence, but deliberately avoid mandating a specific analytic pathway (Wong et al., 2016). As a result, researchers are exposed to a multiplicity of approaches, frameworks, and tools, often leading to confusion about where to begin (Renmans & Pleguezuelo, 2022; Shearn et al., 2017). In practice, a range of methods have emerged for identifying, structuring, and refining [programme] theories from interest group holder engagement and interviews to theory-driven realist reviews (Fick & Muhajarine, 2019; Peters, 2024). This pluralism is a strength and a challenge. Without clear exemplars, it can be difficult for researchers to navigate the early stages of realist inquiry with confidence and rigour.

To understand the importance of [programme] theory development within realist inquiries, it is essential to consider their philosophical and methodological groundings. Rooted in a realist philosophy of science, the realist approach assumes that reality exists independently of our perceptions, where unobservable generative mechanisms produce observable outcomes (Mingers & Standing, 2017; Sankey, 2021). It views causation as contingent and context-dependent, occurring through the interaction of mechanisms with enabling or constraining contextual conditions (Greenhalgh & Manzano, 2021). The aim of realist inquiries is to both describe what works, and to explain how and why it works (or does not), for whom, and in which contexts (Howe et al., 2025; Pawson et al., 2005). These explanations are commonly articulated through context–mechanism–outcome (CMO) configurations, which serve as building blocks for the development of realist [programme] theories (Dalkin et al., 2015; De Weger et al., 2020).

In realist inquiries, there are different types of theories, and developing the Initial [Programme] Theory (IPT) is the crucial first step. IPTs are a set of initial ideas and assumptions about how and why a programme is intended to work, and for whom, under what circumstances (Hounsou et al., 2025). All programmes, defined as structured resources and/or the facilitation of providing those resources, be it material or immaterial (e.g., knowledge), whether explicitly stated or not, possess an underlying theory that researchers are tasked with making explicit (Goodridge et al., 2015; Trevisan & Walser, 2015). The IPT functions as a central component of realist research, aiming to provide plausible explanations for an intervention’s success or failure in specific contexts. Realist research is theory-driven, shifting the evaluative focus from merely “does it work” to the more complex question of “what works, for whom, in what contexts, and how” (Greenhalgh et al., 2015). By sketching the terrain to be investigated, the IPT helps refine the scope of the realist inquiry and guides subsequent research design and data collection methods (Flynn et al., 2020). The IPT is then empirically tested and progressively refined throughout the research process with the help of key interest group holders through iterative consultations, ultimately contributing to the development of transferable middle-range theories (Mukumbang & Wong, 2025). IPTs are thus fundamental to generating insights into how outcomes are produced and are a crucial step in the ongoing process of explanation and theory development within realist research (Pawson, 2003).

This process of explanation requires reasoning strategies that move beyond description. Retroduction, the central realist-informed inference-making approach, allows researchers to work backward from observed outcomes to infer the mechanisms that could plausibly have produced them (Jagosh, 2020; Mukumbang, 2021). Similarly, realist analysis is guided by a depth ontology that distinguishes between the empirical (what is observed or reported in the data), the actual (what occurs in practice, whether observed or not), and the real (the underlying causal mechanisms or structures that generate these events) (Bhaskar, 1978). For example, in health services research, participants’ accounts of acting under pressure may be treated as empirical data, recurring patterns of practice interpreted at the actual level, and then explained at the real level by inferring underlying mechanisms such as professional obligation or fear of failure. Retroductive thinking supports realist inquiry and addresses the complexity and non-linearity typical of health and social interventions (Moore et al., 2018; Skivington et al., 2021).

It is this complexity that leads to the need for explicit analytical strategies in realist inquiries. For example, Nielsen and Lemire (2024) argue that realist analyses should be transparent, theory-led, and clearly aligned with the aims of the project from the outset. This means intentionally selecting which contexts, mechanisms, and outcomes will be investigated and why; articulating how reasoning strategies such as induction, abduction, and retroduction will be employed; and ensuring that theory construction remains visible and iterative throughout the analysis process. Some studies have already provided exemplar cases that offer methodological guidance at the IPT stage and beyond (Eastwood et al., 2022; Flynn et al., 2020; Hounsou et al., 2025; Smeets et al., 2021). While such calls and examples have improved the methodological clarity of later phases in realist research, there remains a relative lack of practical guidance showing how analytical techniques can be combined and applied specifically during the initial phase of [programme] theory elicitation (Shearn et al., 2017).

This paper responds to that gap by proposing and demonstrating a structured, phased approach to realist theory elicitation and analysis that combines two complementary methods: linked coding and realist thematic analysis. While the broader project is informed by Pawson and Tilley’s (1997) realist evaluation principles, the analytic approach deliberately draws on the strengths of both scientific and critical realism. Specifically, linked coding and the use of CMO heuristics support pragmatic, theory-driven analysis, while the depth ontology and explanatory focus of critical realism strengthen causal reasoning. Together, this integration advances methodological practice across both realist traditions (Mukumbang et al., 2023). Drawing on a recent study focused on developing a [programme] theory for integrated mental health care for refugees in Canada (Clark et al., 2026), we outline a five-phase analytical strategy: (1) preparation and team alignment; (2) linked dyad and triad coding; (3) realist thematic analysis across experiential, inferential, and dispositional themes; (4) Intervention-Context-Actor-Mechanism-Outcome (ICAMO) configuration development through abductive synergy; and (5) initial [programme] theory development via retroduction. The aim of this paper is to demonstrate that the strengths of established realist-informed approaches can be productively combined to develop a rigorous, transparent, and theory-led analytical strategy. We thus offer a replicable pathway for other researchers seeking to generate and refine [programme] theory from primary qualitative data, with a focus on conceptual clarity, collaborative reasoning, and analytic depth.

Project Background

This paper details the robust realist theory-building. Three interlinked knowledge translation activities were implemented between 2022 and 2024: a realist orientation workshop, sector-specific online deliberative dialogues, and an in-person World Café event for a participatory realist project titled Promoting Integrated Mental Health Care Services and Supports for Refugees in British Columbia. Participatory realist is the overall framework guiding the work which includes interest group holders’ identification of a policy and practice problem. In realist research, researchers must engage interest group holder perspectives on drivers of change, privileging their knowledge and enhancing rigor of findings. Interest group consultation throughout the processes is necessary for knowledge uptake (i.e., barriers and facilitators to uptake) of that knowledge (Jagosh et al., 2012). This project seeks to understand how integrated mental health care services for refugees operate across diverse services in British Columbia by identifying and analysing [programme] theories through the iterative creation and refinement of ICAMO configurations.

Methods

This section outlines how we combined the strengths of Jackson and Kolla’s dyads and triads (2012), Wiltshire and Ronkainen’s (2021) realist approach to thematic analysis and Mukumbang et al.’s (2018) ICAMO configurations as part of a qualitative primary data collection and analysis strategy to reach a refined IPT. ICAMO was selected for its ability to explicitly articulate Intervention and Actor elements, which are critical in this refugee mental health services context. Steps of the analytical process are outlined below (see Figure 1). The examples presented are fictitious but preserve the real analytical logic of the study to protect participant confidentiality while demonstrating our methods. The study received ethics approval from the University of British Columbia and the University of Victoria, BC REB Number: H22-03195) and all participants provided informed consent.

OPEN IN VIEWER

Figure 1.

Clark et al. (2026) analysis

Preparation Phase: Analytical Alignment

Before initiating the first phase, the team (AAB, NC, MH and FCM) met twice via Microsoft Teams to code an extract from one of the deliberate dialogues together. Two key strategies guided this refinement (1) bi-weekly team meetings provided a space to discuss emerging findings in relation to their original context (e.g., specific dialogue sessions) and begin identifying generative mechanisms and (2) theory validation involved assessing the plausibility and explanatory power of emerging ICAMO configurations. These processes were informed by the team’s previous realist experience, (AAB has 3 years of realist research experience, including leading and contributing to several completed and ongoing realist projects and undertaking training and collaborations with realist researchers. MH has 2 years of realist research experience, informed by professional and lived motivations related to structural discrimination and health inequities), in situ practitioner knowledge (particularly that of NC, who brings 30 years’ experience in community mental health nursing, including 15 years working specifically in refugee mental health services and with refugee communities and participatory realist expertise), methodological guidance and supervision (FCM brings extensive realist methodological expertise to evaluate mental health services and policies), congruence with existing literature, and resonance with frontline experiences. A consensus document was used to assess and validate dispositional statements and refine the ICAMO logic. This allowed the team to clarify questions related to coding and extraction, and to ensure alignment in the realist process. It also helped establish roles, expectations, and timelines.

Table 1 presents a checklist of guiding questions designed to support pre-analysis planning in realist research projects. This list was developed retrospectively, based on a review of the elements that proved most critical for achieving team consensus. While not exhaustive, these questions were instrumental in promoting a shared understanding and preparing the team for the complexities of realist analysis.

OPEN IN VIEWER

Table 1.

Preparation Phase Alignment Checklist

Question checklist for preparation phase	Goal
1. Are we clear on what we mean by realist analysis?	Create space for discussion, promotes shared understanding, and ensures alignment in expectations and methodology.
2. Can we define what we mean by ICAMO?	Establish a shared definition across ICAMO to prevent ambiguity and ensure consistency throughout the analysis.
3. How does retroduction differ from other reasoning approaches and what does it look like?	Establish unified understanding on how retroduction differs from conventional qualitative methods or reasoning and how retroduction will be applied during data interpretation.
4. What are the timeline expectations for everyone?	Set realistic expectations regarding the time needed for analysis, reflection, and revisiting the data to ensure rigorous outputs.
5. What are everyone’s roles, and what is expected of each other?	Clarify individual roles and expectations helps to leverage each member’s strengths across a diverse team.
6. What system will we be using to share files and progress?	Agree on a shared platform (e.g., Microsoft Teams, Google Drive) to ensure transparency, version control, and coordinated access to project materials.
7. Does anyone have any questions?	Address uncertainties early to support smoother collaboration and minimise disruptions.

Phase 1: Jackson and Kolla (2012) Dyads and Triads

The deliberate dialogue transcripts were coded independently by two researchers (AAB and MH) using the dyad and triad method, a realist-informed approach to data analysis that focuses on identifying and refining causal explanations by examining dyads (two linked elements, such as context–mechanism or mechanism–outcome) and triads (full context–mechanism–outcome configurations). By iteratively analysing these relationships, researchers can build, test, and refine [programme] theories in a structured yet flexible way that accommodates complexity and partial data. A key strength of this approach is that it avoids “cataloguing” i.e., coding context, mechanism, and outcomes separately without configuration (Pawson & Manzano-Santaella, 2012). This phase enabled the team to think retroductively throughout the analysis and begin theory building and refinement from the outset.

Table 2 presents a worked example to illustrate Phase 1. The example, while fictitious, is based on a plausible scenario involving a refugee mental health and substance use caseworker, referred to as Alex. Alex operates largely independently in the absence of institutional backing, which compels Alex to “take matters into their own hands,” driven by a strong personal responsibility towards their clients, who would otherwise be underserved. The coded data thus demonstrates a dyad, that is, the relationship between contextual conditions (e.g., lack of systemic resources) and the underlying mechanism (i.e., personal responsibility) that drives action, even in the face of adversity. Notably, at this stage of analysis, the Intervention and Actor components are not explicitly specified, as the original dyads and triads focus primarily on identifying key CMO configurations.

OPEN IN VIEWER

Table 2.

Linked Coding Example From Fictitious Transcript

“Most days, it feels like I’m the only one holding things together. There’s no real support from the system, no proper referral pathways (C), just me trying to do what I can (M). I can’t just sit back and let these clients fall through the cracks. Someone has to step in, I am responsible (M).”

Note. The highlighted extract includes a CM dyad, in this case, the contextual conditions of lack of systemic resources and the personal responsibility activated despite the challenge.

Phase 2.1: Realist Thematic Analysis – Experiential Themes

Wiltshire and Ronkainen’s (2021) realist thematic analysis integrates realist philosophy with thematic analysis by organising themes across three ontological levels: experiential (what participants say and feel), inferential (what we can plausibly interpret or generalise), and dispositional (underlying causal powers or mechanisms). This approach employs multiple reasoning strategies (including inductive, abductive, deductive and retroductive thinking) to generate explanatory themes and aims to bridge the gap between qualitative depth and realist causal explanation, offering a methodologically pluralist alternative to reflexive and codebook TA (Braun & Clarke, 2006). Table 3 highlights the definition of inductive, abductive and retroductive reasoning as applied in this study.

OPEN IN VIEWER

Table 3.

Definition of Inductive, Deductive, Abductive and Retroductive Reasoning

Term	Definition	Reference
Inductive reasoning	Deriving general patterns or propositions from observed data or experiences	Hayes et al. (2010)
Deductive reasoning	Applying existing theories or propositions to examine specific cases or data	Bergdahl & Berterö (2014)
Abductive reasoning	Generating plausible explanatory ideas by reinterpreting data using new or existing concepts	Karlsen et al. (2020)
Retroductive reasoning	Reasoning backward from observed outcomes to infer the underlying mechanisms or structures that could have produced them.	Jagosh (2020); Mukumbang (2021)

Experiential themes were initially developed by AAB and MH using the dyads and triads identified in Phase 1, by collectively reading and discussing the data to identify themes that closely represented participants’ experiences. NC provided contextual knowledge to enhance understanding of service settings and policies. AAB and MH met weekly to discuss progress across each transcript, with all experiential themes included in a master list (and visual Miro board) to observe similarities and differences within and between cases (Pawson, 2024). Deductive reasoning was used to identify recurring themes and new data. At this stage, quotes were extracted as evidence of the experiential theme and dyads and triads were used to evidence the existence of the mechanism and context relationship.

Table 4 presents the format used in Microsoft Word by the two analysts in this study to document the development and progression of the analysis. At this stage, early dyads and triads were translated directly into the analytical table to reflect emerging interpretations. For the purposes of this example there are two quotes and theoretical support. Yet, with more data, an experiential theme could cluster ideas to reflect “many participants suggested…” instead. It is expected, and methodologically appropriate, for the experiential themes and selected quotes to differ slightly between analysts at this point. These initial discrepancies reflect individual perspectives and were addressed through team discussion, adjudication and iterative refinement in subsequent stages of the analysis.

OPEN IN VIEWER

Table 4.

Realist Thematic Analysis Experiential Theme Example

Analyst	Experiential themes	Illustrative quotes and theoretical support
1	The participants in this dialogue described how the absence of systemic support motivated them to take initiative and advocate for their clients, ensuring they received the mental health care they needed.	“Most days, it feels like I’m the only one holding things together. There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible.” – Alex, Lines 346-356
		“It’s exhausting, yeah. But it still feels like my responsibility. These are people who need care, and I can’t just step back because the system’s struggling.”- Alex, Lines 542- 546
		There is theoretical support from moral vocation and care ethics frameworks helped interpret this sense of responsibility as an identity-based moral obligation rather than a situational response (Dik & Duffy, 2007)
2	The participants in this dialogue described how limited resources, both within the system and through referral pathways, did not discourage them, but instead reinforced their sense of responsibility to provide care and take action despite the challenging circumstances.	“There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible. I need to act now before it’s too late” – Alex, Lines 350 – 359
		“These are people who need care, and I can’t just step back because the system’s struggling. Someone has to act, and I rather do it myself than wait until it is too late”- Alex, Lines 544- 548
		Moral vocation theory supports this as a moral obligation rather than motivated by the working environment (Weber, 1978).

Phase 2.2: Realist Thematic Analysis – Inferential Themes

After experiential themes and quotes were developed and extracted, the analysis moved beyond surface-level accounts to develop inferential themes. These themes identify patterns or explanations that could plausibly apply beyond the dataset. AAB and MH worked independently on each transcript using inductive reasoning and abductive reasoning. This helped connect participants’ expressions to wider ideas, enabling deeper exploration and adjudication between potential explanations.

Table 5 presents examples at the level of inferential themes, where both analysts independently explored potential explanations for the observed phenomena. In this case, the focus was on understanding what drives caseworkers’ motivation despite the absence of systemic support and limited resources. At this stage of analysis, the analysts are encouraged to move beyond surface-level observations to infer possible explanations that contribute to the patterns identified at the experiential level.

OPEN IN VIEWER

Table 5.

Realist Thematic Analysis Inferential Theme Example

Analyst	Experiential themes	Illustrative quotes and theoretical support	Inferential theme
1	The participants in this dialogue described how the absence of systemic support motivated them to take initiative and advocate for their clients, ensuring they received the mental health care they needed.	“Most days, it feels like I’m the only one holding things together. There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible.” – Alex, Lines 346-356	It is plausible to claim that workers operating in high-pressure environments feel compelled to act rather than dwell on systemic shortcomings. Their sense of responsibility drives them to do what they can, often perceiving that without their intervention, the system would fail those in need. This perceived necessity serves as a motivating force enabling them to persist despite challenging circumstances.
		“It’s exhausting, yeah. But it still feels like my responsibility. These are people who need care, and I can’t just step back because the system’s struggling.”- Alex, Lines 542- 546
		There is theoretical support from moral vocation and care ethics frameworks helped interpret this sense of responsibility as an identity-based moral obligation rather than a situational response (Dik & Duffy, 2007)
2	The participants in this dialogue described how limited resources, both within the system and through referral pathways, did not discourage them, but instead reinforced their sense of responsibility to provide care and take action despite the challenging circumstances.	“There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible. I need to act now before it’s too late” – Alex, Lines 350 – 359	It is plausible to claim that caseworkers have become desensitised to the lack of resources, viewing systemic inadequacies as an inherent and enduring feature of their working environment. As a result, their professional identity and sense of responsibility often override potential frustration, as they recognise that both their role and their clients’ well-being depend on their continued commitment.
		“These are people who need care, and I can’t just step back because the system’s struggling. Someone has to act, and I rather do it myself than wait until it is too late”- Alex, Lines 544- 548
		Moral vocation theory supports this as a moral obligation rather than motivated by the working environment (Weber, 1978).

Phase 2.3: Realist Thematic Analysis – Dispositional Themes

The final stage of the analysis aimed to arrive at dispositional themes. These explain the underlying causal powers or mechanisms that must exist for observed patterns to occur. This stage used retroductive reasoning, asking: What must be true about the world for this pattern to make sense? We distinguished inferential themes from dispositional themes using a simple decision rule. Inferential themes summarise what is happening in the data, whereas dispositional themes explain what makes those patterns possible. Inferential themes therefore involved conceptual explanations that remained close to participants’ accounts, while dispositional themes involved retroductive propositions about underlying causal powers, mechanisms, or structures that could generate the observed patterns and may not be directly evidenced in the data. AAB and MH worked independently to arrive at dispositional themes. Due to their more abstract nature, AAB and MH met weekly to validate and question the themes, linking them back to theory and the purpose of the work. This reflexivity between the analysts helped unearth, interpret and adjudicate between theories (Downey et al., 2024). This phase formed the core of the analysis by examining data across different ontological levels and explicitly articulating generative causation through retroductive reasoning.

Table 6 presents potential dispositional themes generated independently by Analysts 1 and 2 during the analytical process. At this stage, validation and triangulation across datasets become critical, as the analysis involves a higher degree of abstraction and extrapolation. This includes reasoning beyond what is explicitly stated in the data, drawing on plausible but unobservable explanations, or shifting from granular observations to broader conceptual themes. Regular team discussions were essential to ensure that all dispositional themes remained grounded in the lived experiences of participants and interest group holders, supported by the expert perspective of NC.

OPEN IN VIEWER

Table 6.

Realist Thematic Analysis Dispositional Theme Example

Analyst	Experiential themes	Illustrative quotes and theoretical support	Inferential theme	Dispositional theme
1	The participants in this dialogue described how the absence of systemic support motivated them to take initiative and advocate for their clients, ensuring they received the mental health care they needed.	“Most days, it feels like I’m the only one holding things together. There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible.” – Alex, Lines 346-356	It is plausible to claim that workers operating in high-pressure environments feel compelled to act rather than dwell on systemic shortcomings. Their sense of responsibility drives them to do what they can, often perceiving that without their intervention, the system would fail those in need. This perceived necessity serves as a motivating force enabling them to persist despite challenging circumstances.	The inferred phenomenon is dependent upon the existence of a moral duty activated through ones own professional identity. This ironically can mask systemic problems, as individual heroics maintain minimal function in broken systems.
		“It’s exhausting, yeah. But it still feels like my responsibility. These are people who need care, and I can’t just step back because the system’s struggling.”- Alex, Lines 542- 546
		There is theoretical support from moral vocation and care ethics frameworks helped interpret this sense of responsibility as an identity-based moral obligation rather than a situational response (Dik & Duffy, 2007)
2	The participants in this dialogue described how limited resources, both within the system and through referral pathways, did not discourage them, but instead reinforced their sense of responsibility to provide care and take action despite the challenging circumstances.	“There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible. I need to act now before it’s too late” – Alex, Lines 350 – 359	It is plausible to claim that caseworkers have become desensitised to the lack of resources, viewing systemic inadequacies as an inherent and enduring feature of their working environment. As a result, their professional identity and sense of responsibility often override potential frustration, as they recognise that both their role and their clients’ well-being depend on their continued commitment.	The inferred phenomenon is dependent upon the existence of an adaptive workforce with an internalised responsibility that take on the emotional and ethical burden of the broken system, which in turn avoids reform.
		“These are people who need care, and I can’t just step back because the system’s struggling. Someone has to act, and I rather do it myself than wait until it is too late”- Alex, Lines 544- 548
		Moral vocation theory supports this as a moral obligation rather than motivated by the working environment (Weber, 1978).

Phase 3: Abductive Synergy, Explanatory Statements and ICAMO Statements

At this stage, AAB and MH came together to reach consensus on the analysis. The findings from each researcher were placed side by side, with themes and retroductive theorising maturing throughout. A whole-team consensus was then reached on the final themes and ICAMO/explanatory statements. ICAMO statements were developed by identifying the Interventions, Actors, Contexts, Mechanisms and Outcome apparent at each stage of the analysis. NC and FCM provided iterative feedback and refinements on all ICAMO statements through discussion, comments in Microsoft Word, and bi-weekly meetings. AAB and NC met frequently to collaboratively build the theory and co-design the dispositional themes, jointly refining the ICAMO statements over several meetings by drawing on NC’s contextual expertise. Phase 3 consolidated the final configurational propositions, which were then developed into the full visual IPT and validated with interest group holders.

Table 7 presents the synergy table developed by Analysts 1 and 2 to compare their interpretations and work toward a shared understanding. While their conclusions were broadly aligned, they diverged slightly in emphasis across the different levels of analysis. This stage of the process calls for triangulation, adjudication and consensus building to refine and consolidate interpretations. Once a final interpretation is agreed upon for a given experiential theme, the final column of the table can be used to translate these insights into an explanatory statement as the final step of Wiltshire and Ronkainen’s realist thematic analysis approach (see Table 8). Notably, the explanatory statement also aligns closely with the ICAMO (or similar heuristic) configuration, with each element (context, mechanism and outcome) clearly identifiable throughout the explanatory statements. The intervention and actor components can then be explicitly added to the final ICAMO statement to clarify the focus of the intervention and the individuals involved, in this case, refugee mental health and substance use caseworkers (see Table 9).

OPEN IN VIEWER

Table 7.

Synergy Example

Analyst	Experiential themes	Illustrative quotes and theoretical support	Inferential theme	Dispositional theme	Consensus	Final
1	The participants in this dialogue described how the absence of systemic support motivated them to take initiative and advocate for their clients, ensuring they received the mental health care they needed.	“Most days, it feels like I’m the only one holding things together. There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible.” – Alex, Lines 346-356	It is plausible to claim that workers operating in high-pressure environments feel compelled to act rather than dwell on systemic shortcomings. Their sense of responsibility drives them to do what they can, often perceiving that without their intervention, the system would fail those in need. This perceived necessity serves as a motivating force, enabling them to persist despite challenging circumstances.	The inferred phenomenon is dependent upon the existence of a moral duty activated through one’s own professional identity. This ironically can mask systemic problems, as individual heroics maintain minimal function in broken systems.	Both analysts highlight how, in the face of systemic failure, caseworkers do not retreat but instead lean into a strong sense of moral responsibility. Initially, this manifests as urgent action to protect clients, but over time, it becomes a form of emotional adaptation, normalising dysfunction as part of the job. What unites both perspectives are the recognition that this internalised duty sustains care delivery while inadvertently shielding systemic flaws from scrutiny.	In high-pressure, under-resourced environments, caseworkers develop a deeply internalised sense of moral and professional responsibility that drives them to persist despite systemic failures. This persistence, rooted in identity and ethical duty, both enables client care and paradoxically sustains broken systems by masking the urgency for structural reform.
		“It’s exhausting, yeah. But it still feels like my responsibility. These are people who need care, and I can’t just step back because the system’s struggling.”- Alex, Lines 542- 546
		There is theoretical support from moral vocation and care ethics frameworks helped interpret this sense of responsibility as an identity-based moral obligation rather than a situational response (Dik & Duffy, 2007)
2	The participants in this dialogue described how limited resources, both within the system and through referral pathways, did not discourage them, but instead reinforced their sense of responsibility to provide care and take action despite the challenging circumstances.	“There’s no real support from the system, no proper referral pathways, just me trying to do what I can. I can’t just sit back and let these clients fall through the cracks. Someone has to step in. I am responsible. I need to act now before it’s too late” – Alex, Lines 350 – 359	It is plausible to claim that caseworkers have become desensitised to the lack of resources, viewing systemic inadequacies as an inherent and enduring feature of their working environment. As a result, their professional identity and sense of responsibility often override potential frustration, as they recognise that both their role and their clients’ well-being depend on their continued commitment.	The inferred phenomenon is dependent upon the existence of an adaptive workforce with an internalised responsibility that take on the emotional and ethical burden of the broken system, which in turn avoids reform.
		“These are people who need care, and I can’t just step back because the system’s struggling. Someone has to act, and I rather do it myself than wait until it is too late”- Alex, Lines 544- 548
		Moral vocation theory supports this as a moral obligation rather than motivated by the working environment (Weber, 1978).

OPEN IN VIEWER

Table 8.

Explanatory Statement Example

Explanatory statement one – Persistence through fragmented systems

In part because of the existence of a deeply internalised sense of moral and professional responsibility among caseworkers

There is a tendency for individuals to persist in delivering care despite systemic failures, often normalising chronic under-resourcing as part of their professional reality

This manifested in our study, which showed that workers not only stepped in where systems fell short, but also adapted emotionally over time, allowing broken systems to function through their continued commitment.

OPEN IN VIEWER

Table 9.

ICAMO Statement Example

ICAMO #1 – Persistence through fragmented systems

In the context of a fragmented and under-resourced mental health system with poor referral pathways (C), refugee settlement mental health caseworkers (A) are expected to deliver support to vulnerable clients. Providing mental health care within these settlement services (I) takes place in an environment where structural support is lacking. Despite this, caseworkers continue to act, driven by a deeply internalised sense of moral and professional responsibility (M). This inner commitment overrides their frustration and leads them to fill the gaps left by the system. As a result, mental health care is still delivered, at least partially (O).

Phase 4: Initial [Programme] Theory Elicitation

All ICAMO statements were compiled into a separate Word document, with each component visually highlighted for clarity. This helped distinguish between the different elements and facilitated abstraction at the context and mechanism levels. The ICAMOs were analysed for patterns, including overarching contextual conditions for integrated mental health care and similarities and differences at the mechanism level. At this stage, triangulation was used to assess each ICAMO configuration by balancing interview data with the broader evidence-based literature and collective but diverse expertise of the analysists (AAB, NC, MH, FCM), relevant substantive theory, and analytic interpretation. Competing explanations were initially adjudicated by AAB and MH based on explanatory power, empirical adequacy, and plausibility in accounting for outcome variation, rather than theoretical preference. These judgements were then reviewed with NC and FCM, with team consensus reached when an ICAMO provided a coherent and evidence-based explanation suitable for interest group holder validation. Demi-regularities, or, semi-predictable patterns of behaviour, were identified within the scope of the project and re-described using abductive reasoning, drawing on theories from outside the scope of the project and retroductive reasoning to uncover causal powers (Lawson, 1997). Broader explanatory principles were then developed, synthesising the key [programme] theories to reach an IPT. Although retroduction is employed in these later stages, the entire analytical process explicitly leveraged retroductive reasoning at all stages. From the Preparation Phase and Phase 1’s initial linked coding, abduction, the inventive thinking to imagine hidden mechanisms, was employed to identify patterns and initial interpretations, laying the groundwork for deeper inquiry. Phase 2 directly embodied retroduction by progressing through the themes, which explicitly involved abstract, retroductive reasoning to unearth causal powers. This unearthing of mechanisms is central to retroduction’s aim of understanding why things are the way they are. Subsequently, Phase 3 and Phase 4, involving the development of ICAMO statements and eliciting the IPT, deepened this retroductive understanding by linking key components and continuously refining causal pathways, aiming for closer approximations of the truth about underlying generative processes. Thus, retroduction, in tandem with abductive thinking, was a continuous and fundamental mode of inference throughout the project, enabling the move from observed phenomena to understanding the deeper, unobservable mechanisms that produce them. This final phase produced the final IPT through systematic triangulation of data and theory.

Table 10 presents a ICAMO example, designed to help quickly visualise emerging demi-regularities across configurations and identify broader contextual patterns. This approach supported the identification of recurring links between mechanisms and outcomes within specific contexts. It also facilitated abductive and retroductive reasoning, allowing for deeper interpretation of the data while working toward an appropriate level of abstraction necessary for developing an IPT. Guided by CBPR principles, a consensus-building online webinar using the collaborative online tool Miro was undertaken with our interest group holders to present our findings, as well as to jointly develop priorities for action. Following the presentation of the IPT, most participants agreed with the proposed enabling and hindering mechanisms believed to influence behaviours that support integrated mental health care.

OPEN IN VIEWER

Table 10.

ICAMO for Demi-Regularity and Initial [Programme] Theory Formation

ICAMO

In the context of a fragmented and under-resourced mental health system with poor referral pathways (C), settlement mental health caseworkers (A) are expected to deliver support to vulnerable clients. The intervention, providing mental health care within these settlement services (I), takes place in an environment where structural support is lacking. Despite this, caseworkers continue to act, driven by a deeply internalised sense of moral and professional responsibility (M). This inner commitment overrides their frustration and leads them to fill the gaps left by the system. As a result, mental health care is still delivered, at least partially (O), though this reliance on personal responsibility allows systemic inadequacies to persist unchallenged.

INTEREST GROUP FEEDBACK

“This explanation strongly reflects my experience working in refugee mental health services. I would also emphasise that a supportive team environment and strong community partnerships help sustain moral commitment and make challenges more manageable. That said, I agree with this explanation as it stands and would not suggest any changes”

Figure 2 provides a visual representation of the IPT analysis process and Figure 3 provides the final IPT created from the analytical process.

OPEN IN VIEWER

Figure 2.

Initial [programme] theory process

OPEN IN VIEWER

Figure 3.

An initial [programme] theory diagram

Discussion

Despite the central role of [programme] theory elicitation in realist inquiries, there are currently few examples that explicitly combine the strengths of two established realist analytical tools to demonstrate the reasoning behind [programme] theory development. This paper aims to address that gap by contributing to ongoing discussions about how [programme] theory is constructed in realist work. Specifically, it offers a step-by-step, replicable process that researchers can apply when working with qualitative primary data. This approach is intended to support transparency, rigour, and practical application in realist analysis.

This study adopted a novel approach by combining the strengths of linked coding techniques with realist thematic analysis. Dyad and triad coding have been widely used in realist research, with existing literature highlighting their value in CMO configurations from narrative data, exploring connections early in the analysis and supporting cross-case or cross-unit comparison (Caló et al., 2019; Steiner et al., 2021; Van Der Weide et al., 2025; Willis et al., 2018). Similarly, realist thematic analysis has been recognised for its contribution to producing causal explanations and uncovering underlying generative mechanisms, aligning well with a depth ontology (Fryer, 2022; Marsh et al., 2024). By integrating these two approaches, researchers can strengthen causal inquiry. Specifically, this combination enables the early identification of CMO patterns while also engaging with the deeper, often hidden structures that may not be readily accessed through standard thematic analysis alone. This dual approach supports a layered and theoretically rich understanding of how and why outcomes occur in particular contexts. However, combining methodological strengths does not eliminate the limitations of each individual approach. For example, this type of integration can be cognitively demanding and resource-intensive, particularly in settings with limited time or capacity. Additionally, there can be challenges in clearly distinguishing between thematic patterns and the inferential or dispositional elements associated with mechanisms, which may lead to some conceptual ambiguity. Despite these challenges, we argue that future realist research should explore and consider combining compatible or synergistic methods for [programme] theory elicitation. Doing so can expand the methodological repertoire available to the realist community. Such decisions should always be deliberate and guided by the ontological and epistemological underpinnings of the project, ensuring that any combination of methods remains coherent, purposeful, and aligned with the realist logic of inquiry.

Although the process presented above follows a sequence, as with all realist analysis, the phases were iterative and evolving. For instance, some dispositional themes were adapted based on evolving thinking on inferential themes at later stages of the project. This process presents both analytical and logistical challenges (Gilmore et al., 2019; Manzano, 2016). Analytically, the process of combining linked coding with realist thematic analysis required moving back and forth between coding structures and emerging realist themes. This iterative movement was valuable in identifying and refining underlying mechanisms, as our understanding of the data and emerging theories evolved throughout the project. However, the continuous nature of this development made transparent documentation essential. To manage this, document control and version tracking using Microsoft Teams played a critical role in maintaining consistency and clarity (Kanza & Knight, 2022). From a logistical perspective, employing a combined approach demands considerable time and resources. While this is not unique to realist research, it requires careful balancing of the day-to-day practicalities of project management with the need to preserve ontological and theoretical depth. Time pressures must not compromise the integrity of the analytical process. As such, maintaining an audit trail and ensuring transparency in the documentation of decisions and changes becomes especially important. Realist research must aim to make its processes visible and replicable.

The use of commonly available tools proved sufficient to conduct this project with rigour. This is a noteworthy observation, as it highlights that accessible and widely used software can effectively support the demands of realist analysis. At the outset of the project, the team agreed on a shared approach, and this decision proved beneficial throughout the research process. These tools supported the work in two key ways: first, by allowing collaborative annotation and tracking of analysis through shared working documents online; and second, by enabling visual mapping of codes and conceptual links using Miro. This demonstrates the value of combining different tools to support the analytical strategy, rather than relying on any one platform in isolation. Other researchers have also employed similar software in realist projects (Johansson et al., 2025; Jones et al., 2024), reinforcing the idea that methodological rigour does not necessarily require complex or specialised tools. In sum, accessibility, when paired with clarity and transparency, can be just as effective, if not more so, than adopting a multitude of sophisticated software options. In a research environment where numerous digital tools are readily available, it is important not to overcomplicate the process unnecessarily (Isangula et al., 2024). A practical, accessible, and rigorous approach is often more sustainable and effective.

While the worked example presented attempts to bring the analytical process to life, it has limitations that should be taken into account. First, the analysis was based on a relatively small primary dataset, specifically, six deliberate dialogues, each approximately 90 minutes in length. It is not yet clear whether the same analytical approach would be equally applicable or effective in the context of realist surveys or secondary data collection in a realist synthesis. Second, while the study focuses on the Canadian context, the identified ICAMO configurations have transferability potential; multi-site validation is planned for future phases initially through a realist review and then in various locations across Canada. Third, the research team involved in the example had prior experience in refugee mental health and expertise in realist approaches. This may have made the process smoother and more efficient than it might be for those attempting it for the first time. For this reason, we advise readers, particularly those new to realist research, to proceed with caution. Although we have tried to present the process in a simple and transparent way, it is essential that researchers clearly articulate their realist assumptions and reasoning throughout the process. Future research should consider RAMESES III standards (in development) for realist reporting. Fourth, this work was carried out within a collaborative research team, and the timings and processes described reflect that team-based setting. PhD students or independent researchers working primarily with their supervisor may need more time and support to adapt the process to a smaller-scale or more individualised research context. This analytical process required approximately 200 team hours over 6 months, highlighting its resource-intensive but valuable for robust theory-building. Finally, there was a high degree of alignment among team members in terms of how the analysis was approached and interpreted. This alignment helped the process flow more smoothly. We found that ensuring shared understanding from the onset was particularly helpful. However, in other cases, analysts may diverge more in their interpretations. If this happens, techniques such as data triangulation and interest group holder involvement can help resolve discrepancies. These strategies are consistent with realist principles through theory adjudication and support the rigour of the analytical process (Pawson, 2006; Rycroft-Malone et al., 2012).

Conclusion

In this article we have argued that combining established realist analytical methods to elicit [programme] theory can be a valuable and effective approach. By doing so, researchers can demonstrate how [programme] theory can be developed and refined throughout the research process. A key lesson emerging from this work is that integrating different approaches, when done transparently and with clear justification, can enhance the depth and flexibility of realist analysis. Importantly, the article emphasises the need for early alignment within the research team regarding the theoretical and analytical direction of the project. Establishing a clear and shared analytical strategy from the outset is essential for ensuring consistency and coherence throughout the study. This proposal is not intended to promote a single, standardised method or suggest a universal solution. Rather, it offers an additional option, another analytic process in the realist researcher’s toolkit, that may help in understanding complex social phenomena. The aim is to support methodological innovation while remaining grounded in the core principles of realist inquiry.