Abstract
Systems thinking is a paradigm well-suited to complex social and health fields such as socioeconomic approaches to wellbeing and health intervention research. Yet despite increasing calls for more general application of systems thinking in these fields of research, the paradigm remains poorly understood and systemic analysis is rare, particularly in qualitative research. This paper aims to address these issues. It starts by providing a primer to systems thinking clarifying what systems thinking is as a research philosophy, and describing core systemic principles such as complex causality and emergence. It then introduces the novel method of Inductive Systemic Analysis. Combining the methodology of grounded theory with the principles and key concepts of systems thinking, this method was specifically designed to make systemic analysis both relatable and accessible to a wide range of qualitative researchers. Utilising the case example the method was developed with, the paper works through each step of the research process. It explains preparing for systemic research through boundary considerations and developing a systemic research question; and goes on to describe the four-step analytical process of “zooming in” and “zooming out”, exploring the data in different ways. Ultimately, the method aims to support researchers in finding a balance between reflecting the complexity of social situations with ease of comprehension and applicability.
Keywords
Research is moulded by the lens of our personal worldview, shaped by factors such as culture, values and norms, language, socioeconomic status, upbringing, gender, education and life experiences. It is also moulded by the broader scientific worldview with associated values and norms shaped by scientific history and culture. The scientific worldview shapes which research philosophies, ontologies, epistemologies and methodologies we follow; what problem or topic we choose to research; from whom and how we gather and analyze our data; determination of causal processes; and what types of research are considered “best practice” with associated desirable outputs and outcomes. And so, ultimately, our worldview both as individual researchers and as a research community, determines what we know and how we know it. (Helfgott, 2017; Jackson, 2006, 2006; Midgley, 2003, 2006; Ulrich, 2012).
Worldviews are, however, generally situated in silence; i.e., they are tacit and rarely reflected on, and this is also common in the research context (Hitlin & Piliavin, 2004; Marini, 2000). We unthinkingly use research practices based on historically produced scientific norms and values, with no consideration of alternative and potentially more constructive ways of developing new research processes, creating new theoretical understandings, and discovering new types of knowledge and interventions. Ultimately, in the scientific research community, unthinking acceptance of dominant research worldviews may prevent us from querying the very purpose and situatedness of research in an increasingly complex and changing world (Checkland, 1993; Midgley, 2003).
The Current Viewpoint: Reductionism as the Dominant Scientific Philosophy
The worldview of Western science has its roots in ancient Greece where philosophers such as Hippocrates, Plato and Aristotle introduced explanation and prediction through rational argument, rather than through religion and magic, and developed basic scientific ideas such as observational science, mathematical representation, empiricism (truth from careful observation and testing) and classification by function (Checkland, 1993; Pelham & Blanton, 2007). The scientists of the Middle Ages further shaped modern Western science. For example, Roger Bacon established experimentation as a way of determining truth and argued for inductive analysis from collected data; Newton et al. (2016) cemented science’s move to experimentation and mathematical expression; and Descartes argued for scientific explanation to be based on deductive chains of reasoning, with the aim of showing how composite natures could be reduced to simple natures (Checkland, 1993).
It is because of these and other scientists, and the successful application of their thought processes to the classical scientific disciplines of physics and chemistry, that Western science and academic research continues to be overwhelmingly based on the scientific philosophy of reductionism. In this paradigm, ontologically, beliefs are based on the idea of reality consisting of a minimal number of parts or substances (Ruse, 2005), with prevalence of linearity of thought and mechanistic explanations of associations (Checkland, 1993; Ruse, 2005). Although this is undoubtedly a broad generalization, associated tools and concepts such as experimentation, quantification, replication, internal and external validity, and independent observation, still dominate scientific research endeavors. Many of these tools are intended to show mathematical strength of causation through separating variables from their context and determining linear causality, with the gold standard seen to be the double-blind randomised control trial (Greenhalgh & Fahy, 2015). As Flyvberg (2001, p. 21) observes, Western science continues to be overwhelmingly driven by “episteme” research: “[concerning] universals, and the production of knowledge that is invariable in time and space and achieved with the aid of analytical rationality”.
A Different Viewpoint: Systems Thinking as an Alternate Scientific Philosophy
Systems thinking is increasingly being put forward as an alternative to the reductionist viewpoint. Sometimes mistakenly considered a theory, systems thinking is a philosophy of science, combining both ontology and epistemology (Byrne & Callaghan, 2014; Hammond, 2005). In contrast to reductionism, ontologically a systems thinker sees systems as the basic entities of the world (Mingers, 2011; Rousseau et al., 2018) and views the world in terms of logical groupings of relationships, incorporating clusters of variables and multiplicities of interconnections with ensuing potentiality for complexity, non-linearity and unpredictability.
Systems thinking aligns closely with critical realism (Mingers, 2011) arguing that while systems may be intangible, they are nevertheless real (Pickel, 2011) and that this is authenticated through enduring and commonly found patterns of behaviors and outcomes (Mingers, 2015). However, systems thinkers also realise that due to the fundamental interdependence and interrelatedness of all things, it is impossible to comprehend and understand the “whole system” (Checkland & Scholes, 1990; Midgley, 2006; Ulrich, 2005). Thus, a systems ontology, whilst arguing that systems are real, recognises that knowledge is unavoidably partial and perspective-driven.
Rejecting the notion of structure and agency as distinct and simple containers in which one occurs prior to the other, in systems thinking structure and agency are considered in constant relationship with one another, each influencing the other (Byrne & Callaghan, 2014; Mingers, 2015). Structures in systems thinking are understood as the networks of relationships that create behavior. They can occur at all levels, from the individual social actor who can themself be seen as a system, to the societal. Interactions between structure and agency can lead to both emergence and unpredictability (Byrne & Callaghan, 2014; Mingers, 2015). Boundary judgements influence how we seek to understand and change both structures and agency, for example in the wider social fabric, such as politics and legislation, cultural values and social norms; or through the personal values and norms, education, etc., of social actors (Byrne & Callaghan, 2014).
The Working Definition and Components of a System
Having outlined some of the ontological principles of systems thinking, the obvious question is then: “From this position, what actually is a system?” While there is no concise and generally agreed-upon definition (Williams & Hummelbrunner, 2010), at its most basic, a system can be described as: “…an interconnected set of elements that is coherently organized in a way that achieves something” or “a set of things – people, cells, molecules or whatever - interconnected in such a way that they produce their own pattern of behavior over time” (Meadows, 2008, p. 3). Armson’s (2011) working definition is similar: “A system is a collection of elements connected together to form a purposive whole with properties that differ from those of its component parts” (p. 134). These basic definitions cover the fundamental properties of a system: elements, interconnections, process and purpose.
Elements
The elements in a system can be envisioned as the items in a system and can cover a broad set of entities. The most obvious elements in a system are physical objects. In a school, for example, the elements are the people, the buildings, the books and computers, classes and classrooms, etc. However, elements in a system can also include intangibles such as ideas, concepts, activities and data. The more elements identified in a system, the more complex the system (Armson, 2011; Cioruța et al., 2020; Meadows, 2008).
Interconnections
Interconnections are the key defining component of a system (Meadows, 2008; Zhang & Ahmed, 2020) and are seen as “the relationships that hold the elements together” (Meadows, 2008, p. 13). A systems thinker perceives reality and the world around them as a complex web of interconnections, rather than sets of elements. As the physicist Heisenberg (cited in Mingers, 2011) eloquently describes: …the world is not divided into different groups of objects but rather into different groups of relationships…. The world thus appears as a complicated tissue of events, in which connections of different kinds alternate or overlap or combine and thereby determine the texture of the whole. (p. 306)
Examples of interconnections in mechanical systems are the flow of fluids, electrical currents or chemical reactions. Biologically, interconnections can be the physical flows and chemical processes that drive metabolic processes (Meadows, 2008). Interconnections in human systems can be more difficult to identify. This may be partially due to the fact that interconnections can be conceptual in nature, for example, cause and effect (Armson, 2011). It may also be because we are used to thinking of interconnections as elements rather than as a means of connecting. For example, information can be viewed as an element, whereas Stacey (1995) argues that information is actually an interconnection. Lastly, it may also be that whether a particular identified factor is an element or an interaction becomes a subjective perception, dependent on the research question, the context it is set in and the perspective of the researcher.
Processes
Together, elements and interconnections form a whole that is, in effect, a process. As Hammond (2017, p. 12) emphasizes “it is important to understand that a system is not so much a ‘thing’ as a process”. In other words, a system is dynamic (Abdul-Rahman, 2024), in a constant “process of becoming, rather than static states of being”, with “a continual flow of matter, energy, and information” (Hammond, 2017, p. 12). With such an understanding, an event can be seen as a “slice” of change. The shorter the timeframe, the more aspects of the event, or situation of interest, can be viewed as fixed; the longer the timeframe the more aspects become variable and change (Mingers, 2011).
Purpose
The final basic component is the purpose or function of the system. The elements and interconnections together as a whole, achieve something (Armson, 2011; Cioruța et al., 2020). Whilst theoretically systems thinkers see the outcome of a system as the purpose of the system, social systems often have a stated purpose different to the outcome achieved. For example, the generally understood purpose of a school is to give its students the skills and knowledge to help them successfully complete schooling, with the longer-term purpose of giving them the tools and skills they need to be self-sufficient as adults. However, research is clear that for many students, the actual outcome of the system does not match its intended purpose. As such, to avoid confusion, the authors suggest that when investigating social systems, differentiating between the purpose (the social system’s intended function); and outcomes (what the social system is actually achieving) may improve clarity.
The purpose of a system is important, for in a world where everything is connected, the boundaries of systems are defined by their purpose. Purpose is of particular interest in social systems as different perspectives of a system may attribute different purposes to that system. Take again a school as an example. Generally, the purpose of a school is seen as education-focussed, as described above. But schools can also be seen as places of employment (for staff), government-funded day care centres (for working parents), a place to meet with your friends (for students), or a business (for school administrators). Each of these purposes depends on the perspective of the stakeholder in the system, and each is quite legitimate. However, understanding different interpretations of the purpose of the system can clarify how and why certain groups within the system (sub-systems), behave the way they do, and why the system itself may seem to be “pulling” in multiple directions, and as a whole failing to achieve its stated purpose.
Together, these four basic components of a system, elements, interconnections, process(es) and purpose(s)/outcome(s), can lead to systemic behaviors such as emergence or feedback loops, that may be important for understanding the situation under investigation. As such, awareness of these principles is important.
Systems Principles
Causality (Aetiology)
In essence, the core idea of systems science is that the outcome produced is a function of the whole system and thus, when seeking improvement, it is essential to look at the entire system when implementing change (Midgley, 2006). As a result, causality in systems incorporates consideration of both linearity and non-linearity and can include: • Multi-causality: the more elements and the more interconnections, the more influences must be considered. • Indirect causality: as one variable shifts, that variable, through a chain of relationships, can influence other variables it is not directly connected to. These indirect links often lead to the phenomenon of time lags in system outcomes. • Circular causality, often referred to as feedback mechanisms or feedback loops.
This last type of causality is of particular interest in systems thinking. Feedback mechanisms are considered key to either strengthening a system’s outcome through self-reinforcing chains of interactions (positive feedback loops); or regulating a system’s outcome through a balancing pattern of interactions (balancing feedback loops). A common example of a positive feedback loop is seen when successful organizations accumulate resources, allowing them to become even more successful. At a societal level this equates to a pattern of the rich becoming richer and the poor, poorer: i.e., increasing socioeconomic inequity. Over time, and without a correlating balancing loop in the system, reinforcing loops are, as Meadows (2008, p. 155) explains, “sources of growth, explosion, erosion and collapse in the system”. Balancing feedback loops, on the other hand, are self-correcting, leading to “autopoiesis” of a system, where a particular pattern of relationships in effect maintains a behavioral status quo. A simple example of this is a home heating system where the thermostat acts to keep a house within a certain temperature range by increasing or decreasing the heat it produces to maintain a set temperature.
Additional to the complexity of interconnections within a system and particularly relevant to social systems, are two added complications. Changes in system behaviors can be instigated by factors outside of the system, for example through new legislation. Changes can also be driven by internal agency within the system, as people strive to make the system exhibit the outcome they would prefer. This complexity can lead to the systems characteristics of non-linearity and unpredictability, in effect making knowledge a local concept (Nicolis, 1995), with limited capacity to be generalized and/or predictable. Outcomes for a systems thinker must be understood as the result of a particular constellation of elements and interconnections across a particular segment of time. As such, systems thinking is experientially and pragmatically based and systems thinkers see the value of the theory in how it helps us to understand the world, rather than it merely being a schematic application (Byrne & Callaghan, 2014).
Emergence
Perhaps the most well-known systems characteristic is that of emergence: the oft quoted “the whole is bigger than the sum of its parts”. Emergence in systems thinking is understood as a phenomenon where interconnections between elements can give rise to unexpected structures that exhibit properties quite different to those of each element (Zhang & Ahmed, 2020). The combination of hydrogen and oxygen is a chemical example of emergence. Separately, both are gases with their own properties. Oxygen is a colorless, odorless gas, and many stable substances become more flammable in highly oxygenated environments. Hydrogen is lighter than air and embrittles exposed metals. Linked together, however, they form water, a colorless liquid crucial to life. The properties of these two molecules, connected, are different to the properties of the two molecules as separate entities (Armson, 2011).
In complex systems it is not only the interconnections between elements that can give rise to emergence. Emergence can result from interrelationships between elements themselves, between groups of elements, and between elements, groups of elements, and the system as a whole (Newman, 1996). These different levels of interactions lead to the epistemological concept of investigating systems as both wholes and parts; what Armson (2011) describes as “zooming in” and “zooming out” (p. 33).
Systems Thinking as Both Whole and Parts
Systems thinking examines the whole system, separate factors, and groupings of factors in the system. The research lens zooms in to understand the separate elements and interconnections in a system and then zooms out again to understand how these “fit” together to influence the overall function of the system (Arnold & Wade, 2017). Epistemologically, systems thinking is therefore neither holistic nor reductionist in nature but incorporates elements of both.
Boundary Judgements and Systems Configurations
As mentioned previously, systems thinkers are aware that as humans in a complex universe, we are only capable of understanding small parts of this complexity. Consequently, systems thinkers consciously make boundary judgements. However, defining what constitutes a boundary, and thereby what is defined as the system, is both complex and observer-dependent (Mingers, 2011). Consideration of boundaries, and reflection on the utilisation of systems thinking for analysis, is particularly relevant for researchers, as setting boundaries will shape the theory produced from analysis, and as the below example will show, analysis through a systems lens will create a systems-natured theory. Practicably, from an analytical perspective, consideration of three types of boundaries is suggested.
Boundaries in Perspective
Awareness that all viewpoints, and therefore understandings, are incomplete encourages researchers to consciously consider their own and others’ boundaries in perspective (Abercrombie et al., 2015; Midgley, 2003): those who are included, those who are excluded, whose knowledge is labelled as “truth”, and what this means for the research being produced (Churchman, 1979). For example, mental health research has been overwhelmingly conducted through the lens of a Western highly individualistic culture, a reductionist philosophy, and a medical lens; with as consequence interventions predominantly focussed on change at the individual level and applied in a predetermined “evidenced” way. This is in contrast to First Nations cultures who understand the mental health of an individual as inextricably and deeply interconnected with, and reciprocal to, family, community, culture and country (National Aboriginal Health Strategic Working Party, 1989).
Boundaries in Situation
Inevitably as researchers, we place a boundary around what we are investigating, the situation of interest. As a systemic researcher, it is important to consciously recognize and explicate this. For instance, in the case example, the boundary of the research had been placed at the school level, with the project hoping to achieve change at that level, moving beyond individual students, but not so far as to include students’ families and communities, or education policies.
Boundaries in Time
Systems thinkers are aware that they are only investigating and assessing a particular situation across a particular time frame, and that it is only that time frame they can account for. Knowing that social systems are dynamic in nature and being aware of the potential for unseen time lags, systems thinkers do not presume that findings at one time will necessarily have the same outcomes if investigated across a different time frame. For example, over the course of the 5-year case example, it became increasingly clear that any changes the research project was able to influence, were highly dependent on contextual factors that were in a constant state of flux. In short, a conscious investigation of boundaries in various ways drives reflection of the potential limitations and biases of the research process itself, as well as associated claims of “knowledge”.
Systems Configurations
The placing of boundaries is also informed by the basic configurations that systems exhibit. Perhaps the most well-known of these is the nested system, depicting a hierarchical order in the complexity under investigation (Cioruța et al., 2020). Systems can also be seen as overlapping, where a part of one system is also part of another system. Finally, systems can be perceived as distal – seemingly separate - having no direct elements or interconnections in common. Indeed, depending on the boundaries placed, a system can encompass all of these configurations. Analysis and depiction of system configurations is flexible, based on awareness and judgement of what type of configuration best depicts the data and suits research objectives and aims.
The Axiology of Systemic Thinking
Lastly is consideration of axiology – the values underlying systems thinking. Systems thinking is rooted in philosophical pragmatism, seeking practical solutions to complex issues (Arnold & Wade, 2017; Midgley, 2003; Ulrich, 2012). While from its inception, the core idea of systems science has been that improvement is a function of the whole system (Midgley, 2006; Stroh, 2015), further axiological advances were made in line with the broader social sciences. The rise of phenomenology and constructivism during the 1970s led to the development of the so-called “soft systems”, epistemological and methodological frameworks with both a participatory and action orientation that specifically included those people involved in the situation (Checkland, 1993; Hammond, 2005; Ulrich, 2004). Throughout the 1980s and 1990s, growing recognition of a need for emancipatory approaches saw the rise of critical systems thinking (Mingers, 2011), addressing issues of power relationships in organisations and adopting a more overtly equity-driven orientation of both research and practice (Mingers, 2011; Pickel, 2011). This approach particularly emphasised the need for conscious exploration of boundaries and underlying factors such as social power, relationships and worldviews, understood to be central to the perpetuation of complex social problems (Mingers, 2011; Ulrich, 2004). These further axiological developments led to the recognition that “facts”, and the decisions and actions arising from these, are likely to serve some stakeholders better than others; and that there is no single right way to decide how complex social issues should be addressed (Gates, 2017; Ulrich, 2012).
While there have been increasing calls within social and applied health research to utilise systems thinking, (Carey et al., 2017; Greenhalgh & Papoutsi, 2018), uptake of this approach is slow. Multiple years of investigation in peer-reviewed systems-related publications by the first author suggests that to some extent, this may be due to academic-based systems-based research becoming a highly specialised field in itself, with difficult to understand terminology; complex and/or prescriptive methods; and theories that are, or seem, challenging to apply to local situations. The novel method Inductive Systemic Analysis (ISA) was developed to overcome some of these barriers.
Introducing Inductive Systemic Analysis
ISA combines grounded theory methodology with the basic framework, concepts and principles of systems thinking. This combination was a purposeful design, aimed at addressing some of the barriers (as identified above) slowing the uptake of systems thinking in the social sciences. By combining a widely familiar and flexible methodology that aims to create useful situational theories from inductive analysis (Charmaz, 2014) with the basic tenets of systems thinking, ISA aims to make analytical systems thinking as accessible, adaptable and practicably applicable as possible. This combined intent of both method accessibility and practicable usefulness sets it apart from many other qualitative systems thinking methods and tools.
The Benefits of Grounded Theory as a Systemic Methodological Process
As a methodology, grounded theory is well-suited to a qualitative systemic data analysis (Bainbridge et al., 2019). It is an inductive methodology designed to explore the nature of complex social phenomena, systems and processes. It is flexible, open to further development, and can be used for a wide range of topics. It is also well-known for its rigor of analysis, utilising both constant comparisons across all levels of analysis (Birks & Mills, 2015; Charmaz, 2012); and theoretical sampling, a process of concurrent data collection through purposive sampling, data analysis and further sampling (Mills et al., 2006). Important for a systemic analysis, it has the capacity to take multiple perspectives into account, investigate dynamic processes, explore for similarities and differences, and incorporate contextuality into analysis (Bainbridge et al., 2019; Birks & Mills, 2015; Charmaz, 2008, 2012), all critical aspects for building a systemic understanding of a research topic. With the capacity to apply both a deep and wide analytical lens comparing different perspectives, different contexts and different levels, a grounded theory analysis can construct local and practical situational theories and constructs: “what works, for whom, and why”.
However, while grounded theory is indeed a methodology appropriate for analysis through a systems thinking lens, much of grounded theory work remains anchored in a reductionist philosophy. Utilization of grounded theory for analysis and interpretation of data often continues to separate findings into component parts, lists and dominant themes; looks for direct linear links; and searches for a singular “core process” as mid-level theory. There are however clear signs of change towards a more systemic approach. For example, both Situational Analysis (Clarke et al., 2017) and Transformational Grounded Theory (Redman-MacLaren & Mills, 2015) explore social worlds and different positioning of perspectives. Additionally, Bainbridge et al. (2019) explore the need for change across systems and at multiple levels. With the development of the ISA method, this transformation takes another step forward.
Looking at the Grounded Theory Process Through a Systems Thinking Lens
Systemic analysis explores both the large and the small – “the forest and the trees”- with an emphasis on interconnections and relationships. As such, a grounded theory analyst working in a systems thinking paradigm intentionally looks for elements and interconnections, processes, groupings, boundaries, purpose and outcomes. Analysis investigates how elements are connecting and, just as importantly, where and why they are not connecting. Investigating how elements and interconnections are shaping processes and groups in different ways, with what associated outcomes, helps improve understanding of structures within and between different contexts and levels. For those working in intervention or applied social science, it is important to understand that from a systemic perspective, the context is the system, and intervention or applied research is attempting to influence that context through adding to, or adapting, a pre-existing constellation of elements and interconnections. With this understanding comes the need to appreciate the “fit” between intervention and system; i.e., there is a conscious embracing of the intervention as part of the system it is situated in.
A systemic paradigm encourages theoretical and methodological heterogeneity and researcher creativity in the research process (Boyd et al., 2004; Jackson, 2006), and there are many systemic-based theories, frameworks and models that are worth exploring to aide in this, and to gain a deeper understanding of the many ways to look through the systems thinking lens (Abdul-Rahman, 2024; Cabrera & Colosi, 2008; Monat & Gannon, 2015). For example, complex-adaptive systems theory which emphasises the capacity of systems to adapt to change through self-organisation, learning and reasoning; chaos theory which investigates the relationship between order and chaos, (Cordon, 2013); and the Iceberg Model (Kim, 1999) which uncovers the drivers underlying social structures.
The following section describes the ISA method, starting with consideration of participants, data collection, boundaries, research question, and associated objectives. Data analysis involves four processes aimed at developing a systemic understanding of the research topic through exploring the complexity of social situations in different ways. A worked exemplar is presented throughout utilising the case example for which ISA was initially developed (Van Beek, 2024).
The Case Example
The case example (Van Beek, 2024) was a qualitative investigation of the Resilience Study, a participatory research project taking a socio-ecological approach to supporting and enhancing the mental health and wellbeing of remote-living First Nations students attending secondary boarding schools in Queensland, Australia (McCalman et al.; Redman-MacLaren et al., 2017). ISA was created to explore the complexity underlying the project: the challenges underpinning the psychological distress of students; the resources needed to support students’ mental health and wellbeing (the “resilience ecosystem”); the implementation of social change in boarding schools; and how participatory research was able to support this complex approach to mental health and wellbeing.
Working through the Inductive Systemic Analysis Method
Participants
Participant selection follows grounded theory methodology, starting with purposive sampling, selecting those participants perceived as being able to answer the research question. Coding of initial data will improve understanding of who the various stakeholders in the situation are, guiding further data collection efforts, i.e., theoretical sampling (Charmaz, 2008). While systems thinking argues that the best way to gain understanding of a system is to gather as much information as possible from as many stakeholders in the system as possible, practical limitations such as research funding, time and stakeholder interest, is likely to set boundaries around this ambition. This, and the complexity of most social situations, makes it important that, where feasible and as grounded theory advocates, participants also provide feedback on draft models and theories.
Data Collection
ISA was purposively developed to be flexible to research conditions and as a grounded theory method, holds to the maxim that all is data (Glaser, 1998). For instance, whilst data collection in the case example included interviews with participants, it also incorporated historical recordings from meetings and group forums; various papers and artefacts produced throughout the project and notes from personal observations and conversations.
Boundary Framings of the Research
Boundaries are considered in a number of ways. First is consideration of the researcher’s own boundaries. In the case example this was particularly pertinent. As a non-Indigenous researcher working in Indigenous research, the first author needed to consider what worldview, with associated boundaries, values and paradigm, she brought to the research project, both as a “native systems thinker” and a non-Indigenous woman.
An important step is elucidating the boundaries of perspectives collected. Whose voice is present or missing? How strong are the voices present in the data? For example, whilst data collection in the case example was extensive, the dominant voice remained that of school staff. This was fully acknowledged as a limitation to the research and influenced both boundary framings as well as exploration of data in contexts. A stronger voice from students and families would have added considerable understanding of, and depth to, a systemic understanding of the issue, potential solution, and change effort required.
The researcher also needs to consider the boundary of time: the time across which data was collected, the perspectives of relevant stakeholders in regard to time, and their own boundaries in time. Finally, consideration of the research question and associated objectives also play an important role in determining placement of boundaries.
The case example also highlighted the benefit of boundary flexibility. For example, analysis of the first objective, exploring the challenges underlying students’ psychological distress, found 3 main systems: boarding schools, students and students’ home communities. However, on examining the next objective, the resources and support students required to maintain their wellbeing, boundaries expanded to include the wider education system, the health system and scholarship system(s). Similarly, “playing” with boundaries can be illuminating. In the case example, this was done when exploring intentional social change in schools through the lens of the purpose of various groups (i.e., sub-systems) in the school setting. How do different sub-systems with different purposes influence social change in the system? Such explorations can be powerful ways of understanding influence within the system and outcomes being seen; as well as determine the choice of systems boundaries and configuration throughout analysis.
Research Question
In line with the definition of a system, a systemic question is by nature a “what (elements and interconnections), how (processes) and why (purpose/outcome)” question (Armson, 2011). Therefore, a systemic research question theoretically should be phrased so as to uncover these components of a system. For example, the research question developed for the main author’s thesis was: “What are the elements and interconnections that influence the capacity of participatory research to support intentional social change aimed at improving First Nations students’ mental health?”
The purpose (the “why”) of the research was to unpack the complexity underlying how and to what extent participatory research is able to support social change, utilizing a case example of participatory research focussed on First Nations student wellbeing. The “how” underlying that intent were the processes involved in the participatory research project. From a systems thinking perspective, processes are formed by the collective activity of the elements and interconnections (the “what”) in the system. Framing the research question to incorporate a “what, how and why” (theoretically at least) should enable a systemic view of the data and support development of a systemic-natured theory. More application of the method is required to further assess this.
Research Objectives
Research objectives are developed to uncover the interconnected complexity underlying the situation of interest. Objectives can be broken down into different stages, processes, dimensions, settings, or other features that may be of particular interest. In the case example, following the original intent of the Resilience Study, objectives were framed as “stages of change” and aimed to build a systemic understanding of the complexity underlying: (1) the problem issue: the challenges underlying students’ psychological distress; (2) the theorised outcome: the resources and support that make up a resilience ecosystem for remote Indigenous students attending boarding school; (3) stage one of the intervention: engagement in the Resilience Study; (4) stage two of the intervention: creating social change in boarding school settings; and (5) potential outcomes and impact of the Resilience Study.
The case example had two final objectives: (6) identifying the interconnectors of intentional social change; and (7) understanding whether ISA was able to find a useful balance between complexity and comprehension.
Data Analysis
Systems thinking sees data separated from context as raw facts and figures that provide limited capacity for understanding. The goal of data analysis in ISA then is to shape data provided by participants into information: data analysed and presented within the context(s) shared by participants. This aligns well with grounded theory, where analysis is seen as a “sense-making” practice, rather than seeking evidence of a discrete outcome. Given the emphasis of ISA on creating understanding of, and social change in, complex social situations, evaluation of analytical outcomes follows the lead of constructivist grounded theory (Charmaz, 2014), emphasising the importance of aligning evaluation with purpose and context. This then leads to the evaluation criteria of credibility, originality, resonance and usefulness (Charmaz, 2014).
The Four Key Steps of Systemic Analysis
The four steps of analysis in ISA provide means to explore data in context in different ways. In doing so, a deeper and more nuanced understanding of the situation of interest is built: how this looks through the perspectives of different participants; how stakeholders connect to one another and to the situation of interest; what their role in the system is; their various characteristics; as well as other factors that are influencing stakeholder and system behaviours. It should also include how other internal and external factors are influencing the system; how change in one part of the system will influence other parts of the system; and how all of these factors together are driving the outcomes seen.
The four key steps together form a zooming in and out process of investigation, asking the researcher to dive deep into the data to understand the smaller details, and then pulling the investigative lens back out to place detailed understandings into a “big picture” lens. Each step can be revisited as many times as required to develop a systemic understanding layer by layer. Understandings found for each objective can then be utilized to develop a more generalised theory answering the overall research aim. Figures 1–3 showcase how data analysis involved building models and frameworks to deepen insight into interconnectivity and complexity in the data. Examples are taken from objective 2 of the case example which aimed to build understanding of what a resilience ecosystem for students looked like. More examples of how modelling supported analysis of objectives can be found in the original case study (Van Beek, 2024). The Systems Framework for a Resilience Ecosystem Exploring Systems’ Behaviours for the Resilience Ecosystem The Core System Underlying a Resilience Ecosystem Supporting the Positive Mental Health and Wellbeing of Remote-Living Indigenous Students Attending Boarding School
Zooming In: An In-Depth Analysis of the Narrative
Analysis begins with zooming in to investigate the data using an in-depth narrative of each objective in order to gain an understanding of the many factors influencing that particular objective. The aim is to gain a deep familiarization with the data and start to get a sense of how multitudes of key factors are working together (or not). Open coding is used to find the main factors and processes present in the data, as well as illuminating key systems and potential sub-systems. Cross comparisons are used to understand similarities and differences across various stakeholders, within participating organizations, between individuals, between different levels and within and between different systems. In the case example, cross-comparisons were made between different participating schools, between staff roles, and between different student challenges, support methods and outcomes. Experience so far with the method shows that colour coding data is a particularly useful way to explore the data.
Zooming Out: Building the Systems Framework
Here, the analytical process zooms out to visualize a systems framework (big picture view), incorporating the main systems identified in the detailed analysis. Concepts such as systems mapping, stakeholder analysis and boundary conceptualization are utilized to identify different systems and construct different systems configurations, depicting the identified systems as nested, overlapping, distal, or singular, or indeed all of these. For instance, when exploring various objectives throughout the case example, students were depicted as an overlapping system with both boarding school and home communities; as a subsystem in boarding schools; and as a system at the individual level.
Identified factors from the in-depth analysis are then placed to best fit into the system framework. The combination of in-depth description, systems framework and attempts to place identified factors in systems, provides opportunity to explore the many factors at play, how these are shaping systems configurations and the overall outcomes being seen. For example, factors may be working together in different ways to make up different groups, they may be connecting at different levels; or they may not be connecting at all. Working through this process for objective 1, the challenges underlying students’ psychological distress, it became clear that student challenges often “originated” in different social arenas, for example in boarding school with Western ways of education less suitable to First Nations cultures; or First Nations communities experiencing poverty. However, it also showed that many challenges stemmed from trying to fit two dissimilar systems together, i.e., attempts to “combine” remote living First Nations students and Western boarding schools. The systemic exploration focused on student challenges found that overall, this combination of systems was exhibiting the systemic outcome of “disconnection”: the boarding school system and the student system were in a constant state of disconnecting from one another.
Alternately, a focus on the second objective, a theorized resilience ecosystem supporting students’ wellbeing (Figure 1), found that this scenario required schools, families and communities to work closely together to ensure good outcomes for students, depicted through substantial overlap between the systems. Data analysis also showed that the systems framework needed to expand beyond those identified in objective 1 to include the broader education system, as well as student scholarship and health systems.
Zooming In: Exploring Systems’ Behaviors
The analytical lens at this step zooms in to explore how various key factors identified are working together to form small patterns of systemic behaviors. Here, basic systemic concepts, such as complex causality, are used to dig deeper into the data, exploring various constellations of interactions, and how these constellations are driving particular associated outcomes. Concepts that are particularly important to look for are feedback loops, signs of emergence, and both repeating and unexpected outcomes. By exploring different constellations of factors, a better understanding of larger processes in the system is gained. These larger processes are often the systemic patterns of behavior a system is exhibiting repeatedly, ultimately directing the outcome of the system. Cross-comparisons can be utilized to understand how different contexts are working differently, as well as how commonalities might be driving similar outcomes in different contexts. Findings from applied projects show that it is often the same factor, but with a different characteristic, that is driving different outcomes in different settings, for example supportive or unsupportive leadership.
Figure 2 shows one such exploration of systems behavior, exploring how two particular support measures identified in the analysis could work together to contribute to a resilience ecosystem. Exploration of the measures “student empowerment” and “increasing cultural activities”, theorising how these support measures can create small chains of interconnected outcomes, built awareness of how multiple measures together can create ongoing “ripples” of positive change throughout the whole school for the benefit of students.
Zooming Out: Building the Core System
Here the lens zooms out, making an abductive leap to construct a small systemic situational theory: a core system. This step is similar to finding a core process; however, here the focus is on how the whole “system of interest” is shaping the outcome being seen. The core system shows the main systems with relevant characteristics and key processes that together are driving the outcome(s) being seen or theorised. The example here (Figure 3) again focusses on the resilience ecosystem, showing the associated core system. Main systems are shown here as block arrows, with other systems identified where relevant. Within each main system are the key characteristics identified by participants as significant for achieving the sought outcome of student wellbeing. Note that in the families and community system, the last two characteristics identified (community, family and individual healing and wellbeing, and improved socio-economic conditions) were characteristics schools had no capacity to influence, however a lack of these had been acknowledged as contributing to students’ distress. The key processes contributing to the system’s outcome are shown as interconnected circles interacting with each other, with the outcome (students’ positive mental health and wellbeing) depicted as a circle in the centre.
The goal of a core system is to find a balance between complexity, comprehension and contextual applicability as these models can be used to provide guidance for improvements adaptable to local contexts. Depicting “streamlined complexity”, objectives’ core systems have also proven important to understand further objectives, as well as for developing higher level theory. In the case example, whilst consideration of connections formed an important part of the analysis throughout each step, investigation of the identified key processes in the core systems of the first five objectives was particularly significant for finding the social interconnectors (objective 6). Figure 4 shows how these identified interconnectors of sharing emotions, distributing social power, building knowledge together, and sharing a common purpose, were depicted as the interconnected “energies” that drive change within a social system. The Interconnectors of Social System Improvement: The Entwined Energies of Intentional Social Change
Similarly, further reflection on the core systems together with identification of the social interconnectors, led to the final higher level theory: a systemic theory of intentional social change for positive mental health and wellbeing. Figure 5 shows this theory with accompanying explanation of terminology (Table 1). This theory depicts positive mental health as one part of a large multi-part process, arising from, sustained by, and contributing back to the system around it. This theory aligns well with the holistic and relational view of health and wellbeing of First Nations Australians (National Aboriginal Health Strategic Working Party, 1989). The Final Theory: A Systemic Theory of Intentional Social Change for Positive Mental Health and Wellbeing Accompanying Table of Terms for the Case Example’s Final Theory of Intentional Social Change for Positive Mental Health and Wellbeing
Conclusion
Understanding social issues and implementing and sustaining the complex change required to improve these, necessitates a different paradigm to that of reductionism. We argue that systems thinking is a paradigm well suited to such endeavors, able to “unpack and repack” the complex and dynamic nature of social reality whilst supporting both comprehension and practical, adaptable application (Diez Roux, 2011; Frerichs et al., 2016; Moore et al., 2015). By combining the well-known methodology of grounded theory with the basic principles and tenets of systems thinking, the novel method ISA is put forward as an accessible and flexible method into a paradigm that is often seen as perplexing and difficult to apply. This inductive method supports an easy and constructive exploration of social complexity, and provides potential for finding balanced, operable and more sustainable answers to so-called “wicked problems”.
Footnotes
Acknowledgements
Ros Calder edited the manuscript.
Author Contributions
AvB conceptualized the content of the article and drafted the manuscript and revised subsequent drafts. JM and VS provided critical feedback for revision. All authors approved the final version of the article.
Funding
This research forms part of a PhD dissertation, which was funded by Central Queensland University through the Commonwealth Research Training Program Stipend; as well as further supported by the CQUniversity Centre for Indigenous Health Equity through the Scholarship Program.
Declaration of Conflicting Interests
The authors declare no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.