Georg Baselitz, Aus Gelbrotorange wird Blaudunkel

Changing theories

Thomas Aston
8 min readOct 15, 2023


You may, or may not, be surprised to hear that many theories of change lack what we might generally understand as a theory.

I’ve argued in the past that at the heart of theories of change are “if, then… because statements.” What is so often absent from those theories is the BECAUSE part. This is usually related to supposed “mechanisms” of change and assumptions underpinning change. But, even with these, there’s often something missing. Theories of change may have logical power, often enough, but they regularly lack explanatory power. This is where theory is supposed to help us.

There are, of course, many definitions of what a theory is. But, most dictionary definitions refer to a supposition or system of ideas intended to explain a phenomenon, event, or fact.

Several years ago, Annette Brown made the argument that we need to bring back theory to theory of change. Brown’s argument is that theories of change go beyond arrows and assumptions and speak to underlying theories that explain (or predict) whether (and how) the programme will work. I think she’s right that logic is insufficient, and we should be thinking about whether there are specific theories underlying the supposed causal chains (or other forms of causal explanation). As Brown puts it:

“Theory is what allows us to learn cumulatively from one programme to another and from one evaluation to another.”

Theory thus speaks to the transferability of potential mechanisms of change.

In my view, there are two general, complementary, types of theories lurking behind theories of change: (1) process theories, and (2) behavioural theories. Both of these relate to slightly different understandings of the generative logic which underpin theories of change. Process theories are focused on trajectories of change or pathways of change. They tend to be about sets of events in time and how these are assumed to be connected. Behavioural theories are generally focused on stakeholders’ supposed reasoning, or psychological processes which underpin the decisions taken to pursue a particular course of action.

To this end, I gave a talk for the Centre for Development Impact (CDI) a few years ago explaining some of the differences in mechanistic understanding between Process Tracing and Realist Evaluation. See 11:00–21:00 in particular.

At the heart of Realist Evaluation, among other contextually mindful approaches, is the search for middle-range theories. In reality, there are various levels of “middle,” as I explain in the talk, but one of the main reasons middle-range theory is potentially helpful is that we shouldn’t have to reinvent the wheel; someone somewhere has probably thought about change in a similar way before.

Identifying those different ways of thinking about change can help us to communicate with others, rather than get lost in translation. The University of Antwerp actually has a mechanisms database which sets out 20 common mechanisms which are derived from middle range theories in social science. The database is imperfect and incomplete. But, it’s nonetheless a helpful reference point for theory gleaning.

In the sectors I work in (governance, accountability, advocacy, and research), several theories come up recurrently. For example, two of the sessions I attended at the UK Evaluation Society Conference this year (both on “bricolage”) mentioned “multiple-streams” theory by John Kingdon. This theory is more commonly known by the expression windows of opportunity. As Daniel Béland and Michael Howlett explain, this theory has been widely used since the publication of Kingdon’s book Agendas, Alternatives and Public Policies in 1984.

One mention of the theory came to a recent paper by Boro Douthwaite et al. on Outcome Trajectory Evaluation, another came from my co-presenter Giovanna Voltolina on an evaluation by Itad for the Fleming Fund. Itad’s evaluation outlined that prioritisation of an agenda happens when policymakers concurrently: (1) understand the problem, (2) have a viable solution available and are; (3) (politically) convinced of the need for action.

Itad, 2021

The theory appears in slightly different guises and in more and less formal ways. For instance, windows of opportunity also came up in some recent research by Florencia Guerzovich, María Soledad Gattoni, and Dave Algoso on anti-corruption reforms for the Open Society Foundations (OSF). It was also one of the most common (implicit) theories underpinning a Qualitative Comparative Analysis (QCA) I did for CARE International on their advocacy work and Oxfam found the same for their advocacy work. It thus struck me how often the same theory appeared to show up.

One of the most useful attempts to bring together middle range process theories is Sarah Stachowiak’s Pathways for Change: 10 Theories to Inform Advocacy and Policy Change Efforts. These 10 theories are derived from several social science disciplines. Most of these are process theories, but, there are also several behavioural theories included which sit at a slightly different level of abstraction. It’s a very useful reference paper, especially for those who haven’t read a lot of social science theory.

I thought I’d highlight which (implicit or explicit) theories I most often perceive in my work. I can identify 7 clusters. These are not necessarily mutually exclusive; they are often seen combination. I clustered similar theories under umbrella terms to give a sense of commonalities. Needless to say, this therefore loses the nuance of individual theories. Nonetheless, I’ve signalled several variations and bifurcations. My aim here is not perfect fidelity to the original conceptualisation or wording of each theory, but to provoke a discussion on common theory areas and to illustrate a number of those you may wish to consider, and to provide some real-world examples which either claim to represent, or at least approximate, these theories.

  • Windows of Opportunity: Harness a perceived (policy) “moment” to push through or accelerate change. Change is expected to happen during a “window of opportunity,” particularly if stakeholders pursue multiple streams of engagement to agenda setting (see Kingdon, 1984). Despite different formal origins, this theory is often implicitly interwoven with the theory of “critical junctures.” Dave Algoso et al. further argue that windows might sometimes be conceived as cyclical rather than fleeting. Opportunities may come and go, but they can also reappear. So, we can prepare for them.
  • Critical Junctures: With the right conditions, change can happen in sudden, large bursts which represent a significant departure from the past. There are several variations on critical junctures. It’s sometimes referred to as a “punctuated equilibrium” model (see Baumgartner and Jones, 1993), but there are various different conceptualisations which fit under this umbrella. Notably, Ruth Berins Collier and David Collier’s Shaping the Political Arena provided a five step template of Antecedent Conditions → Cleavage or Shock → Critical Juncture → Aftermath → Legacy (see also Kuhn, 1962 on for a discussion on paradigm shifts in scientific revolutions). And we can find combinatorial theories. Guerzovich et al’s findings on anti-corruption reforms, for example, cut across windows of opportunity and critical junctures.
  • Advocacy Coalitions: Strength comes in numbers and through unity; collective action can be more effectively achieved through a shared agenda for change. Change is expected to happen through coordinated activity among a range of individuals and organisations, often with tactical and time-bound alliances for a particular purpose. It’s believed that coalitions can more forcefully push for a proposed change together than as individual entities. The Asia Foundation’s recent paper on coalition building initiaves in south and east Asia provides a good summary of the rationale for coalitions. There are plenty of theoretical varieties of advocacy coalitions (see Paul Sabatier and Hank Jenkins Smith’s popular Advocacy Coalitions Framework, 1999 and Jonathan Fox, 2010 for a discussion on coalitions and networks). We find examples of insider coalition strategies such as in the State Accountability and Voice (SAVI) programme in Nigeria as well as hybrid (insider-outsider, collaborative-confrontational) coalition strategies such as those expressed in some cases of the Fox’s “Sandwich Strategy.”
  • Creative Destruction: Transformative change processes are inherently disruptive and destructive; phoenixes rise from the ashes. New, disruptive, innovations, or socio-political orders replace existing ones that are rendered obsolete. The most famous elaboration of creative destruction comes from Joseph Schumpeter (Schumpeter, 1942), building on the work of Karl Marx, but we see renovations of the theory in explanations of how neoliberalism spread (see Harvey, 1995). Creative destruction often happens hand-in-hand with perceived windows of opportunity. We see this in “big bang” approaches to anti-corruption, and what David Jackson labels as the “activist tribe” to anti-corruption programming has a strong flavour of this phoenix narrative. A possible variation of such a theory is what Guerzovich et al. (I am one of the al.) termed a “resistance” pathway to scaling up social accountability (see the Human Resources for Health Campaign example in Uganda), which overlaps with some versions of the aforementioned “Sandwich Strategy.”
  • Institutional Layering: Institutional change is a gradual, incremental, and adaptive; efforts search for appropriate fit and layer on previous foundations. Gradual institutional transformation happens through a process in which new elements are attached to existing institutions and gradually change their status and structure. Guerzovich et al. reconfigured layering as a “resonance” pathways to scaling up social accountability, and David Jackson’s category of the realist “tribe” in anti-corruption programming broadly fits under this umbrella as well (see Thelen and Mahoney, 2010 and Van der Heijden, 2013 for a discussion, and see my blog on layering in the Dominican Republic’s education sector).
  • Diffusion of Innovations: Innovations progressively diffuse across a system or population and are sustained once a critical mass is reached and convinced. Change happens when a new idea for a programme or policy is communicated to a critical mass who progressively adopts the innovation. Everet Rogers famously refers to several categories of adopters — innovators, early adopters, early majority, late majority, and laggards (Rogers, 1962, see Greenhalgh et al. 2004 for a discussion). There is a lot of this kind of theory in agricultural research such as in The Consultative Group on International Agricultural Research’s (CGIAR) efforts in Ethiopia. There are, of course, variations of this theory which speak to the adaptation of innovations, and re-innovation (see Garb and Friedlander, 2014 for a discussion).
  • Evidence-based Policy: “Rigorous” evidence is what changes minds. While this pretends not to be a theory at all, evidence-based policy is descendent of rational choice theory which suggests that stakeholders are rational, self-interested, largely apolitical, utility maximisers who just need to be shown the evidence and then they will adopt the policy, innovation, tool, etc. “Rigour” is commonly, and misleadingly, equated with evidence from Randomised Control Trials (RCT), as described in a recent blog by Daniel Handel on Vox (see Cartwright and Hardie, 2012 for a powerful critique). Guerzovich et al. articulated this as a “best practice” pathway to scaling up social accountability (see the Malawi example from CARE International), and it fits with David Jackson’s presentation of the practical innovators “tribe” in anti-corruption programming. An adaptation of evidence-based policy is nudge theory, as personified by the “Nudge Unit” or in the new Australian Center for Evaluation (ACE). See my critique of the Nudge Unit’s logic in “Randomista mania.” See Andrew Leigh’s unconvincing justification for ACE and method hierarchies and see Patricia Rogers’ forceful critique of ACE’s foundations in her recent blog.

Hopefully, outlining some of these common theories offers further food for thought. There are various different ways in which change happens. We need to carefully take (political) context into account, the opportunity structures in front of us, and the comparative advantages we may (or may not) have at our disposal. Reflecting more seriously on what kinds of theories underpin our theories of change is one way to help us get there. If nothing else, they should help refine social science theories with more contact with the real world, and hopefully that can be more of a two-way street.

Thanks to Florencia Guerzovich for discussing through some of these variations with me.



Thomas Aston

I'm an independent consultant specialising in theory-based and participatory evaluation methods.