reflecting on being radical: integrating theories of change as practice

last week, craig valters published new work on theories of change. he calls not for a new tool but for a more careful approach to practicing and engaging in development. that is, changing the state of the world for someone. and learning from it. and, ideally, communicating that learning. (craig is pessimistic that we are near actually ushering in a ‘learning agenda’ to replace the ‘results agenda.’ on this, i hope he is wrong.)

.

in this post, i aim to echo and expand on the idea of theories of change as allowing “space for critical reflection” (p. 4) and to push back slightly on two of the outlined ‘key principles’ of a theory of change approach: being ‘locally led’ and thinking ‘compass, not map.’

.

i should make two disclaimers, given points raised both in the paper and in suvojit’s follow-up blog. the first is a musing, though i have adopted the ‘theory of change’ language along with the herd. i wish we could still revise it to ‘hypotheses of change’ or ‘ideas of change’ or ‘stuff that might matter because we thought hard about it, looked at what had been done before, and talked to people about what could be done now.’ or something else catchier but far more tentative and humble than ‘theory.’ alas.

.

the second is a confession. i really like boxes and arrows. not as the definitive product associated with a theory of change but as some means of organizing ideas that people can stand around, look at, point to, and say, “have we learned anything about how this arrow really works?” while i wouldn’t want to foist the need for a visual on anyone, especially if it is just going to end as a bad flowchart, i feel i should at least lightly advocate that a visual can be a useful tool for learning and may be more friendly to revisit than a lengthy narrative. and be pretty.

.

spaces for critical reflection

i strongly agree with craig’s focus on process and learning, emphasizing the theory of change as an approach, not just a product. it requires a change in practice, not deliverables. lukewarm commitment will relegate theory of change products to templates and checklist items to complete: as valters warns against, suvojit worries about, and duncan green and others have questioned in the past. instead, the approach requires a physical and mental and temporal space actively held open to think and reflect and revisit.

.

to possibly only reword what has already been said: theories of change products will only matter for learning if the approach requires them to be as important at the beginning of the project as they are at the end. ‘doing’ theories of change requires thinking about the end at the beginning — but also to look backwards later.

.

theory of change products (narratives, diagrams) can and should be used to help organize learning along the way — whether revisited as part of the program diaries (great idea!) or as other structured (partly by the product), purposive ways of stopping the good work and reflecting on how we initially thought things would be working and how they are actually progressing. for me, with a goal of generating lessons from which others can learn, documentation is important. some may see this as a hindrance to engagement, so that remains an open question for each organization. in any case, i like craig’s questions related to ‘prioritizing learning,’ which encourage organizations (and donors) to be both explicit and transparent about learning goals (what, for whom, to make what decisions) from the outset.

.

one idea is to think about, as products, an ex ante theory of change and an ex post version. (the process of) producing these would, done well, encourage learning around where the two versions differ and would serve as a way to help organize learning about those differences. by “done well,” i emphasize my own principle of a theory of change approach, which is that a theory of change is never one person’s all-nighter, assembling something to submit as part of a report. to the extent that donors may play a role in encouraging grantees to reflect on how they thought change would happen and how it actually did (if, indeed, it did), they should. this means not just telling grantees to do this but facilitate in terms of finances and time and and convening power and guidance, as needed. active reflection is needed; sometimes external pressure can bring this about.

.

being locally led

by pushing back on being “locally led,” i am obviously not going to suggest that local voices and opinions shouldn’t be central to the process of figuring out what kind of change needs to brought about and how to go about it. but i would like to bring in some nuance as well generally advocate being led by multiple perspectives — top-down and bottom-up — rather than a singular idea of ‘the local.’ craig suggests some mild discomfort with this as well in his footnote that warns us against assuming there is an archetypal and wise end-user with all the answers.

.

first, local-ness is diverse (which almost goes without saying). iteratively, this diversity needs to be reflected in a theory of change process — and theories of change need to reflect this diversity. (with an evaluation hat on, what i am partly saying is that sub-groups in which heterogeneous treatment effects are expected should be visible early in the theory of change process.) beyond the tired statement that ‘the local’ — as possible recipients of a program/policy — is not a homogeneous mass, a key distinction is between ‘intended beneficiaries’ and ‘front-line implementers.’ both are local in important ways and the views and motivations of both should be reflected in theory of change processes and products (as noted by pritchett et al in considering structured experiential learning.)

.

second, local thinking can be constrained precisely by local-ness. researchers and practitioners are important conduits for transferring in good ideas from ‘there’ to consider whether they may work ‘here.’  being ‘locally rooted’ or ‘locally grounded’ may be slightly closer to my sense of an ideal than being ‘locally led,’ in part because…

.

third, i’d like to think there is still a role for theory (or Theory) in program theory and theories of change. again, practitioners and researchers play a role in helping to negotiate an intelligible space between (1) these ideas, (2) what cannot be done in a given local context (because of lacking physical, human, financial resources), and (3) what people don’t know if can be done locally because it has not been tried here in a particular way. ‘the local’ can best describe the various ways in which things have been happening and are critical voices in determining what should happen and generating ideas about what could happen — but it is not the only source of ideas and inspiration and energy and convening power.

.

thinking compass, not map

i understand this impulse for this principle — a reaction to over-design and over-planning — but i am going to push back a little anyway. or at least advocate for a compass plus a really amazing postcard* from the destination. my feeling is that organizations spend a lot of time talking about what they are going to do but not in visualizing or describing what the changed state of the world will look like and how people will behave differently in it if change actually happens. this (considering individuals, here and here) should be specific exercise within the approach, asking for different stakeholders: what do i do now when i wake up and go about my day and how will i do things differently if the intended good change comes about?

.

the discussion about how things will be different — and how we’ll know they are different — should be central to conversations about how change may happen. this can help to uncover some of the assumptions that are so important for theories of change. in addition, a detailed description of what change or success will look like is fairly important for understanding whether we’ve gotten to where we want to go (and if not, why not).

.

knowing where you’re going is different from saying you’re certain about how you will get there and when. over-planning and over-design reflect, i think, (donor-forced?) over-confidence about the route and mode and timing of transport. this is distinct from over-clarity about where you are trying to go, which is too often lacking. by all means, “go west;” carry with you some way of telling whether you’ve made it to oregon or not.

.

in sum

to briefly conclude, a big thank you to craig, who has done a great job laying out some of the promises and concerns of embracing theories of change as an approach to practicing and doing change. i echo the call to emphasize learning (with a distinction between failure to learn and the failure of a project to deliver) and add a few ideas about how the theory of change (as a product) might structure the learning process and its outputs. i explicitly call for the product to play a renewed role after the project launch, as a way of encouraging the process. i push back a bit on being ‘locally led,’ towards being ‘locally grounded’ and/but ‘multi-perspectival.’ i also push back on ‘compass, not map’ in favor of ‘compass and postcard,’ with explicit intent to encourage practitioners and researchers to have a good idea where they are going and to draw and annotate a (not the) map along the way.

*for jim gaffagan fans, this would be a postcard that goes beyond saying “this city has big buildings, i like food, bye” (around minute 2:40).

Advertisements

a small point on thinking about IEs by sector

in looking at how the accumulated impact evaluation evidence in the social sciences is distributed — perhaps with an eye toward making the case for where to concentrate new funding — there is a tendency to categorize studies by sector. with this lens, it is clear that the evidence base remains dominated by health, social protection, agriculture, sanitation.

.

this does not reflect the actual research questions or topics asked by these studies. for example, there are many ‘health’ studies — distributing health products, generally — but very few on health systems. the line is, of course, not clear: would vouchers to encourage entering the health system in terms of institutional/hospital-based delivery of babies fall in the distributing ‘stuff’ or building ‘systems’ category? such parsing would take further and careful thought.

.

nevertheless, for those looking at where the bulk of the literature falls — and where new research is needed — it may be helpful to move beyond sector codes (which are easier to find and seem to dominate the classification strategies of IEs) and start teasing apart ‘stuff’ from ‘systems’/institutions work. to do otherwise is fairly lazy thinking.

.

a small addition, added on 31 october 2015, based on a new paper called ‘when will we ever learn… to change policy?”

In terms of topical focus, a majority of evaluations in health focus on prevention, control and treatment of specific diseases or on the effects of specific health technologies while less research focuses on the health systems that would support these interventions at scale.4 There are critical evidence gaps, for example, regarding front line health workers (Frymus et al., 2013) as well as managers in health systems (Rockers and Bärnighausen, 2013). An ongoing, wide-ranging systematic review of strategies to improve healthcare provider performance rated only 14% of 490 included studies as having a “low” risk of bias (Rowe et al., 2014).

.

This relative lack of quality impact evaluations in health systems research may be due in part to a lack of funding as donors in health prefer to support other types of health research that is seen as more prestigious or science-driven (Hafner and Shiffman, 2013). Health systems are also complex, involving multiple elements that affect ultimate service delivery (Hafner and Shiffman, 2013). This relative lack of rigorous evaluations of approaches to strengthen health service delivery in developing countries is not for a lack of policy interest or resources devoted to this issue (Hafner and Shiffman, 2013, Willey et al., 2013).

.

Strengthening health systems is one of the top priorities of the WHO and numerous multi- and bi-lateral organizations have been working in partnership with ministries of health in efforts to improve health service delivery (Hafner and Shiffman, 2013, Willey et al., 2013)… Learning from existing efforts to strengthen health delivery, particularly those at scale, requires collaboration between researchers and implementing agencies…