P O S T E D B Y A L B E R T
1. Minding the nature of cause and effect. In philanthropic and nonprofit work, we model causes and effects when we plan a program, evaluate an intervention, articulate a theory of change, sketch a logic model, or, more generally, attempt to give an account of what the heck it is we’re trying to do.
2. Causal attribution. At times, after making sober pronouncements about our fundamental inability to identify our work as the cause of a desired effect, we will still gladly take credit for the latter. This is what’s known in the biz as the “problem of attribution,” which, as many of you know, is really no problem at all.
3. Representing causal attribution. We use pictures to represent the connection between our intervention (i.e., the cause, labelled C in Figure 1, below) and its intended effect, dutifully labelled E.
One troubling feature of this cause and effect diagram is its simplicity. This is not what we pay our consultants top dollar for. Prima facie, it fails to represent the fact that it usually takes more than one step to get from our intervention to our intended effect, so that instead of the situation pictured in Figure 1, we require something more like:
Our initial intervention, C1, causes effect, E1, which is itself a cause, C2, of another effect, E2, and so on, until we arrive finally at our intended effect. Here at least we’re getting more causal connections for our money!
Unfortunately, even Figure 2 will not do for any theory of change worthy of the name. As we know, effects can—and in fact almost always do—have multiple causes, and their causes have causes, and so on. The situation for most social interventions is more like that depicted in Figure 3:
This is called a fishbone diagram1, for obvious reasons, and of special interest to us is the fish’s spine, the primary chain of causes and effects (here represented by a dark arrow) that leads to our hoped-for outcome. This diagram faithfully captures the fact that each cause has its own causes (labelled “secondary”), and that these causes too have their causes (labelled “tertiary”), etc.
Judging by the shape of this particular fish, I would guess that it’s a flounder, which, if you think about it, is not a bad metaphor for what many of us do in philanthropy.
4. Playing Pachinko with people’s lives. Are we there yet? No, not even close. In spite of our flounder’s impressive skeletal structure, it turns out we might be barking up the wrong fish. Figure 3 represents only one possible chain of events arriving at our final destination, the “Effect” in the rightmost box. But as we know from experience, just about any intervention we dream up can lead by multiple paths to our desired outcome.
Represented in Figure 4, for example, is a theory of change for a program that gives tenants representation on the management board of their public housing project, with the intended effect of improving the condition of the property.2 I've excluded all possible secondary and tertiary causes to simplify the exposition, but you’ll notice immediately that there is not one, but several, paths that lead to our desired goal.
An extraordinary feature of the theory of change depicted in Figure 4 is that despite its overwhelming complexity and despite all the vagaries of human action, causes and their effects tumble like Pachinko balls toward their lowest state of potential energy, here represented by the box labelled “Improved condition of property.”
Is this reasonable?
To answer this question, it’s worth pausing to consider the reason why some grantmakers require mind-bending theories of change from their grantees. Theories of change, like the one depicted in Figure 4, are intended to make explicit our beliefs about the causal paths that lead plausibly from our social interventions to our desired outcomes. But therein lies their greatest weakness: Our theories of change are limited primarily by our imaginations. As I’ve argued elsewhere, we can imagine a thousand reasons, none of them pictured in Figure 4, why the placement of residents on the management board of a public housing project might lead ultimately to an improvement in the property. One of the resident managers, for example, might have a brother-in-law whose contracting company does excellent work for half the price; or one of the resident-managers might be very successful at organizing volunteer work crews that dedicate one Saturday each month to repairs and upkeep; or the very presence of residents on the management board might inspire a local donor to endow and thereby double the public housing project’s maintenance budget; etc.
It’s also possible that events might take one or all of the paths pictured in Figure 4 and the condition of the property not improve one bit!
Once we introduce human actors into our causal chains we move outside of the art and science of philanthropy [sic] and into the realm of fortune telling (with some exceptions that I describe below). Requiring grantees to construct detailed theories of change for their interventions is the kind of cruelty that only a grantmaker can think to inflict.
5. Gettin’ our loops on. Reality is much more complicated than even Figure 4 suggests. In the world of organic and inorganic beings, certain anticipated effects “loop back” to reinforce, or, in some cases, impede other elements of our program design. Consider, for example, the diagram below, in which various lines of causes and effects lead ultimately to “failed states.”3
In this simplified exposition, notice how GLOBAL WARMING is part of a reinforcing loop that includes RISING SEA LEVEL, SALTING OF COASTAL AQUIFERS, SPREADING WATER SHORTAGES, PUMPING FROM DEEPER WELLS, and INCREASING ENERGY FROM FOSSIL FUELS. Feedback loops like this are the rule rather than the exception for many if not most social interventions. The great pioneers of system dynamics were quick to acknowledge their importance and model their effects in complex social phenomena. We ignore them at our peril.
Notice also the mixing of physical causes (e.g., RISING TEMPERATURES) with human causes (“increased demand for food”), another perplexing feature common to many theories of change.
6. Felicity conditions. To complicate matters further, even the simple cause and effect connection pictured in Figure 1 conceals enormous conceptual difficulties. The figure appears to assert that cause C is inevitably followed by effect E. But this cannot be what it depicts.
Suppose C is my request that you pour me a glass of wine, and E is your pouring me a glass of wine. You might decide to pour me a glass of wine without my asking you (and for which, by the way, thank you). But if I ask you to pour me a glass of wine and, moments later, you do so, we can safely assume that my request was the cause of your kind action.4
If, however, you have no more wine, or I ask you impolitely and you decide to deny my request, or a piano falls on your head before you have a chance to reach for the bottle, the intended effect E will never materialize. The cause C leading to the effect E depends critically on the presence of certain conditions (wine in the bottle) and the absence of others (falling pianos).
Shouldn’t these felicity conditions somehow be part of our theory of change? If so, where do we put them? Even more importantly, is there enough time before our sun goes nova to adduce all the felicity conditions that would make possible your pouring me, upon request, a nice glass of Merlot?
7. The full monty. Once we take into account the inevitable feedback loops, the required felicity conditions, the multiple branches emanating from a simple intervention, the secondary and tertiary causes, etc., here is the complete theory of change for the simple act of mailing of a letter:
8. Some preliminary conclusions/observations. Grantmakers fall into error when they engage in armchair speculation about effects that lie at the distal ends of long causal chains. If our intervention depends on a tightly choreographed march of causes and effects, we will almost certainly be disappointed.
Are we therefore to abandon trying to pinpoint causes and effects in social interventions? Are we to declare the causal nexus an impenetrable mystery? Not at all. Causes and their effects are as much at work in the social as in the physical realm and we need to understand them as best we can. As for felicity conditions, most times when I ask you to pour me a glass of wine, there will be no menacing pianos about.
Our problems begin when we overreach. Our arrogance is on full display when we ask grantees to engage in exercises beyond the abilities of Stephen Hawking. Moreover, a great deal of what’s been written and discussed about theories of change and other jetsam from the world of “strategic philanthropy”—this post included—is a bunch of arm-waving that will continue to enrich technocrats like me but fail to advance the cause of justice, the latter having much more to do with truth and repentance than with logic models. But I digress ...
9. Keeping it real. That said, even in the domain of human behavior, even with all the quirks of human personality, there are surprising regularities we can exploit in order to do good for the world. I’ll describe some of the more esoteric examples in subsequent posts. For now we can at least acknowledge that philanthropy is filled with simple, effective interventions, supported by equally simple theories of change. For example:
[1] I make a grant to Organization XYZ to pay the light bill → Organization XYZ uses my grant to pay the light bill → Organization XYZ’s lights stay on
Nice. Elegant. Too simple, perhaps, to garner you a Genius Award, but arguably the kind of intervention most critical to the health of our republic.
In the Spring 2010 issue of the Stanford Social Innovation Review5, Paul Brest, then-president of the William and Flora Hewlett Foundation, took two of my colleagues, William Schambra and Bill Somerville, to the rhetorical woodshed. He criticized them for their skepticism about theories of change, for advocating that we make grants to extraordinary nonprofit leaders without requiring them to spell out the causal chains that lead to their good outcomes. “These skeptics,” Brest argued, “are implicitly analogizing grantees to idiots savants—individuals who are able to do complex calculations … in their heads without knowing, let alone being able to explain, how they do it.”5
Putting aside the issue of the misplaced idiot savant analogy6, I believe that Schambra and Somerville have it exactly right. Their idea of providing funding to great nonprofit leaders who have succeeded in the past is supported by a compelling theory of change. It looks something like this:
[2] I make a grant to an organization run by a great leader who consistently delivers great outcomes year after year → The organization delivers great outcomes
Some of you will note the similarity between this theory of change and the following familiar example:
[3] The rotation of the earth has caused the sun to rise every day for the past five billion years → The sun will rise again tomorrow
There are no guarantees in the theory of change outlined in [2] just as there is no guarantee that the sun will rise tomorrow. But you can hardly do better than this as a grantmaker.
An organization’s “theory of change” can, in most cases, be directly inferred from what it proposes to do. Instead of agonizing over confusing and manifestly inaccurate cause-and-effect diagrams, simply apply the plausibility test: Is it plausible that the grantee’s actions might lead to the desired outcome? How plausible is it and how do we know?
10. Complaints department. The reader likely has many questions and objections. I’ve listed only those that show my argument to best advantage:
Objection: Sixty-eight percent of our grantees report benefitting from the construction of theories of change.
Reply: I’m not here denying that theories of change, and the process of constructing them, have helped many people in the nonprofit world. I’ve been assured that they have. But while I myself find it helpful to discuss my work plans with my dog, Gracie, I would never dream of requiring this of my grantees. We can perhaps settle the matter as follows: Divide your grantees into two groups. Ask the first group to construct theories of change for their interventions. Ask the second group to describe their work to their pets. Now poll the two groups to find which of them claims to have benefitted most from the exercise. I’ll publish your results on this blog. This, I believe, is what’s called “data-driven decision-making,” the alternative being, I suppose, to ignore the evidence of our senses altogether.
Objection: Figure 4 is overly complicated. It should include only those branches that are most relevant to our intervention.
Reply: Which branches are those? Are some of them perhaps missing from Figure 4? How do you know? What are your criteria for including some branches and excluding others?
Objection: Struggling with complicated theories of change makes our grantees stronger, more able to cope with the vicissitudes of grantmakers even more confused than you or I.
Reply: Point well taken.
-----
Notes:
1. From the RFFlow5 website, available at https://www.rff.com/fishbone.htm.
2. From Carol Weiss’s, “How Can Theory-Based Evaluation Make Greater Headway?” Evaluation Review, Vol. 21 No. 4 (1997).
3. From Lester R. Brown’s “Could Food Shortages Bring Down Civilization?” in the May 2009 issue of Scientific American.
4. This was the argument of the philosopher Donald Davidson in the essay “Actions, Reasons, and Causes,” presented at the 1963 meeting of the American Philosophical Association and published in the Journal of Philosophy 60 (1963). I’m not making this up.
5. Paul Brest, “The Power of Theories of Change” in the Stanford Social Innovation Review, Spring 2010. 6. Ibid., p. 48.
6. One of the interesting characteristics of idiots savants is that the complex calculations they perform in their heads are often correct. If the nonprofit world harbored the analogues of idiots savants—executive directors, for example, who could correctly intuit the interventions required for lifting whole communities out of poverty—I would strongly encourage them rather than force them to give an account of their methods. Despite their inability to explain how they arrived at their program designs, I would prefer that they changed the world the way they do, rather than the way most of us don’t.
Coming up: Human motions and human actions; cause and effect in the law; Donald Davidson and naïve belief-desire psychology; rational actor theory; Pavlov, Skinner, and Duhigg; the Law of Large Numbers; crowdsourcing; leadership development and other empowerment strategies; movement building; arts corridors and anchor institutions; tipping point arguments; and structural change.
Recent Comments