“A leads to B because C”: how do I put this into a Theory of Change?

Just got asked an interesting question by Randy Puljek-Shank: How do we deal with “because” in Theories of Change? His example was this (I’ve simplified it slightly):

If we provide employment for these ex-combatant youths, then we will reduce the likelihood of them being recruited into these particular militant groups, because we know in general it’s unemployed youths who are the most likely to be recruited into militant groups

This argument has this structure:

If A, then B (because in all cases like A, things like B happen)

… so, A falls under “cases like A”, and B falls under “cases like B”. We are arguing from the general to the specific, appealing to a more general fact to derive a more specific one. (This kind of logical structure is related to the concept of “mechanism” in the Realist Evaluation tradition, although there it is used more specifically for decisions made by people.)

We do this kind of reasoning all the time when we construct Theories of Change.

So Randy’s question is, how can I add this explanation into a Theory of Change?

The short answer is: nowhere. You don’t have to put the evidence for (any part of) a Theory of Change into the Theory of Change. Each mini-Theory which makes up a larger Theory of Change is anyway an empirical claim which is more or less right or wrong. You should be prepared to defend and give evidence for any particular link in your Theory of Change, whether that argument is from the general to the specific or some other type of argument. But we don’t normally write the evidence as part of the actual Theory diagram because it would make it too verbose and hard to read.

(Just to clarify: this single step isn’t on its own a Theory of Change - for a start, there are no Objectives, or, as we say in Theorymaker, “valued Variables”; it is merely a fragment of one.)

Sure, you can add arguments defending a particular link into the diagram (or links, as in general there might be more than one link influencing a Variable). You might particularly want to do that when the evidence for some particular link is a bit weak or not obvious.

Most simply, and in the case where there is just one incoming link, I’d just annotate the arrow:

reduced likelihood of youths being recruited into militant groups

 (Evidence - unemployed youths are the most likely to be recruited into militant groups, see study 1 by Mujo and Bloggs)provide employment for ex-combatant youth

If there is more than one incoming link and they interact with one another (i.e. no linear superposition), you’re probably better annotating the downstream Variable.

reduced likelihood of youths being recruited into militant groups//Evidence: unemployed youths are the most likely to be recruited into militant groups, but only if they also adopt alternative value structures, see study 2 by Bloggs and Durakovic

 provide employment for ex-combatant youth

 youth adopt alternative value structures

Finally, if the same evidence applies to several mini-Theories within the same Theory, you might want to supply a surrounding box and put the annotation in the box.

As Randy also points out, these “because” cases may overlap with statements about “assumptions” in Theories of Change. The logic that goes with assumptions is a bit different because they refer to additional empirical factors influencing a Variable which for one reason or another we have not actually included in the Theory (although we could have done). Whereas “because” statements are not usually about missing factors which could or should be added to the Theory of Change; as we have seen, they refer us to background evidence for the links in the Theory.