Sunday, August 22, 2021

Reconciling the need for both horizontal and vertical dimensions in a Theory of Change diagram


In their usual table form, Logical Frameworks are strong on their horizontal dimension but weak on their vertical dimension. On the horizontal dimension is the explanation of what kind of data will be collected and used to measure the changes that are described. This is good for accountability. On the vertical dimension is the explanation of how events at one level will connect and cause events at another level. This is good for learning.  But unfortunately LogFrames often simply provide lists of events at each level, with relatively little description of which event will connect to which, especially where multiple and mixed sets of connections might be expected between events. On the other hand diagrammatic versions of a Theory of  Change tend to be much better at explicating the various causal pathways at work, but weak on the information they provide on the horizontal dimension - on how various events will be observed and measures. Both of these problems reflect both a lack of space to do both things and different relative priorities pursued within those constraints.

The Donor Committee for Enterprise Development (DCED) has produced a web page based Theory of Change to explain its way of  working, which I think points the way to reconciling these conflicting needs. At first glance here is what you see, when you visit this page of their website.

The different causal pathways are quite visible, more so than within a standard LogFrame table format. But another common weakness of diagrammatic versions of Theories of Change is the lack of explanation of what is going on within each of these pathways. The DCED addressed this problem by allowing visitors to click on a link and be taken to another web page, where visitors get a detailed text description of the available evidence, plus any assumptions, about the causal process(es) that connect the events connected by the arrow.

The one weakness in this DCED ToC diagram is the lack of detail about the horizontal dimension- how the various events described in the diagram will be observed./ measured and by who and when and where. But this is clearly resolvable by using the same approach with the links: enable users to click on any event and be taken to a web page where this information is provided for that specific event. As shown below: