52 weeks of BetterEvaluation: Week 10: Having a theory in the theory of change

Patricia Rogers's picture 8th March 2013 by Patricia Rogers

There is increasing recognition that a theory of change can be useful when planning an evaluation. A theory of change is an explanation of how activities are understood to contribute to a series of outcomes and impacts. It might be called a program theory, an intervention logic, an outcomes hierarchy, or something else. It is usually represented in a diagram called a logic model, which can take various forms.

You’ll find more information about different ways to develop and represent a theory of change on the BetterEvaluation site at http://betterevaluation.org/plan/define/develop_logic_model

This week we are sharing information from a workshop held last week at the Community of Evaluators Evaluation Conclave, Kathmandu, where we discussed strategies for avoiding three common mistakes in using a theory of change for evaluation - having no actual theory, having a weak theory, and not really using the theory if change to inform the evaluation. We're discussing the first of these this week and will cover the other two in future weeks. .

Do you have other strategies to suggest? Examples to share? Or other common mistakes to discuss? Register and login to add your comments.

1. No actual theory of change

Some so-called theories of change don't actually explain how change comes about. The diagram can be read as "We do stuff, and stuff happens." and the only explanation is an arrow between boxes.

Clearly identifying a theory of change can help the evaluation to focus on what should be measured, guide analysis, and help with translating findings to new situations. Here are three strategies for ensuring there is an explicit theory of change.

Identify a causal mechanism theory and an action theory

A theory of change ideally has two components – a causal mechanism theory, which explains how change occurs, and an action theory, which explains how the program activities trigger the change process .  For example a public health program might have a component which seeks to change handwashing behaviour.  Its causal mechanism theory might be about people receiving credible new information about risks from dirty hands and the action theory might be about health workers providing information leaflets in local languages during routine health visits.  Another program might have the same causal mechanism theory but its action theory might involve peer to peer information in mothers’ groups.  A quite different program might be based on a different causal mechanism, such as  increasing access to clean water and soap, through providing free soap and water to households.

Draw on research-based theories of cha​nge

Research of various kinds can provide potentially relevant theories about how change occurs for individuals, families, households, and communities. 

For example, from a systematic review of theories of behaviour change for health promotion, Susan Michie and her team identified three clusters of theories, relating to capabilities (skills and knowledge), motivation (positive and negative, intrinsic and extrinsic) and opportunities. 

The COM-B system - A Framework for Understanding Behaviour. Michie et al. Implementation Science 2011 6:42   doi:10.1186/1748-5908-6-42

Use an outcome hierarchy format for the logic model rather than a pipeline

A third strategy to make it easier to articulate an actual theory of change is to use a different format for representing it in a diagram. In particular, it is often useful to show it as a chain of outcomes rather than as a pipeline diagram of inputs-processes-outputs-outcomes-impacts.

In a previous 52 weeks post, we’ve explored different ways of representing theories of change in logic model diagrams. http://betterevaluation.org/blog/drawing_logic_models


52 Weeks of BetterEvaluation. Click here to view past features

A special thanks to this page's contributors
CEO, BetterEvaluation.


rickjdavies's picture
rick davies


I strongly recommend yED (http://www.yworks.com/en/products_yed_about.html), free diagramming software, which I use a lot. It It has various layout options, including one which will generate a hierarchical structure of the kind referred to above. It will also import and export diagrams in a range of network formats. It is very flexible and has many bells and whistles

regards, rick davies

Anonymous's picture

No actual theory: a big problem. But getting a theory is not just cooking up a few words that strung together make grammatical sense; it needs to start with a structured enquiry to understand the setting and its circumstances (or environment), examine interaction pathways and pressures (aka influence factors), evaluate their contribution to the current picture, than devise interventions - ideally with the 'actors'. This follows the Check-Plan-Do cycle that John Seddon uses to good effect. https://beyondcommandandcontrol.com/library/systems-principles/check-pla...

SimonDavies's picture
Simon Davies

Dear Daffy

There has been some good conversation regarding theories of change in our Peregrine discussion group recently. I would encourage you to join the group and engage with the conversation there. Here's the link for you to join: Peregrine Discussion Group - Home

Warm regards


Anonymous's picture
Rick Davies

COM-B looks a lot like a recycling of Means, Motive and Opportunity as used in crime investigations (at least on TV). It seems that the field of evaluation loves reinventing the wheel! :-) 

Add new comment

Login Login and comment as BetterEvaluation member or simply fill out the fields below.