Wednesday, May 21, 2014

Resources on Impact Evaluation

This post consolidates a list of impact evaluation resources that I usually refer to when I am asked about impact evaluations. 

This cute video explains the factors that distinguishes impact evaluation from other kinds of evaluation, in two minutes. Of course randomization isn't the only way of credibly attributing causes and effects - and this is a particularly hot evaluation methodology debate.  For an example of why this is sometimes an irrelevant debate - see this write up on parachutes and Chris Lysy's cartoons on the topic.

Literature on the Impact Evaluation Debate

The Impact Evaluation debate flared up after this report, titled "When will we ever learn" was released in 2006. In the States there also was a prominent funding mechanism which required programmes to include experimental evaluation methods in their design, or not get funding (from about 2003 or so).

The bone of contention was that Randomized Control Trials (RCTs) and Experimental methods (and to some extent Quasi Experimental Designs) were held up as the "gold standard" in evaluation. Which, in my opinion, is nonsense. So the debate about what counts as evidence started again. The World Bank and big corporate donors were perceived to push for Experimental Methods, Evaluation Associations (with members committed to mixed methods) pushed back saying methods can't be determined without knowing what the questions are. And others pushed back saying that RCTs are probably applicable in only about 5% of the cases in which evaluation is necessary.

The methods debate in Evaluation is really an old debate. Some really prominent evaluators decided to leave the AEA because they embarked on a position that they equated with "The flat earth movement" in geography.  Here is a nice overview article, (The 2004 Claremont Debate: Lipsey vs. Scriven. DeterminingCausality in Program Evaluation and Applied Research: Should ExperimentalEvidence Be the Gold Standard?) to summarise some of it.
The Network of Networks in Impact evaluation then sought to write a guidance document, but even after this was released, there was a feeling that not enough was said to counter the "gold standard" mentality.  This document, titled "Designing impact evaluations, different perspectives" provides a bit more information on the "other views".

Literature on Impact Evaluation Methods
 If you are interested in literature on Evaluation Methods, look at Better Evaluation to get a quick overview.

I like Cook, Campbell and Shadish to understand experimental and quasi experimental methods, but this online knowledge base resource is good too.
For some resources on other more mixed methods approaches to impact evaluation, you need to look at Realist Synthesis, General Elimination Method, Theory Based Evaluation, and something that I think has potential, the Collaborative Outcomes Reporting approach. 

The South African Department for Performance Monitoring and Evaluation's guideline on Impact Evaluation is also relevant if you are interested in work in the South African Context.

No comments: