Title: An introduction to Impact Evaluation (IE) for HIV/AIDS Programs
1An introduction to Impact Evaluation (IE) for
HIV/AIDS Programs
Léandre Bassolé ACTafrica, The World Bank
2Background
- Initial response to AIDS epidemic responding to
a crisis. - Build political commitment
- Establish institutions
- Scale up coverage of HIV prevention, treatment
and mitigation services - Now we need to know what really works?
3Traditional ME and Impact Evaluation
- ME monitoring process evaluation
- Is program being implemented efficiently?
- Is targeted population being reached?
- Are outcomes moving in the right direction?
Descriptive analysis
- What was the effect of the program on outcomes?
- How would outcomes change under alternative
program designs? - Is the program cost-effective?
Causal analysis
4Why does impact evaluation matter?
- To know if the program had an impact and the
average size of that impact - Assess if policies work
- Assess the net benefits/costs of the program
- Assess the distribution of gains and losses
5The IE problem
- What do we mean by impact evaluation ?
- Impact the difference between the relevant
outcome indicator with the program and that
without it. - However, we can never observe someone in two
different states of nature at the same time. - While a post-intervention indicator is
observed, its value in the absence of the program
is not, i.e., it is a counter-factual. - So all IE deals with overcoming the issue of
missing data. Requires counterfactual analysis.
6What we need
Given the problem of missing data (individuals
have only 1 existence) we can compare 2 groups
we need A counterfactual a control/comparison
group that will allow us to attribute any change
in the participant group to the intervention
(causality) what would have happened
without the program
7Common IE practices
- 1. Before and after 2. Participants-non-partici
pants - BUT, its difficult to assess the
- TRUE AVERAGE CAUSAL EFFECT
- How to solve the FUNDAMENTAL PROBLEM OF
EVALUATION?
8Comparison group issues
- Two central problems
- Programs are targeted
- Program areas will differ in observable and
unobservable ways precisely because the program
intended this - Individual participation is (usually) voluntary
- Participants will differ from non-participants in
observables and unobservable ways.
9Tools to identify a Counterfactual
- Randomized Designs
- Quasi-experimental Designs
- Matching
- Instrumental variables
- Regression discontinuity
10Some general principles to consider when planning
an IE
- Government ownershipwhat matters is
institutional buy-in - Relevance and applicabilityasking the right
questions - Flexibility and adaptability
- Horizon matters
11Summing upMethods/Practicalities
- Randomization is the gold standard
- Be flexible, be creative use the context
- IE requires good monitoring and monitoring will
help you understand the effect size
12Summing upMethods/Practicalities
- Making IE works for you may require a change in
the culture of project design and
implementation..that is to maximize the
evidence-based upon which policy decisions can be
made to improve the chances for success - Impact evaluation is more than a tool it is an
analytical framework for policy development
13THANK YOU