Sampling

Impact assessment studies are often affected by selection bias and misrepresentation of their target clients. In order to overcome these problems and make these evaluations more reliable, an appropriate approach to sampling is required.

This paper aims to explain:

• What sampling is and why we need it;
• How bias can be taken into account by adequate sampling techniques;
• The initial requirements necessary for the different sampling methodologies;
• How to adapt sampling to the purpose of the analysis and link it to practice.

In conclusion, a list of main recommendations for designing and assessing sampling design is provided:

• Clarifying about the specific purpose of any particular part of an impact assessment is important to guarantee its credibility;
• Combining statistical, qualitative and participatory sampling methodologies is the most cost-effective and credible way of conducting an impact assessment;
• Ensuring representation of different stakeholders in both the assessment process and decision-making in order to link impact assessment with practice and policy;
• Countering the inherent biases against the poorest and most disadvantaged individuals and communities in order to reach reliable conclusions about policies for poverty elimination.