Survey Measures of Corruption in Evaluation

Session Number: 2975
Track: Democracy and Governance
Session Type: Poster
Tags: measurement development
First Author or Discussion Group Leader: Daniel M. Sabet [Senior Impact Evaluation Advisor - Social Impact]
Time: Nov 08, 2017 (07:00 PM - 08:30 PM)
Room: Poster 53

Theme: Learning to Enhance Evaluation
Audience Level: All Audiences

Session Abstract (150 words): 

Evaluators constantly struggle with the challenges of survey measurement, and many concepts that we are interested in measuring risk strong social desirability bias.  Few concepts are as subject to this bias as the issues of corruption. In this paper, I explore the challenges of measuring corruption through surveys and analyzing corruption related survey data. Using two recent corruption focused surveys from an evaluation in Honduras, including a survey of 1,700 public employees and a survey of 800 firms who sell goods and services to the government, the paper first explores the impacts of question ordering and question design on responses. The paper also includes the results of a list experiment and considers the pros and the cons of using survey experiments to measure changes in corruption over time.  Finally, the paper explores pitfalls in data analysis and consider best practices in the analysis and presentation of corruption related survey data.