Comparability of Corruption Measures in Cross-National Surveys

How to best measure corruption, the concept that is secretive in its nature? Despite the growing academic interest in cross-country corruption research, the question remains open. Many survey researchers propose to ask about perceptions as opposed to corruption incidents, which respondents may be unwilling to declare. However, the new developments of crime victimization research show that respondents are less reticent in answering questions on illegal behavior than it was previously assumed. It motivated for the raise of a new approach of measuring corruption, where actors are asked directly about their recent corruption experiences. Both measures - corruption perception and corruption experience – aim to best estimate the unobserved real corruption. Yet the relation between the two remains disputable, as well as the understanding which measure is more accurate, if any. The presentation offers the analysis of main benefits and limitations of the two currently dominant corruption measures in cross-national surveys. On the example of corruption measures, I raise the question of assessing equivalence and increasing comparability on the statistical analysis phase of a study, as opposed to design or implementation stages. The analysis is based on the Global Corruption Barometer, the European Quality of Government and the Life in Transition surveys, which cover the same 30 European countries in 2010. The results aim to open the discussion on assessing measurement quality in already existing survey data for the purpose of substantive analysis and will offer an illustration of practical challenges in substantive research with different measures of the same construct. I also intent to present two new research tools created as the result of my dissertation – a publicly available common file with the documentation of corruption variables in comparative survey projects and the harmonized dataset created for the purpose of the study of corruption in education sector in Europe.