Measuring gender differences in elite behavior through surveys versus observation: what does the comparison reveal?

Surveys constitute the main method of studying elite behavior. A concern with survey data is that they reflect what elites report they do – not what elites actually do. Alternative, process-oriented approaches such as direct observation can help to address this concern. Examining elected representatives’ engagement in debates concerning service provision in the local councils of Albania, I show that observation data – opposed to survey data – portray a picture that is less positively biased towards men and less divided along male and female domains of representation. Findings underscore the importance of using measures that tap into how the behavior of political elites unfolds in practice.