Surveying Bureaucrats: Best Practices to Reduce Attrition and Non-response Bias

• In Event: Bureaucrats in Local Government

Abstract

Fielding surveys with civil servants (a hard-to-reach population) presents inherent challenges, including small sample sizes, high attrition rates, non-response bias, and limited statistical representativeness. This research addresses these issues through a novel survey of 5,243 bureaucrats in the Brazilian federal government, emphasizing a methodological perspective. Our approach aims to uncover insights into survey design while identifying factors contributing to attrition.
Our survey incorporates a diverse range of questions, spanning sociodemographic characteristics, professional ethics, and political preferences. Employing varied formats, including Likert scales, open-ended queries, and a multiple-task conjoint experiment, we discern the impact of question types on attrition and explore individual-level characteristics influencing dropout likelihood.
Our study’s findings are particularly relevant for scholars specializing in survey research on bureaucrats. By addressing challenges in retention and offering recommendations on questionnaire design, our research contributes to refining best practices in this field. Our methodological advancements, thus, aim to enhance the robustness and representativeness of survey research within the public administration domain.

Authors

• João V. Guedes-Neto, Fundação Getúlio Vargas

• Gabriela Lotta, Getulio Vargas Foundation

• Luciana Cingolani, Hertie School of Governance

• Alexandre Gomide, Institute for Applied Economic Research — Ipea

• Mariana Costa Silveira, Fundação Getulio Vargas’s Sao Paulo School of Business Administration (FGV EAESP)