Cross-nationally equivalent data accuracy? Item nonresponse analysis of the generations and gender surveys and implications for international comparative research

Andrej Kveder, Netherlands Interdisciplinary Demographic Institute (NIDI)
Nicole Hiekel, Netherlands Interdisciplinary Demographic Institute (NIDI)
Julia Mikolai, Utrecht University

Is data accuracy equivalent across countries and cultures? Would differential patterns of data quality have an impact on the results of international comparative research? The aim of this paper is to present a comprehensive analysis of item nonresponse patterns and its determinants and to critically evaluate the implications for comparative research. Generations and Gender Surveys (GGS) from eight countries will be used for the analysis. GGS is a panel survey of a nationally representative sample of 18-79 year-old resident population with at least three panel waves and an interval of three years between each wave. Only first wave data (due to current availability) will be used for the present analysis. V-known meta-analytical models will be used to estimate simultaneous logistic models of item nonresponse propensities using respondent and questionnaire characteristics as the two levels of explanatory variables. The item nonresponse models will be initially run on the pooled dataset of all eight countries. In the second stage of the analysis the models will be separately estimated for each country and results will be compared across countries. The results from both estimation stages will contribute to a better understanding of the survey response process and of the validity of international comparative research.