We Have Come a Long Way and We Have a Long Way to Go: A Cross-Survey Comparison of Data Quality in 16 Arab Countries in the Arab Barometer vs the World Values Survey

Abstract

With the launch of the Arab Barometer (AB) project and the incorporation of Arab countries in the World Values Survey (WVS) in the 2000s, public opinion scholars have increasingly turned their attention to the Arab region. However, remarkably little is however known about the quality of these data. To our knowledge, Arab surveys have never been scrutinized in a systematic empirical cross-survey study. Therefore, this study compares sixteen surveys from the AB with sixteen from the WVS concerning four attitudes widely studied by substantive scholars: generalized and institutional trust and gender equality in education and in politics. We assess the comparability of their univariate distributions and their predictors in multivariate models. Our results show considerable diversity across and even within surveys in quality, indicating that blanket statements on Arab surveys’ (lack of) quality are inappropriate. In a minority of tested cases (17%), the conclusions of scholars on what predicts trust or gender equality depend completely on the chosen data source. We also test whether often-heard reasons for Arab surveys’ supposed lack of quality explain the diversity in survey quality. Our results show that neither sample differences nor enumerator fraud drives discrepancies, but there might be some influence of socially desirable answers

Keywords

• data quality
• survey comparison
• comparative data analysis
• gender equality
• trust
• Arab Barometer