Satisfaction Revealed An Evaluation on Non- Response in the Afrobarometer Survey

Abstract

Satisfaction with democracy signifies an important indicator of democratic progress, yet has not been scrutinized for inaccuracy due to non-response. This paper evaluates the non-response bias for a satisfaction with democracy question in Round 3 of the Afrobarometer Survey. A causal model explains both the probability of satisfaction as well as the likelihood of response. A Heckman two step equation method identifies the presence of a strong selection bias. The results provide useful insights for the study of non-response as well as satisfaction with democracy in SubSaharan Africa. The Heckman corrected estimates of satisfaction are used to reevaluate a previous study of electoral quality and satisfaction with democracy.