A Bayesian network analysis of survey climate and response intention in official statistics

Declining response rates and growing concerns about nonresponse bias threaten the quality of official statistics. Although prior studies have identified various factors associated with survey cooperation, such as privacy concerns, response burden, and political trust, most have examined these determinants independently without modeling the dependency structure among them. This study applied Bayesian networks to data from a web-based survey conducted in Japan to explore the probabilistic dependency structure linking broader social orientations, survey attitudes and perceptions, and response intention in official statistical surveys. The analysis identified relationships in which perceived social fairness was associated with trust in official statistics, which, in turn, was associated with survey response intention. Perceptions of others’ honesty in answering sensitive income questions were linked to individuals’ willingness to respond to such items, which was in turn associated with overall participation intention. Community participation was associated with opinions on social issues, which were in turn linked to response intention toward official statistical surveys. These findings suggest that intention to cooperate with official statistical surveys is embedded in broader social orientations. Understanding these dependency structures may help generate hypotheses about variables that could be examined in future studies of actual nonresponse and nonresponse adjustment.