View article

In face-to-face interviews, both unit nonresponse and item nonresponse can be considered as a negative reaction of the respondent. Given this basic common element, we test the hypothesis that the same factors are responsible for unit and item nonresponse and that both kinds of nonresponse are related to each other. To investigate the relationship between item nonresponse and unit nonresponse a panel survey is used. With panel data the amount and pattern of item nonresponse realized during the first wave of a panel can be used to predict the unit nonresponse in a second wave of a panel. The results show that±in the context of a panel survey and controlling for some relevant respondent characteristics±several kinds of item nonresponse during the first interview are indeed related to the unit nonresponse of a second interview.