Item nonresponse and fuzzy logic

As compared to the problem of unit nonresponse in standardized interviews, the topic of item nonresponse is less well explored. Hence this article attempts to explain item nonresponse by the fuzziness of the related social situation, which is supposed to lead to a failure in the formation of attitudes about the respective interview-item. The author assumes that interview-items have for the respondents a fuzzy truth, which varies on a continuous scale between 0 = “false“ and 1 = “true“. By this conceptualization, some interview-items may have a fuzzy truth in the middle between the mentioned poles, i.e. at the level 0.5 = “indeterminate“. Consequently, the main hypothesis of the present study postulates for this situation an increased level of item nonresponse. The empirical part of the article tests this hypothesis with data of the European Values Study (EVS 2008) about self-reported religiosity. For this purpose it assumes a correspondence between item response- and fuzzy truth-functions. It turns out, that the main hypothesis of the article holds true for the full sample of the EVS.