Pseudo-opinions (commenting on fictitious questions, even though they should not really have an opinion): Up to 69% of the respondents give a substantive opinion on the fictitious questionnaires; the better educated sin more.

Wolter F., Junkermann J. (2019) Antwortvalidität in Survey-Interviews: Meinungsäußerungen zu fiktiven Dingen (=approx. Response validity in survey interviews: expressions of opinion on fictitious things). In: Menold N., Wolbring T. (eds) Qualitätssicherung sozialwissenschaftlicher Erhebungsinstrumente. Schriftenreihe der ASI - Arbeitsgemeinschaft Sozialwissenschaftlicher Institute. Springer VS, Wiesbaden. December 30 2018 DOI https://doi.org/10.1007/978-3-658-24517-7_11

Abstract (automatic translation): The article examines the extent and factors influencing the expression of pseudo-Opinions. By this is meant the often documented phenomenon, according to which respondents also comment on fictitious, fictitious questions, even though they should not really have an opinion. The relevance arises on the one hand from the assumption that some respondents do not know real existing question objects, but still express an opinion. The results of such surveys would be distorted. On the other hand, the study of pseudo-Opinions allows to study the process of socially desirable responses with respect to the extent and determinants of a response bias. In addition to sociodemographic influencing factors and measures for incentives through social desirability, in particular the response reaction time as a proxy for the degree of cognitive elaboration is examined for its influence. This is done on the basis of theoretical considerations on respondent behavior, i.a. from the frame selection theory. In the CATI study (N = 499) conducted in Mainz, respondents were asked three fictitious sights in Mainz. It turns out that the amount of response bias through pseudo-Opinions is considerable; Up to 69% of the respondents give a substantive opinion on the fictitious questionnaires. In addition, the propensity to distorting responses varies according to simple socio-demographic characteristics such as age, gender, and education. An effect of the response response time can only be determined for one of the items. Here the latency acts negatively, i. Thinking longer when answering the question leads to fewer pseudo-opinions and thus less distorted data.