Intervieweffekte bei der Erhebung des Körpergewichts: die Qualität von umfragebasierten Gewichtsangaben

Abstract

While surveying measured weight is widely unpractical in national samples, self-reported weight is a simple and inexpensive method of collecting data. This paper deals with data quality of reported body weight in the German Socio-Economic Panel Study (SOEP). Previous research shows that data on reported body weight are plagued by systematic misreporting. This bias is said to be the consequence of the sensitive nature of information on body weight. Numerous studies on survey response suggest that certain modes of data collection are more conducive than others for probing sensitive information. This paper investigates the effect of the anonymity of the interview setting, characteristics of the interviewer and respondents' familiarization with the SOEP, as an indicator of the trust in the relevance and the confidentiality of the survey, as factors that may impinge on reported body weight. Findings of this paper show that refusals of the reported body weight occur infrequently (in less than 1% of the cases). Moreover, characteristics of interviewers account for only a small fraction of the variance in reported body weight (roughly 1 %). Yet the hypothesis that the absence of an interviewer in self administrated interviews increases reported body weight can be confirmed. This interview effect, however, occurred in men only. On average, male respondents in anonymous interview settings report a body weight which is 1 kg more than they would report in other settings. The repeated participation of respondents in the SOEP increases their reported body weight, a finding which suggests a positive panel effect on respondents' willingness to disclose sensitive information.