(a) Relevance & Research Question:
The results of a number of studies comparing data collection via personal interviewing (CAPI) and self administered web surveys (CAWI) suggest that online data collection leads to more honest answers (e.g., Bronner/Kuijlen 2007; Duffy et al. 2005; Hoppe/Lamp 2001; Loosveldt/Sonck 2008; Marta-Pedroso et al. 2007). This is usually attributed to the absence of an interviewer in the online realm and the therefore lower tendency of respondents to answer questions in a manner that will be viewed favorably by others. However, to date no study has used a standardized scale to measure social desirability on an individual respondent level and assess its influence of on data quality in face-to-face and web surveys.
(b) Methods & Data:
In an experimental study respondents were either interviewed by an interviewer in person (n=402) or with a self-administered web questionnaire over the internet (n=387). Respondents for the face-to-face survey were recruited via quota sampling, the web survey was administered to members of an online panel using the same quotas. Among others respondents were asked about their attitudes towards different cultures, countries, and their citizens. Additionally, an adapted version of Paulus’ (1984) scale measuring impression management was used to assess the influence of social desirability on data quality.
(c) Results:
The results confirm that social desirability is less pronounced when respondents answer the questions in a self-administered web survey compared to an personal interview.
(d) Added Value:
The use of a standardized scale allows to test for the potential of social desirability as a moderator for mode effects.