Many textbooks on social research advise researchers to guarantee complete anonymity to the respondents of their surveys. Allegedly, this should reduce the social desirability of the responses. But, does it? And aren’t there any side-effects?
In an ingenious series of studies Lelkes, Krosnick, Marx, Judd, and Park (2012) demonstrate that allowing college students to answer questions completely anonymously sometimes increased reports of socially undesirable attributes, but consistently reduced reporting accuracy and increased survey satisficing.
In a first study, 73 undergraduates of the University of Colorado, spent 45 minutes alone, researching a topic on the internet. After that, about half of the participants are required to write their name, student ID and e-mail address on the first page of a paper questionnaire. The other half was told not to write any identifying marks on the questionnaire, “since it is very important that your responses be completely anonymous”. The number of times a participant gave a socially desirable response to one of the seven target-questions, e.g., “I have sometimes explored pornographic sites on the Internet” was counted. An accuracy measure was also computed. This is the discrepancy between the number of sites actually visited and the number of sites reported visited. The actual sites could be traced through a spyware program that was installed on the computers. At last, a measure of non-differentiation was calculated. Therefore the respondents had to answer four batteries of questions (in total 38 questions) on topics that were controversial enough to differentiate people, e.g. “parents should monitor and/or restrict their children’s Internet”. Normally, these questions should give rise to a wide range of answers.
In the results, there was a slight hint that identifiable respondents give more socially desirable answers than anonymous respondent. The difference was very small. Identifiable participants however were more accurate. They reported the number of relevant websites visited more precisely than did the completely anonymous participants. Finally, identifiable and anonymous participants did not differ on the first two non-differentiation batteries but the identifiable subjects continue to give differentiated answers to the last two batteries, whereby the anonymous group started to engage in non-differentiation, consistent with the idea that they were putting less thought and effort into the questionnaires as they grew tired.
In the next two studies, participants were given the opportunity to eat candy (M&M) while performing the task of filling in a paper questionnaire. Later they were asked how much candy they had eaten. In study 2, the participants received the candy together with the anonymity manipulation. In study 3, the writing down of the personal information occurred after the eating of the candy. In both studies, a question at the end asked the participants to report how much candy they’d eaten. Once again, students who answered anonymously were less accurate about how much they’d indulged. This was the case whether anonymity was promised before or after the opportunity to eat the snacks.
According to the authors, these studies suggest that researchers should not automatically assume that promising participants complete anonymity will increase the quality of the collected responses. The quality of a questionnaire is of course also influenced by several factors; see for example the post The art of survey design.
REFERENCE
The post Anonymity in surveys appeared first on Cogpsy.info.