An Exploration of Social Desirability Bias in Measurement of Attitudes toward Breastfeeding in Public.

BACKGROUND: Measurement of attitudes toward breastfeeding has been based on self-report, which may be subject to social desirability. Increasing the perceived anonymity of questionnaires may reduce social desirability bias, producing more accurate results. OBJECTIVE: We compare a standard questionnaire (SQ) with the unmatched count technique (UCT) to understand the effect of increased perceived anonymity on self-reported attitudes toward breastfeeding in public. METHODS: Measures of attitudes toward breastfeeding in public were adapted from existing questionnaires, subjected to expert review, and pilot tested. A web-based survey was then constructed to compare the UCT and the SQ technique. Participants were recruited online and randomly assigned to either the SQ or the UCT condition. RESULTS: In the overall sample (N = 1477), the UCT condition had significantly higher endorsement for the statement, "Breastfeeding in some public settings should be against the law" [χ2(1, n = 1455) = 9.58, P = .002]. Women more frequently endorsed that item in the UCT condition (15.6%) than in the SQ condition (7.1%) [χ2(1, n = 1025) = 18.27, P < .001]. In contrast, among men, rates of endorsement did not vary between experimental and control groups for that question. CONCLUSION: Perceived anonymity may have influenced responses to some questions about attitudes toward breastfeeding in public. The effects of perceived anonymity may operate differently within demographic sectors. The direction of the effects was not always consistent with hypotheses, and future research is needed to fully explore the various dimensions of attitudes toward breastfeeding. The UCT method shows promise for improving the accuracy of reporting attitudes toward breastfeeding.