Web surveys, public opinion, and police performance

Opinion surveys are frequently used to assess public perceptions about the police. However, observers have questioned the reliability of such information, particularly with regard to whether respondents are representative of the target population. This article investigates such concerns using a survey commissioned by a Canadian police force. Two groups of participants completed a similar questionnaire either during a telephone interview (n = 403) or by using a Web-based self-administered survey (n = 474). The analysis uncovered several differences related to the method used to obtain information. Web respondents reported opinions that were less favourable to the police and were more likely to select moderate choices while telephone respondents, adjusted for average effect size, were more likely to inflate their actual opinions by selecting extreme choices rather than the more moderate choices also offered. These results suggest that comparisons over time for a single organization as well as comparisons between organisations require more careful interpretation than is generally provided.