G M McCarthy1, J J Koval, J K MacDonald. 1. Division of Oral Biology, Faculty of Medicine and Dentistry, University of Western Ontario, London, Canada. gmccarth@julian.uwo.ca
Abstract
OBJECTIVE: This study investigated late response and nonresponse bias in an HIV-related survey of dentists. METHODS: Questionnaires with ID numbers were mailed to all dentists in Ontario (N = 5,997) with additional mailings four and seven weeks later. RESULTS: Proportionately more respondents who returned questionnaires less than four weeks after the first mailing reported that they knowingly treated (P < .05) or were willing to treat HIV-infected patients (P < .05); that they had an accurate perception of risk of HIV infection after a needlestick injury (P < .01), and preferred not to refer HIV-infected patients (P < .01). Linear extrapolation of cumulative percent responses indicated nonresponse bias in terms of attitude and knowledge items; however, the magnitude was low. CONCLUSIONS: The effects of late response and nonresponse bias on the results of this study were small. However, these results cannot be generalized beyond the study population, and obtaining high response rates and testing for nonresponse bias in surveys of attitudes related to HIV are recommended.
OBJECTIVE: This study investigated late response and nonresponse bias in an HIV-related survey of dentists. METHODS: Questionnaires with ID numbers were mailed to all dentists in Ontario (N = 5,997) with additional mailings four and seven weeks later. RESULTS: Proportionately more respondents who returned questionnaires less than four weeks after the first mailing reported that they knowingly treated (P < .05) or were willing to treat HIV-infectedpatients (P < .05); that they had an accurate perception of risk of HIV infection after a needlestick injury (P < .01), and preferred not to refer HIV-infectedpatients (P < .01). Linear extrapolation of cumulative percent responses indicated nonresponse bias in terms of attitude and knowledge items; however, the magnitude was low. CONCLUSIONS: The effects of late response and nonresponse bias on the results of this study were small. However, these results cannot be generalized beyond the study population, and obtaining high response rates and testing for nonresponse bias in surveys of attitudes related to HIV are recommended.