Mail Versus Internet Surveys

The authors evaluated determinants of response to Internet-based surveys in a sample (n = 5600) of Texas healthcare professionals. Participants were given the option of responding by mail or over the Web (response, 66%). Internet respondents were younger (p < .001), had worked fewer years in healthcare (p < .001), and were more likely to be male (p < .001) and to work in a hospital (p = .007). Missing questionnaire items were significantly higher among Web responders with regard to age, sex, race, body mass index, and smoking (p < .001). In the final multivariate logistic regression, only male gender (odds ratio [OR] = 2.09, 95% CI = 1.56-2.80) and younger age remained significantly associated with response over the Internet. Age quartile and responding electronically were inversely associated. Taken together with a priori knowledge of the demographic and professional profile of a study population, these findings can be useful in planning and implementation of surveys among healthcare workers.