Introduction: Web-based epidemiologic surveys are being widely used, but still present lower response rates compared to traditional methods. Their design can influence survey response rates. Objective: Analyse the influence of questionnaire length, frequency of reminders, and the interaction between them, on the response rates of five web-based questionnaires. Methods: This is a 2×2 factorial study. Participants registered into the "coortesnaweb" platform (n=1,277) were randomly assigned to respond to short or long questionnaires, and to receive high or low frequency of reminders. We analysed the influence of these factors on the response rates of five web-based questionnaires applied in a longitudinal manner. The relative risk of responding to an additional questionnaire was also analysed. Results: The mean response rate was 54.3%. Sending reminders more frequently was positively associated with the response rates for the first questionnaires. Questionnaire length did not influence response rates. We found no interaction between questionnaire length and frequency of reminders. Women and highly educated participants had, respectively, 13.0% and 28.0% increased probability of responding an additional questionnaire. Conclusions: We obtained high response rates for the first questionnaires. Sending reminders more frequently and providing conditional incentives should be employed. Long questionnaires did not jeopardize response rates of web-based questionnaires.