Response rate and content differences between web and mail surveys

Abstract

AbstractThe Florida Fish and Wildlife Conservation Commission (Commission) conducts an annual harvest survey of licensed wild turkey (Meleagris gallopavo) hunters to determine the proportion of license holders that hunt and the estimated total harvest. The harvest survey is traditionally conducted with a mailed postcard, which is increasingly costly and time‐consuming to administer. The Commission has considered using a less costly and less protracted web‐based survey. From 2012 through 2014, we conducted web and mail surveys of licensed turkey hunters to compare response rates and reported rate of hunting participation. The average response rate differed between mail and web survey respondents (25.5% and 16.5%, respectively). Rates of hunting participation also differed between the survey modes; 40.5% and 61.4%, respectively, of mail and web survey respondents said that they hunted during the previous spring season. When hunting participation rates were extrapolated to the total number of license holders in the state, the projected average number of hunters across the 3 years was 17,356 (52%) greater from the web survey than the mail survey. Demographic weighting (age class, sex, ethnicity, median household income, or urbanicity) failed to account for the difference in reported hunting participation between survey modes. Our results indicated that respondents of the 2 survey modes differed fundamentally in either their hunting behavior or in their propensity to respond to the surveys, both of which have ramifications when management decisions are based on the results of harvest surveys. Although a mixed‐mode survey may be a feasible option to reduce cost and still obtain accurate harvest estimates, we caution against abruptly switching from mail to web surveys given the vastly different outcomes.