Nonrespondents in homogeneous groups: Implications for mailed surveys

Abstract

An on‐site survey was conducted during the summer of 1978. The population consisted of boaters on dredge spoil islands on the Mississippi River from Guttenberg, Iowa, to Saverton, Missouri. By using a direct contact approach, 94 percent of the users filled out questionnaires concerning recreation on the river. As a treatment to determine differences between respondents and nonrespondents, a follow‐up questionnaire was mailed to a subset of the on‐site survey respondents in the fall of 1979. Forty‐six percent responded. The on‐site survey results were then reanalyzed, using response to the mailed survey as a new variable. Of 31 study variables, 28 showed no difference at the .05 level of significance between respondents and nonrespondents. The results of this study, with support from previous research, indicate that when dealing with homogeneous groups (people with a common interest) high response rates, and therefore extensive follow‐ups, are not necessary to avoid nonresponse bias.