Evaluating Nonresponse Bias in Survey Research Conducted in the Rural Midwest

An understanding of private landowner's perceptions can guide decisions about conservation in rural landscapes. However, mailed surveys that evaluate landowner views are increasingly plagued by falling response rates and nonresponse bias. Using survey data from research conducted in the Midwestern United States in 2007 and 2017, we adapted a framework for testing nonresponse bias across demographics, attitudes, and land-use practices inferred from aerial imagery. We compared respondents and nonrespondents, early and late-respondents, low- and high-interest respondents, and our sample and published data. Across all comparisons, we found little consistent evidence of nonresponse bias, except a possible under-sampling of Amish landowners and conflicting results on gender. Many discrepancies were likely related to the sampling methodology. Because comparisons yielded conflicting results, we recommend that researchers engaged in survey research with rural communities use multiple methods to triangulate nonresponse bias to determine whether subgroups of rural populations are under-sampled.