Using Spatial Autocorrelation Analysis to Guide Mixed Methods Survey Sample Design Decisions

The value of spatial autocorrelation analysis to guide sampling decisions is examined and designs that identify perception clustering are argued for, which can affect qualitative sampling as well as advance the integration of quantitative and qualitative research.

Abstract

Mixed methods researchers share a commitment to knowing their sampling frames and minimizing discovery failure, especially when using surveys. Notwithstanding advances in sampling strategies, the geographic clustering of perceptions has not been fully considered for its relevance to sampling. This article examines the value of spatial autocorrelation analysis to guide sampling decisions. Spatial autocorrelation refers to the clustering of (dis)similar phenomena and signals the likely existence of perception subgroups. Through a spatial autocorrelation analysis of Dallas, Texas, the authors identify sampling frames for collecting data about perceptions of West Nile Virus eradication measures. They furnish some empirical confirmation of the geographic clustering of perceptions and argue for designs that identify perception clustering, which can affect qualitative sampling as well as advance the integration of quantitative and qualitative research.