This research project will examine impacts of the tendency for individuals with similar attributes to cluster in geographic space on the sample size of a qualitative research study. Sample size is a crucial feature in empirical scientific research projects, affecting generalizability and manipulation/control in quantitative studies and transferability and discovery in qualitative studies. Selecting a minimum number of observations helps ensure the adequacy of a sampling strategy. An insufficient sample size can undermine the soundness of research findings and/or waste resources. Because of the tendency for individuals with similar attributes to cluster in geographic space rather than to be distributed randomly, a characteristic known as spatial autocorrelation, special care must be taken in the design of research projects. This project will seek to determine the appropriate sample size for qualitative research in the presence of spatial autocorrelation. The theoretical foundations of this project draw on Thomas Schelling's Nobel Prize-winning work, adapting it for use in a geospatial context. The investigators will use a set of simulation experiments to assess metrics that can be used as guidelines for determining appropriate sample sizes for a set of different sampling approaches. They also will conduct a qualitative sampling project to evaluate the computer laboratory findings with real world field work.
This research project will advance understanding of sample size determination for qualitative research conducted in the presence of spatial autocorrelation. The project will address the controversial issue of qualitative sample size determination in the more general literature, supplement what is known for quantitative sample size determination, furnish numerical evidence for selecting a qualitative sample size, and make qualitative researchers more aware of spatial autocorrelation impacts on their research. The project will provide a new tool to bolster the rigor of qualitative spatial research, thereby helping to make its execution more efficient and its findings more widely acceptable. The project will contribute numerical evidence for determining qualitative research sample size, and it will help qualitative spatial researchers better understand how spatial autocorrelation complicates their data collection and impacts their findings. The project also will yield a real-world example of the use of this new spatially informed qualitative sampling method.
