Appendix: Methodological Considerations for Surveys of Local Election Officials

In this chapter, we provide evidence to support the use of a specific sampling algorithm for drawing random samples of local election officials (LEOs) in the United States. Surveying LEOs creates unique challenges that cannot be resolved with normal probability sampling methods. The enormous diversity of local jurisdictions and the hyperfederalized institutional structure of American elections combine to create methodological challenges to drawing a random sample that allows generalizations both about LEOs and about the American voting experience. The chapter explores the statistical foundations of several unequal inclusion probability sampling methods. We show using simulations that the extremely skewed distribution of jurisdictions (by population size) causes anomalies in the sampling process, resulting in overly variant samples and extreme values for some sampling weights when using the “minimal support” sampling algorithm. We further show that the “random systematic” sampling method is superior for this context, resulting in lower variance estimates, and is just as easy to implement as minimal support.