Broadly speaking, NHST seeks to classify election results as accurate or irregular (although the null hypothesis requires a specific model for how accurate results are generated).
