DNA sequences sampled from a genetic locus within a population are related by a genealogy. If there is no recombination within the locus, each pair of sequences is descended from some ancestral sequence, one of which is the most recent common ancestor of the entire sample. Past demography shapes this genealogy since the branch lengths depend on the size history of the population. For this reason, observed distributions of allelic types carry information about the population's demographic history. Because of their abundance and relative ease of typing, microsatellites, or short tandem repeats, represent a useful class of loci for the study of demography. This thesis investigates the properties of the imbalance index beta, a microsatellite-based statistic constructed for demographic inference. Simulated data sets are used to explore the sampling properties of beta and to compare its performance to that of other statistics available in the literature. Tests based on these statistics are applied to samples of microsatellite loci from human populations, and the results are interpreted in light of recent hypotheses concerning the evolution of modern humans.