Small tests are easier to understand. Small tests are limited in their scope. Small tests most likely test one facet of the behavior. Small tests are less likely to incur coupling. Small tests are easier to diagnose when they fail. Need I say more?
