ABSTRACT
Research applications in public administration, as in social science research more broadly, have been influenced heavily by the increasing application of statistical methods to handle situations in which the outcomes of interest fall into two discrete possible categories. Dichotomous outcomes arise commonly. For example, you might want to determine what explains whether a city government is run by a strong mayor system or by some other form of government, whether students who participate in learning communities are more likely than those not in learning communities to
some cities gain and other cities lose population, or whether a local bond referendum will pass or fail.
