Research techniques that were used during the last 20 years to examine the homeless were investigated to identify the types of homeless people studied and the advantages/limitations of various research strategies. The objectives of this investigation were to: (1) review and critique the type of data collected and the research methods used by investigators to study the homeless; (2) determine how homelessness had been defined and operationalized, and which subgroups of the homeless population had been studied; (3) make recommendations about the type of data that should be reported in studies on the homeless to aid in comparing and interpreting such data; and (4) clarify gaps in the content and methods of past research that apply to policy questions. The findings suggest that much of the research that has been done is descriptive, exploratory, and atheoretical. The authors concluded that studies should provide explicit conceptual frameworks, along with more systematic data on the facilities and subjects under study. The deficiencies in sampling strategy and operational criteria for homelessness were seen as major barriers to the use of existing research in policy development. (Author)