Limitations of NSHAPC Findings
There are some important aspects of the NSHAPC study that readers need to know if the study's findings are to be interpreted correctly. The Study Is Descriptive This study is intended to provide information describing currently homeless and other clients using homeless assistance programs in the United States. There is no intent to infer causes of homelessness from this descriptive information. Statistics are presented as simply as possible, for ease of understanding. Where information is available, the report compares study findings for homeless clients with statistics describing all American adults, all poor adults, or other relevant national populations. This is intended to help the reader understand similarities and differences between poor people or the population in general and clients experiencing homelessness in the fall of 1996. When a statistic looks simple but actually reflects some hidden third factor, the report tries to point this out. For example, the report shows that veterans comprise a higher proportion of street stayers than they do of shelter stayers. But the report also points out that this is because more street stayers are men and almost all of the veterans in the sample are men, not because veterans have a special propensity for sleeping on the streets. The People in the Study Come from Homeless Assistance Programs The people interviewed for NSHAPC are clients of homeless assistance programs. In cities and other communities with many programs, this approach is an efficient and effective way to find and interview a very high proportion of homeless people. However, in communities without many services, this approach will miss many homeless people, and the complete absence of services in a community, and therefore of people interviewed for this study (as happened in two rural sampling areas), cannot be taken to mean that such communities do not have any homeless people. In addition, there may be some systematic biases in the homeless people who are interviewed and those who are missed when a community does not have a full range of homeless assistance services. If soup kitchens are rare in rural areas, typical soup kitchen users will be less likely to appear in rural homeless samples. If suburbs will accept transitional housing programs for families but not for recovering substance users or people with mental health problems, then a service-based methodology will make suburban homeless populations look as if they have higher proportions of families and lower proportions of clients with mental health or substance use problems. A service-based approach to data collection is the most reasonable way that a national study of homelessness could be undertaken and still be statistically meaningful. (See the paper by Tourkin and Hubble, appendix A of the Technical Report, for an explanation of why this study used a service-based design.) Local studies can compensate for gaps in a community's service system, but there is no realistic way for a national study to do so. The reader is therefore advised to use caution in interpreting differences in homelessness between communities of different types, as some of the differences will probably reflect service system variations rather than true differences in homeless populations. NSHAPC Was Designed to Collect Data on Clients Who Use Homeless Assistance Services Many homeless assistance programs serve clients who are not currently homeless. As a result of the study's random sampling of all program clients, some clients in the NSHAPC sample were not homeless at the time they were interviewed. This is particularly true for programs that are not shelters or transitional housing programs. Information collected during the interview indicated that some have been homeless at least once in their lifetime; this report refers to this group as "formerly homeless" clients. The remaining people, who reported never having been homeless, are referred to as "other service users." The study designers wanted to know the characteristics of people using the programs, including information about their living situation. The reader should remember that while the study design produces as close to a nationally representative sample of homeless clients as possible, the same is not true for formerly homeless clients and other service users. As unrepresentative of their larger categories as these two subgroups of the sample may be, information about them is important for service providers. These two groups make up almost half of all clients who use homeless assistance programs, so information about their characteristics can be of considerable help to program managers. All Client Information Comes from the Clients Themselves The study interviewed clients of homeless assistance programs about their experiences, and recorded their responses. No attempt was made to verify or confirm the accuracy of what clients said about themselves. This is especially important for readers to remember when they review information about the clients' health conditions, use of alcohol and drugs, mental health problems, incarceration, victimization, joblessness, and other possibly sensitive subjects. Clients may not actually know some things, such as medical conditions (e.g., hypertension, anemia), if they do not see doctors regularly. They may have forgotten, or wish to downplay, other things that carry some level of social stigma. Furthermore, many questions were asked and left up to the client to interpret, including such critical issues as whether or not they had ever been homeless (no "official" definition was given or imposed on clients). What the Study Does Not Do All studies have limitations, and NSHAPC is no different. This report does not include information on the following issues, because the study was not designed to address them:
It is also important for the reader to remember that this study obtained information about homeless clients and other service users during October-November 1996. As with all information that focuses on homeless people at a single point in time, it will overemphasize people with long episodes of homelessness and underemphasize people with short periods of homelessness and also people who are homeless for the first time. Any characteristics associated with length of a homeless episode will likewise be skewed toward the characteristics of people with longer spells of homelessness.
|