...Several role conflicts that arose for the researcher in this community-based work are also described, and the importance of nurturing long-term personal relationships with various community leaders is emphasized. The evolution of research ideas is documented by highlighting how the findings and problems encountered in earlier studies led to improved methods and new research questions.

**********

Beginning in 1986, while was at the University at Buffalo, and continuing after relocating my research team to Wayne State University in Detroit in 1992, I have devoted most of my professional career to understanding the problems of homelessness and poverty. Over this 20-year period, my research program has involved studies on homeless, poor, and otherwise marginal adults, families, children, and youth. In 23 distinct studies conducted to date (for an up-to-date listing see my web page: http://sun.science.wayne.edu/~ptoro/), my team and I have developed a range of psychometrically sound measures appropriate for use with the homeless, compared the characteristics of the homeless to carefully matched housed groups, rigorously evaluated interventions helping homeless people, followed large samples of initially homeless and other marginal people in longitudinal designs, assessed public opinion and media coverage, and collected multinational data. While doing this research, we have experienced many difficulties as well as some satisfying moments. This article will share some of the challenges and successes, material that seldom appears in our articles on the various studies. I will focus here not on the specific findings but rather on the process of our research in order to provide a personal account of the troubles and triumphs we encountered while carrying it out. This article was purposely written in a free-form style to help enhance the sense of flow in the research process over the past 20 years. Keeping with the emphasis of Merrill-Palmer Quarterly, I also focus on those studies involving homeless children, youth, and families. However, I will also talk about our general research program that involved studies on homeless adults, including many in the 18-29 year range.

Methodological and Practical Problems, as Well as a Few Discoveries

My research team and I had to contend with many methodological and practical problems inherent in studying people who are homeless. Like other social scientists at the time, we were taken by the intense media coverage on homelessness that developed during the 1980s (Buck, Toro, & Ramos, 2004). In the late 1980s when we became interested in homelessness, very few high-quality studies were available. We were, however, impressed with the work being conducted by two research teams, one in Los Angeles (Burnam & Koegel, 1988) and another in St. Louis (Morse, Calsyn, Allen, Tempelhoff, & Smith, 1992). Due to the absence of good studies, we often found ourselves developing methods based on trial and error. Some of our colleagues thought we were crazy to attempt to do research with homeless people. Luckily, at the time, probably due to the great interest in homelessness in the media and among politicians, large amounts of federal grant monies were rapidly becoming available (e.g., through the National Institutes of Mental Health and Alcoholism and Alcohol Abuse).

Many policy makers seemed to view homelessness as a national crisis in need of immediate attention, including sound research. As a young assistant professor with a genuine personal concern about marginalized people, I found that the availability of grant funds made it more tempting to consider this line of risky research.

Sound Measurement and Sampling Methods

We immediately faced two methodological problems: (1) assuring use of psychometrically sound measures appropriate for homeless people; and (2) obtaining representative samples of homeless people. With the obvious exceptions of the work in Los Angeles and St. Louis, no existing studies seemed to pay adequate attention to either problem. Most studies appeared to have been hastily put together, using a simple set of questions, then obtaining a sample at one particular shelter in one particular part of a particular city, and then proceeding to draw conclusions about the problems and needs of homeless people on this basis. For example, very high rates of mental disorders were often reported among studies of homeless adults when, upon closer scrutiny, it was apparent that unreliable assessment tools had been used, that substance abuse had been lumped in as a type of "mental illness," and/or that samples were not representative (Toro, 1998). Our very first study attempted to determine whether homeless adults found at different types of sites (i.e., shelters, soup kitchens, and the streets) yielded samples with different characteristics (they did, at least to some extent; see Toro & Wall, 1991). This initial study also attempted to pull together from the various existing studies measures that seemed to have the best reliability and validity. We also added measures that had yet to be used by other researchers (e.g., ones assessing social support and social networks) and made sure that all measures took into account the context of homeless people's lives. For example, in our research on homeless and matched housed adolescents (McCaskill, Toro, & Wolfe, 1998; Toro, Wolfe, & McCaskill, 2006), we adapted the widely used Diagnostic Interview Schedule for Children (Piacentini et al., 1993) because several items used to determine a diagnosis of conduct disorder explicitly asked about whether the participant had run away or otherwise had experience with homelessness. Including such items would have inflated housed-homeless differences observed.

A useful methodological discovery. Building on the work of Burnam and Koegel (1988), we developed a general sampling method that can be applied across cities in the United States and other developed nations (Toro, Wolfe, et al., 1999). This method first relies on service providers and other experts on the local homeless population to provide prevalence estimates and share potential recruitment locations. Then research staff are sent to identified sites--including in- and outpatient programs and street sites as well as shelters and soup kitchens--to determine the distribution of the homeless population throughout a metropolitan area, correcting for the fact that some homeless people can be double-counted because they use multiple sites over a period of time. In developing such "probability sampling" methods, we came upon one very useful methodological discovery. The media, professional literature, and local advocates often talked about "street people" who are homeless but seldom use shelters or other services. It was generally believed that there were large numbers of street people and that, to obtain good representative samples, researchers would have to do extensive combing of the streets. Our research has found that there were actually extremely few true street people. In two different cities, less than 1% of the total homeless sample surveyed used absolutely no other human services over a year's time. Such findings are particularly heartening to researchers who, due to the lack of resources and concerns about the safety of staff sampling from the streets, may wish to avoid this type of sampling. Furthermore, our study showed that relatively small, unique percentages of the overall homeless population were to be found at a range of different in- and outpatient programs (only 6-10% across the two cities). The researcher sampling a wide range of shelters and soup kitchens could capture the vast majority of the homeless population (89-93% across the two cities).

In our studies on homeless children and adolescents, we found that true "street kids" are even rarer than adults who live on the streets. Homeless children, typically under age 10 and homeless with their mothers, are virtually nonexistent on the streets, perhaps because our society provides the widest range of services to protect them. We have also found that homeless adolescents under age 18 are quite rare, except in certain large urban areas on the East and West Coasts, such as New York City, Los Angeles, San Francisco, and Seattle (see Haber & Toro, 2004; Robertson & Toro, 1999). In our largest study on homeless adolescents in the Detroit metropolitan area, the Housing, Adolescence and Life Outcomes project, we looked carefully for "street kids," with virtually no success. While we have no systematic data...

NOTE: All illustrations and photos have been removed from this article.



More articles from Merrill-Palmer Quarterly
How can basic research on children and families be useful for the poli..., April 01, 2006

Looking for additional articles?
Search our database of over 3 million articles.

Looking for more in-depth information on this industry?
Search our complete database of Industry & Market reports by text, subject, publication name or publication date.

About Goliath
Whether you're looking for sales prospects, competitive information, company analysis or best practices in managing your organization, Goliath can help you meet your business needs.

Our extensive business information databases empower business professionals with both the breadth and depth of credible, authoritative information they need to support their business goals. Whether it be strategic planning, sales prospecting, company research or defining management best practices - Goliath is your leading source for accurate information.