Any researcher understands that there is a plethora of topics on which study is needed. There are "basic" research topics-topics that concern that fundamental workings of the economy, or equilibria, or firms. The purpose of this type of research is to understand how and why things work the way they do. There is no particular eye toward solving a current problem. Rather, it aims to elucidating basic mechanisms underlying an issue. Another category of topic is the applied research topic. The purpose of this type of research is to understand how and why observable events occur the way they do-why wages get set the way they do, why tariffs are negotiated as they are, what effect regulation has on an industry. I would like to write a bit on a subcategory of applied research, namely social policy research. I consider social policy research to be research that directly addresses a public policy issue-what effect do time limits have on the length of time people are on welfare, what effect does a specific social security scheme have on people's retirement behavior, should the government spend more on after-school activities.

One of the fundamental features of policy research that attracted me to the field is that one has chance of having an effect on the way things run. One not only learns more about why the world works the way it does, but one contributes to changing it. By answering questions that policy makers are currently interested in, rather than answering questions that only the researcher thinks is interesting, timely social policy research information can be put into the hands of people who make or set policy. This is not unique of social policy researchers, for many applied researchers address issues that are of policy relevance, such as the research on regulating intellectual property. But what distinguishes much social policy research from other applied research is that the ultimate goal of the research is to provide useable answers and information to the relevant policy makers and program operators in order to effect change, not to publish the research in an academic journal. Thus, social policy research often does not necessarily use the newest econometric technique or generate a new theoretically model of the issues. Rather, good social policy research uses whatever the most appropriate techniques are (ever if they are just old boring OLS). The results are also not written up in a manner that is similar to journal articles. To be effective, social policy research should be communicated in an easily understood non-technical manner.

Writing good social policy reports is not easy. It must be written using language that an interested lay-person could understand but demonstrate to the econometrician or evaluation expert that the research is methodologically sound. Technical appendices and footnotes come in very handy for the more technical discussions.

Well-done, well-written and timely social policy research can have a powerful effect on policy, as was the case with Barbara Devaney's work on the WIC program and my work on youth mentoring. However, it is more common for one's research to have a delayed or indirect impact on policy. Often politicians make their decisions long before the data are in or the analysis is done. For example, transportation assistance is becoming a large part of the welfare reform effort long before data on the P/PV's Bridges to Success demonstration (an experiment that is testing the effects of this type of transportation assistance) is in. Rather, a sage mentor of mine once told me that while your current research rarely has an impact on the initiative you are actually studying, it feeds into the decision about similar initiatives in the future. For example, the negative income tax demonstrations did not affect the welfare reform efforts being considered at that time, but did affect the construction of the food stamps program later on.

What else have I learned over the almost twenty years I have been in the social policy field? First, I discovered that there were two entire areas of knowledge one needs to know to do policy research that economists are not trained in, namely experimental design and primary data collection. Psychologists and sociologists receive this training, but economists do not. In general, economists answer questions using econometrics on already collected data so they do not need to know what issues are involved in collecting primary data. Thus, over the first several years I was in the field I learned as much as I did during graduate school. It was a lot of fun because in my eyes experimental design and issues of primary data collection were reverse econometrics--how do you collect data so that you don't have to use complex econometric techniques to correct for bias or other problems.

Because most policy questions are about the effectiveness of a specific program or policy approach, policy research almost always relies on data collected explicitly for the study. This means you are not constrained by the specific questions available in existing data sets. You can ask all the questions you believe you need to in order to get the best answer. [However, if during the analysis you realize you are missing a piece of data, you have no one else to blame but yourself!]

It is extremely rarely for academic economists to collect their own data. Why? I believe it is because it takes too much time, especially if you need a longitudinal data set to answer your questions. Academic tenure clocks, and the need to publish fairly regularly even after tenure if you are to be respected in the field, pushes academic economists away from a whole area of social policy research. It is a rare academic who is intimately involved in demonstration research; those few that do exist are tenured faculty and are able to do so through long-term consulting arrangements. It is a shame because many of brightest minds are not considering critical policy issues.

Another thing I learned--and this one more slowly--was that one could not really satisfactorily answer most policy questions with economics alone. The full answer must be obtained using qualitative, as well as quantitative methods; and should address concerns outside those traditionally examined by economists, such as organizational behavior and practices, political sustainability, mental health of the participants, or child development. Many policy initiatives have impacts that go beyond economics. Policy makers will and should be concerned all of them. For example, take the question, "Do the lives of inner-city job searchers improve if they participate in a tailored 18-month mobility/jobs program that provides them transportation and support services to enable them to obtain a job in the job rich suburbs?" This is the question being asked in the currently operating Bridges-to-Success demonstration. Such a program could affect the participant's earnings trajectory (the clear purview of an economist), but also their children since the parent must commute long hours but is potentially bringing more money (an issue a developmental psychologist is more able to answer). Issues to racial attitudes could also play a powerful role in determining who participates, who lasts the 18 months, and who is hired. To generate information that will be truly useful to policy makers and practitioners, one must either be willing to broaden one's own training into new disciplines and/or work collaboratively with researchers in other disciplines. [And by collaborative, I don't mean begrudgingly letting other researchers answer the uninteresting parts of the research, but rather integrating the economic research agenda with the agendas of researchers in other fields.] Knowing if a program increases earnings or education level is only one small part of what practitioners and policy makers need to know if they are to devise appropriate and feasible policy.

Field-based social policy research is a very labor intensive and time-consuming endeavor. It rarely provides its participants with top-rated journal articles. But only this type of research can answer many of the country's most pressing social questions. How should we restructure the employment and training system? What is to be done with public housing? Is federal after-school funding really worth it? The gratification for doing this type of research comes in knowing that better policy is being pursued because of your work. And every once in a while, one does have a chance to publish a piece of work that has a direct impact on the world-a worthwhile program is saved from the budget-cutters block or a good policy is expanded citing your work. Few accomplishments are as rewarding as that.