RESEARCH

Research Experience

The bulk off each scholars time is spent on a program of research, focused on their developing and carrying out an individual research project/program, but also working collaboratively with faculty and other scholars on research relevant to their scientific and professional goals. The University of Michigan has vast resources for research and research mentoring, and we expect Scholars to avail themselves of these. Each Scholar develops a working relationship with a primary research mentor, and other secondary research mentors as appropriate. The Scholar and mentor(s) work together to develop and facilitate the Scholars unique research agenda. These agendas will necessarily be very individualized.

One of the unique resources of Michigan is the opportunity for Scholars to gain experience in both primary and secondary data analysis. And we see distinct advantages to Scholars having both kinds of experience, with the mix determined by their prior background and future career and scientific objectives.

Secondary data analysis: With the increasing availability of large public use data sets, there is a great need and opportunity for secondary analysis addressing important scientific questions. Such analyses can also lead more quickly to presentations and publications, which are essential to scholars early career development. Further, any efforts at primary data generation must be informed by a clear sense of how the data will be analyzed. However, if scholars engage only in analysis of secondary data, they gain little or no appreciation of the underlying strengths and weaknesses of the data to which they may apply highly sophisticated analytic tools. These strengths and weaknesses often stem from problems in the data collection process, which may be opaque when looking at a secondary data set. For example, error and bias in survey data arise from problems of sampling, response rates, the nature of questions and the interview process, and data coding and entry procedures, but little of this information is provided with secondary data, or well understood by scholars lacking knowledge and experience of the whole process of survey design and data collection.

Similar issues arise in the use of biomedical, archival, or more qualitative interview data. In our experience, these problems are best understood through participation in process of primary data collection, as well as formal training in them. Finally, secondary analysis constrains scholars to asking and answering questions for which secondary data exist, and then often imperfectly. This constraint is particularly damaging to emerging fields like population health, which have not yet been able to influence heavily the research design and data collection of prior and many ongoing studies.

Primary data generation: Observing and participating in processes of primary data generation allows scholars to acquire the skills necessary to generate new data uniquely tailored to the research questions they and the field of population health need to address, while also better appreciating the strengths and limitations of such data. In the current research environment, most scholars have had extensive experience with secondary data analysis, but only limited experience with and hence appreciation of the promise and problems of primary data generation (and hence also of the data they utilize for secondary analysis). For example, much secondary data derives from respondent reports about their current and past experiences, behavior and states of health, yet there are substantial variations in the degree to which people accurately recall different things. Similarly, physical and biological measures are subject to varying potential errors of measurement or laboratory assay. There is much interest in collecting physical and biological measures in the context of surveys, but the difficulties and problems in doing so must be adequately appreciated if the data are to be utilized appropriately in analysis.

The University of Michigan is uniquely able to offer scholars training in both secondary data analysis and primary data generation and analysis, tailored to their prior background and future career lines. Michigan is a leader in the generation as well as analysis of primary data relevant to issues of population, especially in prospective or longitudinal studies, which are increasingly multilevel in design. The Institute for Social Research (ISR) and the School of Public Health (SPH) have multiple ongoing prospective studies, which include both psychosocial variables and biomedical as well as self-report measures of health. Among these are the Americans Changing Lives and Health and Retirement Studies, the national Survey of African American Mental Health, the Monitoring the Future study of adolescent and early adult substance use and abuse, and others. SPH is the scientific home of the Alameda County study, the Kuopio Ischemic Heart Disease study, the Pitt County study of hypertension in African-Americans, and ongoing community base health studies in Detroit discussed above, as well as other more biomedical studies. Scholars would also have access to a vast array of more clinically based research in the Medical School. These studies provide opportunities for secondary analysis beyond that which can be done even on the public use versions of these data, with ready access to those who designed and collected the data. Equally importantly, the large array of studies assures that primary data collection activities are ongoing in one or more of these studies every year. Thus, it is frequently possible for scholars to engage with both secondary analysis and primary data collection within the same study. Further, Michigan offers an unparalleled array of summer and regular academic year courses in all aspects of primary data generation via the academic Departments and summer institutes of the Department of Epid emiology, Survey Research Center, and Inter-University Consortium for Political and Social Research.

[ back to top ]