Methodology

The Methodology Core provides a central resource for the development and maintenance of MPRC samples, the development of new measurement techniques, data management, and research design and analysis techniques. This core also carries out special methodological studies and activities designed to contribute to MPRC research programs.

Construct Validity and Measurement Development
Creating valid and reliable measures for MPRC research projects is an essential activity for future Center project development. In 1995, MPRC researchers designed and carried out a special Methodology Core study involving a selected subsample of JOBS II participants. The study was specially designed to oversample JOBS II focal respondents who exceeded the cutting score for major depressive disorders on CIDI screening scales, both from the focal respondents and significant others perspective (“cases”). The sample consisted of 222 focal respondents with a 91% response rate and 214 significant others for a 88% response rate. This study is intended to lay the groundwork for a large benefit cost study of economic status, mental health and social functioning.

Statistical Models for Attrition and NonparticipationApplying new statistical techniques to MPRC JOBS data is critical to the evaluation of the trials and to adapting these techniques to MPRC uses. A second set of studies conducted by Linda Yau, MPRC trainee from the Department of Biostatistics, School of Public Health and MPRC faculty member Roderick Little from the School of Public Health, involves the development of new models for estimating selection and attrition in randomized trials. Preliminary data analyses have been conducted using the JOBS I and JOBS II dataset.

Training in Structural Equation Modeling
Continuous improvement of staff and trainee skills is critical to increased MPRC research effectiveness. Amiram Vinokur conducted a series of workshops with MPRC faculty and graduate students on the use of EQS methods to carry out structural equation modeling using the MPRC dataset for teaching examples.