登入選單
返回Google圖書搜尋
Data for DoD Manpower Policy Analysis
註釋To allow analyses of its personnel practices, DoD maintains a multi-mode data collection effort. That effort includes both historical administrative data files and DoD surveys of military personnel. In addition, military manpower analyses make some use of civilian cross-sectional and longitudinal data. This report considers this data collection effort as a system. For data users, it provides a high-level discussion of data strategies to support a variety of analytic goals. For data funders, it considers how the current system might or might not be expanded. Beyound serving as a tutorial for analysts, the document has two main themes. The first theme concerns emerging, but currently underutilized, data-matching strategies. These strategies involving matching the core DoD administrative data files to: (1) civilian administrative data (e.g., Social Security Administration earnings data); (2) DoD survey data; and (3) civilian survey data. Such data-matching initiatives raise important human subjects issues, but these issues can be addressed through interagency cooperation and appropriate research protocols. Recent experience suggests that the potential payoffs in better analysis for DoD-and therefore better policy-are large. The second theme considers expanding the data collection effort. Specifically, DoD has been approached about participating in a National Longitudinal Survey of Youth for 2010 (NLS-Y2010) and about starting a military panel survey. Both of these efforts would be very expensive and, despite the expense, would have relatively small samples. With respect to an NLS-Y2010, if one is implemented, it will likely provide the data necessary for most military manpower analyses whether DoD participates or not. In addition, any such survey is likely to include so few people in the military as to be of minimal use in military manpower analyses. Thus, the incremental contribution of another wave does not seem to justify a large DoD financial contribution.