Enlistment decisions in the 1990s : b evidence from individual-level data / c M. Rebecca Kilburn and Jacob A. Klerman

This work updates previous estimates of individual enlistment models, investigating the relationship between family, individual, local labor market, and other background characteristics and the decision to enlist. The study makes three primary innovations to earlier models. First, it uses data from the early 1990s, while the most recent estimates were from the early 1980s. The data report the enlistment behavior of a cohort of individuals from the National Educational Longitudinal Study (NELS) who were high school seniors in 1992. In general, the authors find that their coefficient estimates are similar to those estimated by earlier models, while the mean levels of the explanatory variables are more often significantly different from those in earlier data. Second, the authors explore the utility of including some additional variables in the model that are more relevant to the 1990s or were not available in early data. These include measures of immigrant status, criminal behavior, drug use, in-state college tuition, and whether parents were in the military. The research finds that immigrant status, criminal behavior, and having parents in the military are significant determinants of individual enlistment decisions. Third, the authors estimate the individual enlistment decision as a three-choice decision-whether to enlist, enroll in college, or work after high school graduation-in contrast to earlier studies, which modeled the enlistment decisions as a two-way choice of whether to enlist or not. The study concludes that the trivariate-choice model dominates the bivariate model because it produces more significant coefficient estimates and yields more insights into the reasons that individuals enlist rather than choosing alternative activities.