Going to War and Going to College: Survey of Veterans

More to the point, the inclusion of the interaction term tends shifts the estimated effect of World War II service somewhat towards 0. These point estimates of the effect of World War II service are between .26 and .36 for years of college completed and between 5 and 9 percentage points for college completion. With the nonlinear specification of the Korean War trend, the effect of World War II service is .28 for years of college completed and 6 percentage points for college completion. The above discussion suggests that the specifications we should prefer are the ones represented by columns (7) and (9).
An alternative to the longer period panel with two endogenous variables is to restrict the estimate of the parameter associated with Korean War service to some plausible range of values using the data for 1923-28. Figure 3 graphically presents results from such a procedure where we have varied the Korean War service effect from 0 (replicating the results in column (1) of Tables 4 and 5) to 0.5 in the case of years of college completed to 0.1 in the case of college completion. The crux of the matter in narrowing these bounds of the veterans effect for those serving in World War II is determining how large the impact of the Korean War was on collegiate attainment. The vertical lines at the right of each panel represent the within cohort estimates of effect of service during the Korean War on the outcome in question 0.37 in the case of years of college completed and 0.077 in the case of college completion – for the 1923-1928 cohorts. On the presumption that there continued to be positive selection among these cohorts and that there were no “spill over” effects, these within cohort estimates would seem to be plausible upper bounds on the effect of service during the Korean War educational attainment.
The 1979 Survey of Veterans provides an alternative fix on the effect of service during the Korean conflict on educational attainment. For cohorts born between 1923 and 1928, Korean War veterans obtained, on average .19 (0.054) years of college after returning to civilian life while approximately 3.6 (0.014) percent obtained a college degree, with the standard errors indicated in parentheses. Some of these men might have continued their education even with the G.I. Bill. Thus these figures would seem to represent upper bounds on the effect of service on educational attainment. (As Appendix Table 1 shows, later cohorts in the Korean conflict were more likely to use the G.I. Bill.) Using these measures as our estimates of the effect of service during the Korean War, we see the implied estimate of the average effect of service during World War II on years of college completed to be 0.26 (.05) while for college completion the implied estimate is about 6.1 (1.3) percentage points, estimates that are roughly in the middle of the range of possible values represented in Figure 3.