Table 6 presents results from linear regressions that use total months spent on aid as a
dependent variable. Though not shown, comparable models were estimated using percent of time in
the NLSY with differently scaled, but essentially equivalent results. Each model in Table 6 includes
the controls that were included in previously presented models. In addition, model 1 includes a
control for whether the welfare recipient attended college while on aid. Model 2 includes this
dummy variable, plus one that indicates whether the welfare recipient attended college at some time
in the NLSY panel outside of her time on aid. It is possible that recipients could have spent time in
college both while on aid and outside of their time on aid. If so, both dummy variables would be
coded as 1. Model 3 includes only the dummy variable indicating college enrollment while on aid
and also includes a dummy variable indicating whether the recipient graduated from the program
within three months of the end of the welfare spell. Model 4 includes all the measures in models 2
and 3, plus one indicating whether the welfare recipient graduated from college at a time when she was not also receiving welfare.