individual-level regressions (Table 8)

[bkudrzycki]

1. add row with reference category
2. be consistent in using "wave" or "endline" as variable term
3. add knowledge model (applicants vs. participants)
4. control for years of training at baseline (and maybe education?)