could the reason be, that the sample size is too small ?
Or may be age group.
I am thinking if free money were given to unemployed college graduates for 2 years, than they would more likely find jobs (even if they pay less than what they were expecting with their particular college degrees)