Table 5

Second Test for Selective Migration

District level regressions
No controls for enrollment
	1	2	3	5	5	6
	log of AP course enrollment	percent 11th and 12th graders taking AB/IB exams	log of graduates	log of SAT/ACT takers	log score above 1100/24 on SAT/ACT	In attend college
Share treated	0.002 [0.084]	5.177 [1.003]^**	0.056 [0.035]	0.036 [0.048]	0.212 [0.074]^**	0.092 [0.044]^*

Controls for the second lag of 10th grade enrollment
	7	8	9	10	11	12
	log of AP course enrollment	percent 11th and 12th graders taking AB/IB exams	log of graduates	log of SAT/ACT takers	log score above 1100/24 on SAT/ACT	In attend college
Share treated	−0.042 [0.084]	5.023 [1.002]^**	0.026 [0.023]	−0.014 [0.038]	0.158 [0.068]^*	0.076 [0.037]^*

Controls for the second lag of 12th grade enrollment
	13	14	15	16	17	18
	log of AP course enrollment	percent 11th and 12th graders taking AB/IB exams	log of graduates	log of SAT/ACT takers	log score above 1100/24 on SAT/ACT	In attend college
Share treated	−0.053 [0.083]	4.928 [1.012]^**	−0.003 [0.019]	−0.045 [0.036]	0.105 [0.063]^*	0.033 [0.037]
Observations	85	85	171	156	156	74

Robust standard errors in brackets. Standard error clustered at the district level.
↵+ significant at 10 percent; * significant at 5 percent; ** significant at 1 percent;
The results are based on a panel of 15 school districts. All regressions include district fixed effects and year fixed effects. Columns 1 through 6 have no other controls, Columns 7 through 12 include the second lag of Grade 10 enrollment, and Columns 13 through IS show results that control for 12 grade enrollment at the district level.