The effect of banning affirmative action on human capital accumulation prior to college entry

Table 3 College board SAT and GPA DDD estimates

	SAT M	SAT V	GPA	SAT M	SAT V	GPA
	(1)	(2)	(3)	(4)	(5)	(6)
Panel 1: 1994–6 (pre) vs 1998–2001 (post)
Post*CA	0.003	0.020***	0.019***	0.001	0.016***	0.018***
	(0.005)	(0.004)	(0.005)	(0.004)	(0.003)	(0.005)
	[-.13,.14]	[-.11,.14]	[-.13,.18]	[-.14,.13]	[-.11,.13]	[-.13,.18]
PostBlackCA	0.022**	0.011	0.069***	0.024**	0.013*	0.067***
	(0.010)	(0.007)	(0.013)	(0.010)	(0.007)	(0.012)
	[-.29,.22]	[-.2,.2]	[-.19,.29]	[-.29,.23]	[-.22,.2]	[ -.18,.28]
PostHispanicCA	-0.026***	-0.029***	0.062**	-0.016***	-0.018***	0.073***
	(0.006)	(0.008)	(0.027)	(0.005)	(0.007)	(0.025)
	[-.15,.32]	[-.18,.37]	[-.13,.3]	[-.16,.3]	[-.17,.36]	[-.13,.28]
Observations	2488390	2488390	2574569	2488390	2488390	2574569
R-squared	0.255	0.239	0.136	0.255	0.239	0.136
Panel 2: 1994–5 (pre) vs 1999–2001 (post)
Post*CA	0.001	0.046***	0.014	-0.001	0.042***	0.013
	(0.008)	(0.008)	(0.009)	(0.008)	(0.007)	(0.009)
	[-.22,.14]	[-.26,.22]	[-.23,.24]	[-.23,.13]	[-.26,.21]	[-.24,.24]
PostBlackCA	0.023	0.004	0.082***	0.025*	0.008	0.080***
	(0.014)	(0.009)	(0.015)	(0.014)	(0.009)	(0.014)
	[-.24,.22]	[-.15,.35]	[-.16,.39]	[-.25,.21]	[-.17,.35]	[-.16,.38]
PostHispanicCA	-0.040***	-0.045***	0.062*	-0.028***	-0.034***	0.073**
	(0.005)	(0.007)	(0.035)	(0.005)	(0.006)	(0.034)
	[-.16,.35]	[-.16,.27]	[-.16,.34]	[-.18,.32]	[-.16,.26]	[-.16,.32]
Observations	1845603	1845603	1907498	1845603	1845603	1907498
R-squared	0.257	0.241	0.138	0.257	0.242	0.138
Demographics	x	x	x	x	x	x
Post*Demographics				x	x	x

Notes: Each column shows the coefficient estimates from a regression of the outcome variable listed at the top of each column on an indictor for whether the student was from California, an indictor for whether they took the SAT in the post period, and an indicator for the student’s race, along with the full set of interactions between these variables. Additional controls for parental education and income, gender, first language spoken, and citizenship status are also included in the regressions. All outcome variables normalized to have mean zero and standard deviation one.
Included time trends are linear state-specific time trends. The excluded racial group includes white, other, and unknown. Excludes Louisiana, Florida, Mississippi, Texas, Washington, North Dakota, and Wyoming. Clustered standard errors shown in parenthesis. Confidence intervals generated by permutation tests displayed in brackets.
Significance indicators associated with clustered standard errors: ***p <0.01, **p <0.05, *p <0.10.