Table 4. Heterogeneous relationships—OLS estimation.

	(1)	(2)	(3)	(4)	(5)	(6)
	Test 2	Test 3	Test 4	Test 2	Test 3	Test 4
Higher achieving school	0.044	0.165***	0.183***	0.060	0.155***	0.171***
	(0.036)	(0.047)	(0.040)	(0.037)	(0.048)	(0.041)
Higher achieving school * Test 1	-0.057	-0.027	-0.074	-0.055	-0.030	-0.071
	(0.040)	(0.052)	(0.045)	(0.040)	(0.053)	(0.045)
Test 1	0.699***	0.613***	0.559***	0.701***	0.613***	0.555***
	(0.029)	(0.037)	(0.032)	(0.029)	(0.037)	(0.032)
Time between test 1 & 2 (in months)	-0.002			-0.007
	(0.024)			(0.024)
Time between test 1 & 3 (in months)		0.050*			0.049*
		(0.027)			(0.027)
Time between test 1 & 4 (in months)			0.034**			0.038**
			(0.017)			(0.017)
Constant	0.089	-0.748**	-0.634**	0.223	-0.410	-0.106
	(0.100)	(0.315)	(0.270)	(0.331)	(0.482)	(0.437)
Observations	1,112	1,112	1,112	1,112	1,112	1,112
Adj. R-squared	0.528	0.366	0.380	0.529	0.368	0.382
Parental background controls	Yes	Yes	Yes	Yes	Yes	Yes
Neighborhood controls	No	No	No	Yes	Yes	Yes

Notes: All test scores are standardized to mean zero and a standard deviation of one. A higher-achieving school is defined as having an above median three year school average CITO score. Parental background controls are the household income and the education level of the father and mother. Neighborhood controls include a set of variables measured at the four digit postal code area. The neighborhood controls are the percentage of households under the social minimum income, the percentage of households with low income (less than €25,100 per year), the percentage of households with high income (more than €46,500 per year) and the percentage of households with at least one child. The data on neighborhood characteristics was collected by CBS Statistics Netherlands. S3 Table shows the full model and reports the coefficients and standard errors for all included controls. Standard errors are in parentheses;

*** p<0.01,

** p<0.05,

* p<0.1.