Table 3.

The impact of the Compulsory Education Law on formal schooling‐DDD analysis (first stage of IV regression)

Dependent variable		Years of schooling				Completing 8 years of schooling
	Column 1	Column 2	Column 3	Column 4	Column 5	Column 6	Column 7	Column 8
treatment*intensity*2008	0.014***	0.009***	0.009***	0.011***	0.003***	0.002***	0.002***	0.003***
	(0.004)	(0.004)	(0.003)	(0.003)	(0.000)	(0.000)	(0.000)	(0.000)
Controls
Ethnicity	No	Yes	Yes	Yes	No	Yes	Yes	Yes
Rural/urban status during childhood	No	Yes	Yes	Yes	No	Yes	Yes	Yes
A child‐gender dummy	No	No	Yes	Yes	No	No	Yes	Yes
The birth order dummies	No	No	No	Yes	No	No	No	Yes
R‐squared	0.779	0.810	0.810	0.826	0.370	0.436	0.436	0.464
F‐statistics	14.97	7.00	7.24	12.40	44.96	34.84	35.33	39.34
Observations	3,339	3,327	3,327	3,327	3,339	3,327	3,327	3,327

Note. Women aged 18–29 in 2003 and 2008 form the sample of analysis. Women aged 18–21 form the treatment group. The intensity is the difference between the 1997 and 1996 government funds distributed for primary school construction at the region of childhood. Robust standard errors in parentheses cluster at the region of the childhood. F‐statistics are the test of the joint significance of the triple interaction term, that is, the instrument (treatment*intensity*2008). The baseline Models 1 and 5 include no control variable. In addition to the controls given in the table, all models include ethnicity, the urban/rural status of the region of residence in childhood (except Models 1 and 5), the region of childhood, year of birth and age of respondent fixed effects, the intensity variable, the interaction of year of birth with gross enrolment rate in the region of childhood, and the interaction of treatment and intensity variables. DDD = difference‐in‐difference‐in‐difference.

^* p < .1.

^** p < .05.

^*** p < .01.