Table 2

OLS and DID Estimates of the Effect of a Fail Inspection on Test Scores (Outcome variable: age 11 (Key Stage 2) national standardized test score)

	Mathematics			English
	1	2	3	4	5	6
Panel A: OLS
Early fail	0.120^** (0.028)	0.127^** (0.028)	0.130^** (0.027)	0.082^* (0.038)	0.090^** (0.029)	0.090^** (0.029)
Student characteristics	No	Yes	Yes	No	Yes	Yes
Age-7 test scores	No	No	Yes	No	No	Yes
R-squared	0.04	0.27	0.49	0.04	0.32	0.53
Observations	16,617	16,617	16,617	16,502	16,502	16,502
Number of schools	394	394	394	394	394	394
Panel B: Difference-in-differences
Post x early fail	0.117^** (0.033)	0.113^** (0.032)	0.117^** (0.030)	0.080^* (0.038)	0.075^* (0.036)	0.072^* (0.036)
Post	0.014 (0.026)	0.038 (0.025)	0.028 (0.024)	0.061 (0.032)	0.085^** (0.029)	0.078^** (0.030)
Student characteristics	No	Yes	Yes	No	Yes	Yes
Age-7 test scores	No	No	Yes	No	No	Yes
R-squared	0.07	0.24	0.49	0.08	0.31	0.54
Observations	33,730	33,730	33,730	33,386	33,386	33,386
Number of schools	394	394	394	394	394	394

↵Notes: Standard errors reported in parentheses; * and ** indicate significance at the 5 percent and 1 percent levels, respectively. Standard errors clustered at the school level. OLS and DID models estimated for schools rated Fail in the years 2006 to 2009. “Early fail” dummy switched on for schools failed September to November; switched off for schools failed mid-May to mid-July. The dummy “Post” is turned on for the year of inspection and off for the previous year. Controls for student characteristics: dummies for female, eligibility for free lunch, special education needs, month of birth, first language is English, ethnic group, and census information on local neighborhood deprivation (IDACI score). Missing dummies included for student characteristics and age-7 test scores. All OLS regressions in Panel A include school controls (math and English attainment, percent free lunch, percent nonwhite, all from year before inspection); DID regressions in Panel B include school fixed effects.