Skip to main content
Latin American Economic Review

Table 7 F tests of differences on treatment effects across partial exams

From: Closing the achievement gap in mathematics: evidence from a remedial program in Mexico City

Null hypothesis

Treated schools

All sample’s schools

F test

P value

F test

P value

Partial 1 = Partial 2

0.900

0.342

0.100

0.747

Partial 2 = Partial 3

4.330

0.037

0.550

0.460

Partial 3 = Partial 4

5.690

0.002

9.760

0.002

Partial 3 = Partial 5

13.550

0.000

13.050

0.000

Partial 4 = Partial 5

1.680

0.195

0.240

0.626

Observations

29,400

110,005

  1. This table presents the F test and P values’ result of testing the null hypothesis of estimated coefficients for the treatment effects of each couple of partial exams to be equal. First two columns for the regressions run for treated schools (reported in Table 5), the last two for those run for all sample’s school (in Table 6)
Back to article page