The teacher incentives experiment of Table 4.1 under block random assignment. Assignment probabilities differ across blocks, so the design needs weights.
Format
A data frame with 40 rows and 9 columns:
- observation
Subject number, matching Table 4.1
- block
Block, one of three religious-composition strata
- X_pretest
Pretest score, strongly predictive of the outcome
- X_pretest_weak
Pretest score, weakly predictive of the outcome
- Z
Realized treatment assignment under blocking
- prob
Probability of assignment to treatment in this subject's block
- wt
Inverse-probability weight
- Y_Z_0
Untreated potential outcome
- Y_Z_1
Treated potential outcome
Source
ISPS Data Archive study D081 doi:10.60600/YU/VUH8KD
, file 42335
(GerberGreenBook_Chapter4_Combined_Table_4_1_4_2). Built by
data-raw/book_tables.R.