ECON 151 Lecture Notes - Lecture 45: Kindergarten, Takers, Equation
Document Summary
At baseline, about same number of winners and losers for males. None of the variables are significantly different (difference column) Variation in the probability of attending promise academy (0. 599) Harlem=attended promise academy and attend any charter school (under schools characteristics) 0. 113 more likely to attend charter school. Measure of takers = 0. 256 and 0. 113. T-value: sd in regression - coefficient in regression. Confidence interval: coefficient - 1. 96 (for 95%) * sd. "no excuse" strategy: strict discipline (like the angrist and abdulkadiroglu papers) First-stage: rct with imperfect compliance (people are offered treatment but can refuse - can refuse to go to harlem children zone) Hcz igt = t + g + lottery ij + xi + zi + igt. For each individual i for each year t (04,05,06) and each grade (3rd to 5th). Lottery takes value one when the child wins. Give the treatment effect when hcz is instrumented by lottery. Y = b0 + b1* lottery + e.