ri (version 0.9)

estlate: Estimation of local average treatment effects under noncompliance

Description

Function for estimating the local average treatment effect (LATE) via variants of Wald/2SLS estimation (taking the ratio of two average treatment effect estimators). Permits regression adjustment for covariates, difference estimation (with a pretreatment measure of the outcome variable), inverse probability weighting and Horvitz-Thompson estimation.

Usage

estlate(Y, D, Z, X = NULL, Ypre = NULL, Dpre = NULL, prob = NULL, HT = FALSE)

Arguments

Y

numeric vector of length N, outcome variable

D

binary vector (0 or 1) of length N, treatment receipt indicator

Z

binary vector (0 or 1) of length N, treatment assignment indicator

X

N-by-k numeric matrix of covariates for regression adjustment

Ypre

numeric vector of length N, pretreatment measure of the outcome variable for difference estimation

Dpre

numeric vector of length N, pretreatment measure of the treatment receipt variable for difference estimation

prob

numeric vector within the (0,1) interval of N-length, probability of treatment assignment, as output by genprob() or genprobexact(). When prob=NULL (by default), assumes 0.5 probability of assignment to treatment

HT

when HT=TRUE, invokes the Horvitz-Thompson (difference-in-totals) estimator. When HT=FALSE, invokes the inverse-probability-weighted regression estimator

Value

a numeric scalar, the estimated LATE

References

Angrist, Joshua D, Guido W. Imbens and Donald B. Rubin. 1996. Identification of Causal Effects Using Instrumental Variables. J. Amer. Statist. Assoc. 91: 444-55.

Gerber, Alan S. and Donald P. Green. 2012. Field Experiments: Design, Analysis, and Interpretation. New York: W.W. Norton.

Horvitz, D.G. and D.J. Thompson. 1952. A generalization of sampling without replacement from a finite universe. J. Amer. Statist. Assoc. 47: 663-684.

See Also

estate

Examples

y <- c(8,6,2,0,3,1,1,1,2,2,0,1,0,2,2,4,1,1) 
Z <- c(1,1,0,0,1,1,0,0,1,1,1,1,0,0,1,1,0,0)
D <- c(1,0,0,0,0,0,1,0,1,1,0,1,0,0,1,0,0,1)

cluster <- c(1,1,2,2,3,3,4,4,5,5,6,6,7,7,8,8,9,9)
block <- c(rep(1,4),rep(2,6),rep(3,8))
probs <- genprobexact(Z,block,cluster) # generate probability of assignment
late <- estlate(y,D,Z,prob=probs) # estimate the LATE; estimated LATE = 9