LATEerror: GMM inference to estimate the LATE for a mismeasured binary...

In tkhdyanagi/LATEerror: Inference on the local average treatment effect for a mismeasured binary treatment

Description

LATEerror implements the GMM inference to correct a measurement error problem due to a mismeasured binary treatment in the LATE inference developed in Yanagi (2017). See the package vignette via vignette("LATEerror") for more details.

Usage

LATEerror(Y, D, Z, V, weight = NULL, optimal = FALSE, equal = TRUE,
  lower = NULL, upper = NULL, control = NULL)

Arguments

Y	vector of the outcome
D	vector of the treatment
Z	vector of the instrument
V	vector of the exogenous variable
weight	weighting matrix for the pilot GMM estimation
optimal	logical whether implementing the optimal GMM estimation
equal	logical whether the misclassification probabilities do not depend on the instrument
lower	vector of lower bounds of the parameters
upper	vector of upper bounds of the parameters
control	list of the control parameters for DEoptim

Value

list that contains the two lists: (i) pilot, (ii) optimal.
pilot contains the following elements.

estimate	vector of parameter estimates
se	vector of standard errors
ci	matrix that contains 95 percent confidence intervals
optim	list that contains the results of the optimization via DEoptim
weight	weighting matrix for the pilot GMM estimation

optimal contains the following elements.

estimate	vector of parameter estimates
se	vector of standard errors
ci	matrix that contains 95 percent confidence intervals
overidentification	vector that contains a value of the test statistic and p-value for the over-identification test
weight	weighting matrix for the optimal GMM estimation
optim	list that contains the results of the optimization via DEoptim

See Also

DEoptim for the elements in control

tkhdyanagi/LATEerror documentation built on May 3, 2019, 9:39 p.m.