emparam: EM Algorithm Estimator for Parametric Model

In kota7/rddsigma: Sharp Regression Discontinuity Design with Running Variable Measured with Errors

Description

Estimate the standard deviation of measurement error in the running variable of sharp RDD. This estimator is constructed under the assumption that true running variable and measurement error follow known parametric distribution.

Usage

emparam(d_vec, w_vec, cutoff, init_sigma = NULL, x_dist = c("gauss"),
  u_dist = c("gauss", "lap"), reltol = 1e-05, maxit = 200L,
  integ_method = c("romberg", "simpson", "trapezoid", "simpson2"),
  integ_reltol = reltol * 0.25, integ_depth = 100L, verbose = FALSE, ...)

Arguments

d_vec	binary integer vector of assignment
w_vec	numeric vector of observed running variable
cutoff	threshold value for assignment
init_sigma	initial values of sigma. If NULL, randomly assigned
x_dist, u_dist	distribution of x and u
reltol	relative tolerance requied
maxit	maximum number of iteration
integ_method	numerical integration method
integ_reltol	relative tolerance for numerical integration
integ_depth	maximum recursion depth for numerical integration
verbose	if true, progress is reported
...	currently not used

Value

object of rddsigma class

References

Kevin M. Murphy and Robert H. Topel (1985), Estimation and Inference in Two-Step Econometric Models. Journal of Business & Economic Statistics, 3(4), pp.370-379

Examples

## Not run: 
set.seed(456)
dat <- gen_data(500, 0.2, 0)
emparam(dat$d, dat$w, 0, x_dist = "gauss", u_dist = "gauss", verbose = TRUE)
emparam(dat$d, dat$w, 0, x_dist = "gauss", u_dist = "lap", verbose = TRUE)

## End(Not run)

kota7/rddsigma documentation built on May 20, 2019, 1:11 p.m.