qr2me: QR with 2-sided measurement error

In bcallaway11/qrme: Quantile Regression with Measurement Error

Description

Estimates QR parameters in the case with measurement error in the outcome and measurement error in a particular continuous "treatment" variable.

Usage

qr2me(
  yname,
  tname,
  xformla,
  tau,
  data,
  xdf = NULL,
  tvals = NULL,
  copula = "gaussian",
  Qyx,
  Qtx,
  retFytxlist = FALSE,
  ndraws = 100,
  messages = TRUE
)

Arguments

yname	name of the outcome in the passed in data
tname	name of the treatment in the passed in data
xformla	a one-sided formula for additional covariates (assumed not to be measured with error)
tau	a vector containing particular quantiles that have been estimated ??
data	a data.frame containing the data used for estimation
xdf	If you want conditional distributions to be returned, pass in the value of the distribution here; otherwise the default behavior is to return a single distribution that averages over all values of X in the dataset
tvals	values of the treatment to compute conditional distribution-type parameters for
copula	which type of copula to use (default is "gaussian")
Qyx	quantile regression estimates (can be adjusted for measurement error) of Y on X
Qtx	quantile regression estimates (can be adjusted for measurement error) of T on X
retFytxlist	whether or not to return the conditional distribution for every value of x in xdf (default is FALSE because this can take up a lot of room in memory)
ndraws	number of draws to use in MH algorithm to estimate first step quantile regressions (default 250)
messages	whether or not to report details of computation as they occur (default is FALSE)

bcallaway11/qrme documentation built on June 30, 2021, 12:52 p.m.