senslm: Sensitivity analysis of a least squared regression estimator

In ivregEX: Create Independent Evidence in IV Analyses and Do Sensitivity Analysis in Regression and IV Analysis

Description

Performs a sensitivity analysis of the least squared regression

Usage

senslm(lmfit, coefmat, variable, delta1 = seq(-0.1, 0.1, 0.01), alpha = 0.05, ...)

Arguments

lmfit	An obeject of class "lm".
coefmat	Optional matrix of class "coeftest" an output of coeftest.
variable	Treatment variable of interest.
delta1	A numeric vector of values of the first sensitivity parameter. Default value seq(-.1, .1, .01).
alpha	Significance level. Default at 0.05.
...	May be ignored, used in sensOLS.

Details

delta1 is the partial correlation of the unobservable and the treatment. It measures the amount of violation in the assumption of no-unmeasured confounding.

Value

Returns an object of class "si".

An object if class "si" is essentially a matrix of two columns for the the upper and lower sensitivity limits corresponding to the sensitivity parameters.

Author(s)

Bikram Karmakar

See Also

sensIvreg, plot.sensIvregEX, plot.si

Examples

	
data("CigarettesSW", package = "AER")
CigarettesSW$rprice <- with(CigarettesSW, price/cpi)
CigarettesSW$rincome <- with(CigarettesSW, income/population/cpi)
CigarettesSW$tdiff <- with(CigarettesSW, (taxs - tax)/cpi)

lmfit <- lm(log(packs) ~ log(rprice) + log(rincome), data = CigarettesSW, subset = year == "1995")

z <- senslm(lmfit, variable = 'log(rprice)')

ivregEX documentation built on May 1, 2019, 10:08 p.m.