bigmatch (version 0.6.4)

findexact: Build an exact match variable given a matrix of covariates ordered by the user according to their importance.

Description

The function is used to build an exact match variable given a matrix of covariates ordered by the user according to their importance. It will select as many important covariates as possible.

Usage

findexact(z, E, ncontrol=1)

Value

miss

The covariates that cannot be exactly matched.

variables

The covariates that can be exactly matched.

NewExact

The constructed exact match variable.

Arguments

z

A vector whose ith coordinate is 1 for a treated unit and is 0 for a control.

E

A matrix or a vector with length(z) rows giving the covariates that need to be exactly matched.

ncontrol

The number of controls to be matched to each treated individual.

Examples

data(nh0506)
attach(nh0506)
# The following example uses all of the variables in the propensity score,propens
ex<-findexact(z,cbind(female,age,education,black,hispanic,povertyr))
head(ex$miss)
head(ex$variables)
table(z,ex$NewExact)

# In order to exact match on age as well
# one option is to divide age into several levels
age_quantile<-as.integer(cut(age,quantile(age,c(0,0.25,0.5,0.75,1))))
ex2<-findexact(z,cbind(female,age_quantile,education,black,hispanic,povertyr,bmi))
head(ex2$miss)
head(ex2$variables)
table(z,ex2$NewExact)
detach(nh0506)