strata_match: Strata Match

In raikens1/BigMatch: Stratification and Matching for Large Observational Data Sets

Strata Match

Description

Match within strata in series using optmatch. Requires optmatch package to be installed.

Usage

strata_match(object, model = NULL, method = "prop", k = 1)

Arguments

object	a strata object
model	(optional) formula for matching. If left blank, all columns of the analysis set in object will be used as covariates in the propensity model or mahalanobis match (except outcome, treatment and stratum)
method	either "prop" for propensity score matching based on a glm fit with model model, or "mahal" for mahalanobis distance matching by the covariates in model.
k	the number of control individuals to be matched to each treated individual. If "k = full" is used, fullmatching is done instead of pairmatching

Value

a named factor with matching assignments

Examples

# make a sample data set
set.seed(1)
dat <- make_sample_data(n = 75)

# stratify with auto_stratify
a.strat <- auto_stratify(dat, "treat", outcome ~ X2, size = 25)

# 1:1 match based on propensity formula: treat ~ X1 + X2
strata_match(a.strat, model = treat ~ X1 + X2, k = 1)
# full match within strata based on mahalanobis distance.
strata_match(a.strat, model = treat ~ X1 + X2, method = "mahal", k = 1)

raikens1/BigMatch documentation built on April 1, 2022, 9:47 p.m.