generateBalanceTable {MultiObjMatch}

Generate balance table

Description

The helper function can generate tabular analytics that quantify covariate imbalance after matching. It only works for the 'Basic' version of matching (produced by distBalMatch).

Usage

generateBalanceTable(
  matchingResult,
  covList = NULL,
  display.all = TRUE,
  statList = c("mean.diffs")
)

Arguments

matchingResult

an object returned by the main matching function distBalMatch

covList

(optional) a vector of names of covariates used for evaluating covariate imbalance; NULL by default.

display.all

(optional) a boolean value indicating whether or not to show the data for all possible matches; TRUE by default

statList

(optional) a vector of statistics that are calculated for evaluating the covariate imbalance between treated and control group. The types that are supported can be found here: bal.tab.

Details

The result can be either directly used by indexing into the list, or post-processing by the function compareTables that summarizes the covariate balance information in a tidier table. Users can specify the arguments as follows: * covList: if it is set of NULL, all the covariates are included for the covariate balance table; otherwise, only the specified covariates will be included in the tabular result. * display.all: by default, the summary statistics for each match are included when the argument is set to TRUE. If the user only wants to see the summary statistics for matches with varying performance on three different objective values, the function would only display the matches with number of treated units being excluded at different quantiles. User can switch to the brief version by setting the parameter to FALSE. * statList is the list of statistics used for measuring balance. The argument is the same as stats argument in bal.tab, which is the function that is used for calculating the statistics. By default, only standardized difference in means is calculated.

Value

a named list object containing covariate balance table and statistics for numer of units being matched for each match; the names are the character of index for each match in the matchResult.

See Also

Other numerical analysis helper functions: generateRhoObj(), getUnmatched()

Examples

## Generate matches
data("lalonde", package="cobalt")
psCols <- c("age", "educ", "married", "nodegree")
treatVal <- "treat"
responseVal <- "re78"
pairDistVal <- c("age", "married", "educ", "nodegree")
exactVal <- c("educ")
myBalVal <- c("race")
r1s <- c( 0.1, 0.3, 0.5, 0.7, 0.9,1, 1.1, 1.2, 1.3, 1.4, 1.5, 1.6, 1.7)
r2s <- c(0.01)
matchResult <- distBalMatch(df=lalonde, treatCol=treatVal, myBalCol=myBalVal,
rhoExclude =r1s, rhoBalance=r2s,
distList=pairDistVal, exactlist=exactVal,
propensityCols = psCols,ignore = c(responseVal), maxUnMatched = 0.1,
caliperOption=NULL, toleranceOption=1e-1, maxIter=0, rho.max.f = 10)

## Generate summary table for balance 
balanceTables <- generateBalanceTable(matchResult)
balanceTableMatch10 <- balanceTables$'10'