r - Using standard deviations in GenMatch to encourage more pairs -

so following example matching package , in particular genmatch example. this continues on previous question

link r package here

following example in genmatch

library(matching) data(lalonde) attach(lalonde)  x = cbind(age, educ, black, hisp, married, nodegr, u74, u75, re75, re74)  balancemat <- cbind(age, educ, black, hisp, married, nodegr, u74, u75, re75, re74,                     i(re74*re75))  genout <- genmatch(tr=treat, x=x, balancematrix=balancemat, estimand="ate", m=1,                    pop.size=16, max.generations=10, wait.generations=1)  genout$matches genout$ecaliper  y=re78/1000  mout <- match(y=y, tr=treat, x=x, weight.matrix=genout) summary(mout) 

we see 185 treated observation paired 270 non-treatment observation.

we can generate table treatment cases , age on left , control case , age on right by:

pairs <- data.frame(mout$index.treated, lalonde$age[mout$index.treated], mout$index.control, lalonde$age[mout$index.control]) 

now, literature weight.matrix generated genmatch cryptic , doesn't explain these values represent. have open question here. lets want relax matching more flexible pairing on age criteria occurs.

we see sd(lalonde$age) gives sd of 7 years our data.

so want weight.matrix account this. want use limit of 1 sd age variable , return more pairs original 185-270.

my guess generate second genmatch function continue code. use:

genout <- genmatch(tr=treat, x=x, balancematrix=balancemat, estimand="ate",                    pop.size=1000, max.generations=10, wait.generations=1,                    caliper=c(2,1,1,1,1,1,1,1,1,1)) 

but not increase number of pairs return.

any hints or solutions going wrong

as nick kennedy describes:

summary(as.logical(lalonde$treat))    mode   false    true    na's  logical     260     185       0  

genmatch match m times each treated case. can potentially drop treated cases, , drops control cases since many don't match, can't generate new treated cases out of thin air: multiple imputation ;-)

if mean, generating more matches per treated case achieved m argument, caution needed, when number of controls close number of treated cases, in lalonde data, since has found best match, , adding additional matches unlikely improve matters, , worsens them. best when number of controls >> number of treated.

you can reconstruct each 'pair' of matches when m > 1 output data, if like, , give number of rows greater 185 in treatment group, of course duplicates.