Dear MatchIt list
Is there any possibility to use exact matching, but in the meantime
specify the pre-defined intervals for accepted matches? For example,
if age is X: keep exact matching, but specify that a match is
acceptable if it is in a +/- 2 interval (age of 25 from the treatment
gets an acceptable match if the control is between 23 and 27 years old).
An alternative specification to this would be to just recode into age
categories such as 18-22, 23-27, 28-32. But in this case two people
who are just 1 year apart (22 and 23 for example) could be considered
a mismatch and people who are 5 years apart (28 and 32) could be
considered a match. I consider the above cited formulation to be
superior but we don't know how to set it up.
Does Matchit allow for these specifications? If yes, how? If not it
would great if such was implemented (or someone could point me to a
relatively strait forward way of automating this). (In case you made
a conscious decision not to implement these, what are the underlying
reasons for this?)
Cheers
Zoltan Fazekas and Levente Littvay
--
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)lists.gking.harvard.edu
Subscribe/Unsubscribe: http://lists.gking.harvard.edu/?info=matchit
MatchIt Software and Documentation: http://gking.harvard.edu/matchit/
Browse/Search List Archive: http://lists.hmdc.harvard.edu/lists/matchit/
Dear All,
I'm running matches on the same sample using nearest-neighbor, optimal, and genetic matching.
With NN matching I use a 5-to-1 ratio and it works just fine.
With OPTIMAL matching, a 2-to-1 ratio works fine but any higher ratio produces the following error message:
"Error in fullmatch(d, min.controls = ratio, max.controls = ratio, omit.fraction = (n0 - : negative 'omit.fraction' with 'min.controls' >= 2 not permitted"
With GENETIC matching, a 2-to-1 ratio works fine but any higher ratio produces the following error message:
"Error: subscript out of bounds"
Am I making a mistake or is 2-to-1 the highest possible ratio in MatchIt for optimal and genetic matching?
Thank you much,
Sinan
--
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)lists.gking.harvard.edu
Subscribe/Unsubscribe: http://lists.gking.harvard.edu/?info=matchit
MatchIt Software and Documentation: http://gking.harvard.edu/matchit/
Browse/Search List Archive: http://lists.hmdc.harvard.edu/lists/matchit/
