This is a test. Please ignore.
--
Erik Ray Unix Systems Administrator
Harvard-MIT Data Center Harvard University
phone: (617) 496-5097 mobile: (781) 710-1162
-
MatchIt mailing list served by Harvard-MIT Data Center
List Address: matchit(a)latte.harvard.edu
Subscribe/Unsubscribe: http://lists.hmdc.harvard.edu/?info=matchit
MatchIt Software and Documentation: http://gking.harvard.edu/matchit/
Dear Dr. Ho,
I am studying the matching method in order to use it for the evaluation of environmental conservation policies. I am just starting, and I found the application "matchit" very useful. However, I have had problems trying to simulate the examples that you present in your article. I found the following problem doing the nearest neighbor propensity score matching:
Assignment model specification:
matchit(formula = list93 ~ mamal + bird + reptil + amph + monotyp + lnleng + ncr93, data = species, replace = TRUE)
Summary of propensity score for full and matched samples:
Means Treated Means Control SD T-stat Bias
Full 0.67 0.25 0.32 19.37 1.29
Matched 0.67 0.67 0.26 0.06 0.01
Sample sizes:
Treated Control Total
Full 242 317 559
Matched 242 84 326
I can't see the frequency with which observation is used as a match. The command that you have on your documentation results in NULL for my data. However, all commands before works perfectly well.
> ps1$weights[ps1$treat==0]
NULL
Please send me a clue to solve my problem. It is real important for my research to analyze the frequency with which observation is used as a match.
Thanks for your time,
Monica Ospina