This post has NOT been accepted by the mailing list yet.
I am using the program psmatch2 to perform propensity score matching and I have several problems I hope to resolve.
1) My dataset contains data on firms separated into Type A, Type B, Type C. I ran psmatch2 to find the propensity score of a firm being of Type A. Then I hope to perform to different set of matching. Matching firms of Type B to Type A, and matching firms of Type C to Type A. I am unsure how to do this as I am only able to perform matching of pooled Type B and Type C firms with Type A.
2) I am using the nearest neighbour matching with only 1 match. I hope to clarify how that works. Is the method simply comparing the pscores and finding the one with the nearest pscore to the firm supposed to be matched to? I was under the impression this is so, but someone told me differently and that some bootstrapping is involved?
3) After a match has been made, I needed to compare the Type A firm and its matched firm on some firm fundamentals. Initially, I did a paired T-test to find out if the difference in these fundamentals are statistically significant, however, I was told that I should not do so. The reason I was given was that some bootstrapping and simulation is involved to get the statistic and that I should get these output as part of the propensity score matching process on STATA. However, I checked the help file on psmatch2 and am not sure this is possible. Is there another program that performs such statistical tests along with propensity score matching?
Thanks for offering any advice on any of my problems!