Hello,
I'm trying to estimate ATT using teffects psmatch in a large sample (approx. 1.9 million observations--1.1M treatment and 0.8M controls), but the program cannot reach a solution; even after waiting 24 hours. I do not get any error messages. However, when I run my code on a smaller (N=1K) sample, I am able to quickly obtain a solution. Can someone tell me if there's a limit on the sample size that teffects psmatch is able to handle? Any ideas as to why this might be happening? Can someone suggest any solutions?
Here is the code I’m trying to run:
teffects psmatch (readmit) (observation covariate_1… covariate_k), atet nneighbor(1) caliper(0.116425)
Thanks so much for any assistance you can provide,
Kafuti
I'm trying to estimate ATT using teffects psmatch in a large sample (approx. 1.9 million observations--1.1M treatment and 0.8M controls), but the program cannot reach a solution; even after waiting 24 hours. I do not get any error messages. However, when I run my code on a smaller (N=1K) sample, I am able to quickly obtain a solution. Can someone tell me if there's a limit on the sample size that teffects psmatch is able to handle? Any ideas as to why this might be happening? Can someone suggest any solutions?
Here is the code I’m trying to run:
teffects psmatch (readmit) (observation covariate_1… covariate_k), atet nneighbor(1) caliper(0.116425)
Thanks so much for any assistance you can provide,
Kafuti
Comment