What is the difference between teoverlap and tebalance density?

[Phuc Nguyen]

Hi all, I am learning propensity matching score. I saw many papers using psmatch2 (a user-written code), but there is teffects psmatch built by STATA (which is more flexible, handle with standard error better but a little bit slower)
I spent the whole day to search for the graph plotting the density of propensity matching score after using teffects psmatch.

Unlike psmatch2 with this code you can get the graph for plotting before and after matching, it is hard to find a document to show that.
After searching and searching I found the code that perform what I want.

I used the dataset from University of Wisconsin as below

use http://ssc.wisc.edu/sscc/pubs/files/psm, replace 

  teffects psmatch (y) (t x1 x2),gen(match)

The post-matching, I found a code from the help file to draw a graph

  teoverlap

I tried it and the results show is

I did some searches and saw that it is the density and propensity pre-match, not post-match

Afterwards, I found another code for plotting after using teffects psmatch

tebalance density

And I got the results

I am wondering why these two graphs are so different. Which one is correct or similar to the graph below from psmatch2

Many thanks and best regards.

[Tara Boyle]

Interesring question, did you find an answer to this ?

[Phuc Nguyen]

Interesring question, did you find an answer to this ?

yes, I did, just change to t=1 in the first graph and thing gets solved