propensity score matching

[Max van Paridon]

Hi,

I am trying to run a model with propensity score matching with the following inputs: teffects psmatch (w_car10102) (Listed DealvaluethUSD_x lnmarketcap dealsize leverage w_profitability CashOnlyDeal multiple_acquisitions1010, probit). When I do this my results give:
there are 7 propensity scores greater than 1 - 1.00e-05
treatment overlap assumption has been violated; use the osample() option to identify the observations

However when I then try to use the osample() option I do not get results. I do not know how to fix this so that I get my output, do I need to drop these 7 propensity scores or how am I able to fix this?

edit: when I remove DealvaluehtUSD_x I do not get this error?

Thanks in advance

[Jeff Wooldridge]

You can estimate the probit model “by hand,” get the fitted values, and check for those very close to one. Some people recommend dropping observations with phat < 0.1 or phat > 0.9.

[George Ford]

I've run into the same problem.

If you're new to this, I recommend you go through this:

https://personalpages.manchester.ac.uk/staff/mark.lunt/propensity_guide.pdf

It will allow you to do what Jeff recommends, and you'll learn a lot.

Anytime I use the teffects commands, I make sure to do it by hand so I know what's going on under the hood.

I thought Scott C's exercise here was really neat:

https://causalinf.substack.com/p/an-amazing-journey-with-claude-35?utm_source=post-email-title&publication_id=306886&post_id=145930221&utm_campaign=email-post-title&isFreemail=false&r=3qhqsh&triedRedirect=true&utm_medium=email