propensity score matching model

[ALKEBSEE RADWAN]

This time I will try to make my request a bit clear.
I have watched many tutorial videos about PSM model in stata but all of them explain examples with ready treatment variable. but my difficulty is how to create a treatment variable.
in the meantime i m confused about the command it self because some analysts use pscore command and some use psmatch2.
Accordingly, i need to learn how to generate the treatment variable. in addition, the right command.
I am so sorry for being not smart enough

[Mike Lacy]

If you want to analyze the effect of a treatment, there must either already be a treatment variable, or somewhere in your data set there must be information so that a treatment variable can be created. It sounds like your situation is the latter. For us to help you, we would need to know what is in your data set and what you want to do with it. We cannot guess that. Please read in the FAQ regarding using -datatex- to post example data.

Regarding pscore vs psmatch2: There are a number of commands that do PSM models, including a built-in -teffects- command. To get started, I would recommend you use whichever one has the documentation you find the most clear. Later, when you get some experience, you might decide one is better than the other for your purposes.

[ALKEBSEE RADWAN]

@Mike Lacy thank you
i investigate the relation between CEO pay and earnings management. my treatment will be high pay vs low pay so i used the median to divide and classify CEO pay and create my treatment ( = 1 if CEOpay over the median 0 otherwise). am i right ?
but i could not use the command. i think i dont understand the way

[Mike Lacy]

Truncating pay into above/below the median is likely not a good idea for substantive reasons.

Setting that aside, the treatment variable you say you have created should work. However, neither I nor other people here can intelligently respond with the limited information you present here and in other posts. I'd encourage you to read in the FAQ about presenting sample data, and about showing exact Stata code and Stata's response that you are seeing as a problem. If you follow that advice, you're more likely to be able to make postings that people will try to answer. With what you have here, you leave people to guess.