Stata 16

[Rabab Al hasni]

Dear Statalist,

I hope this finds you well. I would like to ask regarding the new version of Stata 16. In fact, I used to work on the Stata version 15 and two weeks ago I made an upgrade for the software but I found that the company has sent to me version 16 instead of 15. My concerns regarding the syntax command in the software: are there big differences between the two versions need to be considered while continuing working on my analysis? Do I need to use (xt) for the regression, such as xtheckman especially that in Stata 15 the command is heckprobitt? I am no familiar with using xt for regression analysis, any advice, please.


Many thanks for your reply and your kind cooperation

Kind regards,
Rabab

[Rich Goldstein]

see

help whatsnew15to16

[Maarten Buis]

It is good practice to always start your .do files with the version command to ensure that new versions are not going to break your do files. You can still add that command. If you do that, your do-files should run just fine. Even without the version command your do-files are very likely to run just fine.

The heckprobit command is still around in Stata 16, and it does something completely different from xtheckman, so I don't understand why you seem to think you need to replace heckprobit with xtheckman. So can you tell us more about that?

[Rabab Al hasni]

It is good practice to always start your .do files with the version command to ensure that new versions are not going to break your do files. You can still add that command. If you do that, your do-files should run just fine. Even without the version command your do-files are very likely to run just fine.

The heckprobit command is still around in Stata 16, and it does something completely different from xtheckman, so I don't understand why you seem to think you need to replace heckprobit with xtheckman. So can you tell us more about that?

Dear Maarten

Thank you so much for this clarifications.

Regarding the above syntax command, I thought that they have the same function and objective and just differ in syntax command. In fact, in my work analysis I used Multinomial Logit Model but the reviewer advice me to apply Heckman selection model in order to avoid the issue of selection bias that may exist with former type of the analysis because of IIA (independence irrelevant alternatives). which one please of these two: xtheckman and heckprobit are relevant to my need for the analysis work?

Could you please also advise me on how to avoid selection bias with the multinomial analysis. I have read about Multinomial with selection model but I could not find any materials that would help in the process of performing the analysis in Stata software. I do not know if the multinomial with selection would help in solving the problem of selection bias. May be important to mention that I am using a primary survey and I got 300 respondents. each category has its own questions, wondering how selection bias would exist in this condition.


Appreciate your kind help

Best regards,
Rabab

[Rabab Al hasni]

see

help whatsnew15to16

Dear Rich

many thanks for your advice

Kind regards
Rabab