checking linearity with lack of fit test

[Nikoletta Agnes Palmai]

Hello, I'm running a linearity check (analitical data), and I'd like to run lack of fit test after linear regression. The command maxr2 doesn't run after the regress command. (But in the past, it was ok). I use Stata 15. What is the solution? Thank you for that!

[Rich Goldstein]

maxr2 was originally written when the command was -fit-; I did re-write later and you should check that you have the current version; here is what I have

. which maxr2
/Users/rich/Library/Application Support/Stata/ado/plus/m/maxr2.ado
*! version 3.1.1 4 January 2011 RG (orig STB-9: srd13)

[Nikoletta Agnes Palmai]

Thank you very much for that! Here is what I have:
. *! version 3.1.1 4 January 2011 RG (orig STB-9: srd13)
But I also tried it after fit, and still no result. what can I do?

[Joseph Coveney]

Above you said that it was OK in the past. So either your installation has been corrupted or your dataset is peculiar in some manner that is incompatible with the command. For the former, try re-installing it (from SSC, I guess?). For the latter, you're probably going to have to read the FAQ and act accordingly in order to improve your chances for a helpful reply.

[Rich Goldstein]

except for one point, I agree with Joseph Coveney ; the program (most recent version anyway) is from the Stata Journal; but please follow the FAQ and show exactly, in CODE blocks, what Stata returned after your command (which you need to show also);

added in edit - let me clarify the last point; you need to show all code you entered from your estimation command through the maxr2 command; entering certain commands in between these will overwrite or delete some of the returned values that maxr2 uses and relies on