[Dirk Enzmann]

Both commands only work for lower case variable names.

This is ambiguous: To which commands to you refer to? Using your data (which has upper case letters) works with -omega- and with -omega2-.

[Yapeng Wang]

This is ambiguous: To which commands to you refer to? Using your data (which has upper case letters) works with -omega- and with -omega2-.

Dear Dirk: as you may notice the responses from Joseph and Brian, it seems that -omega- and -omega2- do not work if one has variable names starting with a capital letter. After I renamed the variables, I no longer saw the Stata error message.

[WJ Wang]

Hello! Sorry for busting out this old thread, but does anyone know if this version of -omega2- command could calculate the omega coefficient between two items? I replicated the code successfully with three items (many thanks to Dr. Enzmann for the code!) but when it comes to two items, an error message popped up saying that "convergence not achieved". I'm assuming this is because the code is based on the -sem- function, but would there be any alternatives for omega between two items? Thanks in advance!

[Dirk Enzmann]

A factor model with only two items is underidentified -- you would need to constrain the two loadings to be equal in order to identify it, hence the program will not work with only two items.

[WJ Wang]

A factor model with only two items is underidentified -- you would need to constrain the two loadings to be equal in order to identify it, hence the program will not work with only two items.

Thank you, Dr. Enzmann, for your response! That's exactly what I thought, but since I also saw journal articles examining 2-item dimensions with omega coefficients (using, for example, the -psych- package in R, as mentioned in this article), I was wondering if it is possible to find a Stata equivalent. Anyway, thanks again for your answer!