How to test whether there is a significant difference between two regression coefficients

[James Liu]

Dear all, please give me any references on whether or not we should do t-test or z-test on one significant regression coefficient from a subsample and another insignificant regression coefficient from another subsample? Thank you very much. James

[Andrew Musau]

help suest

[Joseph Coveney]

How to test whether there is a significant difference between two regression coefficients

. . . any references on whether or not we should do t-test or z-test on one significant regression coefficient . . . and another insignificant regression coefficient . . .?

Well, there's this one (preprint here), and some discussion here.

[George Ford]

As noted in #3, if you have a poorly estimated coefficient, then it has a wide range of values in the CI (including perhaps a well-estimate coefficient). It's a common problem, and one you have to mention when discussing the test, especially when the test fails to reject but the coefficients are quite different and you might expect a difference.

[James Liu]

Thank all of you very much.