Statistical tests about the differences in coefficient estimates across the two groups

[Mohsin Javed]

I have been asked to perform statistical tests about the differences in coefficient estimates across the two groups (men and women). My analysis examines the impact of computerisation on the employment outcomes of men and women. I want to determine whether the coefficients of the computerisation variable for men and women are statistically different. For example, I run the following regressions:

ivreg2 d_employment_men (computer_exposure=computer_exposure_iv)
ivreg2 d_employment_women (computer_exposure=computer_exposure_iv)

Any ideas?

[Andrew Musau]

See https://www.statalist.org/forums/for...ferent-samples

[Mohsin Javed]

Andrew Musau: thanks for providing this link. I ran the code from this link on my data and got the following output:

chi2( 1) = 1.80
Prob > chi2 = 0.1802

I believe this implies that the coefficients on variable computer_exposure for men and women are same (not different). Is that right?

[Carlo Lazzaro]

Mohsin:
it seems correct.
I write "it seems" as only the final outcome of the test is visible and I cannot see what your code was.