How do you assess whether two (dichotomous variable) coefficients are statistically significantly different in the same OLS regression model?

Let x1 and x2 be the two variables you want to test for equal coefficients. You can do this in the REGRESSION procedure by including x1 and (x2-x1). You would compute, say, z = x1 - x2 to enter that term.

But a nicer way to do this is with the UNIANOVA command. It has a TEST subcommand. So, you would specify your model and then add this subcommand /TEST= x1 VS x2.

If you are using the dialog box (Analyze > General Linear Model > Univariate), specify the analysis in the dialog box, and then Paste the syntax and add the TEST subcommand (before the period command terminator):

Thank you--I will try your suggestions!