Problems about the difference significance between groups

[Xuan Dong]

Hello everyone,
I am looking at a how income is affected before and after the policy(tcp_universal) between the control and treatment group(group_universal), and I am trying to look at the marginal effect.
I think I am using the right margins command, so the result is the control group had a 8% increase which is significant, and the treatment group only had a 6% increase which is not significant.
My focus is on the difference between the two groups, and whether it is significant.
Is there a ways is see the significance of the difference between the two groups? (I remember once heard my prof mention about a function which uses the interval to calculate but I cannot find that, or there might be any command in stata that might work?)
Thank you very much

[Clyde Schechter]

You don't show the regression that led up to the -margins- commands. But on the assumption that it contains i.group_universal##i.tcp_universal,* you can get what you are asking for from the regression output directly. Look at the 1.group_uninversal#1.tcp_universal row in the regression output table: that's where the difference in differences will be found.

*or, equivalently, if it contains i.group_universal, i.tcp_universal, and i.group_universal#i.tcp_universal.

[Xuan Dong]

You don't show the regression that led up to the -margins- commands. But on the assumption that it contains i.group_universal##i.tcp_universal,* you can get what you are asking for from the regression output directly. Look at the 1.group_uninversal#1.tcp_universal row in the regression output table: that's where the difference in differences will be found.

*or, equivalently, if it contains i.group_universal, i.tcp_universal, and i.group_universal#i.tcp_universal.

Thank you very much, this if of great help! But I do not know why the two equals, and if next time I meet a similar case in a regression which lead me no former step to look for, how should I solve the problem?

Have a nice day!

[Clyde Schechter]

But I do not know why the two equals, and if next time I meet a similar case in a regression which lead me no former step to look for, how should I solve the problem?

Sorry, but I don't understand what this says. What "two equals" are you referring to. And I don't know what "lead me no former step to look for" means.