Differences in means by sub groups with categorical variables

[Mattia Rossi]

I found this analysis on a paper that i wanted to replicate with my dataset.
Basically there are different scores and i want to test the difference in means by groups (like gender) for each score as you can see in the picture.

In don't know(even if i doubt) if author used a single formula and maybe "appended" it for each column, but as I am at entry level i decied to compute them "manually using the t statistic.

So i was testing gender using:

[code]
ttest score, by (gender)


Gender is a dummy so it was not a problem but with categorical variables it obviously gives me an error ar there are too many variables.

ttest FL, by (***_income)

by(): too many variables specified

In this case i have low_income, med_income, high_income.

How can I solve this problem? (If there is a way to compute direcly as in the paper table it would be awesome)

Thanks in advance

[Mattia Rossi]

Forgot to mention that even if computing:

ttest FL, by (low_income)

I would only test low income vs all the other incomes and not low vs med only. How to test the difference between low and med only?