Test of Mean Differences by Categorical Variable

[Louis Kos]

Hello Stata users,

I would like to perform a test of mean differences between two independent variables (q1 and q2), however, I would like to test the mean differences between 5 age categories i.e. test the mean differences in q1 and q2 between age_cat == 1 and age_cat == 2, age_cat == 1 and age_cat == 3, age_cat == 2 and age_cat == 4, etc (just to name a few). In total, this would mean 10 mean difference comparisons.

However, using the ttest function, you cannot perform: ttest q1 = q2, by (age_cat) since the error message gives: r(420): "more than 2 groups, only 2 allowed".

Is there a way to code the testing that what I would like to do?

Thank you in advance!

[Rich Goldstein]

it is not completely clear to me what you want but you should start with -anova- and then use either -contrast- or -pwcompare-; see

help anova
help anova postestimation

[Carlo Lazzaro]

Louis:
can't you -regress- (q1-q2) onto -i.agecat- and then use -test-?

[Marcos Almeida]

You didn't inform the pattern of distribution of q1 and q2, and I'm wondering whether they are Likert scales.