mean-test

#1
Dear colleagues,

I´m trying to run a mean test (ttest) for the variable satisfaction with life by neighborhoods, however I´m considering 3 different neighborhoods (the variable neighborhood assumes three different values: 1, 2 and 3) and the message I`m getting id that :

"more than 2 groups found, only 2 allowed
r(420);"


Is there any other test I can run or is there another way run a ttest?

I heard something about a mvtest, but I´m not sure about the homogeneity (or not) of the covariance matrix.
Could anyone help me?

Thanks, in advance.
 

hlsmith

Not a robit
#2
If you have >2 groups then you should look at ANOVA. If ANOVA is significant then you run pairwise ttests to figure out which groups differ. Though if you run this last followup step you need to correct you level of significance to address possible false discovery from rmultiple tests.

Let us know if you have questions!