How to find ttest at 1% ?

[Arthur Pandart]

Hello.

I want to make a t-test at 5%and one at 1%. I use this code :

ttest d2, by(b7a)

Then, t = 0,4282 <1,96 so it is not significative at 5%. But what about 1% ? Do I have to make an other code (Which one then ?). Or saying 0,4282<2,58 is enough to say that it is not significative at 1% ?

[William Lisowski]

The answer to the question posed in the title of your post is suggested by the output of help ttest to be

ttest d2, by(b7a) level(99)

But with that said, the answer to the question posed in the body of your post is that the p-value for your test (assuming the alternate hypothesis is that the means are not equal) is .6691, so the difference is not significant for any value below 66.91%.

[Marcos Almeida]

A few things to take into consideration. The n for the yes group is quite short (11, against 141). Surprisingly, the SD for the "no" group is more the 40 times larger. No wonder p-values were high. Sattertwhaite adjustment could be an option if p-values were low and we wish to consider the heterogeneity. Hopefully that helps.