left 5% quantile of t-values

[Davide Monchi]

Hello to all,

Does anyone know how to calculate the left 5% quantile of t-tests?

Thanks to all for being so available here

[Bruce Weaver]

If I understand your question, you need to look at the invt() function.

help invt

Apologies if I have misunderstood what you are asking.

[Joro Kolev]

The function - invttail(df,p)- does what you want.

E.g., to find the value that cuts off 5% in the right tail of t-distribution with 7 degrees of freedom, you do

. dis invttail(7,.05)
1.8945786

However, remember that for a two sided test, where you reject when the absolute value of the statistic exceeds 1.8945786, you have overall level of the test of 10%, because you cut off 5% in the right tail, and symmetrically below 5% in the left tail.

So if you want to do a two sided test at 5% level, you need to cut off 2.5% in the right tail, and the critical value is

. dis invttail(7,.05/2)
2.3646243

. dis invttail(7,.025)
2.3646243