Test for survey means in SAS and Stata

Hi I am using survey data for my analysis, and I have a question about the test for difference of means in two groups. Due to my professor's preference, I used both SAS and Stata to the test.

In Stata,

svyset [pweight=x_llcpwt], strata(x_ststr) psu(x_psu)
svy: mean age, over(fs)
lincom [age]0 - [age]1

While SAS runs regression to test F-value

proc surveyreg data=snaped3.elig;
STRATA X_STSTR; CLUSTER X_PSU; WEIGHT X_LLCPWT;
class fs;
model age=fs/noint solution vadjust=none;
contrast 'Ho: 1=0' fs 1 0; run;

They gave me different answers. (One is significant and another is not)

My question is, 1. Which is better to take, t-test or F-test? 2. What is the reason of this difference?

Thank you in advance.