SAS

code
analysis
SAS code
Author

Zhenglei Gao

Published

December 16, 2024

I have once again installed SAS on my computer and renewed the license today.

Here is my test script.

data calculation;
x = 3.0+4.5;
put 'The answer is:' x;
run;

It is incredible that I still need write so many additional text for doing a simple calculation. However, SAS does have many advantages that open source languages, such as R and python, don’t have. For one thing, I really enjoy reading the extensive documentation of SAS, including both methods and implementation.

I still remember back 10 years ago, SAS PROC GLM using residual error in all F tests even when a RANDOM statement is supplied. While PROC GLM also has a random statement, its models are estimated as though all effects are fixed. The random effects are treated in a post hoc fashion after the complete fixed effect model is fit. Standard errors for estimates and LS-means based on the fixed effects model may be significantly smaller than those based on a true random effects model. PROC MIXED instead use true random effects models.