topic Re: Superiority testing and Proc T test in SAS Procedures

Superiority testing and Proc T test

reshmape — Tue, 06 Mar 2018 05:36:21 GMT

Hi All,

I would like to use PROC T TEST for superiority testing. The hypothesis will be tested by one-sided t-test at the alpha level of 0.025. Rejection of the null-hypothesis is consistent with superiority of the investigational treatment to the placebo.

Do you have any idea how can we perform in SAS?

Thanking in Advance

Re: Superiority testing and Proc T test

ballardw — Tue, 06 Mar 2018 23:58:16 GMT

The option SIDES on the Proc Ttest statement is used to indicate 2 L(ower one side) or U(pper oneside) test, default is 2

The option Alpha sets the significance, use Alpha=0.025. Default is 0.05.

Data should have a Class variable to indicate which of the two populations records come from (treatment or placebo), this variable can be numeric or character and is indicated with a CLASS statement.

A variable(s) should hold the measure(s) of interest and must be numeric and would be placed on a VAR statement;

If you are testing one variable against a known value then you only need the variable and you would specify the "known Value" with the h0 option such as h0=30. Translates as: the Null hypothesis is the value is 30.

There are several examples in the online help you should have available.

Re: Superiority testing and Proc T test

reshmape — Wed, 07 Mar 2018 10:22:27 GMT

Thank You for the response.

Re: Superiority testing and Proc T test

Rick_SAS — Thu, 08 Mar 2018 21:06:03 GMT

See the discussion and examples in Castello and Watts (2015), "Equivalence and Noninferiority Testing Using SAS/STAT® Software"