topic Re: Calculation of R square and adjusted R square in PROC REG and PROC SURVEYREG in Statistical Procedures

Calculation of R square and adjusted R square in PROC REG and PROC SURVEYREG

momi — Wed, 21 Dec 2016 21:44:13 GMT

Is there any difference between R square and adjusted R suqare calculaiton between PROC REG and PROC SURVEYREG? We should get the same value from both procedures, correct?

Thanks,

Re: Calculation of R square and adjusted R square in PROC REG and PROC SURVEYREG

Reeza — Wed, 21 Dec 2016 21:50:46 GMT

I'm moving this to the Analytics> Statistical Procedures Forum, you should get an answer there.

Intuitively I want to say no, especially if you've use a WEIGHT or STRATA statement. Then the data that goes into each regression model would differ and the variance would, so therefore the R square values would not be the same.

You should probably wait for an answer from someone more qualifed 🙂

Re: Calculation of R square and adjusted R square in PROC REG and PROC SURVEYREG

Rick_SAS — Sun, 25 Dec 2016 12:01:39 GMT

Show us the code. What options are you using?

As Reeza says, presumably you are using SURVEYREG because you have sampling weights, strata, domains, other features that you are trying to use to provide corrections to the variance calculations.