topic Re: nonormalize for weights in SAS Procedures

nonormalize for weights

PinaHolway — Wed, 15 Apr 2015 21:09:53 GMT

Hello. I am new to SAS. I have read that by default, weights are normalized in SAS. However, I would like to use the weights as they are. I have tried the suggested "nonormalize" option with weights, but I get the following error message:

63 proc surveylogistic;

64 cluster clusterid;

65 strata strataid;

66 weight weightid/nonormalize;

200

ERROR 22-322: Expecting ;.

ERROR 200-322: The symbol is not recognized and will be ignored.

Is my syntax incorrect? Any advice/suggestions would be greatly appreciated. Thank you in advance!

Pina

Re: nonormalize for weights

ballardw — Wed, 15 Apr 2015 22:33:16 GMT

The weight statement in Surveylogistic has no options. The Nonormalize only applies to the High Performance procedures in the ETS module.

I don't believe that the weights are normalized in Surveylogistic or any of the non-High Performance procedures.

Re: nonormalize for weights

PinaHolway — Thu, 16 Apr 2015 00:46:04 GMT

Thank you for your reply - I must have misunderstood. My concern is that when including my weight variable, the -2 log likelihood is extremely high (in the millions) and shouldn't be. When running the analyses without the weight, the value is much more reasonable, so it tells me that something is going on with the weight.