As far as I know, you have to compute R-squared outside of PROC MIXED. ... View more

First, I'd like to make a recommendation on another statistical technique that many find preferable to stepwise regression or best subsets. The problem is that Least Squares just isn't that good a modelling technique when you have many X variables and they are highly correlated with one another. There are many critiques of stepwise and best subsets on the Internet (and elsewhere) in these situations. Here are some. A better technique is Partial Least Squares regression, which in SAS is PROC PLS. If you absolutely have to use PROC REG and want interactions and polynomial terms in the model, first create an X matrix in PROC GLMMOD to represent your main effects, interactions and polynomial terms, and then run that through PROC REG. I will note that I have never had much success doing things this way, hence the PROC PLS recommendation. Message was edited by: Paige ... View more

You are right that in theory, the NIPALS algorithm can be run with missing values for some data. However, I believe that the PROC PLS implementation of NIPALS doesn't allow for missing values. You have to use MISSING=EM or MISSING=AVG, which as you said, fills in estimated values. ... View more

You have to set values for x1, x2, and my favorite, x3, and then generate a predicted value for X. Otherwise, your problem statement seems meaningless to me. ... View more

Seriously, I wouldn't worry about it. Go ahead and use the model you came up with. ... View more

It is deleting any record where the claim number exists in both data sets. (Or, it is keeping records for which the claim number exists in one, but not both, data sets) ... View more

There is no CONTRAST, ESTIMATE or LSMEANS statement in PROC REG. Which leads me to ask, why don't you use PROC GLM or PROC MIXED or PROC GLIMMIX for this purpose? Message was edited by: Paige ... View more

So the p-test is really a p-value? Virtually every SAS statistical procedure produces p-values. Problem solved! PROBNORM is explained in SAS help, along with an example. I don't write code for people on the internet ... if you try to write code and can't get it to work, I'll be happy to provide further assistance. ... View more

Perhaps you could enlighten us as to what a P-test is? A Z-test can be done very simply in a datastep as (Xbar-mu)/sigma, and then you look up the value from the z-table with the SAS function PROBNORM. ... View more

Without seeing your data that it is checking, there is no way to tell what the cause could be. Your last statement is syntactically correct. But here's a guess anyway. Use trim(x) instead of x to eliminate trailing blanks from causing the assignment statement to fail. Message was edited by: Paige ... View more

I don't see anything in your code that could cause the problem that you spoke of originally: "The problem I am having is that if for a country the same quantity is given twice (within the quarter) then PROC MEANS only take one of the two quantities." Your code is not causing missing values to appear in the data. So it must be something in the data. Could you post some sample data? By the way, there is no need for data set test2. It accomplishes little, and could be combined into data set test1. data work.test1; set work.test; if Month in (1,2,3) then m=1; else if Month in (4,5,6) then m=4; else if Month in (7,8,9) then m=7; else if Month in (10,11,12) then m=10; date=mdy(m,01,year); format date date7.; drop m; run; ... View more

As Doc said ... unless we see your code, there really isn't a way for us to help figure out why you are having this problem (if you are having a problem at all). ... View more

If you want to know which week (out of the 52 weeks per year) has the maximum value per year (I hope I am understanding you properly), then try the MAXID option of the OUTPUT statement. ... View more

Cynthia I have been there and done that myself. My recommendation above is because I have never ever found any value to the PROC FREQ/PROC MEANS on everything approach. If the researcher cannot specify in advance some things he might want to know about the data, if its just a huge big fishing expedition, then I don't expect much success other than by random chance. So, if someone wants to compute means, or frequencies, on phone numbers, please go right ahead. I'll pass ... ... View more

I don't see much value in obtaining a "brids eyer view" that consists of the unique values of every variable. I doubt it would help me in this situation. Usually, for us to provide advice, and for you to make meaningful progress in understanding what the data is telling you, you need to formulate actual questions that you want to answer from the data. Data analysis is pointless without some clear question(s) to answer — its like computing the average phone number from a list of phone numbers, of course you can do the calculations, but why bother? ... View more