HomeScience & TechnologyRelated VideosMore From: SAS Software

Understanding the Subject= Effect in SAS® Mixed Models Software

21 ratings | 2747 views
Jill Tao of SAS will help you understand the SUBJECT= effect in SAS mixed models software. SUBSCRIBE TO THE SAS SOFTWARE YOUTUBE CHANNEL ABOUT SAS SAS is the leader in analytics. Through innovative analytics, business intelligence and data management software and services, SAS helps customers at more than 75,000 sites make better decisions faster. Since 1976, SAS has been giving customers around the world THE POWER TO KNOW®. VISIT SAS CONNECT WITH SAS SAS ► SAS Customer Support ► SAS Communities ► Facebook ► Twitter ► LinkedIn ► Google+ ► Blogs ► RSS ►
Get embed code!
Text Comments (6)
redmotherfive (28 days ago)
This is great!
Dave Schneider (2 months ago)
Very helpful. Is there also some video help on understanding the group= parameter?
SAS Software (2 months ago)
There is not a video that explains more on the group= option, but here is related information that may be helpful: GROUP=effect GRP=effect defines an effect that specifies heterogeneity in the covariance structure of. All observations that have the same level of the GROUP effect have the same covariance parameters. Each new level of the GROUP effect produces a new set of covariance parameters with the same structure as the original group. You should exercise caution in properly defining the GROUP effect because strange covariance patterns can result from its misuse. Also, the GROUP effect can greatly increase the number of estimated covariance parameters, which can adversely affect the optimization process. Continuous variables are permitted as arguments to the GROUP= option. PROC MIXED does not sort by the values of the continuous variable; rather, it considers the data to be from a new subject or group whenever the value of the continuous variable changes from the previous observation. Using a continuous variable decreases execution time for models with a large number of subjects or groups and also prevents the production of a large "Class Level Information" table. For example, comparing the following two REPEATED statements in PROC MIXED – Repeated hour / subject=id type=ar(1;          and Repeated hour / subject=id type=ar(1) group=trt; There will be two parameters from the REPEATED statement 1) above – one for the residual variance sigma_2and one for the AR(1) parameter rho. The resulting R matrix is a block-diagonal matrix, each block corresponds to each ID, and within each block, the structure is AR(1) with the estimated value for the residual variance sigma_2 and AR(1) parameter rho. With the GROUP=trt in 2), the number of the parameters will be a multiple of 2. The multiple depends on the number of levels for trt. Suppose there are two trt in the data, then the number of parameters will be 2x2=4. The resulting R matrix is still a block-diagonal matrix, each block corresponds to each ID, and within each block, the structure is AR(1). The difference is that, for those ID’s that take trt #1, the parameter estimates are residual variance sigma_21 and the AR(1) parameter rho1, and for those ID’s that take trt #2, the parameter estimates are residual variance sigma_22 and the AR(1) parameter rho2. So in summary, the GROUP= option allows you to estimate different sets of covariance parameters for different groups identified by the GROUP= option (in this case, trt). It increases the flexibility of your model but also increase the complexity of the model. Whether to include this option in your model depends on your data. You can examine the fit statistics table output to decide whether it is a good idea to have it or not.
SAS Software (2 months ago)
Glad you liked it, thanks for sharing! And, we're looking into some additional resources/information for you, stay tuned!
Enriching Exchanges (4 months ago)
Very helpful!
SAS Software (4 months ago)
We're so glad you enjoyed the content!