Multiple Regression: Model Comparison

[ChaseZachary]

Hello all! I am currently working on a homework assignment. The question presents three models (some nested, others not) and two hypotheses of these models (see below):

Hypothesis 1 – Model B will fit significantly better than Model A

Hypothesis 2 – Model C will fit significantly better than Model B

I am tasked with testing these hypotheses and presenting the output. The class uses SPSS but I am using SAS. In SPSS they can force enter blocks of variables for each of the models and then have the package create a model summary. Is there anything like this in SAS? Do I just have to run each of the models and then compare the stats? Just wondering thank you so much in advance!

[ebowen]

I'm not sure what criteria you're using to measure model fit, but most SAS regression procedures will allow you to output various fit measures. For instance, PROC REG will output r-squared by default. If you include the PLOTS option it will output various diagnostic and fit plots. You can also use the AIC and BIC options for Akaike’s information criterion or Bayesian information criterion respectively. The CRITERIA/CRITERIONPANEL option will give you pretty much anything you could ever need!