With a good grounding in multiple regression analysis there are several related techniques you can explore. A good reference to start with Mertler and Vannatta's Advanced and Multivariate Statistical Methods (Pyrczak Publishing, Los Angeles, 2001) for other types of multivariate statistical analyses while Yaffee's and McGee's Introduction to Time Series Analysis and Forecasting: With Applications of SAS and SPSS (Academic Press Inc., San Diego, 2000) covers forecasting techniques currently in use. Many of these techniques are related to least squares.
One extension of multiple regresssion found in text are logistical regression or "logit models" (see Bowerman and O'Connell and example of logistic or logit regression given on page 611 of that text). The key difference between a logit model and ordinary least squares regression is that the dependent variable is a dummy variable in logistical regressions.
To get SAS to run a a logistic regression with this data set try the following:
data
logit;input
Group Test1 Test2;datalines
;1 96 85
1 96 88
1 91 81
1 95 78
1 92 85
. . .
. . .
datalines can be copies from text or accessed from website.
0 81 84
0 85 78
0 83 77
0 81 71
;
proc
logistic descending;model
group=test1 test2;run
;Part of the output you get will be the following:
Note how similar these results are to least squares results. Again note that the key difference here is that in logistical regression the dependent variable is a dummy variable.
For more back ground on these models and how to interpret SAS results where a logistical regression has been run, click here.
For further background (and how to run logit models using another major package, SPSS) click here
See also chapter 11 of Mertler and Vannatta and chapter 9 of Cody and Smith's Applied Statistics and the SAS Programming Language (Prentice Hall, 4th ed.) For more help understanding logistical regression click here, here, or here.
N.B. This page is under construction. More to be added later.