Information Technology Reference
In-Depth Information
Figure 6.16
The configured
Arithmetic
screen.
6.9 THE OUTPUT OF THE OMNIBUS ONE-WAY BETWEEN-
SUBJECTS ANOVA IN SAS
6.9.1 BRIEF OVERVIEWOF THE GENERAL LINEAR MODEL
Theoutputfromthe
One-Way ANOVA
procedure in SPSS presented in
Section 6.7 was structured in a way that matched our hand calculations.
But the summary tables produced by SAS and SPSS in certain of their
ANOVA procedures contain some entries that we have not yet covered. As
you will see in a moment, the summary table from the SAS
LinearModels
procedure that we have just used to generate our analysis has additional
wording that we need to introduce.
In its
Linear Models
procedure, SAS computes what is called the
reduced, restricted
,or
corrected model
as opposed to the
full
or
uncorrected
model;
SPSS will do likewise in some of its procedures, but there are times
when it will present the results of the full model. We will briefly explain
at a very surface level the differences between the full and the corrected
model here. A more comprehensive description of these models is beyond
the scope of this topic, but more complete descriptions can be found in
Kirk (1995), Maxwell and Delaney (2000), and Myers and Well (1991).
However,weneedtodescribethesemodelstoacertainextentsothatyou
can more fully interpret the SPSS and SAS output.
ANOVA is one of many statistical procedures - multiple linear regres-
sion, discriminate analysis, and canonical correlation are some other
members of this set of procedures - that on the surface may appear quite
different but are all manifestations of the
general linear model
. The general
linear model statistical procedure is designed to formulate a model to pre-
dict for each case the values of the dependent variable from a knowledge of
the level or levels of the independent variable each case experienced. The
