Information Technology Reference
In-Depth Information
Levene's Test of Equality of Error Variances
a
The Levene test yielded an
F
ratio that was not statistically
significant. Thus, our data
have met the assumption of
homogeneity of variance.
Dependent Variable: range 0-60
Sig.1
F
df1
df2
.123
5
24
.986
Tests the null hypothesis that the error variance of the
dependent variable is equal across groups.
a.
Design: Intercept
+
gender
+
reside
+
gender
∗
reside
Figure 8.11
Test of homogeneity of variance.
all statistically significant. Recall our discussion of the full and corrected
models in Section 6.9.1 in the context of the general linear model. What
we see in Figure 8.12 are the results for the full model that includes the
intercept term. For our purposes, we can ignore the intercept and focus
on the corrected model.
The entries in the SPSS summary table relative to what is contained in
Table8.1maybeunderstoodasfollows:
Corrected Model
combines all of the effects attributable to the inde-
pendent variables into one sum of squares value. In this design,
adding the sums of squares associated with the main effect of
gen-
der
, the main effect of
reside
, and the Gender
Reside interaction
yields the corrected model sum of squares. The footnote on the
value in the summary table is used to report the eta squared (called
Rsquared
by SPSS) for the corrected model (4,364.167 divided
by 5,248.167 is .832). We ordinarily compute separate eta squared
values for each of the effects separately as shown in Table 8.1.
Corrected Total
corresponds to the total variance from Table 8.1.
It is based on the “corrected” model that excludes the regression
solution.
×
These are the
probabilities of the
F
ratio occurring by chance
alone for the main
effects and interaction.
Tests of Between-Subjects Effects
Dependent Variable: range 0-60
Type III Sum
of Squares
df
Mean Square
F
Sig.
Source
Corrected Model
Intercept
GENDER
RESIDE
GENDER * RESIDE
Error
Total
Corrected Total
3
4364.167
18500.833
1400.833
1551.667
1411.667
884.000
23749.000
5248.167
872.833
18500.833
1400.833
775.833
705.833
36.833
23.697
502.285
38.032
21.063
19.163
.000
.000
.000
.000
.000
5
1
1
2
2
24
30
29
a. R Squared
=
.832 (Adjusted R Squared
=
.796)
Figure 8.12
Summary table for the omnibus analysis.
