Information Technology Reference
In-Depth Information
Mauchly's Test assesses equality of
group variances and correlations. Because
there is only one correlation for each
within variable, they cannot be compared
to any other. Therefore, the sphericity
assumption cannot be tested. This is why
there are zero
df
and why there are no
computed probabilities (Sig.).
Mauchly's Test of Sphericity
b
Measure: MEASURE_1
Epsilon
a
Approx.
Chi-Square
Greenhouse
e-Geisser
1.000
1.000
1.000
Lower-
bound
Within Subjects Effect
COLOR
TOYTYPE
COLOR * TOYTYPE
Mauchly's W
1.000
1.000
1.000
df
Sig.
Huynh-Feldt
.000
.000
.000
0
0
0
.
.
.
1.000
1.000
1.000
1.000
1.000
1.000
Tests the null hypothesis that the error covariance matrix of the orthonormalized transformed dependent
variables is proportional to an identity matrix.
a. May be used to adjust the degrees of freedom for the averaged tests of significance. Corrected tests
are displayed in the Tests of Within-Subjects Effects table.
b. Design: Intercept
GENDER
Within Subjects Design: COLOR
+
+
TOYTYPE
+
COLOR* TOYTYPE
Figure 15.12
The results of Mauchly's test.
matrices cannot be rejected. Mauchly's Test of Sphericity, the results of
which are shown in Figure 15.12, cannot be performed with only two lev-
els of the within-subjects variable. Levene's test of error variance equality
was not statistically significant (see Figure 15.13) for any of the within-
subjects factors; we therefore cannot reject the homogeneity of variance
assumption for our repeated measures.
The within-subjects summary table is presented in Figure 15.14. We
see from this table that the main effects of
color
, the two-way interaction
of
Gender
∗
Toytype
are statistically significant. With the highest order interaction
(the three-way) being significant, our simple effects analyses need to be
focused at this level.
∗
Toytype
, and the three-way interaction of
Gender
∗
Color
The equal variances assumptions are not
violated; probabilities (Sigs.) are > .05
Levene's Test of Equality of Error Variances
a
F
df1
df2
Sig.
YELHAND yellow
hands on
YELPRET yellow
pretend
BLUHAND blue hands
on
BLUPRET blue pretend
1.871
1
8
.208
2.679
1
8
.140
1.588
1
8
.243
.044
1
8
.839
Tests the null hypothesis that the error variance of the
dependent variable is equal across groups.
a. Design: Intercept
+
GENDER
Within Subjects Design:
COLOR
+
TOYTYPE
+
COLOR*TOYTYPE
Figure 15.13
The results of Levene's test.
