The value of Cronbachs alpha should be greater than 0.6 to consider the test as reliable. But in the
given question the Cronbachs alpha is 0.422. The Cronbachs alpha value wont improve beyond 0.436
even if we exclude some variables as can be seen in below Tables 1,2 and 3.
Reliability Statistics
Cronbach's
Alpha N of Items
.422 20
Table 3
OUTPUT
KMO and Bartletts test of sphericity measures the sampling adequacy. Sampling adequacy or the KMO
value should be greater than 0.6 and the Bartletts Test should have a p value less than 0.05 to consider
it as significant sphericity. The null hypothesis is that the correlation matrix is an identity matrix and in
this case it is rejected. As can be seen in Table 4, both the requirements are met and the sample is both
adequate and significant.
2. Communalities
Communalities explains the variance in the variables that has been used for by the extracted factors. It
is better if values are greater than 0.5. We can see that all the values are greater than 0.5 except for one
which is near to 0.5 (0.490) and hence it is acceptable.
This is a plot of the eigenvalues against all the factors. Only those factors having eigenvalues greater than
1 should be considered significant. From figure 1, we can see that from the factor 8 eigen value falls above
1 which should be considered significant.
5. Component Matrix
From this matrix we can understand that if absolute value of the loading is higher, then the factor contributes
more to the variable. We can see in below table that factor 6 contributes the most to variable 5 with an
absolute value of 0.697.
Table 7: Component Matrix
We can see from this matrix that there is a weak correlation between variable 3 and factor 1 which is -.016.
A correlation of less than 0.3 is regarded as non-significant. We can see that the first factor states that
there is a strong chance that customer care executives are able to handle customer complaints directly and
immediately.
Table 8: Rotated Component Matrix