The correlation matrix waVH[DPLQHGWRLGHQWLI\LWHPVWKDWDUHHLWKHUWRRKLJKO\FRUUHODWHG U RU
not correlated (r < .30) with one another. If items are too highly correlated, there is a problem of
multicollinearity and one or more of the highly correlated items need to be dropped from the analysis.
If the items are not correlated strong enough, there will not be much shared common variance, thus
potentially yielding as many factors as items. There were no items with an r . It can be concluded
that there are no items too highly correlated with each other. Nor is there an item which does not
correlate with any of the other items. There were two negative correlations, EHWZHHQWKHLWHP ³,
focus on long-WHUPREMHFWLYHVZLWK³,FUHDWHDQDWPRVSKHUHLQZKLFKSHRSOHIHHOFRPIRUWDEOH ́DQG³,
UHZDUGP\SHRSOHIRUJRRGSHUIRUPDQFH ́
Prior to the extraction of factors, several tests were performed to assess the suitability of the
respondent data for factor analysis. These tests included the Kaiser-Meyer-Olkin (KMO) Measure of
Sampling Adequacy and the %DUWOHWW¶V7HVWRI6SKHULFLW\(Pett et al., 2003). This resulted in a KMO =
0.795 and a significant result from the %DUWOHWW¶V7HVWRI6SKHULFLW\(p=0.000). Hence the data set
passed the KMO test (>.50) DQG%DUWOHWW¶V7HVW(p<.50) and hence can be considered suitable for factor
analysis (Hair et al., 1998; Tabachnick & Fidell, 2007).
On the basis of the Kaiser criteria (eigenvalues > 1), the factor analysis for this data set resulted in
four distinct factors (3.735, 1.385, 1.308 and 1.006), explaining 62% of total variance. A Varimax
orthogonal rotation resulted in the factor structure shown in Table A.4.
Table A.4: Pattern matrix of factor solution
Items that either loaded strongly (>.40) on several factors that did not load on any factor, or that did
not conceptually fit any logical factor structure were discarded. Traditionally, at least two or three
variables must load on a factor so it can be given meaningful interpretation (Henson & Roberts,