Abstract
M.Com. (Industrial Psychology)
The purpose of this study was to (a) investigate the psychometric properties of the Maree
Career Matrix (MCM) on a sample of South African working adults and, (b) to determine
whether or not Holland’s six RIASEC vocational personality types could be extracted from
the 19 interest categories of the MCM. This study was conducted in response to a growing
need for valid and reliable measures of vocational interests in the South African context. The
MCM has demonstrated good reliability and validity for high school learners. However, no
study to date has investigated the psychometric properties of this measure with a sample of
working adults, nor has any study explicitly investigated the structural validity of the MCM.
A total of 117 adults from several different industries completed the MCM. The reliability of
the 19 MCM interest categories was investigated using Cronbach’s alpha and McDonald’s
coefficient omega.
Predominantly, coefficients were greater than .70. Structural validity of the 19 interest
categories was investigated using principal components analysis and conversion of
rectangular to polar coordinates. The results showed evidence of a circular structure among
the 19 interest categories and that most of the interest categories were located as theoretically
projected. This supported the conclusion that the MCM is structurally valid. RIASEC scales
were obtained by summing interest categories that were located as theoretically expected and
that had better projection into the aforementioned circular space. Reliability coefficients of
the RIASEC scales were mostly greater than .80 and were therefore also mostly satisfactory.
Principal components analysis showed that the RIASEC scales had clear circular ordering
and that the correct RIASEC ordering occurred. The randomisation test of hypothesised order
relations revealed that the circulant ordering model fit the RIASEC correlation matrix
appropriately for this sample of adults, thereby providing further evidence for the circular
structure of interests as measured by the MCM. As a whole these results show that Holland’s
model can indeed be extracted from the MCM interest categories. Overall this study builds on
the promising work of earlier research and shows that the MCM is a potentially reliable and
valid measure of adults’ vocational interests. Implications for theory and practice are
presented.