Quality & Quantity 35: 343–349, 2001.
© 2001 Kluwer Academic Publishers. Printed in the Netherlands.
Testing Holland’s Hexagon: Explanation and
T. P. HUTCHINSON and B. MYORS
Department of Psychology, Macquarie University, Sydney, N.S.W. 2109, Australia
Abstract. A description is given of Hubert and Arabie’s test of whether an empirical matrix of
correlations (between the six measures of vocational interests commonly abbreviated R, I, A, S,
E, and C) supports Holland’s hexagonal (circular order) model. The description makes clear that
the test has severe limitations. Speciﬁcally, the circular ordering RIASEC is only compared with
other circular orderings. Consequently, correlation matrices arising from a model that is qualitatively
different may be rated very highly by Hubert and Arabie’s test; several examples of this are given.
Key words: Hubert and Arabie’s test, personality structure, RIASEC hexagon, vocational interests
Assessing whether someone’s personality and an occupation or a speciﬁc job are
compatible is important both to the individual and the employer. In this context,
a theory of six broad personality types proposed by Holland (1992, for example)
is prominent. By means of a personality inventory, subjects may be scored on six
measures: R, I, A, S, E, and C (realistic, investigative, artistic, social, enterprising,
and conventional). According to the theory, the 15 correlations between pairs of
the six measures should fall into three sets: RS, IE, and AC should be the smallest
correlations, RA, IS, AE, SC, ER, and CI should be intermediate, and RI, IA, AS,
SE, EC, and CR should be the largest correlations. (For brevity, RS has been written
for the correlation r
, and so on.) We can picture R, I, A, S, E, and C (in that
order) as being the six corners of a regular hexagon – the ﬁrst set of correlations
is smallest because each pair of personality types consists of types that occupy
opposite corners of the hexagon, the second set of correlations is intermediate in
size because each pair consists of types separated by one other on the hexagon,
and the last set of correlations is largest because each pair consists of types that
are adjacent on the hexagon (the two types are relatively similar). Table I(a) is an
example of the pattern that is hypothesised.
A researcher who has obtained scores on R, I, A, S, E, and C from a number of
subjects or clients, and has calculated the 15 correlations between pairs of these,
may then be interested in testing for the presence of Holland’s hexagon in the mat-