(hierarchy, hexagon, or circumplex), three-dimensional (sphere),
or hierarchical (Gati, 1991).
These studies by Rounds and Tracey, along with others in the
last five years by Hansen, Collins, Swanson, and Fouad (1993),
Khan, Alvi, and Kirkwood (1990), Prediger, Swaney, and Mau
(1993), Care (1996), and Boyle and Farris (1992), confirm the exis-
tence of and efficiency of a circular structure with occupations
ordered in the manner Holland and colleagues (1993) described.
The structure appears in studies of women and men and has regular
distances between and among types that are characteristic of a cir-
cumplex. It is possible that a three-dimensional sphere (Tracey &
Rounds, 1996) may provide a reasonable approximation of the
structure of interests and would be useful if clients and profession-
als could understand and use the spherical structure. There is some
question about the shape of the circumplex across cultures; appar-
ently, metanalysis does not reveal the circumplex structure in non-
Western cultures. Although there may be problems in sampling,
cross-cultural measurement, and translation, this finding, if repli-
cated, would be of concern to all of us.
Validity of the Theory. Are the theory—its constructs, measures,
and interventions—valid outside of Western cultures and subcul-
tures? Social science models and theories, because of their context-
sensitivity, are particularly sensitive to cultural bias. Thus we should
not presume a theory to be culturally valid or useful until research
results support that assumption. Conversely, we should not con-
clude that a theory is invalid unless evidence concerning its inva-
lidity is presented. Unfortunately, cross-cultural validity is much
more difficult to establish than is generally presumed (Spokane,
1986). Although the SDS, the VPI, and the SII have each been
translated into multiple languages (Shaowen & Yaoxian, 1999),
and cross-cultural studies are available, these studies are not with-
out problems. Samples are rarely comparable across cultures, mea-
sures may not translate well, and even the basic constructs and
HOLLAND’S THEORY 405