Jl. Horn, Comments on integrating person-centered and variable-centered research on problems associated with the use of alcohol, ALC CLIN EX, 24(6), 2000, pp. 924-930
For any sample of m different variables obtained on a sample of n different
persons, there is an identity transformation between classifications of pe
rsons (types) that are defined with person-centered methods and classificat
ions of variables (factors) defined with variable-centered methods: types a
re manifested in factors and factors are expressed in types. Both kinds of
methods analyze the variability-the same variability-in a persons-by-variab
les data matrix Person-centered types can be transformed into variable-cent
ered factors. Thus, at a basic level there is a basis for integrating perso
n-centered and variable-centered research.
But there are many different ways to implement person-centered methods, jus
t as there are many different ways to resolve variables into factors, and g
enerally any model far analysis directed at identifying types will not be t
he complement of a model selected for analysis leading to factors. Thus, wh
ile the results from the two kinds of analysis can be related to one anothe
r, the results they produce in applications can be quite different. Also, i
n practice, one method is used before the other. For example, person-center
ed methods might be used to form types, which are then studied with multipl
e-group, variable-centered methods to test hypotheses specifying invariance
or differences of relationships across types. There are good reasons to th
ink in terms of types as one contemplates analysis in terms of variables. B
ut the reverse also is true.
In typological analysis, it must be assumed that types exist and that sampl
es of subjects and indicator variables are drawn in ways that can reveal th
is. Different typologies indicated by person-centered research are most con
cretely compared if the different studies use comparable sets of persons, i
ndicator variables, and other variables that can indicate correlates.
Strictly speaking, none of these sets were comparable in the three substant
ive studies reviewed here. Nevertheless, at an abstract level it was possib
le to see that the results of the three studies agreed in indicating that t
here is a type of adolescent and young adult that, by and large, does not h
ave- or at least does not report having-problems associated with the use of
alcohol. This type is in the majority, accounting for approximately 60% of
the youth sampled in the reviewed studies. The studies agree also in sugge
sting that types may emerge along a continuum of maladjustment characterize
d by increasing use of alcohol and other drugs, failures in school, problem
s at home, interpersonal difficulties, delinquency, and legal problems. Var
iable-centered dynamic analyses might further describe phenomena of this ki
nd.