Variable-oriented empirical research is based on the proposition that populations are homogeneous. Person-oriented research is based on the propositions that (1) distinct subgroups may exist and (2), if they exist, aggregate-level parameters may contradict parameters estimated for groups or individuals. This article first reviews the main tenets of person-oriented research with respect to developmental research. Three criteria of person-oriented research are proposed: (1) data are analyzed under the assumption that they were drawn from more than one population; (2) attempts are made to establish the external validity of groupings; and (3) groupings are interpreted based on theory. Illustrations use sample studies from the literature, for which both variable-oriented and person-oriented results are presented. Two groups of theorems and propositions of personoriented research are presented. The first group is method-centered and posits that parameter estimates can be group-specific. Thus, aggregation can lead to conclusions that describe none of the individuals or groups in a population validly. The second, new group posits that certain data situations are particularly prone to misleading conclusions after aggregation. Finally, the ecological fallacy is discussed as one of the chief reasons why person-oriented research is necessary. Conditions for proper variable- and person-oriented research are outlined.
von Eye, Alexander and Bogat, G. Anne
"Person-Oriented and Variable-Oriented Research:
Concepts, Results, and Development,"
Merrill-Palmer Quarterly: Vol. 52
, Article 3.
Available at: http://digitalcommons.wayne.edu/mpq/vol52/iss3/3