ABSTRACT
For many years, multivariate statistical techniques have simplified the analysis of complex sets of
data. As a collective group, these techniques enable researchers, evaluators, policy analysts, and others to
analyze data sets that include numerous independent variables (IVs) and dependent variables (DVs). In
other words, they allow researchers to analyze data sets where the participants have been described by
several demographic variables and also have been measured on a variety of outcome variables. For instance,
a researcher may want to compare the effectiveness of four alternative approaches to reading instruction on
measures of reading comprehension, word recognition, and vocabulary, while controlling for initial reading
ability. The most appropriate method of analyzing these data is to examine the relationships and potential
interactions between all variables simultaneously. Relying on univariate statistical procedures would pre-
vent proper examination of these data. Due to the increasingly complex nature of research questions in the
social sciences, and to the advent-and continued refinement-of computer analysis programs (e.g., SPSS®, SAS®, BMDP®, and SYSTAT®), the results of multivariate analyses are appearing more and more
frequently in academic journals.