In fact, two common errors seem to be associated with the failure to accurately
discriminate between univariate and multivariate approaches toward data analysis.
First, many researchers believe that conducting a MANOVA will provide
protection from Type I error inflation when conducting multiple univariate
ANOVAs. Following this erroneous reasoning, for instance, we would first
conduct a MANOVA for the personality data above and, if significant, judge the
statistical significance of the univariate ANOVAs based on their unadjusted
observed p-values rather than their Bonferroni-adjusted p-values. Although such
an analysis strategy is common in the literature, it is not to be recommended
because the Type I error rate will only be properly controlled when the null
hypothesis is true (Bray & Maxwell, 1982), which is an unlikely occurrence in
practice and therefore an unrealistic assumption. Type I error inflation can be
controlled through the use of a Bonferroni adjustment or a fully post hoc critical
value derived from the results of a MANOVA, but the researcher must make the
extra effort to compute the critical values against which to judge each univariate
F-test (see Harris, 2001, and below). To reiterate, simply running a MANOVA
prior to multiple ANOVAs will not generally provide appropriate protection
against Type I error inflation. The extra step of computing the Bonferroni-adjusted
critical values or the special MANOVA-based post hoc critical value must also
be taken.