When people read, hear, or prepare research summaries,
they sometimes have misconceptions about what is or isn't "sound
practice" regarding the collection, analysis, and interpretation
of data. Here are some of these common (and dangerous) misconceptions
associated with the content of Chapter 13.
A two-way ANOVA is better than a one-way ANOVA.
A 2x2x2 ANOVA is a two-way ANOVA.
If there are two independent variables (each having
2 levels) involved in the study, it would be wrong for the researcher
to subject the data to a one-way ANOVA having 4 levels.
The main effect F-values should be examined before
the interaction F-value.
If the interaction from a 2x2 ANOVA turns out to
be significant, and if the four cell means are equal to 20, 20, 20,
and 40, then we can point to the cell having the mean of 40 and say
that it "caused" the interaction to be significant.
If you're given a graph of the cell means from a
two-way ANOVA, you can look at it and determine (with a high degree
of confidence) whether or not the interaction is significant.
A good researcher will not make any comparisons among
the cell means unless the interaction has turned out significant.
A 3x3 ANOVA deserves more respect than a 2x2 ANOVA.
The findings relative to a main effect are predictive
of what would have been discovered if a different and simpler study
had been conducted within which only one independent variable existed,
with the levels of that particular factor compared by a one-way ANOVA.