Many scholars have written
guidelines for extraction of factors, focusing on the strengths of
a particular extraction technique or a particular extraction criterion
(e.g., eigenvalues, scree plots, parallel analysis). Throughout this
chapter we argue that each of these methods provides valuable insight
on the number of factors underlying the data. One method should not
be used as the method to rule them all, and similarly we do not need
consensus among the methods that we do use. It is best to make decisions
based on multiple criteria, and it is essential that those decisions
make theoretical sense. Remember,
EFA is meant to be followed up by CFA—so the decisions that
you make are not set in stone.