1: Actually, our first question is
usually whether the initial EFA got it right in asserting there were
several “first-order” factors, or whether the analysis
should have concluded there were fewer—or one—factors.
[return]
2: This sort
of question can spawn endless debate among scholars, and the literature
is replete with examples of this type of debate. In our opinion, these
(often vitriolic) debates fester because of the exploratory, volatile,
and nonreplicable nature of these analyses. If authors would quickly
replicate or (ideally) move to confirmatory analyses, these debates
would be less apt to erupt as there are clear ways to test competing
hypotheses.
[return]
3: We tend to
recommend Barbara Byrne’s excellent reference on structural
equation modeling (Byrne, 2010) for readers interested in CFA/SEM
and higher-order factor analysis in a confirmatory context.
[return]
4: This is
not, in our mind, a reason to decide to use PCA, by the way.
[return]