The direction of a factor is arbitrary, if all of
the items load negatively just ignore the sign.
What did the authors use for a stopping rule? do they have many
trivial factors?
Substantively, how do the items with negative loadings go with
items with positive loadings? Was this a special groups of
respondents, so that the content would have some dialect (e.g.,
teen-speak difference in meaning)?
If negative loadings were very occasional was the the data entry
verified?
Art Kendall
Social Research Consultants
6:28 AM, Antti Kärnä wrote:
Hi all,
In an exploratory factor analysis of the WISC IV scale I have
recently seen reported strong negative factor loadings (e.g.,
-.70). The expectation naturally is that facets of intelligence
(factors) would be in a positive association with subtest scores
(indicators). The analysts have used the VARIMAX rotation. Are
the negative loadings an anomalous result, or can they somehow
be interpreted just by reversing the sign? I would think that
they are anomalous, but can someone give some possible
explanation for these results? Thanks in advance!
Best wishes,
Antti
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Art Kendall
Social Research Consultants
Locked
