🚀 go-pugleaf

John B. Collins writes.....
>
> The issue of phi coefficients in factoring dichotomous data is a RED HERRING.
> If the data are dichotomous, SAS, SPSS and other packages will yield a matrix
> of PEARSON correlations, NOT phi coefficients.  True, extreme p values will
> skew the data, but that's because of the DATA, not the correlation matrix.
> You'll know if this is a problem in your particular data set if you get
> Heywood cases (loadings greater than 1.00).  But until this becomes an
> actual problem, go ahead and factor.  The issue of scaled loadings reflecting
> differential "skill levels" calls into question when in the research sequence
> the factor analysis is performed. IRT (Hambleton et al) deal effectively with
> this question.   John
>

        One of the concerns with factor analysis on dichotomous data
that I haven't seen brought up yet is the stability of the
correlations.  We must remember that, as with any experiment, 'our'
data set is not THE data set, but rather one of many that could and
should be collected.  If the results of our experiment can not be
reproduced by others then they are not useful (see the kitchen sink
cold fusion example.  :-).

        I have not done anything myself to investigate this.  Are
there any reliable references for this kind of problem?  How unstable
are the correlations and therefore the factors?

charles d. kincaid              STATS:  (904) 392-1941x206
statman@stat.ufl.edu            CIRCA:  (904) 392-2007
NeXTMail much preferred         Statistical Shoe Salesman:: "I think
STOP BENDIN' THE SHAFTS!!       that a t-test would fit you better, ma'am"