Thanks to everyone who responded to my two queries. For recommendations on
multivariate stats texts: Pamela Warner, Alison Fisher, Stanley Wasserman,
and Clare Morris. Two recommended Tabachnik and two recommended other
books. I will do a chi-square test on this result in the morning. :-)
Those who responded to my query on the cross-tabulation problem were, again
in no particular order, were Cor Stolk, Phil McShane, Art Kendall, Todd
Melander, Cliff Lunneborg, Ken Lakhani, and David Smith. The consensus was
for loglinear analysis, which I am equipped to do with SPSS. However, I am
leaning to a two-step approach in which I will do a contingency table
separately for each of the 12 or so dichotomous variables (behavioral
symptom checklist items) against the type of abuse category assigned to the
case. This will give me a smaller set of variables to enter into the
loglinear model, presumably.
Thanks to all,
David Klein
|