Since our purpose in using factor analysis was strictly descriptive we did not think it necessary to employ some of the more sophisticated mathematical techniques which may be used to limit the number of factors. Rather, based on our judgments as to the meaningfulness, interpretability, and comparability of factors in all four voter data sets, and generally following the guidelines set out by Rummel (in Applied Factor Analysis, Northwestern University Press, 1970, chapter 15) we chose to limit the number of factors in our analysis to ten. Factors beyond ten were substantively meaningless or contained only one or two variables loading above .4 and contributed little to the amount of variance explained by the factor analysis.
The SPSS factor analysis program was employed using Varimax factor rotation with the PA1 method of factor extraction. The following table presents the percent of variance explained by the first ten factors in each state.
Presented below are the complete factor matrices for each voter sample. Variables are identified by question number and the exact wording of the questions appears in Appendix A. Variables loading about .4 in the tables are identified by an asterisk.