What multivariate techniques can handle ordinal and nominal data?
Differentiate by dependent vs. non-dependent procedures. Dependent: Discriminant function analysis when the dependent is dichotomous (there is a trichotomous variant). ANOVA when the independents are nominal or ordinal. ANCOVA when some independents are interval and some are nominal or ordinal. Non-dependent Cluster analysis (of units or of variables) Loglinear analysis (Leo Goodman) Homals, Princals, Canals (Gifi group) 14. What is the distinction between exploratory and confirmatory analysis? Answer: In confirmatory analyses, a priori hypotheses are being tested, whereas in exploratory analysis a posteriori hypotheses may be promulgated. Confirmatory research is considered more scientific by Tacq and many others. 15. What is the distinction between R factor analysis and Q factor analysis? Answer: Most forms of statistical analysis, including most factor analyses, focus on variables. That is, most factor analysis is directed at reducing a larger number of manifest variables to a small
