1989 年 16 巻 26 号 p. 31-46
Many contingency tables have ordered response categories. This paper compares two major approaches to ordinal response categories, namely the point representation and the block (interval) representation. Specifically, two methods are compared, each representing each of the two approaches. One is ideal point discriminant analysis (IPDA) and the other the successive categories method (SCM). Similarities and distinctions between the two methods are explicated. The goodness-of-fit (GOF) is compared through AIC using several example data sets. IPDA and SCM were found to provide similar GOF, but IPDA was found to provide a slightly better fit in all the data sets examined.
