Mathematical analysis of content

Insofar as mathematical approaches to content analysis may be a fruitful extension of content analysis in general, further research into the use of methods involving inferred categories cannot be excluded. A preference for a priori categorization is premature. There are several arguments in favor of methods in the field of computer-assited content analysis which may be termed nondictionary computer-assisted content-analysis programs, such as Iker's WORDS program. This approach may be based on an open, unrestricted theory of language, on direct multidimensional scaling of co-frequent units of observation (i.e., of words used in texts to be analyzed), and on criteria for the fit of multidimensional models which may be deduced from the stability or instability of the stress between groups of texts to be compared.