CORRELATIONAL ANALYSIS OF QUALITATIVE DATA.
NAVAL SCHOOL OF AVIATION MEDICINE PENSACOLA FLA
Pagination or Media Count:
The development of computers has made possible the analysis of the complex interrelationships of relatively large numbers of variables. In order to use regression or multiple correlation techniques, the data from these variables must be meaningfully quantified. The raw data in medical and biological studies, however, are often expressed as categories such as diagnoses or in purely qualitative form such as the presence or absence of a symptom. This paper describes a method of converting such categorical or qualitative data into a series of so-called pseudovariables, which permits their inclusion in correlational analysis and prediction studies. Examples illustrate the method as it might be applied to a study of the relationships of occupation and cholesterol level. Extensions of the technique and additional applications are suggested. Author