This paper addresses itself to methods of analysis forcategorical variables of the sort utilized in attitude and human behaviour research. It recommends the adoption of a technique which has been successfully applied to epidemiological, clinical investigation, laboratory and microbiological data. It is known as ridit analysis. After reviewing some general attitude scaling methods and problems of analysis related to them, the ridit method is described. Application of ridit analysis to a recent study undertaken to assess health care service quality in Western North Carolina is presented. The technique is conceptually and computationally simpler than other conventional statistical methods. It is also a distribution free method. Basic requirements and limitations on its use are also indicated.
1.
Introduction:
Measuring attitudes has been the subject matter of many researches in a variety of disciplines such as sociology, management, and political sciences. Attitude measurement generally means finding out where a particular attitude lies along a scale ranging from extreme favourableness to extreme unfavourableness. This is usually carried out through sample surveys where a collection of questions (statements) covering the various aspects of certain attitude are directed to every member of the selected sample. Responses to these questions (statements) are then appropriately scaled and combined into a meaningful whole defining the measured attitude. The applied scales may be of nominal, ordinal, interval, or ratio nature depending on the degree of sophistication the researcher wishes to introduce in his
measurement, Moser and Kalton (1975).