Halim, Z., A. El Bakry, and S. Moussa,
"IDEAL TYPES: AN ALTERNATIVE METHODOLOGICAL APPROACH TO CLASSIFICATION?",
SAJP , vol. 13, issue 3, Sun City, South Africa, pp. 106, 2007.
AbstractBackground. Current Psychiatric classification is guided by a specific methodological approach, the use of operationally defined diagnostic concepts to ensure adequate reliability, the latter being an initial step towards establishing valid diagnostic categories. This approach reflects a particular methodology and philosophy of science: logical empiricism. An alternative approach is the use of diagnostic concepts based on ideal types. Ideal types were originally introduced by Max Weber in the social sciences. Karl Jaspers acknowledged their potential clinical usefulness. More recently, ideal types were proposed as an alternative approach to classification in psychiatry.
This paper attempts to evaluate two different methodological approaches to classification: one that is empirically derived, based on operationally defined concepts, and another which is phenomenologically derived, based on ideal type concepts.
To achieve this purpose, the classification of subtypes of schizophrenia was compared, using the Diagnostic and Statistical Manual of Mental Disorders (third edition revised)(DSM-III-R) and a classification specifically designed to be based on ideal types.
To achieve this purpose, we investigated the reliability and validity of ideal types. Then we investigated the differences between the two approaches in other areas pertinent to classification such as coverage and patterns of discrimination between different subtypes within and across both classifications.
Method. The sample consisted of 94 male subjects from the services at the department of psychiatry, Cairo University. The subjects selected had an equivocal diagnosis of schizophrenia and represented as much as possible the different DSM- III-R subtypes of schizophrenia.
Tools. Each subject was rated by at least two raters. Each rater was provided with a manual describing the ideal types, which were designated types A, B and C, as well as guidelines for rating. The outcome measure for reliability was inter-rater reliability. Discriminant validity was used as an outcome measure of validity for both classifications.
Discriminant validity was assessed using components derived by factor analysis of scores on Positive and Negative Syndrome Scale. It was administered to each subject by an independent rater.
Statistical procedures included: kappa statistic for inter-rater reliability and
discriminant function analysis for discriminant validity. In addition, regression with a dummy dependent variable and analysis of variance were used to investigate further differences between the two classifications. Principle Component analysis was used to identify components and their variance order in the whole sample.
Results. Ideal types demonstrated adequate levels of reliability and validity.
Comparison between the two approaches showed differences in coverage and in the way in which corresponding types relate to the clinical profile of the whole sample.