Classification of frequency distributions of diagnostic criteria scores in twelve personality disorders by the curve fitting method

Abstract The purpose of the present study was to investigate whether each personality disorder (PD) has a different frequency distribution of diagnostic criteria, and to classify the distribution of each PD on the basis of shape by the curve fitting method. A total of 4740 male subjects in early adulthood completed the Korean version of the Personality Diagnostic Questionnaire (4th revision; PDQ-4+). The frequency distribution of scores in each PD was specified to the best fitted model by the curve estimation procedure. Twelve distributions of PD criteria were fitted to three types of curve models (quadratic, cubic, and logistic). The quadratic model included obsessive–compulsive and histrionic PD. The cubic model included borderline, narcissistic, avoidant, passive–aggressive, paranoid, and schizotypal PD. The logistic model contained dependent, depressive, schizoid, and antisocial PD. The results suggest that each PD has a different type of distribution and each distribution of PD might be classified to a specific curve model. Also, the results suggest that the issue of continuities between PD and the normal group could not be generalized to all PD, but should be considered individually for each PD.