The Personal Structure of Personal Need for Structure

Abstract

Abstract. This study investigated the responses of N = 1,789 participants to a set of 12 Likert-type items for the assessment of personal need for structure (PNS). Mixture-distribution Rasch models were used to analyze the homogeneity of the response format across items and the homogeneity of the item parameters and category parameters across persons. Model selection yielded a two-class rating scale model as the favorite model. This model contains the assumptions that the Likert response scale is used in a constant way for all items but that the item or category parameters differ between two latent subpopulations. The parameter estimates revealed large differences in the threshold parameters for the response categories between the two subpopulations. While the larger subpopulation showed a tendency to avoid extreme response categories, the smaller subpopulation used the whole range of the response scale. The different response styles identified by the mixture-distribution Rasch analysis were validated by significantly higher Extraversion scores for participants in the smaller subpopulation that showed more extreme and impulsive rating behavior. The results confirmed that PNS reflects quantitative interindividual differences, and they also showed that the total score of the 12 PNS items forms a combination of the latent PNS trait and response style.