New methods for mapping response patterns
Tommaso Lando, Lucio Bertoli Barsotti · Aisberg (University of Bergamo) · 2013
In the Rasch model, the probability of obtaining a particular item score is explained by the ability of a person and a characteristic of the item – so called item difficulty. It is well known that if a person answers according to the Rasch model, his response pattern is expected to follow a scheme similar to a Guttman pattern. Besides, unexpected (“aberrant”) response patterns reveal a certain amount of lack of fit, at the individual level. The aim of this paper is to provide a description of an individual's performance on a set of items in a graphical –that is, number free– two-dimensional format. The two dimensions can be interpreted as ability (that is, overall test performance) versus non-ability, and aberrancy versus non-aberrancy (to be intended as a measure of the similarity with respect to the Guttman pattern). In the present study we make use the recently available VOSviewer mapping software. Two different similarity measures are presented and compared. The empirical analysis confirms the effectiveness of the proposed application of mapping to item response patterns.