Applying the thurstonian item response approach to force choice scales: the factory structure of the kolb learning styles questionnaire

Scoring and analysis of Forced Choice(FC) scale has caused a great deal of controversy among researchers. Most of the discussions are related to the superficiality of the obtained results due to the ipsative nature of FC scale datas.

The aim of this study was to use the Thurstonian Item Response (T-IRT) approach in scoring and examining the factor structure of FC scales.

The statistical population of this study was students of Allameh Tabataba'i University of Tehran. The T-IRT model was used to score and analyze the data obtained from the Kolb Learning Styles Questionnaire (LSI), which is a Forced Choice approach.

The results of the T-IRT model in examining the factor structure of the LSI showed that the four-factor model fit well with the data and also, although the correlation structure between the factors in the orthogonal bipolar bifactor model was consistent with the proposed Kolb structure; but this model did not fit well with the data. The results of Multinomial logistic regression model in examining the relationship between learning styles of gender and field of study of students showed that the gender variable has no Share in predicting learning styles (CH:4.87, P>0.05); but the model related to the field of study was significant (CH:305.26, P<0.001). Also, the results of beta coefficients and significance of the model showed that students of psychology (Sig=0.01, β=-1.69 ) and law (Sig=0.01, β=1.48 ) have more Divergent and Assimilator learning style.

These findings indicated that the T-IRT approach works well in solving problems related to scoring and analyzing FC scales, and this could expand the use of these scales. Also, the results of studying learning styles indicate that in teaching and education, more attention should be paid to students' learning styles and teaching methods appropriate to their learning styles should be used.