Abstract
In the present study, the robustness of the Emotional Intelligence Quotient (EQ) Test was examined against response bias, using paired comparison methods expressed by structural equation modeling. The EQ test was then standardized. In Experiment 1, paired-comparison and Likert-type EQ tests were administered to 140 participants in recruitment and non-recruitment situations, in order to evaluate the robustness of the test. Standard deviations, effect size, and convergent validity showed that in both situations, the paired comparison test was more robust against response bias than the Likert-type test. In Experiment 2, in order to stabilize the scoring, data obtained from another set of 383 participants were added to the data from Experiment 1. Items on the test were then divided into 10 subsets, yielding parallel tests, and the correlation of scales among those tests was calculated. The parallel tests were showed to have concurrent validity.
