The accuracy of the E, N and P trait estimates: an empirical study using the EPQ-R

This study analyzed the responses of 1175 university students to the E, N and P EPQ-R scales and used Item Response Theory (IRT) (1) to estimate the `trueʹ distribution of the E, N and P traits in this population, and (2) to assess the measurement accuracy of the E, N and P scale scores over a range of trait values. For (1), the results suggested that the distributions of the three traits are unimodal and roughly bell-shaped. The estimated distributions for E and N tended to be symmetrical, but the distribution of P was clearly asymmetrical. These results were related to previous studies based on direct scores. For (2) the results suggested that the precision of the E and N scores is acceptable across a wide range of trait values. However, the P scores provide poor measurement for a large number of subjects in a non-pathological population. At the points at which accuracy is maximal, E, N and P scores are reasonably precise according to IRT standards. However, there is a high degree of uncertainty in the individual estimates which must be taken into account when interpreting test scores.