The accuracy of EAP method based on the summed score in IRT ability parameter estimation was
investigated through a simulation study. Also, the results were compared with those by the traditional
EAP method based on the item-response pattern. The former has an advantage in that it can produce
more convincing scale scores to the public, by providing ability estimates which have one-to-one
correspondence with the number-correct scores. At the same time, the former seems to have a problem
ignoring the information regarding item-response patterns, then the accuracy in recovering true parameters
can be lower than that of the latter. In this study, the performances of the two EAP methods were
compared at the contexts of various simulation conditions. Under the use of the 1PM, the difference
between both methods barely could be found. As the models became more complicated, however, the
latter tended to show better recovery results. Nevertheless, such differences between the two methods in
terms of estimation accuracy were so insignificant. This kind of finding is implying that the EAP method
based on the summed score can be a promising alternative under the operation of an actual testing
program.