Construction and analysis by the Rasch model of two computerized recognition memory tests

The objectives of this study were to develop items for a computerized version of an associative recognition memory test and to develop items for the construction and computerization of an individual recognition memory test, using the Rasch model. This study also aimed to test the hypothesis related to the item difficulty levels and to search for validity evidences for the instruments investigating if there are differences between male and female performance. Two hundred and seventeen college students, male and female, with average age of 25 years, participated in the study. Results indicated that the tests fit the Rasch model, presented adequate reliability index and did not present differential item functioning (DIF). The expected difficulty order of item categories from the individual subtest as well as a large correlation between the pairs expected difficulty order and the pairs obtained order were found. Women presented significantly higher performance than men and it is not due to DIF.