Introduction

Every year, the Spanish Ministry of Health, Social Services and Equality convenes the examination for access to specialized medical health training. The Qualification Commission of the exam is responsible for approving the test questionnaire and withdrawing the questions they consider to be inappropriate. The present work proposes a methodology to help detect questions that might be withdrawn based on the metrics of the Classical Test Theory and the Item Response Theory.

Subjects and methods

For the accomplishment of the present work, the answers given by a total of 13,984 examinees to the questions of the MIR exam have been analyzed, using the exams between 2010 and 2015.

Results

The authors propose the removal of 26 questions which, whilst not having been withdrawn by the Qualification Commission, would be susceptible to withdrawal based on the values of their point-biserial correlation coefficients and on discrimination in the two-parameter logistical model.

Conclusions

There are a number of questions that have not been withdrawn whose poor psychometric quality slightly reduces the overall quality of the MIR test. Given that this test is done in order to classify physicians who want to gain a specialized training place, from the authors’ point of view it would be advisable for the Qualification Commission to withdraw all questions that present poor psychometric quality, even at the cost of making the final number of valid questions slightly fewer than 225.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )