Influences of deep learning, need for cognition and preparation time on open- and closed-book test performance

Abstract

The ability to master discipline-specific knowledge is one of the competencies medical students must acquire. In this context, 'mastering' means being able to recall and apply knowledge. A way to assess this competency is to use both open- and closed-book tests. Student performance on both tests can be influenced by the way the student processes information. Deep information processing is expected to influence performance positively. The personal preferences of students in relation to how they process information in general (i.e. their level of need for cognition) may also be of importance. In this study, we examined the inter-relatedness of deep learning, need for cognition and preparation time, and scores on open- and closed-book tests. This study was conducted at the University Medical Centre Groningen. Participants were Year 2 students (n = 423). They were asked to complete a questionnaire on deep information processing, a scale for need for cognition on a questionnaire on intellectualism and, additionally, to write down the time they spent on test preparation. We related these measures to the students' scores on two tests, both consisting of open- and closed-book components and used structural equation modelling to analyse the data. Both questionnaires were completed by 239 students (57%). The results showed that need for cognition positively influenced both open- and closed-book test scores (beta-coefficients 0.05 and 0.11, respectively). Furthermore, study outcomes measured by open-book tests predicted closed-book test results better than the other way around (beta-coefficients 0.72 and 0.11, respectively). Students with a high need for cognition performed better on open- as well as closed-book tests. Deep learning did not influence their performance. Adding open-book tests to the regularly used closed-book tests seems to improve the recall of knowledge that has to be known by heart. Need for cognition may provide a valuable addition to existing theories on learning.