This study investigated whether design guidelines for computer-based learning can be applied to computer-based testing (CBT). Twenty-two students completed a CBT exam with half of the questions presented in a split-screen format that was analogous to the original paper-and-pencil version and half in an integrated format. Results show that students attended to all information in the integrated format while ignoring information in the split format. Interestingly, and contrary to expectations, they worked more efficiently in the split format. A content analysis of the ignored information revealed that it was mostly not relevant to answering the questions, unnecessarily taxed students' cognitive capacity and inefficiently increased the mental effort they expended. Further comparisons of different mental effort measures indicate that mental effort had an explicit (ie, self-reports, explicit utterances) and an implicit component (ie, silent pauses in thinking-aloud, eye tracking parameters). Consequently, when designing CBT environments, not only the design of the tasks but also the content of the given information and their effect on the different aspects of mental effort must be considered.