Seventy-three students learned central concepts about signal detection theory with an introductory text followed by a computer simulation. Each learner was randomly assigned to one of the three treatment groups: text boxes, scrollbars or drag and drop opportunities to modify parameters in the simulation. These three interactive elements were used for linking the different multiple representations in the simulation dynamically. Retention and transfer test, evaluation of the instructional material as well as time spent with the simulation were used as dependent measures. Learners who had to use scrollbars or had to modify parameters per drag and drop performed better on transfer, but not on retention, than did learners who had to use text boxes. Learners who had to modify parameters per drag and drop also evaluated the instructional material better than did learners who had to use text boxes. These results are consistent with the split-attention principle. Copyright © 2009 John Wiley & Sons, Ltd.