Do it twice! Test-taking fosters repeated but not initial study of multimedia instruction

Are practice tests only helpful when having studied before taking them? We investigated this question in a multimedia learning scenario. Participants (N = 85) were randomly assigned to one of two groups, in which they did either not study or studied before taking a test. Afterwards, all participants (re-)studied with the same materials, and took the same test again. Participants had generally higher scores in the second than in the first test; however, taking the first test without having studied before did not improve subsequent learning. Only with repeated study and test-taking, performance in the first test predicted subsequent mental effort. Hence, test-taking fostered repeated but not initial study of multimedia materials, presumably because performance in the first test informed about the effectiveness of the initial study to which learners could adapt during restudy (cf. indirect testing effect). Knowing the test (high test expectancy) alone was not sufficient to foster multimedia learning.