Practical revision tests, which are common in classrooms, have been widely considered to be an effective tool for facilitating learning. However, students and teachers have expressed the legitimate concern that carrying them out can generate anxiety during assessment - compromising academic performance - which would explain why they are not widespread and why it is recommended to minimize their use in teaching contexts.
A meta-analysis recently published in the Educational Psychology Review explored this question. More specifically, the authors set out to determine the effect of these practice tests on so-called test anxiety and to explore potential moderating factors of this effect. They evaluated five possible factors: research setting, type of anxiety measurement, test format, performance and publication status. It is worth noting that most studies on the subject had mainly aimed to assess the effects of practice tests on learning and academic performance, but not on anxiety. The question of factors was also still little explored (or even unexplored). After a systematic literature search, the research team gathered 24 eligible studies, from which 25 effects were extracted, based on data from 3,374 participants.
The results showed strong statistical indications that practice tests can reduce anxiety about assessments, with a significant average effect. To explain this effect, the study proposed two hypotheses. On the one hand, that of improved learning, according to which tests reduce anxiety thanks to improved student performance. More specifically, the motivation generated by the test leads students to study more and prepare better for future assessments, reducing worry and therefore anxiety. On the other hand, the hypothesis of reduced uncertainty. Since tests inform students about the content that will be assessed, the format in which the assessment will be carried out and the difficulty of the test, they can reduce students' uncertainty about upcoming exams and alleviate anxiety. The results of all the analyses to detect publication bias - the tendency to publish positive, significant or unpublished research results - also showed that the included studies only suffered minimally from this bias. And easy practice tests tended to be more effective at reducing anxiety than difficult ones, but less effective at improving learning.
These findings support the recommendation that teachers include practice tests in their subjects. However, teachers also need to recognize that tests are per se an anxiety-producing learning activity, so they should find ways to mitigate it. To this end, previous studies have recommended that teachers make tests ungraded and gamified, while still allowing students the possibility of retaking them several times.
The small number of studies in this area is one of the major limitations of this meta-analysis, as is the small set of effects included, which compromises the statistical power of the analysis of moderating factors. In fact, of the five potential factors evaluated, only test performance had a significant effect on anxiety. Future studies should also investigate important potential factors such as corrective feedback and explore individual differences between students that influence the effectiveness of practice tests in reducing anxiety.