Mixed Practice Help

When Does Mixed Practice Help?
The study I worked on with Marta Mielicki was based off the published study, The Shuffling of Mathematics Problems Improves Learning, by Dough Rohrer and Kelli Taylor of the University of South Florida. The purpose was to study the way that mathematics problems are distributed or organized affects the learning of new material. The researchers from the article observed that most math textbooks tend to follow a standard “blocked” structure: a lesson is explained along with several worked examples, then students are given practice problems to test on the subject immediately after. However, some textbooks have used what they call an “interleaving” distribution. All the lessons for a certain chapter or topic are given

Subjects were given problems based on finding the volume of certain geometric prisms; spheroids, spherical cones, wedges, and half cones. Participants were split into two conditions, “blocked” and “mixed.” The Blockers would see a worked example for one type of shape, followed by practice problems corresponding to the same shape. Mixers would receive examples for various prisms sequentially, then be given a mixed grouping of practice problems. After one week, both groups would receive an eight-question test of mixed problems. In the paper’s study, both groups would have two practice sessions, spaced a week apart, with the test following another week-long interval. Mielicki’s study only had one practice

Their practice performance put Mixers at 60% accuracy and Blockers at 89% accuracy. Test performance resulted in 63% accuracy for the Mixers and a low 20% accuracy for Blockers. In Mielicki’s replication experiment (Experiment 1), those results were confirmed. Mixers had 51% of accuracy on the practice, while the Blockers had 85%. In the test portion, Blockers had only 9% of accuracy, while Mixers showed 29%. Although Mixers did better than the Blockers, neither group performed particularly well. Yet in Experiment 2, with the permutation problems, it was demonstrated that the mixer group did worse in both the practice and testing sessions. The practice portion yielded 41% accuracy for the Mixers and 76% percent for the Blockers, and the test accuracy was 38% for the Mixers and 46% for the Blocked. It was suggested that the benefits of the interleaved system might only be attainable if participants reached a certain level of understanding of the coursework during