Pan, Lovelett, Phun, and Rickard (2019) Systematic and random interleaving for second language learning (preprint)

Abstract

Repeatedly switching between a series of to-be-learned topics, also called interleaved practice (interleaving), can improve learning over traditional, one-topic-at-a-time blocked practice (blocking). We investigated whether interleaving’s benefits for foreign language learning are facilitated by random schedules, wherein training trials follow unpredictable patterns, or systematically alternating schedules, wherein trials are predictably sorted. Students learned to conjugate Spanish verbs in the preterite and imperfect tenses and then took a 48-hr. delayed verb conjugation test. A consistently random (Experiment 1) or systematically alternating schedule (Experiment 2) did not improve learning versus blocking. However, the combination of both types – systematic alternation for study trials and randomization for practice trials – enhanced learning (Experiments 3-4). Thus, neither interleaving schedule alone appears to be sufficient; for verb conjugation skills and likely other materials involving study followed by problem-solving practice, both are needed. Interleaving’s benefits are therefore impacted by the alignment between training schedule and task type.

The synergistic benefits of systematic and random interleaving for second language grammar learning.

“…Additionally, if training involves randomized practice problems and especially if there is a constraint against any problem category appearing multiple times in succession, then (c) a strategy of reusing solutions across successive trials is rendered ineffective, as is (d) the use of any feedback from a given trial as a hint for the next (Lee & Simon, 2004). Consequently, randomly interleaved practice exercises require more memory retrieval and may yield better learning (Carpenter & Mueller, 2013; Pan, 2015; Rohrer, 2012). However, some researchers have also suggested that randomization can be too confusing and generates cognitive overload, particularly in novice learners (Wulf & Shea, 2002).…”