Why working memory capacity can change what a lesson does
Working memory is the mental scratchpad that holds information while you process it. If a lesson asks learners to juggle more at once, those with a smaller scratchpad have less room to spare. In a study of 184 five-year-olds, children learned more when the instructional materials carried low intrinsic load, meaning the material itself was not too complex, and learned less when extraneous load was added, meaning unnecessary demands were layered on top [1]. Children with high working memory capacity showed higher learning effectiveness than those with low capacity, and across different teaching plans, learning effectiveness varied depending on the children's working memory capacity [1]. That last point is the key one for this question: the same lesson did not work equally well for everyone, and measuring capacity revealed who was helped and who was not.
A similar pattern appears in older students. Among 509 middle-school students learning Science and Technology, computer-assisted instruction outperformed inquiry-based learning after controlling for prior knowledge and other factors, and the advantage was particularly pronounced for students with higher working memory capacity [5]. The authors frame this as a reason to adapt computer-assisted instruction for students with poorer working memory [5]. In both cases, the instructional method and the learner's capacity interacted: the method's benefit was not a single number but a range that depended on who was learning.
When good design makes working memory capacity matter less
The picture changes when the instruction is built to keep demands manageable. In an experiment with 62 university students watching educational video podcasts about lightning, working memory did not significantly predict learning outcomes and did not moderate the effects of added elaborative or seductive details [3]. There were also no differences across conditions in measured cognitive load or situational interest [3]. The researchers suggest that well-designed multimedia environments may stabilize learning outcomes across differences in working memory capacity [3]. In other words, when narration and visuals are aligned, the lesson may do some of the organizing work for the learner, shrinking the advantage that a larger working memory would otherwise provide.
One detail from that study is worth keeping in view. In free recall, the control condition outperformed the seductive-details condition, but that difference was not significant once working memory was controlled for, and transfer did not differ across conditions [3]. Students in the added-detail conditions also reported feeling overwhelmed even though their cognitive load ratings were similar [3]. So the absence of a working memory effect does not mean the added features were harmless; it means the measured capacity did not explain the differences in this sample.
Capacity can track not just how fast people learn, but how much they keep
Working memory capacity is often treated as a predictor of initial learning. A multi-site classroom experiment on second-language vocabulary found that even when target words were embedded in a highly supportive context with both input and output activities, working memory was positively associated with vocabulary learning outcomes on immediate and delayed posttests [4]. The association remained at the delayed posttest, which the authors say suggests working memory may affect retention as well as learning rate [4]. That matters for the original question because it means a capacity measure can pick up differences that show up later, not only on the day of the lesson.
The studies differ in how directly they manipulated instructional load, and that difference shapes how much weight each result can bear. The preschool study deliberately built six types of materials with different intrinsic and extraneous loads [1], and the multimedia study deliberately added elaborative or seductive details [3]. The vocabulary study, by contrast, placed words in a supportive context rather than varying load as its main manipulation [4], and the computer-assisted versus inquiry comparison varied the instructional method rather than a load parameter [5]. So the strongest evidence that measuring capacity changes the effect of a load manipulation comes from the preschool and multimedia experiments, and they point in opposite directions: capacity mattered in one and not the other.
What a teacher or designer should take from this
The practical reading is not that working memory testing should become routine, but that the value of knowing a learner's capacity depends on how demanding the instruction is. When materials pile on unnecessary load, capacity differences show up as learning differences, as in the preschool study where added extraneous load lowered effectiveness and high-capacity children did better [1]. When materials are well designed, those differences can narrow, as the multimedia experiment suggests [3]. The middle-school result adds a further wrinkle: a method that helped on average helped high-capacity students most, which is a reason to check whether a method's average gain is hiding a gap [5].
What remains genuinely open is the mechanism behind these mixed results. The multimedia authors propose that well-designed environments may stabilize outcomes across capacity differences [3], but that is a suggestion rather than a demonstrated cause. The preschool findings show that capacity and teaching plan jointly shape effectiveness [1], yet they do not establish why some plans are more capacity-sensitive than others. And the vocabulary result suggests capacity may reach into retention [4], which raises the question of whether load manipulations would show their capacity-dependent effects more clearly on delayed tests than on immediate ones. Until studies manipulate load and measure retention together, the safest answer is that measuring working memory can change how a load manipulation looks, but whether it does depends on the instruction being tested.
Sources used in this answer
Impact of cognitive load and working memory on preschoolers’ learning effectiveness
Preschoolers learned more with low intrinsic load materials and less when extraneous load was added; high working memory capacity children showed higher learning effectiveness, and effectiveness varied by capacity across teaching plans.
Evaluation of children’s cognitive load in processing and storage of their spatial working memory
In a spatial working memory task, processing demands and memory load reduced storage and increased subjective load, and different load indicators predicted memory performance under low versus high complexity conditions.
Cognitive Load and Working Memory in Multimedia Video Podcasts: Effects of Elaborative and Seductive Details
Working memory did not significantly predict learning or moderate the effects of added details in video podcasts, and the control condition outperformed the seductive-details condition on free recall but not after controlling for working memory.
The Effect of Working Memory Capacity on Instructed Vocabulary Learning
Working memory capacity was positively associated with second-language vocabulary learning on both immediate and delayed posttests even in a highly supportive classroom context, suggesting an effect on retention as well as learning rate.
Computer-assisted instruction versus inquiry-based learning: The importance of working memory capacity
Computer-assisted instruction was more effective than inquiry-based learning, and the benefit was particularly pronounced for students with higher working memory capacity.