Chunking to Reduce Working Memory Load¶

Strategy · All strategies
Evidence · 1 claim (1 for) · 3 studies (2 review, 1 causal), q2–q3 · 0 of 3 report an effect size

Description¶

Working memory can hold only a small number of discrete items at once — classically estimated at seven plus or minus two Miller, 1956 [+S], with more recent estimates closer to three to five for novel material Cowan, 2001 [+S]. Chunking recodes individual elements into larger, meaningful units so that a single chunk occupies one slot regardless of how many elements it contains. Because chunks are stored in long-term memory and retrieved as units, chunking effectively expands functional capacity without changing the underlying limit Gobet et al., 2001 [+S].

Design Implications¶

Chunking is a core mechanism of Cognitive Load Management Chunking reduces working memory load by grouping information into fewer, more meaningful units. [+S]: presenting material in coherent, meaningful groups keeps intrinsic load within working memory limits, whereas unstructured streams of elements overwhelm learners and force premature forgetting Sweller et al., 2011 [+S]. The effect depends on the chunk being meaningful to the learner — arbitrary groupings of unrelated items do not reduce load, because the learner must still hold each element separately.

Context¶

Requirements¶

• Content organized around meaningful boundaries (semantic categories, procedural steps, rule-based groupings) rather than arbitrary length limits
• Learners with sufficient prior knowledge to recognize the grouping structure — or explicit instruction in the grouping scheme
• Segmenting of continuous media (video, animation) at natural conceptual boundaries

Constraints¶

• Chunks must map to the learner's existing knowledge; a grouping that is meaningful to an expert may be several independent elements to a novice [~S]
• Over-chunking or chunking at too coarse a grain can obscure the relationships between chunks, harming integration of the whole [~M]
• Chunking reduces load but does not eliminate it; if element interactivity is inherently high, additional Cognitive Load Reduction measures are needed [-W]
• Expertise reversal: highly chunked, segmented presentation becomes redundant and slows learners who have already formed the chunks [~S]

Implementation Variability¶

• Segmenting multimedia into learner-paced parts (e.g., Khan Academy videos, Mayer's segmenting principle)
• Categorical grouping of lists or facts into semantic clusters
• Procedural chunking of multi-step instructions into sub-goals, as in Part-Task Practice
• Schema-based chunking where learners are taught the expert's grouping structure directly

Target Learners¶

• Novices, who lack the long-term memory structures to chunk spontaneously and must be given the grouping Sweller et al., 2011 [+S]
• Learners with low working memory capacity benefit disproportionately from external chunking [~M]
• Experts already chunk automatically; imposed chunking can interfere with their fluent processing [~S]

Target Learning Goals¶

• Retention of multi-item factual content (vocabulary, lists, terminology)
• Procedural learning where steps must be held in mind simultaneously
• Schema construction: building larger knowledge units that themselves serve as future chunks

Instructions¶

1. Analyze the material for natural boundaries. Identify where the content divides on conceptual, procedural, or categorical lines; chunk on those seams rather than on a fixed item count (Advance Organizers).
2. Size each chunk to novel-element count, not length. Count how many genuinely new, interacting elements a chunk contains — three to five for unfamiliar material — rather than how many minutes or slides it occupies.
3. Name the chunk. Give each unit an explicit label, so the group can be retrieved as one item instead of reconstructed element by element.
4. Present chunks one at a time. Segment continuous media and multi-step procedures at the boundaries identified in step 1, and let learners pace the transitions where the medium allows (Part-Task Practice).
5. Practise within, then across. Have learners work each chunk to fluency before requiring them to integrate chunks, so integration draws on consolidated units rather than loose elements (Spaced Repetition).
6. Make the relationships between chunks explicit. Once the parts are secure, show how they compose into the whole — chunking that never reassembles leaves learners with fragments.
7. Fade the chunking as expertise grows. Withdraw the imposed segmentation once learners chunk the material themselves, since retaining it becomes redundant and slows them down.

Related Strategies¶

• Chunking Content — the same strategy stated in delivery terms: segmenting lessons, videos, and texts into coherent units
• Segmenting — chunking applied to continuous media
• Advance Organizers — supply the framework that makes chunks meaningful
• Spaced Repetition — consolidates chunks into long-term memory so they free working memory capacity

Examples¶

• Phone numbers are grouped into 3–4 digit segments precisely because ungrouped 10-digit strings exceed working memory capacity.
• Chess expertise: master players recall board positions almost perfectly by chunking them into familiar patterns, but show no advantage for randomly arranged pieces — evidence that chunking depends on stored patterns, not raw capacity Gobet et al., 2001 [+S].
• Khan Academy videos are short and single-concept, segmenting instruction so each segment fits within working memory limits.

Key Sources¶

• Miller, G. A. (1956). The magical number seven, plus or minus two: Some limits on our capacity for processing information. Psychological Review, 63(2), 81–97. doi:10.1037/h0043158
• Cowan, N. (2001). The magical number 4 in short-term memory: A reconsideration of mental storage capacity. Behavioral and Brain Sciences, 24(1), 87–114. doi:10.1017/S0140525X01003922
• Gobet, F., Lane, P. C. R., Croker, S., Cheng, P. C.-H., Jones, G., Oliver, I., & Pine, J. M. (2001). Chunking mechanisms in human learning. Trends in Cognitive Sciences, 5(6), 236–243. doi:10.1016/s1364-6613(00)01662-4
• Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive Load Theory. Springer. doi:10.1007/978-1-4419-8126-4