Cognitive Load Estimator

Cognitive load theory rests on an architectural claim: long-term memory is effectively unlimited, working memory is not, and everything new has to pass through the narrow channel. George Miller's 1956 paper put the span at seven plus or minus two chunks; Nelson Cowan's 2001 reanalysis argued the effective limit for novel material is closer to four. The theory splits demand into three parts, and this estimator scores each on a 0 to 10 scale before normalizing the sum. Intrinsic load comes from element interactivity, the number of items that must be held simultaneously because they only make sense in relation to each other, scaled by learner expertise. Extraneous load comes from presentation choices that add processing without adding learning, which is why it is modelled here from instructional complexity and the number of competing media channels. Germane load is the productive effort spent building schemas, and it is treated as inversely related to prior knowledge because a learner with strong schemas already has much of the structure built. The practical value is diagnostic rather than absolute. Only extraneous load is freely reducible: split-attention effects are fixed by physically integrating a diagram with its labels, redundancy effects by deleting narration that duplicates on-screen text, and search costs by replacing early problem-solving with worked examples. Intrinsic load can only be staged, typically by pre-training vocabulary or isolating elements before combining them. Expertise reverses these prescriptions, which is the expertise reversal effect documented by Slava Kalyuga and colleagues: scaffolding that helps a novice becomes redundant processing for an expert. Treat the numeric output as a relative comparison between two versions of the same lesson, not a measurement. Empirical work uses subjective rating scales, notably the nine-point mental effort scale introduced by Fred Paas in 1992, plus secondary-task and pupillometry methods.

John Sweller developed the theory at the University of New South Wales while investigating why conventional problem-solving practice produced weak learning. His experiments in the early 1980s on means-ends analysis showed that students who solved many problems often could not state the underlying rule, because the search process consumed the very capacity needed to build a schema. The consolidated account appeared as Cognitive Load During Problem Solving: Effects on Learning in Cognitive Science in 1988, and the worked example effect it described became the theory's first practical export. Through the 1990s Sweller worked with Paul Chandler on the split-attention and redundancy effects, with Fred Paas on measurement of mental effort, and with Slava Kalyuga on expertise reversal, and the three-way division into intrinsic, extraneous and germane load was formalized in a 1998 Educational Psychology Review article by Sweller, Jeroen van Merrienboer and Paas. Richard Mayer's parallel program at the University of California, Santa Barbara, built the cognitive theory of multimedia learning on Alan Baddeley's dual-channel model of working memory, yielding the modality, contiguity and coherence principles that govern instructional video and slide design; Mayer's book Multimedia Learning appeared in 2001. The germane load category was later revised, with Sweller arguing after 2010 that it is better treated as working memory resources redirected to intrinsic processing than as a separate additive source, a refinement that most current instructional design guidance now reflects.