Perspectives from the CHDS Media Hub

Create Blueprint

Cognitive Theory of Multimedia Learning

Jacob B Waxman, Ed.M., A.L.M.
Sue J. Goldie, MD, MPH
Center for Health Decision Science, Harvard T.H. Chan School of Public Health

Cognitive Theory of Multimedia Learning

Applying cognitive load theory to the design of educational multimedia
Introduction The CTML, extensively described by 2. Germane cognitive load
The cognitive theory of multimedia Mayer (2009, 2011, 2014), is based on CLT Germane cognitive load refers to
learning (CTML) draws together various in an environment of multimedia intrinsic cognitive load that is
theories of learning and cognition- learning. CLT is a theory of learning that associated with more difficult learning
particularly Paivio’s dual encoding is based on what we know about how tasks, more complex learning material,
theory (Clark & Paivio 1991) and learners process new information and and occurs when learners engage in
cognitive load theory (CLT) (Sweller et construct knowledge in long-term “deep cognitive processing” (Mayer
al. 1998) – and applies them to the memory (Sweller 1998, 2020). See our 2008) and quite related to a learner’s
process of learning through multimedia monograph on “Cognitive Load Theory effort. Framed in terms of cognitive
instruction. At its core, the theory and Instructional Design” (Goldie and processing, Mayer describes the
asserts that people learn more Waxman 2021) for a review of the “making sense of new material by
effectively from words and graphics theory, assumptions, and implications reorganizing it into a coherent
than from words alone. While numerous for instructional design. structure and integrating it with
investigators have contributed to this relevant prior knowledge,” as
work, we have found Mayer’s (2009) What is cognitive load? “generative processing” (Mayer 2009).
empirically derived principles of Cognitive load is a central concept
multimedia design, to be the most critical to cognitive load theory and 3. Extraneous cognitive load
relevant and useful for instructional highly relevant to learning and
Unlike intrinsic and germane cognitive
designers and educators. instructional design (Sweller 1994, 1999).
load, which are both essential for
Cognitive load refers to the “cognitive
learning, extraneous cognitive load
What is multimedia instruction? processing demands” placed on a
refers to information processing that is
It is not a video! Multimedia instruction learner in the context of the limitations
not needed or irrelevant to comprehend
(also referred to as “educational of working memory. There are three
the new material in a lesson plan. This
multimedia”) refers to teaching and general types of cognitive load that have
might include (a) unrelated
learning material that contain words been described: (1) intrinsic cognitive
information, such as visual elements
(e.g., spoken or printed text) and load, (2) germane cognitive load, and (3)
with no purpose, (b) extensive details
images (e.g., illustrations, charts, extraneous cognitive load.
that are not necessary for the particular
photos, animation, video) (Mayer 2009).
1. Intrinsic cognitive load learning task, (c) an instructional
Multimedia instruction therefore has
Intrinsic cognitive load refers to the format that introduces distraction, such
relevance well beyond a video-based
cognitive processing that is needed to as animation for “animation” sake.
lesson – it can inform our teaching over
comprehend new information. Intrinsic Framed in terms of cognitive
a broad range of visual and audio
cognitive load will depend on the nature processing, Mayer refers to the
modalities, from what we sketch on a
of the material, the learner’s prior draining of limited cognitive processing
chalkboard while lecturing to how we
knowledge, the learner’s affect, and capacity without contributing to
construct a diagram we create for a
contextual features of the learning learning as “extraneous processing”
handout in class, or the design of a
environment (Clark 2006). Framed in (Mayer 2009).
power-point slide to use in class.
terms of cognitive processing, Mayer
describes the selecting of relevant Note that all three types of cognitive
Where did the cognitive theory of
information and organizing it in working load need to be considered in the
multimedia learning come from?
memory as “essential processing” context of (a) learner expertise, (b)
(Mayer 2009). content complexity, and (c)
 Cognitive Theory of Multimedia Learning (cont.)

instructional methods and pedagogical The Signaling Principle

strategies. “People learn more deeply from a 1. Reduce Extraneous
multimedia message when cues are added Mapping
Processingthe
How can we design educational multimedia that highlight the organization of essential Learning Ecosystem
that contributes to effective learning material” (Mayer 2009). Figuring out what • Coherence: Delete
through attention to cognitive processing? to pay attention to counts as cognitive load! extraneous material
Signaling what’s important, through visual • Signaling: Highlight
In a nutshell, the goal is to make the cognitive
organization and cues (e.g., outline, essential material
processes associated with learning as
headings, highlighting) can reduce the
effective and efficient as possible, using a • Redundancy: Don’t add
attentive and cognitive load on the learner
three-pronged strategy: (1) Reducing onscreen captions to narrated
and free up working memory space for
extraneous processing, (2) Managing graphics
intrinsic or germane cognitive processing.
essential processing (3) Fostering generative
• Spatial contiguity: Place
processing. Just as Mayer adapted the
The Spatial Contiguity Principle printed words near
cognitive load theory to multimedia learning,
“People learn more deeply from a corresponding part of graphic
he contextualized the goals above using
multimedia message when corresponding
principles that were established based on • Temporal contiguity:
printed words and graphics are presented
empirical data (Mayer 2004, 2008, 2009, Present spoken words at same
near rather than far from each other on the
2012) We expand on these three strategies time as corresponding
page or screen” (Mayer 2009).
below. graphics
Presenting information close to its media
counterparts reduces the amount of time • Image Principle: People
1. Reducing Extraneous Processing don’t necessarily learn better
and cognitive effort the learner must
Learners have a limited WM capacity; when the speaker’s image is
expend to separate out and connect the
therefore, we want to be efficient with how visible.
relevant pieces of information, freeing up
we utilize it! Intrinsic and germane cognitive
capacity for intrinsic or germane cognitive
loads have pedagogical value, while
processing. 2. Manage Essential
extraneous cognitive load has none-
therefore we should aim to first reduce any Processing
The Temporal Contiguity Principle
extraneous load on the learner. Mayer (2009) • Segmenting: Break lesson
“People learn more deeply from a
identified five principles, based on empirical into learner-paced parts
multimedia message when corresponding
evidence, to reduce extraneous overload in
graphics and narration are presented • Pre-training: Present
the context of educational multimedia: (a)
simultaneously rather than successively” characteristics of key concepts
Coherence, (b) Redundancy, (c) Signaling, (d)
(Mayer 2009). Sensory and WM are limited before lesson
Spatial contiguity, (e) Temporal contiguity.
in terms of both time and capacity! We can
• Modality: Use spoken words
only hold information in these memories for
The Coherence Principle rather than printed words
mere seconds. When complimentary pieces
“People learn more deeply from a multimedia
of information are presented near each
message when extraneous material is excluded
other in time, the learner can spend more of
rather than included” (Mayer 2009). 3. Foster Generative
that time capacity working to organize the
Eliminating distracting and unnecessary Processing
new information and assimilate knowledge.
material (e.g., humorous graphics or
irrelevant animations) results in better • Personalization: Put words
The Image Principle in conversational style rather
learning. Simplify learner workflows and
“People do not necessarily learn more deeply than formal style
eliminate unnecessary content.
from a multimedia message when the
• Voice: Put words in human
speaker’s image is visible on screen (Mayer
The Redundancy Principle voice rather than machine
2009). Every instant of a multimedia
“People learn more deeply from graphics and voice
message is information that must either be
narration than from graphics, narration, and
on-screen text” (Mayer 2009). Learners may
assimilated or rejected by WM- the • Embodiment: Have
teacher’s face isn’t always integral to onscreen agent use human-
focus on the printed words rather than the
learning the material and doesn’t like gestures and movements
relevant portions of the graphics and can
necessarily result in greater learning. This (Mayer 2001)
expend limited processing capacity in WM by
should be weighed with affective
trying to reconcile the two verbal streams of
considerations.
information.
2
Cognitive Theory of Multimedia Learning (cont.)

2. Managing Essential Processing can be used for the information that References
Essential overload occurs the cognitive would have been in the adjacent text.
capacity of the learner is exceeded by Clark, J. M. & Paivio, A. (1991). Dual
the amount of essential processing a 3. Fostering Generative Processing coding theory and education.
new piece of information requires. Principles aimed at fostering Educational Psychology Review, 3(3),
Mayer (2009) identified three principles, generative processing “use social cues 149-170.
based on empirical evidence, to manage to prime the learner’s motivation to Mayer RE (2001). Multimedia learning.
essential processing in the context of exert effort to make sense of the New York: Cambridge University Press.
educational multimedia: (a) Segmenting, material” (Mayer 2009). Mayer (2009)
(b) Pre-training, and (c) Modality. identified three principles, based on Mayer RE (2008). Applying the science
empirical evidence, to foster of learning: Evidence-based principles
The Segmenting Principle generative processing in the context for the design of multimedia
“People learn more deeply when a of educational multimedia: (a) instruction. Cognition and Instruction
multimedia message is presented in Personalization, (b) Voice, (c) 19, 177-213.
learner-paced segments rather than as a Embodiment. Mayer RE and Moreno R (2003). Nine
continuous unit” (Mayer 2009). ways to reduce cognitive load in
Segmenting allows learners to fully The Personalization Principle multimedia learning. Educational
process one “chunk” of the learning “People learn better from multimedia Psychologist 38, 43-52.
experience, before having to move onto lessons when words are in
the next one. If we chunk our content conversational style rather than formal Miller, G. (1956). The magical number
into manageable chunks (defined by the style ” (Mayer 2009) seven, plus or minus two: Some limits
content complexity and learner level), An informal conversational style on our capacity for processing
learners can more efficiently manage established an environment of “social information. The psychological review,
essential cognitive processing in WM. proximity” and “social presence” 63, 81-97.
between educator and learner. This Paivio, A. (1986). Mental
The Pre-training Principle both supports the necessary Representations. New York: Oxford
“People learn more deeply from a “essential” cognitive processing but University Press.
multimedia message when they have also contributes to enhancing learner
learned the names and characteristics of motivation to engage meaningfully Sweller J (1988). Cognitive load during
the main concepts” (Mayer 2009). and participate in active learning. problem solving: Effects on learning.
Providing students with definitions and Cognitive Science 12, 257-285.
main characteristics of key elements in a The Voice Principle Sweller J, van Merrienboer JJG, Paas F.
lesson, frees up cognitive capacity for “People learn better when the Cognitive architecture and
more important essential cognitive narration in multimedia lessons is instructional design. Ed Psychol Rev.
tasks, such as understanding the spoken in a friendly human voice rather 1998;10(3):251–296.
relationships between pieces of than a machine voice” (Mayer 2009).
information, integrating new Similar to the personalization principle, Sweller J (1989). Cognitive technology:
information into existing schema, and a human voice contributes to “social Some procedures for facilitating
creating new schema. proximity” and “social presence” learning and problem-solving in
between educator and learner. mathematics and science. Journal of
The Modality Principle Educational Psychology 81, 457-466.
“People learn more deeply from a The Embodiment Principle Sweller J (1994). Cognitive load theory,
multimedia message when the words are “We learn more effectively when the learning difficulty, and instructional
spoken rather than printed” (Mayer instructors on screen in a multimedia design. Learning and Instruction 4,
2009) Spreading complimentary lesson demonstrate movement, eye 295-312.
information across the verbal and visual contact, facial expression and convey
channels can be more efficient than genuine human connectedness through Thomas M, et al. Annual Research
using just one of them. Consider a physical gestures” (Mayer 2009). Review: Educational neuroscience:
diagram with adjacent text versus voice This eye contact and physical progress and prospects. Journal of
narration. With voice narration, the expression of “connectedness” Child Psychology and Psychiatry 60:4
visual channel can be used to process contributes to “social presence.” (2019), pp 477–492.
the diagram, while the verbal channel

3
 Cognitive Theory of Multimedia Learning (cont.)

The figure below “represents memory stores as rectangles: sensory memory, which temporarily holds incoming
images and sounds; working memory, which allows for mentally manipulating a small amount of the incoming visual and verbal
material; and long-term memory, which is the learner’s permanent storehouse of knowledge” (Mayer 2009).

“Cognitive processing includes (1) selecting, which transfers some of the incoming images and sounds to working memory for
additional processing; (2) organizing, which organizes the images into a pictorial model and the words into a verbal model in
working memory; and (3) integrating, which connects the models with each other and with relevant knowledge activated from
long-term memory” (Mayer 2009).