Instructional Visualizations, Cognitive Load Theory, and Visuospatial Processing

Abstract

There are basically two formats used in instructional visualizations, namely, static pictures and dynamic visualizations (e.g., animations and videos). Both can be engaging and fun for university students in the fields of health and natural sciences. However, engagement by itself is not always conducive to learning. Consequently, teachers, lecturers, and instructional designers need to utilize the cognitive processing advantages of visualizations as well as engagement to achieve full instructional effectiveness. A cognitive processing focus has outlined many ways in which instructional visualization can be optimized. Specifically, cognitive load theory and the cognitive theory of multimedia learning are two research paradigms that provide several methods for directing the design of visualizations by considering how learners process visuospatial information. In this chapter, we describe five methods based on these cognitive theories: (a) the split attention effect and spatial contiguity principle, (b) the modality effect, (c) the redundancy effect and coherence principle, (d) the signaling principle, and (e) the transient information effect. For each of these effects, examples of applications for education in health and natural sciences are provided, where the influence of visuospatial processing is also considered. We end this chapter by discussing instructional implications for science education and providing future directions for research.

Full text

111© Springer Nature Switzerland AG 2019
J. C. Castro-Alonso (ed.), Visuospatial Processing for Education in Health and
Natural Sciences, https://doi.org/10.1007/978-3-030-20969-8_5
Chapter 5
Instructional Visualizations, Cognitive
Load Theory, andVisuospatial Processing
JuanC.Castro-Alonso, PaulAyres, andJohnSweller
There is ample evidence showing that instructional visualizations, both in static
(e.g., illustrations and photographs) and in dynamic formats (e.g., animations and
videos) can be engaging and fun for university students in the disciplines of health
and natural sciences. For example, in reviewing studies of the health sciences, Houts
etal. (2006) reported that large positive effects could be achieved by combining
instructional texts with static or dynamic visualizations. The review showed that
adding these instructional visualizations to the textual information increased par-
ticipants’ attention, comprehension, recall, and adherence (behavioral change).
With regards to static images, Hosler etal. (2011) investigated the effectiveness
of a comic book in delivering content and engaging 98 undergraduates (61%
females) in the biology topics of vision and evolution. Results showed increases in
both knowledge of the topics and positive attitudes toward biology. With regards to
dynamic visualizations, Jaffar (2012) asked 91 medicine and surgery undergradu-
ates to rate their opinions about the instructional effectiveness of a YouTube™ chan-
nel showing videos of human anatomy. The videos included plastic models,
cadaveric dissections, radiographs, PowerPointTM presentations, and surgical proce-
dures. Most participants (92%) agreed or strongly agreed that the style of the video
channel was helpful to learn anatomy. Particularly, they valued the properties of
increasing understanding (98%), creating memorable images (96%), and all-day
availability (94%).
These examples show that studies employing both static and dynamic visualiza-
tions provide evidence that students enjoy these materials. Nevertheless, students’
emotions and opinions are not always related to actual learning. For example,
Mahmud etal. (2011) reported a study with 287 medicine and surgery undergraduates
J. C. Castro-Alonso (*)
Center for Advanced Research in Education, Universidad de Chile, Santiago, Chile
e-mail: jccastro@ciae.uchile.cl
P. Ayres · J. Sweller
School of Education, University of New South Wales, Sydney, Australia

112
(69% females) investigating the instructional effectiveness of dissection videos
shown during classes and made available for later restudy. After the treatment of
approximately 6weeks, students’ opinions on the videos were enthusiastic, but the
anatomy test scores were not signicantly altered. In other words, liking did not
translate into learning.
Therefore, teachers, lecturers, and instructional designers need to be aware that
emotions and motivation do not necessarily lead to more learning through instruc-
tional visualizations. A variable that must be considered for effective learning under
these conditions is working memory processing. Although the instructional visual-
ization literature provides examples of emotional and motivational factors affecting
science learning and working memory processing (cf. Fraser etal. 2014), the focus
of this chapter is not on this relationship. Instead, we will focus on working memory
(cognitive) processing, specically visuospatial processing, without including emo-
tional or motivational inuences. In particular, this chapter has three goals: (a) to
show that instructional visualizations can optimize cognitive processing, and thus
be effective tools for learning about health and natural sciences; (b) to describe
cognitive methods for increasing the effectiveness of these instructional visualiza-
tions; and (c) to portray how visuospatial processing impacts on science learning
through visualizations.
5.1 Science Learning Optimized ThroughInstructional
Visualizations
There is a long tradition in the educational psychology literature (e.g., Calkins
1898; Shepard 1967) showing that visualizations tend to be easier to memorize than
verbal information (spoken and written words). This cognitive advantage of instruc-
tional visualizations has been recognized by instructional designers who add visu-
alizations to textual passages, as a way to improve learning. In fact, this combination
of visual and verbal representations has generally been found to improve learning.
The review by Vekiri (2002) outlined two research avenues that support the use
of this picture plus text combination. On the one hand, the visual argument hypoth-
esis predicts that visualizations can involve less search for the relations between the
learning elements, as compared to equivalent textual information. As such, visual-
izations require less working memory processing, so any additional resources can
be allocated to understanding the texts. In contrast, pure textual information would
not leave as many resources for in-depth processing.
On the other hand, dual coding theory (see Clark and Paivio 1991) proposes that
there are two distinct working memory systems: the visuospatial system deals
mostly with visualizations and the verbal system deals mostly with texts and narra-
tions (see also Castro-Alonso and Atit this volume, Chap. 2). Notably, dual coding
theory argues that these systems are interconnected, and fostering these associations
produce more effective learning. In this case, combining visualizations and text
J. C. Castro-Alonso et al.

113
makes it easier to process and memorize information than purely providing visual-
izations or text alone. This effect has been termed the multimedia principle (e.g.,
Mayer 1989; see Butcher 2014), in that two different modes are better than one.
Particularly for science learning, the combinations of visuospatial and verbal
representations can help learners understand more effectively complex cause and
effect systems. For example, visualizations can be scaffolds to avoid misinterpreta-
tions that could happen when learning science systems from verbal information
only (e.g., Eitel et al. 2013; see also Hegarty 2011). In more detail, Mayer and
Gallini (1990) described two types of scaffolds that visualizations can convey to
help learning a science system: system topology and component behavior. System
topology shows the components of the system and their locations within the overall
structure. Component behavior shows the changes in the components, and how
these changes affect other components and the overall systemic mechanism. As the
authors observed in three experiments with a total of 300 university students, both
system topology and component behavior must be conveyed by visualizations to
boost learning. In fact, for novice learners studying brakes, pumps, and electric
generator systems, it was observed that when only one of these scaffolding func-
tions was provided, the outcomes for conceptual recall and problem solving were
equivalent to not providing visualizations. Hence, both scaffolds explaining the
topology and the behaviors were required to produce signicant gains for visualiza-
tions compared to text-only information.
Altogether, these ndings support the argument that visualizations can be useful
assets for learning about health and natural sciences. However, as observed by
Vekiri (2002), for visualizations to be effective, they must help to process the textual
learning material. For example, by scaffolding the verbal information, visualiza-
tions can aid in processing the learning content. On the contrary, visualizations not
designed for this scaffolding purpose can be ineffective or even counterproductive
for learning. How to design instructional visualizations, in order to facilitate visuo-
spatial processing and thus optimize science learning, can be informed by cognitive
load theory, as described next.
5.2 Cognitive Load Theory
Cognitive load theory (see Sweller etal. 2011; see also Sweller et al. 2019) is an
instructional theory based on the knowledge of the human cognitive architecture.
The theory, broad enough to include many learning situations, has dealt with visual,
spatial, verbal, and auditory learning materials on a number of diverse topics.
However, in this chapter, we will focus on visuospatial materials about health and
natural sciences. As outlined below, there are different examples of cognitive load
theory being applied to instructional visualizations about the health sciences (see
Issa etal. 2011; Wilson 2015; see also Fraser etal. 2015) and the natural sciences
(see Ginns 2005; Schneider etal. 2018).
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

114
5.2.1 Science asaRelevant Category ofKnowledge
There are many ways of categorizing knowledge with most schemes having limited
or no instructional consequences. One scheme that has profound instructional con-
sequences was provided by David Geary (Geary and Berch 2016; Sweller 2016).
Geary divided knowledge into biologically (or evolutionary) primary and secondary
knowledge. He argued that humans have evolved to acquire various forms of pri-
mary knowledge, such as learning to speak a native language, or recognizing ges-
tures (see Castro-Alonso etal. this volume-b, Chap. 7). Generic-cognitive skills,
such as general problem-solving skills, also provide a large class of biologically
primary skills (Tricot and Sweller 2014). It is argued that we have evolved to acquire
these skills naturally and therefore they are learned easily, automatically and with-
out instruction.
Unlike biologically primary knowledge, we have not evolved to acquire specic
biologically secondary knowledge. We can acquire secondary knowledge but do so
consciously, with effort, and often assisted by explicit instruction (Geary 2002,
2007). Almost everything that is taught in training and education institutions is bio-
logically secondary (Sweller 2015). We invented these institutions because, without
them, the cultural knowledge that constitutes secondary skills tends not to be
acquired (Geary 2002, 2007). Learning to read and write provides an obvious exam-
ple of secondary knowledge, and in direct contrast to learning to listen, speak, and
gesture (primary knowledge).
While generic-cognitive skills commonly provide examples of primary knowl-
edge, domain-specic skills commonly provide examples of secondary skills. For
example, the educational syllabi for the health and natural sciences are composed
mostly of secondary skills. Learning science concepts and skills needs to be taught
explicitly because they are domain-specic skills that we have not evolved to
acquire automatically (Tricot and Sweller 2014). In contrast, learning a general
problem-solving skill such as using mean-ends analysis (cf. Larkin etal. 1980) to
solve a wide variety of problems cannot be taught because it is learned automati-
cally as a biologically primary skill (unless specic applications are guided; see
Youssef etal. 2012). Cognitive load theory is primarily concerned with the acquisi-
tion of biologically secondary rather than primary knowledge. Consequently, the
theory is highly relevant for science instruction, including the use of instructional
visualizations. The theory is described next.
5.2.2 Human Cognitive Architecture
As stated above, cognitive load theory mainly describes the acquisition of biologi-
cally secondary knowledge. It only is concerned with primary knowledge to the
extent that primary knowledge can assist in the acquisition of secondary knowledge
(Paas and Sweller 2012). The cognitive architecture used to acquire secondary
J. C. Castro-Alonso et al.

115
knowledge and skills mimics the procedures used by biological evolution (see
Sweller and Sweller 2006). It can be described by the following ve basic principles
of cognitive load theory, outlined by Sweller and Sweller (2006):
• The information store principle. In order to function, human cognitive architec-
ture requires a substantial store of information. Long-term memory provides that
large store. For example, visual long-term memory has shown a signicant
capacity to store details about pictures of objects (e.g., Brady et al. 2008). A
major purpose of instruction is to assist learners in acquiring this large long-term
store of biologically secondary, domain-specic information. In consequence, a
chief purpose of science education is to assists learners to acquire scientic
information to incorporate to long-term memory.
• The borrowing and reorganizing principle. Most biologically secondary infor-
mation held in long-term memory is obtained from other people, by listening to
what others tell us, watching what they do, or observing what they show. Our
ability to communicate information between us is a biologically primary skill
that we have evolved to acquire. Accordingly, communication between us pro-
vides the primary means by which we obtain the substantial amounts of second-
ary information held in long-term memory. For example, we have incorporated
information from teachers and instructional visualizations to enhance our knowl-
edge of health and natural sciences.
• The randomness as genesis principle. Sometimes, we are required to generate
novel information. As it is new information, no one else has it, so it cannot be
obtained from others. Under those circumstances, information can be generated
during problem-solving by using a generate-and-test procedure. We can ran-
domly generate novel information and test it for effectiveness with effective
information retained and ineffective information discarded. These processes take
place in working memory, and if they involve visualizations, they occur in the
visuospatial processor of working memory (and the central executive, see Castro-
Alonso and Atit this volume, Chap. 2).
• The narrow limits of change principle. Only limited amounts of novel verbal or
visuospatial information can be processed at any given time. Working memory is
used to manage novel information and working memory is severely restricted in
both capacity and duration when processing novel information (see Cowan
2001). Furthermore, it is subject to additional depletion after use (and recovery
after rest, see Chen etal. 2018). When investigating the limits of visuospatial
processing in working memory, Luck and Vogel (1997) reported that approxi-
mately four features could be processed correctly when memorizing displays of
squares that changed in color, orientation, and size. More recently, Oberauer and
Eichenberger (2013) replicated these ndings and showed that there is a trade-off
in visuospatial processing between the number of visual elements, number of
features, and detail precision. These studies show that working memory process-
ing is severely limited when dealing with novel visuospatial information, which
can include science visualizations.
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

116
• The environmental organizing and linking principle. Once information is stored
in long-term memory, it can be transferred to working memory following envi-
ronmental signals to generate action appropriate to the environment. Information
transferred from long-term to working memory has none of the limitations
affecting novel information. There are no known capacity or duration limitations
when working memory processes familiar information transferred from long-
term to working memory. It is at that point that the full advantages of education
manifest themselves. For example, although novice students may struggle when
processing one visual representation of molecules, more knowledgeable learners
can process several representations simultaneously (cf. Stull etal. 2018), show-
ing that chemistry education can overcome initial visuospatial processing
limitations.
Cognitive load theory uses the cognitive architecture described by these ve
principles to devise effective instructional procedures (see Paas and Sweller 2014).
Those procedures are based on the assumptions that learners must acquire biologi-
cally secondary, domain-specic knowledge stored in long-term memory for use in
an appropriate environment. This general goal of cognitive load theory includes the
specic learning scenario of studying science topics through visualizations, as
described next.
5.3 Cognitive Load Theory Effects forScience Visualizations
University students from the areas of health and natural sciences must acquire bio-
logically secondary scientic knowledge to succeed in their elds. Cognitive load
theory can be employed to produce more effective instructional visualizations for
science learning. In order to do this, the limits of visuospatial processing in working
memory must be considered and circumvented (cf. Castro-Alonso and Uttal this
volume, Chap. 3).
There are basically two tracks to elude the limitations of visuospatial processing:
(a) reducing unnecessary visuospatial processing, and (b) increasing total visuospa-
tial processing. As predicted by cognitive load theory, following one or both of these
tracks will lead to more effective instructional visualization. In the remainder of the
chapter, we will describe ve methods that follow these paths. Cognitive load theory
describes them as effects (see Sweller etal. 2011). A related theory, the cognitive
theory of multimedia learning, describes them as principles (see Mayer 2014a).
These methods to optimize science visualizations and their names in both theories
are shown in Table5.1.
A further description of the cognitive load theory effects in these examples is
provided next, where we describe the ve effects and also consider the inuence of
visuospatial processing. Details of most of the visuospatial processing abilities and
instruments described in this chapter can be found in Castro-Alonso and Atit (this
J. C. Castro-Alonso et al.

117
volume, Chap. 2), and also other versions of similar tests are described in Castro-
Alonso etal. (this volume-a, Chap. 8) and in Castro-Alonso etal. (2018a).
5.4 Split-Attention Effect andSpatial Contiguity Principle
According to cognitive load theory, a split-attention effect occurs when a multime-
dia presentation is designed in such an ineffective manner that the visuospatial con-
tents are separated in the display, so learners have to mentally integrate them in
working memory (see Ayres and Sweller 2014; see also Fraser etal. 2015).
For example, an instructional page of human anatomy could depict the skeletal
system of the hand as shown in Fig.5.1a. As the bones are exhibited at the top left
and the explanatory legend is shown at the bottom right, this design would require
students to look up and down continuously and would produce a split-attention
effect, counterproductive to learning. A solution would be to move the texts from
the legend into closer proximity to the bones, as shown in Fig.5.1b. This physical
integration would increase spatial contiguity between text and images. Such an inte-
grated format would require less visuospatial processing due to searching and
matching, and thus be more useful for learning, as it has been shown since the semi-
nal study of this effect (Tarmizi and Sweller 1988) and many others which followed
(e.g., Chandler and Sweller 1991; Makransky etal. 2019; Purnell etal. 1991).
The meta-analysis by Ginns (2006), which considered a total of 37 effect sizes
for the spatial contiguity principle, showed an overall effect size of d=0.72. A more
current meta-analysis by Schroeder and Cenkci (2018) included additional 21 inde-
pendent comparisons. In total, it analyzed 58 independent effect sizes (n= 2,426
participants), the majority (60%) concerning post-secondary education. The overall
effect size of this updated meta-analysis was of g=0.63. According to the bench-
marks of effect sizes by Cohen (1988), these overall magnitudes represents medium
to large sizes. Next, we give examples for health and natural sciences visualizations
being optimized when considering this effect.
Table 5.1 Methods to optimize visualizations and examples for visuospatial information
Cognitive load theory
Cognitive theory of multimedia
learning Example of solution
Split attention effect Spatial contiguity principle Physically integrate the visuospatial
information
Modality effect Modality principle Present some information auditorily
Redundancy effect Coherence principle Delete unimportant visuospatial
information
Signaling principle Signal important visuospatial
information
Transient information
effect
Avoid fast-paced visuospatial
information
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

118
5.4.1 The Effect inOptimization ofScience Visualizations
The meta-analysis by Schroeder and Cenkci (2018) described above, also analyzed
the split-attention effect in different disciplines. For anatomy/medicine, 8 compari-
sons (n=368) showed an overall effect size of g=0.52. For biology/earth science,
12 effect sizes (n=557) showed an overall g=0.56. These effects correspond to
medium sizes.
An example of university biology study showing the split attention effect is given
by Cierniak etal. (2009), in which the learning topic was the structure and function
of the kidney’s unit, known as the nephron. The authors compared learning of 98
university students (64% females) randomly allocated to either a split or an inte-
grated format between nephron diagrams and textual labels. Students were assessed
in both simple tests (terminology and labeling), and more demanding tasks (com-
plex facts and inferences). It was observed that the spatial contiguity condition out-
performed the split-attention condition in three out of the four measures, as there
were no signicant differences in the complex inference tasks.
Similar results were obtained by Erhel and Jamet (2006) with multimedia mod-
ules showing the topics of the human heart and the replication of a virus. In the
experiment with 72 psychology undergraduates (85% females), the participants
were randomly assigned to one of three conditions: (a) separated, where the images
and the blocks of texts were spaced apart on the screen; (b) integrated, where the
images and the blocks of texts were near to each other, thus following the spatial
contiguity principle; and (c) pop-up texts, where the images and texts also followed
the spatial contiguity principle, but in this case an interaction was required from the
Fig. 5.1 Example of (a) separated format that produces split attention, and (b) integrated format
that follows the spatial contiguity principle
J. C. Castro-Alonso et al.

119
learners (see also Castro-Alonso and Fiorella this volume, Chap. 6). For the four
assessed measures (two retention and two transfer tests) it was observed that the
integrated and pop-up texts conditions outperformed the separated design.
Huff etal. (2012) presented a stereoscopic vexing-image technology to manage
split-attention, in an experiment with 80 university students (80% females). The
task was to notice failures when contrasting two visualizations of mechanical pen-
dulum clocks. Randomly, a group of split design was compared to a group with this
novel technology that avoids eye movement. As predicted, the stereoscopic technol-
ogy, due to avoiding split-attention, produced a faster and more accurate learning
performance. Next, we describe how the split-attention effect and spatial contiguity
principle are inuenced by high versus low visuospatial processing.
5.4.2 Visuospatial Processing Inuencing theEffect
Wiegmann etal. (1992, Experiment 2) investigated 34 psychology undergraduates
learning about the human autonomic nervous system from concept maps. Randomly,
the participants were allocated to one of two map design conditions: (a) an inte-
grated format showing a single large map, or (b) a split attention format where six
interconnected smaller maps showed the information. Also, the visuospatial ability
of eld independence was measured with the Group Embedded Figures Test (see
discussion of these abilities and test in Castro-Alonso and Atit this volume, Chap.
2). Results showed that, for the students with higher eld independence, the split
attention design was more effective for learning than the integrated format. In con-
trast, those with lower scores in the visuospatial processing task were not benetted
by either design. Similarly, Fenesi etal. (2016, Experiment 1) used a dual working
memory task to measure working memory capacity of 76 undergraduates (59%
females) required to learn the topic of visual memory from slideshows. According
to the design of each slide, the participants were randomly allocated to a split atten-
tion or an integrated condition. Linear regression analyses showed that working
memory capacity predicted learning for the split attention group but not for the
integrated condition.
Huff and Schwan (2011) measured three-dimensional (3D) mental rotation of 84
university students (68% females), employing the common instrument called the
Mental Rotations Test. In the study, the learning task involved relating structural
information from proteins shown in different animations. For this biochemistry
task, results showed that high mental rotators outperformed low mental rotators
when the different animations were presented in a split attention format. However,
in the integrated presentation, where the animations followed the spatial contiguity
principle, mental rotation scores did not inuence achievement.
Bauhoff etal. (2012) investigated 44 university students (50% females) attempt-
ing to notice missing pieces and other failures in two comparable visualizations of
mechanical pendulum clocks. Different degrees of split attention designs were
investigated by changing the distance between both clock visualizations. The par-
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

120
ticipants were also assessed on visual working memory with the Visual Patterns
Test. Results showed a split attention effect, as more separated visualizations
involved longer processing intervals. In addition, participants with lower visual
working memory had to rely on their knowledge of the clock’s mechanisms to cope
with the demanding split attention formats.
Using a more abstract task, Dutke and Rinck (2006) studied 96 university partici-
pants (61% females) memorizing arrangements of either two or ve images of tools,
musical instrument, and animals. They also measured the students’ scores in a dual
visuospatial task of working memory. Results showed that students with lower dual
visuospatial scores could not memorize the arrangements with ve images as ef-
ciently as that with only two pictures. In other words, the split attention with ve
different depictions was not as manageable as the split attention with two depic-
tions. In contrast, high visuospatial students could manage both arrangements of
ve and two images efciently. In conclusion, and consistent with the predictions of
the split attention effect and the spatial contiguity principle, separated formats tend
to be counterproductive to learning for students with less visuospatial processing
capacity. In contrast, the split attention formats are not as problematic for students
with greater visuospatial resources or abilities, who are more capable of coping with
these demanding visualization designs.
5.5 Modality Effect or Modality Principle
A modality effect (e.g., Mousavi etal. 1995) is observed when instructional visual-
izations are less effective when supplemented with written text as compared to spo-
ken text (see Fraser etal. 2015). In a meta-analysis of 43 effect sizes and more than
1900 students, Ginns (2005) concluded that this principle presented a medium to
large effect size for instruction (d= 0.72). Moreover, when analyzed by instruc-
tional discipline, the meta-analysis revealed a large effect for science topics
(d=1.20).
We consider two non-mutually exclusive explanations for this effect or principle.
The rst explanation, described in Low and Sweller (2014) and consistent with the
cognitive theory of multimedia learning (see Mayer 2014b), is based on certain
separability when processing visuospatial and auditory information (see also
Castro-Alonso and Atit this volume, Chap. 2). As reviewed by Penney (1989), this
separability can be observed in experiments showing double dissociations, in which:
(a) visuospatial information is selectively interfered by new visuospatial informa-
tion, but not as much by new verbal information, and (b) verbal information is
selectively interfered with by new verbal information, but not as much by new
visuospatial information (e.g., Brünken etal. 2002, Experiment 1).
The modality principle calls to employ this degree of separability to increase the
total working memory capacity allocated to learning. For instance, when learners
study visualizations, visuospatial working memory processing is devoted to the
visual learning elements. In the example of Fig. 5.2, visuospatial processing
J. C. Castro-Alonso et al.

121
resources will be employed in studying the anatomical relationship between the
bones of the hand. If the names of the bones are written, then extra visuospatial
processing would be required for reading those names and matching them to the
bones. Thus, cognitive load would be added to the visuospatial processor of work-
ing memory. This can be particularly problematic if the written information is far
from the depicted learning information, generating split attention between the learn-
ing elements and their names (see Fig.5.2a). In contrast, if the names of the bones
are spoken, this narration can be processed somewhat separately, and thus visuospa-
tial processing is not overloaded, and a possible split attention does not occur (see
Fig.5.2b).
A second explanation can also be provided for this effect. As noted in Sect. 5.2.1
(see also Ong 1982; Paas and Sweller 2012), since listening to spoken texts is a
biologically primary knowledge, it can involve less working memory processing
than the biologically secondary knowledge of reading written texts. Examples of the
research about modality effects on science education are described next.
5.5.1 The Effect inOptimization ofScience Visualizations
We provide four examples of experiments showing a modality effect for instruc-
tional visualizations about science or technical topics. First, Kühl et al. (2011)
investigated 80 university students (79% females) learning the biology topic of sh
locomotion from computer presentations. Results showed that narrated multimedia
outperformed written multimedia in both retention and transfer tests. Second, the
study by Moreno and Mayer (1999, Experiment 1) concerned 132 university
Fig. 5.2 Example of (a) separated format that produces split attention, and (b) format with some
of the information in auditory working memory, fostering a modality effect or modality principle
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

122
students learning the topic of lightning formation from different designs of dynamic
visualizations. When comparing the animations supplemented with spoken versus
written texts, it was observed that the spoken format presented signicantly higher
scores on tests of retention and transfer. Similarly, Schmidt-Weigand etal. (2010,
Experiment 1) investigated 90 university students (58% females) learning lightning
formation from visualizations and found that the participants studying with supple-
mentary auditory explanations outperformed those presented with supplementary
on-screen texts. Last, Kalyuga etal. (1999, Experiment 1) assessed rst-year trade
apprentices learning to interpret a fusion diagram for soldering. After randomly
allocating a group to a diagram supplemented with written texts versus a group
given the diagram and auditory texts, it was observed that the narrated condition
self-rated less cognitive load and obtained higher performance scores in this engi-
neering task.
The modality effect or modality principle can also be inuenced by the students’
visuospatial processing ability. Although this inuence has not often been directly
investigated for this effect, the following two examples provide some evidence.
5.5.2 Visuospatial Processing Inuencing theEffect
In Experiment 2 reported in Seufert et al. (2009), 78 university students (74%
females) were presented computer static pictures and texts about the structure and
function of the enzyme ATP Synthase. For this biology topic, half of the participants
received on-screen texts and the other half was given the textual information as nar-
rations. Also, an aggregated score of visuospatial processing was calculated by
averaging the scores in two common tests: the 2D mental rotation instrument termed
the Card Rotations Test, and the mental folding instrument called the Paper Folding
Test (see Castro-Alonso and Atit this volume, Chap. 2). Learning was measured
with recall, comprehension, and transfer tests. Results showed a modality trend for
the three tests, being only signicant for recall performance. Hence, recalling facts
about the enzyme was better when the static visualizations included narrated rather
than on-screen texts. It was also observed that visuospatial processing was a signi-
cant predictor in comprehension and transfer performance. In other words, for visu-
alizations supplemented with either narrated or visual texts, higher visuospatial
processing was benecial for comprehension and transfer achievements. In short,
although a modality effect was observed, and visuospatial processing abilities were
inuential, there were no interactions between the format of the texts and visuospa-
tial processing of the students.
Similarly, Lee and Shin (2012) investigated 72 adult participants attempting the
procedural task of replacing a printer cartridge. In the study, visuospatial processing
of the participants was measured by aggregating the scores of the 3D mental rota-
tion Cube Comparisons Test and the mental folding Paper Folding Test. With the
combined scores, a median split separated high vs. low visuospatial processing par-
ticipants. For task performance, comparisons were made between high and low
J. C. Castro-Alonso et al.

123
visuospatial subjects, given either written or auditory instructions for the proce-
dures. For the tasks shown as static images, ndings showed that high visuospatial
participants performed signicantly better than lower spatial scorers. In contrast to
this visuospatial processing effect, there were no differences in attempting the task
after reading or hearing the instructions. In other words, no modality effect was
observed, but only that visuospatial processing was benecial to learn from both
written and auditory supplementary texts.
Although the two examples of this section investigated the effects of visuospatial
processing on the modality effect, they did not show interactions where a low versus
a high degree of visuospatial processing would affect differently studying science
visualizations supplemented with either written or spoken texts. In other words,
there is a research gap in studies investigating the effects of visuospatial processing
on the modality effect or principle.
5.6 Redundancy Effect andCoherence Principle
The redundancy effect (see Kalyuga and Sweller 2014; see also Fraser etal. 2015)
is observed when there is more than the essential visuospatial information provided
to learn, so students have to process both essential and non-essential information,
leading to a greater reduction of visuospatial working memory available for learn-
ing. For the same anatomy example (see Fig.5.3a), if the texts of the bones are
already integrated into the anatomy images, having a legend at the bottom right
would be unnecessary, and it might produce a redundancy effect. To optimize
Fig. 5.3 Example of (a) redundant format, and (b) non-redundant format that follows the coher-
ence principle
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

124
learning, deleting the legend would be preferable, in a design that follows the coher-
ence principle (see Fig.5.3b), and also the spatial contiguity principle.
In a meta-analysis about redundant information, Rey (2012) collected 34 effect
sizes (N=3,535 participants) for retention tests, and 21 effect sizes (1,634 partici-
pants) for transfer tests. The analyses showed an overall small to medium effect size
for retention measures (d = 0.30) and an overall medium effect size for transfer
assessments (d=0.48). Examples related more exclusively to visualizations depict-
ing health and natural sciences are described next.
5.6.1 The Effect inOptimization ofScience Visualizations
In a study with 49 rst-year medicine students (33% females), Garg etal. (1999)
compared the instructional effectiveness of virtual visualizations showing either
few or multiple rotational views of carpal bones. As there was no signicant differ-
ence between both views, the visualization with more images was not more produc-
tive than that with fewer images for learning these anatomy structures, suggesting
redundancy. For the natural science topic of electrical light circuits, Chandler and
Sweller (1991, Experiment 2) provided the rst evidence for the redundancy effect
in cognitive load theory. In the study with 28 trade apprentices, two redundant con-
ditions were tested: a split attention or separated condition (the diagram far from the
redundant texts) was compared to an integrated condition (the diagram near redun-
dant texts). It was observed that both conditions did not differ in learning outcomes.
The authors concluded that, as opposed to integrated formats with only relevant
information (presented in Sect. 5.4), integrated formats with redundant information
were not useful learning tools.
In the related area of meteorology, in an experiment with 74 university partici-
pants studying the topic of lightning formation, Harp and Mayer (1997) compared
the instructional effectiveness of booklets with or without redundant information
that was also seductive. The between-subjects study investigated four groups,
according to a 2 (Seductive texts: Yes vs. no)×2 (Seductive illustrations: Yes vs. no)
design. It was observed that the best learning condition was that with no extra
seductive adjuncts, arguably because this seductive information distracts from the
relevant passages and illustrations to understand the chain of events in lightning. In
a later study of four experiments, totaling 357 undergraduates studying the forma-
tion of lightning, Harp and Mayer (1998) consistently observed that extra informa-
tion (texts and visualizations) hindered performance on retention and transfer tests
(see also Eitel etal. 2019).
Investigating this effect further, Mayer etal. (2008) reported two experiments
studying the inuence of level of interestingness in redundant texts. The novelty of
this approach is that all conditions included off-topic texts of comparable lengths,
but these redundant texts had been rated before as highly interesting vs. highly
uninteresting. In Experiment 1, with 89 psychology undergraduates (66% females),
the learning task was understanding how the cold virus infected the human body.
J. C. Castro-Alonso et al.

125
Randomly, half of the participants learned this health sciences lesson including low
interestingness redundant texts, and the other half studied under the high interest-
ingness condition. It was observed that the low interestingness outperformed the
high interestingness (seductive) groups in the transfer tests.
Experiment 2, with 53 psychology undergraduates (75% females), extended
these ndings to slideshow presentations about the process of swallowing during
digestion. As both low and high interestingness texts were matched in the quantity
of redundant information, both experiments showed that more interesting irrelevant
supplements (seductive) could be more damaging for understanding the main learn-
ing concepts than less interesting redundant additions. The inuence of visuospatial
processing on the redundancy effect and the coherence principle is described next.
5.6.2 Visuospatial Processing Inuencing theEffect
In the experiment by Levinson etal. (2007), 118 psychology undergraduates (75%
females) were randomly assigned to multimedia showing either key views or mul-
tiple views of the brain. In the key views conditions, only four images were shown
(anterior, inferior, lateral, and superior views of the brain photographed digitally).
In contrast, the multiple views groups studied 24 images (digital photographs at 30°
increments around the brain model). A redundancy effect was observed, as the mul-
tiple views were signicantly less effective instructional visualizations than the key
views. Hence, employing visually demanding anatomical depictions, this study pro-
vided a replication to the ndings (see above) by Garg etal. (1999) with simpler
carpal bone images. In addition, the study by Levinson etal. (2007) also measured
student’s performance on the 3D instrument Mental Rotations Test. Notably, it was
observed that the disadvantages of the multiple views were larger for students with
lower mental rotation. As such, for low visuospatial processing students, perfor-
mance after multiple view study was approximately 30% lower than outcomes after
studying the key views.
Korbach etal. (2016) investigated 108 university students (74% females) learn-
ing about the structure and function of the enzyme ATP Synthase from 11 multime-
dia slides showing images and texts. Randomly, the students were allocated to a
group without redundant information versus a group with redundant information in
4 of the 11 screens. Visuospatial processing was also assessed by averaging the
scores of two paper instruments, respectively measuring 2D mental rotation and
mental folding. As predicted, an overall redundancy effect was observed, in which
the redundant condition was outperformed in retention and comprehension tests by
the non-redundant (coherence) design. Also, the effect was more signicant for low
visuospatial processing students.
Not only the visuospatial processing of working memory but also total working
memory can inuence the redundancy effect. For example, Fenesi etal. (2016,
Experiment 2) measured the working memory capacity of 71 undergraduates (63%
females), employing a dual instrument that involved a memory and a processing
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

126
task. Then, participants were shown slideshows depicting the topic of visual mem-
ory. Randomly, the participants observed these slideshows in either a redundant or
an essential design for each slide. Linear regression analyses showed that working
memory capacity predicted learning for the redundant condition but not for the
essential group. Consistent with the prediction of cognitive load theory, the redun-
dant design hindered learning in those students with lower working memory capac-
ity, but it was manageable by those with more cognitive resources. Although the
working memory test used in this experiment was not exclusively visuospatial, the
results show that total working memory (including verbal working memory) helps
to process redundant information such as non-essential illustrations.
In conclusion, as predicted by cognitive load theory and the cognitive theory of
multimedia learning, redundant visualizations hinder learning because they require
unnecessary processing in understanding the main science concepts. Redundant
visuospatial information is less problematic for high visuospatial processing stu-
dents, as they can manage in working memory both essential and redundant infor-
mation. In contrast, low visuospatial processing students are less able to process
both essential and non-essential or redundant visuospatial information.
5.7 Signaling Principle
The signaling principle (see van Gog 2014) incorporates visual cues to signal the
essential learning elements and their relationships, so learners know where their
main focus should be. We found three meta-analyses for the signaling principle. The
most recent of these analyses were conducted by Schneider etal. (2018) on 145
comparisons (from 103 studies and N=12,201 participants). The analysis showed
a positive effect of signaling for retention scores, as 117 out of 139 comparisons
revealed benecial signaling representing an overall medium effect size (g=0.53).
Transfer performance showed that 55 out of 70 comparisons were positive for sig-
naling, in an overall small to medium effect size (g=0.33). For cognitive load, 19
out of 27 effect sizes were positive and showing an overall small size (g=0.25),
indicative that signaling reduced perceived cognitive load on learners. Also, 20 out
of 27 comparisons were negative for learning time, representing an overall small to
medium size (g=−0.30), which denoted that materials with signals involved more
learning time. The study by Xie etal. (2017) included three meta-analyses: (a) for
retention scores, the meta-analysis of 25 studies (N=2,910) revealed a small to
medium effect (d= 0.27); (b) for transfer scores, the meta-analysis of 29 studies
(N=3,204) revealed a small to medium effect (d=0.34); and (c) for lowering per-
ceived cognitive load, the meta-analysis of 32 studies (N=3,597) revealed a small
effect (d= 0.11). The third analysis, conducted by Richter etal. (2016), reported
that 38 out of 45 comparisons (from 27 studies) showed positive effects for signal-
ing, and that this represented a small to medium overall effect size (r=.17). These
meta-analyses, which included some common sources, supported the effectiveness
of the signaling principle in increasing retention and transfer scores, while lowering
J. C. Castro-Alonso et al.

127
perceived cognitive load. These positive effects of visual signaling are also observed
for topics in health and natural sciences, as described next.
5.7.1 The Effect inOptimization ofScience Visualizations
Based on the review by de Koning etal. (2009), signaling can aid learning from
visualizations by highlighting (a) importance and (b) relationships. The rst goal of
highlighting importance is the most common. It signals the main learning elements.
As redundant or non-essential information is given less precedence, signaling pro-
duces less cognitive load that could impede learning. The second goal of highlight-
ing relationships has been less researched. It involves showing relations between the
learning elements, which makes it easier to group and memorize them together.
To attain both goals, signaling techniques can be broadly organized in two
groups, as described in Castro-Alonso etal. (2014a; see also de Koning etal. 2009):
(a) signaling with added elements, and (b) signaling without added elements.
Signaling with extra elements include pointing devices (e.g., arrows, ngers, hands,
and lines), frames, alphanumeric characters, labels, among others. Signaling with-
out these elements comprise blurring, lighting, transparencies, ashing, zooming,
colors, contrasts, and combinations. An example of these two types of signaling
techniques is provided in Fig.5.4, which shows a hydrogen bond between water
molecules depicted without signaling (Fig5.4a), using signaling with added ele-
ments (Fig5.4b), and using signaling without added elements (Fig5.4c).
An example of effective signaling with added elements, in this case, the added
depiction of red arrows, is provided by Lin and Atkinson (2011). A group of under-
graduates given static images with signaling was faster in learning the rock cycle
topics than the group given static images without signaling. It was observed that this
positive effect for signaling on static images of the rock cycle was not presented on
equivalent animations. Hence, consistent with many other domains, an important
variable to consider in signaling studies is whether the visualization is static or
animated.
Fig. 5.4 Hydrogen bond between two molecules of water. Example of (a) no signaling, (b) signal-
ing with added elements (arrow), and (c) signaling without added elements (transparency)
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

128
In contrast, we provide three examples showing effective signaling without
added elements. Color coding was used by Skulmowski and Rey (2018) on an anat-
omy task. In the experiment, 108 adults (78% females) had to memorize the parts of
a ctional 3D bone model. Following a 2 Realism (high vs. low)× 2 Signaling
(colors vs. no signaling) between-subjects design, four groups were compared.
Showing an interaction, color signaling was only helpful with the high realistic
depiction of the bone model but was counterproductive on the depiction with fewer
details. In other words, when more visuospatial processing was required due to
redundant (unnecessary) realistic details, signaling was a needed asset to reduce this
cognitive load and produce better memorization of the anatomical parts.
Amadieu etal. (2011) conducted an experiment employing zoom-in to focus on
the main elements of a neurobiology phenomenon related to synapsis. The study
investigated 36 psychology undergraduates (83% females) learning from an anima-
tion with or without this signaling technique. It was observed that only the signaled
animation with zooms helped to increase the comprehension scores when watching
more than once the dynamic visualization. This means that even repeating the ani-
mation was not helpful in the conditions without signaling.
The technique of spotlight cueing is reported by de Koning etal. (2010) in an
experiment with 76 psychology undergraduates (74% females) studying animations
of the human cardiovascular system. A non-signaled condition was compared to a
signaled condition where the main elements kept their luminance, and the less
important visuals were obscured. When comparing both conditions, it was observed
that the spotlights were effective for retention, inference, and transfer outcomes.
Although there are examples of successful signaling with added elements, such
as the mentioned study with red arrow signals (Lin and Atkinson 2011), it was sug-
gested by de Koning etal. (2009) that this type of signaling could be less effective
than that without added elements. This is because not adding elements could be
more effective, as it keeps constant (and ideally low) the number of visuospatial
elements to process. The inuence of visuospatial processing on signaling is
described next.
5.7.2 Visuospatial Processing Inuencing theEffect
Münzer etal. (2009) investigated 94 university participants (77% females) studying
a multimedia module about the biological topic of synthesis and structure of the
ATP molecule. A non-signaled static pictures condition was compared to a signaled
static pictures format that included motion arrows. Also, a global score of visuospa-
tial processing was calculated by averaging the scores of the Card Rotations Test of
mental rotation and the Paper Folding Test of mental folding. Results showed that
visuospatial processing was a signicant predictor of learning with the signaled
format but not with the non-signaled depictions. Thus, in the signaled condition,
visuospatial processing may have helped to deal with the extra information of the
arrows, and this information was used to achieve better learning. In contrast, low
J. C. Castro-Alonso et al.

129
visuospatial processing students could not cope with both the visualizations and the
extra information of the arrows and were not beneted by these extra signals. With
the non-signaled statics, both low and high visuospatial processing showed similar
low performance, as there were no cues to produce positive signaling effects.
Imhof etal. (2013) investigated 71 university students (65% females) learning
about sh locomotion patterns from three types of computer static visualizations,
either presenting or not presenting static arrows signaling the sh motions. Also,
mental rotation with 3D gures and mental folding were measured with the Mental
Rotations Test and the Paper Folding Test, respectively. Results showed that adding
signaling arrows to multiple visualizations was less effective than adding the arrows
to a single visualization. In other words, the multiple visualizations already pre-
sented useful information of the sh movements, so adding a signaling with added
elements was redundant and harmful for learning. Also, mental rotation and mental
folding did not interact with the signaling effects. Presenting higher visuospatial
processing scores predicted higher achievements in all conditions, independent of
the type of signaling and visualization.
Lee and Shin (2011) reported similar ndings in their study with 63 adult partici-
pants (44% females) learning about the four-stroke internal combustion engine. The
participants were randomly allocated to one of three instructional visualization con-
ditions: (a) static pictures, (b) static pictures signaled with the extra element of
motion arrows, and (c) animations. Also, a composite visuospatial processing score
was calculated with the results of a 3D mental rotation test (the Cube Comparisons
Test) and a mental folding instrument (the Paper Folding Test). The aggregated
score was used for a median split to compare high vs. low visuospatial students. It
was observed that higher visuospatial processing students outperformed their lower
counterparts in all three types of visualizations, including signaling or non- signaling
static picture conditions. In other words, including arrows to signal the relevant
movements in the engine system was not helpful for low visuospatial processing
students, because they performed poorly in both signaling or non-signaling static
picture conditions, signicantly lower than the high visuospatial processing partici-
pants. These examples suggest that visuospatial processing may be more inuential
than signaling with added elements to boost learning from science visualizations.
5.8 Transient Information Effect
Our last effect based on cognitive load theory is the transient information effect (see
Ayres and Paas 2007b). This occurs when videos or animations show information
that leaves the screen too rapidly for learners to cope with the pace and process it in
visuospatial working memory. The effect predicts that highly transient information
will be less effective for learning than less transient information (e.g., Castro-
Alonso etal. 2014b, 2018b). A similar negative impact of long narrations has also
been investigated (e.g., Singh etal. 2012; Wong etal. 2012).
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

130
Castro-Alonso etal. (2014a) described two general techniques to reduce the tran-
siency of instructional visualizations: pacing control and segmenting. When follow-
ing the pacing control technique, the students are given controls to adjust the pace
of the animation or video, such as pause and rewind buttons (see Mayer 2008).
Dynamic visualizations with features of pace control will contain less transient
information and thus be more effective than visualizations where pacing cannot be
controlled (see Ayres and Paas 2007a). When following the segmenting guideline,
shorter animations or videos are given to students, instead of a whole dynamic visu-
alization. By providing pauses between observations, students do not get cognitive
overloaded by accumulated transient information, and consequently, these shorter
segments are more efcient than a longer visualization (see Ayres and Paas 2007a).
In addition to the segmenting effectiveness due to lowering transient informa-
tion, there is also another benet of this technique: The segmenting of longer anima-
tions introduces pauses between visualizations, and these pauses can be used to
include additional learning activities (e.g., answering a short question, see Cheon
et al. 2014). Figure5.5 shows the pacing control and segmenting techniques to
avoid the problematic transient information in an animation.
Although both pacing control and segmenting reduce the transient information
of dynamic visualizations, Castro-Alonso etal. (2014a) reported a critical differ-
ence between them (see also Spanjers etal. 2010). This difference is in the agent
who segments the animation, as the student enacts pacing control, but the instruc-
tional designer produces segmenting. As such, segmenting could be more effective
than pacing control because it is an expert instructional designer who chooses to add
pauses in the best places for a meaningful presentation of the contents (see Spanjers
Fig. 5.5 Animations showing hydrogen bonds between molecules of water. Example of (a) whole
transient animation, (b) pacing controlled animation, and (c) two segmented (shorter) animations
J. C. Castro-Alonso et al.

131
etal. 2010). In contrast, relying on novice learners to add the pauses could halt the
continuity of the multimedia and lead to visualizations that are difcult to process.
Support for segmenting over pacing control is provided in the meta-analysis by
Rey etal. (2019), which included results of retention and transfer tests. For the pac-
ing control method, results revealed a non-signicant effect for retention tests
(d= 0.19), but a small to medium effect size for transfer tests (d=0.45). For the
segmenting technique, small to medium effect sizes were observed for both the
retention (d=0.42) and transfer (d=0.35) tests. Hence, both techniques were effec-
tive for transfer, but only segmenting was also effective for retention assessments.
Examples of these two techniques for science learning are provided next.
5.8.1 The Effect inOptimization ofScience Visualizations
An example of pacing control being effective for health and natural science educa-
tion is observed in Stiller etal. (2009), who reported a study of 110 university par-
ticipants (76% females) studying a multimedia presentation about the structure of
the human eye. Among the groups given the multimedia with explanatory on-screen
texts, a pace control condition was compared to a group showing an animation with-
out speed controls. The results revealed a higher performance of the students con-
trolling the pace of the multimedia screens. Importantly, this effect was not observed
among the groups given explanatory text as narrations, consistent with the modality
effect. In other words, the pace control effect was only observed in the conditions of
high visuospatial processing (learning depictions plus visual text) but not in the
conditions of lower visuospatial requirements (depictions plus auditory text). Only
in high processing conditions did learners need to control the pace of the multime-
dia in order to manage its cognitive load; in low processing conditions, learners
could manage the extra load imposed by the pace of the multimedia. These ndings
are also consistent with the modality effect.
Another positive example of pacing control is in the study by Höfer and
Schwartz (2011) with 82 university students (68% females) learning the chemical
process of dirt removal from a surface. In this experiment, the groups provided with
pause, rewind, and fast-forward features outperformed and self-reported less cogni-
tive load than those without these pacing controls.
The segmenting technique has also shown effectiveness in science education. For
example, Biard etal. (2018) investigated 68 occupational therapy undergraduates
(87% females) learning a medical hand procedure from videos in three formats. The
group assigned to the segmented videos outperformed the students in the two other
groups, namely: (a) whole videos that allowed pausing, and (b) whole videos with-
out segments. Thus, in line with the prediction by Spanjers etal. (2010) describe
above, the segmented videos were more effective than longer videos, either with or
without a pause feature.
Hasler etal. (2007) reported a study with 72 male primary students watching a
narrated animation about the causes of day and night. A condition of a whole con-
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

132
tinuous animation was compared to a segmented format with shorter animations.
For the difcult questions, it was observed that students learning from the seg-
mented visualizations surpassed those given the continuous animations. Hence,
when more working memory processing was needed to deal with difcult questions,
the segmenting method was most effective.
The segmenting technique has also been employed to investigate animations vs.
still pictures. In this case, the comparison is not made between longer and shorter
animations, but between animations and static pictures. Visuospatial processing has
also been inuential in these comparisons, described next.
5.8.2 Visuospatial Processing Inuencing theEffect
The transient information effect predicts that animations without pacing control will
be too demanding and thus less instructionally effective than static pictures, which
do not contain transient but permanent visuospatial information. Consequently, high
visuospatial processing students should be better prepared to deal with transient
animations, as compared to low visuospatial students, and the differences should be
less manifested in static picture conditions. This prediction has been investigated for
educational visualizations of both the health and natural sciences.
Regarding the health sciences, Nguyen etal. (2012) studied 60 adult participants,
among university students and staff (52% females), who learned the anatomical
structures of the esophagus, trachea, and mediastinal aorta. The dynamic condition
studied a video of the anatomical parts continuously rotating around the axes,
whereas the static condition studied six static standard views of the structures. The
participants were also measured in 3D mental rotation ability using a computerized
Mental Rotations Test. Results on the anatomy test showed that only in the dynamic
group high visuospatial processing students outperformed their lower peers. For the
statics condition, no differences were reported between high and low scorers in
mental rotation.
In an experiment with 29 university students, Loftus etal. (2018) used the same
visualization of the esophagus, trachea, and aorta as Nguyen etal. (2012). Loftus
and colleagues also incorporated a new visualization of ankle and foot bones. Both
anatomical models were shown rotating, in animations with no pacing control. A
median split from the scores on the paper Mental Rotations Test was used to divide
high and low visuospatial participants. Results showed that higher mental rotation
students outperform their lower peers in learning tasks with these highly transient
anatomical models. The largest effects favoring higher visuospatial learners were
observed in tasks demanding greater mental rotations of the models.
In a study with 49 rst-year kinesiology undergraduates (18% females), Berney
etal. (2015) randomly allocated students to either a static or a dynamic visualization
condition to learn topics about human shoulder anatomy. Also, the Mental Rotations
J. C. Castro-Alonso et al.

133
Test was used to measure mental rotation with 3D gures, and the Group Embedded
Figures Test was employed to assess eld independence. Considering the ve anat-
omy tasks investigated, mental rotation was a broader predictor, as it predicted per-
formance in three tasks, whereas eld independence predicted performance in two
tasks. This was observed for both static and animated versions.
Concerning natural science education, Aldahmash and Abraham (2009) investi-
gated 142 students in a university organic chemistry course. The topic focused on
reactions of nucleophilic substitution and elimination, and it was presented in either
animations or static pictures. Before students were randomly allocated to one of
these two visualization conditions, they were assessed in 3D mental rotations (the
Purdue Visualization of Rotations) and eld independence (the Find A Shape
Puzzle). A score of total visuospatial processing was calculated by averaging the
scores of both instruments. Results showed that the animated condition was more
effective than the static visualizations, and that this benecial learning effect of
animation was more pronounced in students with high visuospatial processing.
In an experiment with 198 university students (76% females), Kühl etal. (2018)
compared learning after watching a static picture versus an animation about the
velocity of planets orbiting the sun. The animation conditions outperformed the
static groups. In addition, the effect of mental folding, measured with the Paper
Folding Test, was also investigated. Notably, students lower in mental folding pre-
sented lower results in the astronomy topic when they learned it from the static
pictures, but presented similar results to higher mental folders with the animation.
Hence, the mental folding visuospatial processing was not as necessary for dynamic
visualizations as it was for statics.
In a more abstract task, Castro-Alonso etal. (2018b) investigated 104 university
students (50% females) memorizing positions and colors of symbols placed on the
screen. Every participant attempted the task in both a static format (symbols shown
simultaneously) and a transient format (symbols shown consecutively). As pre-
dicted by the transient information effect, the outcomes were higher in the static as
compared to the transient condition. Also, mental rotation with 2D shapes was mea-
sured with the Card Rotations Test and spatial working memory was measured with
a computer Corsi Block Tapping Test. Multiple regression analyses showed that (a)
mental rotation was a signicant predictor of achievement in both the static and the
transient tasks, and (b) spatial working memory was a close to signicant predictor
of performance in the transient but not in the static task.
Summarizing these results of health and natural sciences, we can conclude that
visuospatial processing is a benecial asset for comprehending science visualiza-
tions and multimedia. We can also draw conclusions about the prediction of the
transient information effect, forecasting that visuospatial processing would be more
important to deal with transient visualizations than with static pictures. The studies
presented here show that this prediction needs further investigation, as both dynamic
and static visualizations were supported by visuospatial abilities.
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

134
5.9 Discussion
In addition to being engaging and fun, visualizations can convey a processing
advantage. For example, adequately designed instructional visualizations can help
to learn from science texts, especially if the concepts are complex and need such
learning scaffolds. As mentioned, visualizations need to be properly designed to
accomplish these instructional benets. We believe that proper instructional designs
are those that follow an understanding of the human cognitive architecture applied
to learning. Cognitive load theory and also the cognitive theory of multimedia learn-
ing can provide this understanding. Moreover, as these two cognitive theories inves-
tigate domain-specic biologically secondary skills, they are especially pertinent in
designing visualizations for science education. In these learning situations, there are
working memory limitations affecting visuospatial processing to deal with the
visual and spatial information of instructional visualizations.
We considered ve effects identied by these two cognitive theories, which can
be applied to optimize static and dynamic instructional visualizations. These effects
or principles are: (a) the split attention effect and spatial contiguity principle, (b) the
modality effect, (c) the redundancy effect and coherence principle, (d) the signaling
principle, and (e) the transient information effect. We presented examples for health
and natural sciences instructional visualizations for each of these ve methods. We
also considered the role of visuospatial processing abilities, measured in standard
tests that assess mental rotation, mental folding, and eld independence, among
others. In general, the higher the visuospatial processing ability of the learner, the
greater the learning from instructional visualizations.
5.9.1 Instructional Implications forHealth andNatural
Sciences
A rst implication for learning about the health and natural sciences is that visual-
izations should be designed according to cognitive principles rather than just on
engagement variables. In other words, the processing advantages of instructional
visualizations, such as providing scaffolds for textual information, should be
utilized.
A second instructional implication is that applying the guidelines from cognitive
load theory (and similar cognitive theories) can help optimize learning from science
visualizations.
A third instructional implication is that measuring visuospatial processing can
also help predict the most effective learning from instructional visualizations,
including static and dynamic formats. The general trend is that higher visuospatial
processing scores result in greater learning from instructional visualizations, but the
design of the visualizations can affect this trend.
J. C. Castro-Alonso et al.

135
A fourth implication, concerning the split attention effect, is that visualizations
should not be designed with their information separated but integrated spatially. For
example, visual elements and their text labels should follow the spatial contiguity
principle and be placed near each other. Also, low visuospatial processing students
are particularly challenged by the separated placement of visuospatial information
in visualizations.
A fth implication, considering the modality effect, is that computer visualiza-
tions should be designed with auditory descriptions, rather than too much informa-
tion on-screen. By providing narrations, the auditory processing is used and
visuospatial processing is left to manage the key visual elements and not also writ-
ten explanations. Narrations should be short, to avoid an adverse transient informa-
tion effect.
A sixth implication concerns the redundancy effect. Extra information that is not
fundamental for the learning topic should be deleted from the visualizations. This is
especially important for interesting redundant supplements, as these can be more
disruptive than less interesting additions. Moreover, low visuospatial processing
students are particularly challenged by unnecessary visuospatial information, as it
interferes more with the critical learning information.
A seventh implication, considering the signaling principle, is that visualizations
could be more effective by including elements to cue the most important parts.
Using signals that do not add extra visual elements, such as color, zooming, or trans-
parency differences, could be the most effective.
The last implication deals with the transient information effect. Sometimes, long
dynamic visualizations whose pace cannot be controlled should be changed to pace
control formats, shorter versions, or static visualizations. Similarly, long narrations
should be shortened.
5.9.2 Future Research Directions
We note some future directions that research into science visualizations and visuo-
spatial processing might follow. The rst direction goes beyond the emphasis of this
chapter and of cognitive load theory. When learning through visualizations, there
are factors besides cognitive processes to be considered, such as behavioral, motiva-
tional, and emotional inuences (cf. Fraser etal. 2015). For example, future research
could include emotions as variables that inuence perceived cognitive load when
learning from science visualizations or simulations (e.g., Fraser etal. 2014).
A second direction is based on the point that visualization and multimedia
researchers have commonly measured the visuospatial processing abilities of men-
tal rotations with 3D gures and mental folding. However, there are many other
visuospatial tasks, such as those measured in tests of 2D mental rotation, eld inde-
pendence, and other visuospatial working memory instruments (see Castro-Alonso
and Atit this volume, Chap. 2; see also Castro-Alonso etal. this volume-a, Chap. 8).
Future research could investigate the effects of these other visuospatial abilities in
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

136
learning from science visualizations optimized by cognitive load theory effects. An
analogous direction for further research involves studying the effects of visuospatial
tests on particular scientic visual and spatial tasks (see Castro-Alonso etal. 2019a).
Third, concerning the modality and signaling principles, there are not many
examples of visuospatial processing inuencing these principles for science visual-
izations. Thus, future investigations could study how visuospatial processing inter-
acts with the benecial effects of the modality and the signaling principles when
learning through visualizations depicting health and natural science topics.
A fourth direction, regarding the transient information effect, is that the predic-
tion that visuospatial processing would be more important for animations and vid-
eos than for static pictures, was not supported by the research reported here. On the
contrary, the studies reviewed in this chapter suggest that visuospatial processing is
benecial for both dynamic and static visualizations. Further investigations about
static vs. dynamic visualizations can help reach a stronger prediction whether visuo-
spatial processing is more helpful for static, dynamic, or both formats of visualiza-
tions. These efforts would also be benetted by future investigations tackling
moderating variables for static vs. animation research, such as gender (e.g., Castro-
Alonso etal. 2019b), the design of the dynamic or static images (see Castro-Alonso
etal. 2016), and the strategies learners employ when studying visualizations (see
Ayres etal. 2019).
5.9.3 Conclusion
Properly designed instructional static and dynamic visualizations can be useful
assets to health and natural science learning. To be effective, the designs should fol-
low guidelines of cognitive load theory and other approaches that investigate the
human cognitive architecture for learning. Furthermore, the limitations of visuospa-
tial processing when managing novel visual and spatial information should also be
considered in order to design effective visualizations. In this chapter, we described
ve guidelines, based on well-researched effects, to optimize visualizations, includ-
ing the inuence of visuospatial processing, and providing applications for health
and natural sciences education. The guidelines were based on: (a) the split attention
effect and spatial contiguity principle, (b) the modality effect, (c) the redundancy
effect and coherence principle, (d) the signaling principle, and (e) the transient
information effect. Regarding education for health and natural sciences, often these
guidelines were most important for low visuospatial processing students.
Acknowledgments This research was partially supported by funds from PIA–CONICYT Basal
Funds for Centers of Excellence, Project FB0003, and by a 2018 research grant from the School of
Education, University of New South Wales. The rst author is thankful to Mariana Poblete for her
assistance.
J. C. Castro-Alonso et al.

137
References
Aldahmash, A.H., & Abraham, M.R. (2009). Kinetic versus static visuals for facilitating college
students’ understanding of organic reaction mechanisms in chemistry. Journal of Chemical
Education, 86(12), 1442–1446. https://doi.org/10.1021/ed086p1442.
Amadieu, F., Mariné, C., & Laimay, C. (2011). The attention-guiding effect and cognitive load in
the comprehension of animations. Computers in Human Behavior, 27(1), 36–40. https://doi.
org/10.1016/j.chb.2010.05.009.
Ayres, P., & Paas, F. (2007a). Can the cognitive load approach make instructional animations more
effective? Applied Cognitive Psychology, 21(6), 811–820. https://doi.org/10.1002/acp.1351.
Ayres, P., & Paas, F. (2007b). Making instructional animations more effective: A cognitive load
approach. Applied Cognitive Psychology, 21(6), 695–700. https://doi.org/10.1002/acp.1343.
Ayres, P., & Sweller, J.(2014). The split-attention principle in multimedia learning. In R.E. Mayer
(Ed.), The Cambridge handbook of multimedia learning (2nd ed., pp. 206–226). NewYork:
Cambridge University Press. https://doi.org/10.1017/CBO9781139547369.011.
Ayres, P., Castro-Alonso, J.C., Wong, M., Marcus, N., & Paas, F. (2019). Factors that impact on
the effectiveness of instructional animations. In S.Tindall-Ford, S.Agostinho, & J. Sweller
(Eds.), Advances in cognitive load theory: Rethinking teaching (pp. 180–193). New York:
Routledge.https://doi.org/10.4324/9780429283895-15.
Bauhoff, V., Huff, M., & Schwan, S. (2012). Distance matters: Spatial contiguity effects as trade-
off between gaze switches and memory load. Applied Cognitive Psychology, 26(6), 863–871.
https://doi.org/10.1002/acp.2887.
Berney, S., Bétrancourt, M., Molinari, G., & Hoyek, N. (2015). How spatial abilities and dynamic
visualizations interplay when learning functional anatomy with 3D anatomical models.
Anatomical Sciences Education, 8(5), 452–462. https://doi.org/10.1002/ase.1524.
Biard, N., Cojean, S., & Jamet, E. (2018). Effects of segmentation and pacing on procedural
learning by video. Computers in Human Behavior, 89, 411–417. https://doi.org/10.1016/j.
chb.2017.12.002.
Brady, T.F., Konkle, T., Alvarez, G.A., & Oliva, A. (2008). Visual long-term memory has a mas-
sive storage capacity for object details. Proceedings of the National Academy of Sciences,
105(38), 14325–14329. https://doi.org/10.1073/pnas.0803390105.
Brünken, R., Steinbacher, S., Plass, J.L., & Leutner, D. (2002). Assessment of cognitive load in
multimedia learning using dual-task methodology. Experimental Psychology, 49(2), 109–119.
https://doi.org/10.1027//1618-3169.49.2.109.
Butcher, K.R. (2014). The multimedia principle. In R.E. Mayer (Ed.), The Cambridge handbook
of multimedia learning (2nd ed., pp.174–205). NewYork: Cambridge University Press. https://
doi.org/10.1017/CBO9781139547369.010.
Calkins, M. W. (1898). Short studies in memory and in association from the Wellesly College
Psychological Laboratory. Psychological Review, 5(5), 451–462. https://doi.org/10.1037/
h0071176.
Castro-Alonso, J.C., & Atit, K. (this volume). Different abilities controlled by visuospatial pro-
cessing. In J.C. Castro-Alonso (Ed.), Visuospatial processing for education in health and natu-
ral sciences (pp. 23–51). Cham: Springer.https://doi.org/10.1007/978-3-030-20969-8_2.
Castro-Alonso, J.C., & Fiorella, L. (this volume). Interactive science multimedia and visuospatial
processing. In J.C. Castro-Alonso (Ed.), Visuospatial processing for education in health and
natural sciences (pp. 145–173). Cham: Springer.https://doi.org/10.1007/978-3-030-20969-8_6.
Castro-Alonso, J.C., & Uttal, D.H. (this volume). Science education and visuospatial processing.
In J.C. Castro-Alonso (Ed.), Visuospatial processing for education in health and natural sci-
ences (pp. 53–79). Cham: Springer.https://doi.org/10.1007/978-3-030-20969-8_3.
Castro-Alonso, J.C., Ayres, P., & Paas, F. (2014a). Dynamic visualisations and motor skills. In
W.Huang (Ed.), Handbook of human centric visualization (pp.551–580). NewYork: Springer.
https://doi.org/10.1007/978-1-4614-7485-2_22.
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

138
Castro-Alonso, J.C., Ayres, P., & Paas, F. (2014b). Learning from observing hands in static and
animated versions of non-manipulative tasks. Learning and Instruction, 34, 11–21. https://doi.
org/10.1016/j.learninstruc.2014.07.005.
Castro-Alonso, J.C., Ayres, P., & Paas, F. (2016). Comparing apples and oranges? A critical look
at research on learning from statics versus animations. Computers & Education, 102, 234–243.
https://doi.org/10.1016/j.compedu.2016.09.004.
Castro-Alonso, J. C., Ayres, P., & Paas, F. (2018a). Computerized and adaptable tests to mea-
sure visuospatial abilities in STEM students. In T.Andre (Ed.), Advances in human factors
in training, education, and learning sciences: Proceedings of the AHFE 2017 International
Conference on Human Factors in Training, Education, and Learning Sciences (pp.337–349).
Cham: Springer. https://doi.org/10.1007/978-3-319-60018-5_33.
Castro-Alonso, J.C., Ayres, P., Wong, M., & Paas, F. (2018b). Learning symbols from permanent
and transient visual presentations: Don’t overplay the hand. Computers & Education, 116,
1–13. https://doi.org/10.1016/j.compedu.2017.08.011.
Castro-Alonso, J.C., Ayres, P., Wong, M., & Paas, F. (2019a). Visuospatial tests and multimedia
learning: The importance of employing relevant instruments. In S.Tindall-Ford, S.Agostinho,
& J. Sweller (Eds.), Advances in cognitive load theory: Rethinking teaching (pp. 89–99).
NewYork: Routledge. https://doi.org/10.4324/9780429283895-8.
Castro-Alonso, J.C., Wong, M., Adesope, O.O., Ayres, P., & Paas, F. (2019b). Gender imbalance
in instructional dynamic versus static visualizations: A meta-analysis. Educational Psychology
Review,31(2), 361–387. https://doi.org/10.1007/s10648-019-09469-1.
Castro-Alonso, J. C., Ayres, P., & Paas, F. (this volume-a). VAR: A battery of computer-based
instruments to measure visuospatial processing. In J.C. Castro-Alonso (Ed.), Visuospatial pro-
cessing for education in health and natural sciences (pp. 207–229). Cham: Springer.https://
doi.org/10.1007/978-3-030-20969-8_8.
Castro-Alonso, J. C., Paas, F., & Ginns, P. (this volume-b). Embodied cognition, science edu-
cation, and visuospatial processing. In J. C. Castro-Alonso (Ed.), Visuospatial processing
for education in health and natural sciences (pp. 175–205). Cham: Springer. https://doi.
org/10.1007/978-3-030-20969-8_7.
Chandler, P., & Sweller, J.(1991). Cognitive load theory and the format of instruction. Cognition
and Instruction, 8(4), 293–332.
Chen, O., Castro-Alonso, J. C., Paas, F., & Sweller, J. (2018). Extending cognitive load the-
ory to incorporate working memory resource depletion: Evidence from the spacing effect.
Educational Psychology Review, 30(2), 483–501. https://doi.org/10.1007/s10648-017-9426-2.
Cheon, J., Crooks, S., & Chung, S. (2014). Does segmenting principle counteract modality prin-
ciple in instructional animation? British Journal of Educational Technology, 45(1), 56–64.
https://doi.org/10.1111/bjet.12021.
Cierniak, G., Scheiter, K., & Gerjets, P. (2009). Explaining the split-attention effect: Is the reduc-
tion of extraneous cognitive load accompanied by an increase in germane cognitive load?
Computers in Human Behavior, 25(2), 315–324. https://doi.org/10.1016/j.chb.2008.12.020.
Clark, J., & Paivio, A. (1991). Dual coding theory and education. Educational Psychology Review,
3(3), 149–210. https://doi.org/10.1007/bf01320076.
Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale:
Erlbaum.
Cowan, N. (2001). The magical number 4in short-term memory: A reconsideration of mental
storage capacity. Behavioral and Brain Sciences, 24(1), 87–185. https://doi.org/10.1017/
S0140525X01003922.
de Koning, B.B., Tabbers, H.K., Rikers, R.M. J.P., & Paas, F. (2009). Towards a framework for
attention cueing in instructional animations: Guidelines for research and design. Educational
Psychology Review, 21(2), 113–140. https://doi.org/10.1007/s10648-009-9098-7.
de Koning, B.B., Tabbers, H.K., Rikers, R.M. J.P., & Paas, F. (2010). Learning by generating vs.
receiving instructional explanations: Two approaches to enhance attention cueing in animations.
Computers & Education, 55(2), 681–691. https://doi.org/10.1016/j.compedu.2010.02.027.
J. C. Castro-Alonso et al.

139
Dutke, S., & Rinck, M. (2006). Multimedia learning: Working memory and the learning of word
and picture diagrams. Learning and Instruction, 16(6), 526–537. https://doi.org/10.1016/j.
learninstruc.2006.10.002.
Eitel, A., Scheiter, K., Schüler, A., Nyström, M., & Holmqvist, K. (2013). How a picture facili-
tates the process of learning from text: Evidence for scaffolding. Learning and Instruction, 28,
48–63. https://doi.org/10.1016/j.learninstruc.2013.05.002.
Eitel, A., Bender, L., & Renkl, A. (2019). Are seductive details seductive only when you think
they are relevant? An experimental test of the moderating role of perceived relevance. Applied
Cognitive Psychology, 33(1), 20–30. https://doi.org/10.1002/acp.3479.
Erhel, S., & Jamet, E. (2006). Using pop-up windows to improve multime-
dia learning. Journal of Computer Assisted Learning, 22(2), 137–147. https://doi.
org/10.1111/j.1365-2729.2006.00165.x.
Fenesi, B., Kramer, E., & Kim, J.A. (2016). Split-attention and coherence principles in multi-
media instruction can rescue performance for learners with lower working memory capacity.
Applied Cognitive Psychology, 30(5), 691–699. https://doi.org/10.1002/acp.3244.
Fraser, K.L., Huffman, J., Ma, I., Sobczak, M., McIlwrick, J., Wright, B., etal. (2014). The emo-
tional and cognitive impact of unexpected simulated patient death: A randomized controlled
trial. Chest, 145(5), 958–963. https://doi.org/10.1378/chest.13-0987.
Fraser, K. L., Ayres, P., & Sweller, J.(2015). Cognitive load theory for the design of medi-
cal simulations. Simulation in Healthcare, 10(5), 295–307. https://doi.org/10.1097/
sih.0000000000000097.
Garg, A. X., Norman, G., Spero, L., & Taylor, I. (1999). Learning anatomy: Do new com-
puter models improve spatial understanding? Medical Teacher, 21(5), 519–522. https://doi.
org/10.1080/01421599979239.
Geary, D.C. (2002). Principles of evolutionary educational psychology. Learning and Individual
Differences, 12(4), 317–345. https://doi.org/10.1016/s1041-6080(02)00046-8.
Geary, D.C. (2007). Educating the evolved mind: Conceptual foundations for an evolutionary
educational psychology. In J.S. Carlson & J.R. Levin (Eds.), Psychological perspectives on
contemporary educational issues (pp.1–99). Charlotte: Information Age Publishing.
Geary, D.C., & Berch, D.B. (2016). Evolution and children’s cognitive and academic develop-
ment. In D.C. Geary & D.B. Berch (Eds.), Evolutionary perspectives on child development
and education (pp.217–249). Cham: Springer. https://doi.org/10.1007/978-3-319-29986-0_9.
Ginns, P. (2005). Meta-analysis of the modality effect. Learning and Instruction, 15(4), 313–331.
https://doi.org/10.1016/j.learninstruc.2005.07.001.
Ginns, P. (2006). Integrating information: A meta-analysis of the spatial contiguity and temporal
contiguity effects. Learning and Instruction, 16(6), 511–525.
Harp, S.F., & Mayer, R.E. (1997). The role of interest in learning from scientic text and illus-
trations: On the distinction between emotional interest and cognitive interest. Journal of
Educational Psychology, 89(1), 92–102. https://doi.org/10.1037/0022-0663.89.1.92.
Harp, S.F., & Mayer, R.E. (1998). How seductive details do their damage: A theory of cognitive
interest in science learning. Journal of Educational Psychology, 90(3), 414–434. https://doi.
org/10.1037/0022-0663.90.3.414.
Hasler, B.S., Kersten, B., & Sweller, J.(2007). Learner control, cognitive load and instructional
animation. Applied Cognitive Psychology, 21(6), 713–729. https://doi.org/10.1002/acp.1345.
Hegarty, M. (2011). The cognitive science of visual-spatial displays: Implications for design.
Topics in Cognitive Science, 3(3), 446–474. https://doi.org/10.1111/j.1756-8765.2011.01150.x.
Höfer, T. N., & Schwartz, R. N. (2011). Effects of pacing and cognitive style across dynamic
and non-dynamic representations. Computers & Education, 57(2), 1716–1726. https://doi.
org/10.1016/j.compedu.2011.03.012.
Hosler, J., Boomer, K.B., & Kalumuck, K. (2011). Are comic books an effective way to engage
nonmajors in learning and appreciating science? CBE Life Sciences Education, 10(3), 309–
317. https://doi.org/10.1187/cbe.10-07-0090.
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

140
Houts, P.S., Doak, C.C., Doak, L.G., & Loscalzo, M.J. (2006). The role of pictures in improv-
ing health communication: A review of research on attention, comprehension, recall, and
adherence. Patient Education and Counseling, 61(2), 173–190. https://doi.org/10.1016/j.
pec.2005.05.004.
Huff, M., & Schwan, S. (2011). Integrating information from two pictorial animations: Complexity
and cognitive prerequisites inuence performance. Applied Cognitive Psychology, 25(6), 878–
886. https://doi.org/10.1002/acp.1762.
Huff, M., Bauhoff, V., & Schwan, S. (2012). Effects of split attention revisited: A new display tech-
nology for troubleshooting tasks. Computers in Human Behavior, 28(4), 1254–1261. https://
doi.org/10.1016/j.chb.2012.02.008.
Imhof, B., Scheiter, K., Edelmann, J., & Gerjets, P. (2013). Learning about locomotion patterns:
Effective use of multiple pictures and motion-indicating arrows. Computers & Education, 65,
45–55. https://doi.org/10.1016/j.compedu.2013.01.017.
Issa, N., Schuller, M., Santacaterina, S., Shapiro, M., Wang, E., Mayer, R. E., et al. (2011).
Applying multimedia design principles enhances learning in medical education. Medical
Education, 45(8), 818–826. https://doi.org/10.1111/j.1365-2923.2011.03988.x.
Jaffar, A. A. (2012). YouTube: An emerging tool in anatomy education. Anatomical Sciences
Education, 5(3), 158–164. https://doi.org/10.1002/ase.1268.
Kalyuga, S., & Sweller, J.(2014). The redundancy principle in multimedia learning. In R.E. Mayer
(Ed.), The Cambridge handbook of multimedia learning (2nd ed., pp. 247–262). NewYork:
Cambridge University Press. https://doi.org/10.1017/CBO9781139547369.013.
Kalyuga, S., Chandler, P., & Sweller, J.(1999). Managing split-attention and redundancy in multi-
media instruction. Applied Cognitive Psychology, 13(4), 351–371.
Korbach, A., Brünken, R., & Park, B. (2016). Learner characteristics and information processing
in multimedia learning: A moderated mediation of the seductive details effect. Learning and
Individual Differences, 51, 59–68. https://doi.org/10.1016/j.lindif.2016.08.030.
Kühl, T., Scheiter, K., Gerjets, P., & Edelmann, J.(2011). The inuence of text modality on learn-
ing with static and dynamic visualizations. Computers in Human Behavior, 27(1), 29–35.
https://doi.org/10.1016/j.chb.2010.05.008.
Kühl, T., Stebner, F., Navratil, S.C., Fehringer, B.C. O.F., & Münzer, S. (2018). Text information
and spatial abilities in learning with different visualizations formats. Journal of Educational
Psychology, 110(4), 561–577. https://doi.org/10.1037/edu0000226.
Larkin, J., McDermott, J., Simon, D. P., & Simon, H. A. (1980). Expert and novice perfor-
mance in solving physics problems. Science, 208(4450), 1335–1342. https://doi.org/10.1126/
science.208.4450.1335.
Lee, D.Y., & Shin, D.-H. (2011). Effects of spatial ability and richness of motion cue on learning
in mechanically complex domain. Computers in Human Behavior, 27(5), 1665–1674. https://
doi.org/10.1016/j.chb.2011.02.005.
Lee, D. Y., & Shin, D.-H. (2012). An empirical evaluation of multi-media based learning of a
procedural task. Computers in Human Behavior, 28(3), 1072–1081. https://doi.org/10.1016/j.
chb.2012.01.014.
Levinson, A.J., Weaver, B., Garside, S., McGinn, H., & Norman, G.R. (2007). Virtual reality
and brain anatomy: A randomised trial of e-learning instructional designs. Medical Education,
41(5), 495–501. https://doi.org/10.1111/j.1365-2929.2006.02694.x.
Lin, L., & Atkinson, R.K. (2011). Using animations and visual cueing to support learning of
scientic concepts and processes. Computers & Education, 56(3), 650–658. https://doi.
org/10.1016/j.compedu.2010.10.007.
Loftus, J.J., Jacobsen, M., & Wilson, T.D. (2018). The relationship between spatial ability, cere-
bral blood ow and learning with dynamic images: A transcranial Doppler ultrasonography
study. Medical Teacher, 40(2), 174–180. https://doi.org/10.1080/0142159X.2017.1395401.
Low, R., & Sweller, J.(2014). The modality principle in multimedia learning. In R. E. Mayer
(Ed.), The Cambridge handbook of multimedia learning (2nd ed., pp. 227–246). NewYork:
Cambridge University Press. https://doi.org/10.1017/CBO9781139547369.012.
J. C. Castro-Alonso et al.

141
Luck, S.J., & Vogel, E.K. (1997). The capacity of visual working memory for features and con-
junctions. Nature, 390(6657), 279–281. https://doi.org/10.1038/36846.
Mahmud, W., Hyder, O., Butt, J., & Aftab, A. (2011). Dissection videos do not improve anatomy
examination scores. Anatomical Sciences Education, 4(1), 16–21. https://doi.org/10.1002/
ase.194.
Makransky, G., Terkildsen, T.S., & Mayer, R.E. (2019). Role of subjective and objective measures
of cognitive processing during learning in explaining the spatial contiguity effect. Learning
and Instruction, 61, 23–34. https://doi.org/10.1016/j.learninstruc.2018.12.001.
Mayer, R.E. (1989). Systematic thinking fostered by illustrations in scientic text. Journal of
Educational Psychology, 81(2), 240–246. https://doi.org/10.1037/0022-0663.81.2.240.
Mayer, R.E. (2008). Research-based principles for learning with animation. In R. K. Lowe &
W.Schnotz (Eds.), Learning with animation: Research implications for design (pp.30–48).
NewYork: Cambridge University Press.
Mayer, R.E. (Ed.). (2014a). The Cambridge handbook of multimedia learning (2nd ed.). NewYork:
Cambridge University Press.
Mayer, R. E. (2014b). Cognitive theory of multimedia learning. In R. E. Mayer (Ed.), The
Cambridge handbook of multimedia learning (2nd ed., pp. 43–71). New York: Cambridge
University Press. https://doi.org/10.1017/CBO9781139547369.005.
Mayer, R.E., & Gallini, J.K. (1990). When is an illustration worth ten thousand words? Journal of
Educational Psychology, 82(4), 715–726. https://doi.org/10.1037/0022-0663.82.4.715.
Mayer, R.E., Grifth, E., Jurkowitz, I.T. N., & Rothman, D. (2008). Increased interestingness of
extraneous details in a multimedia science presentation leads to decreased learning. Journal of
Experimental Psychology: Applied, 14(4), 329–339. https://doi.org/10.1037/a0013835.
Moreno, R., & Mayer, R.E. (1999). Cognitive principles of multimedia learning: The role of
modality and contiguity. Journal of Educational Psychology, 91(2), 358–368. https://doi.
org/10.1037/0022-0663.91.2.358.
Mousavi, S.Y., Low, R., & Sweller, J.(1995). Reducing cognitive load by mixing auditory and
visual presentation modes. Journal of Educational Psychology, 87(2), 319–334. https://doi.
org/10.1037/0022-0663.87.2.319.
Münzer, S., Seufert, T., & Brünken, R. (2009). Learning from multimedia presentations:
Facilitation function of animations and spatial abilities. Learning and Individual Differences,
19(4), 481–485. https://doi.org/10.1016/j.lindif.2009.05.001.
Nguyen, N., Nelson, A.J., & Wilson, T. D. (2012). Computer visualizations: Factors that inu-
ence spatial anatomy comprehension. Anatomical Sciences Education, 5(2), 98–108. https://
doi.org/10.1002/ase.1258.
Oberauer, K., & Eichenberger, S. (2013). Visual working memory declines when more features
must be remembered for each object. Memory & Cognition, 41(8), 1212–1227. https://doi.
org/10.3758/s13421-013-0333-6.
Ong, W.J. (1982). Orality and literacy: The technologizing of the word. NewYork: Methuen.
Paas, F., & Sweller, J.(2012). An evolutionary upgrade of cognitive load theory: Using the human
motor system and collaboration to support the learning of complex cognitive tasks. Educational
Psychology Review, 24(1), 27–45. https://doi.org/10.1007/s10648-011-9179-2.
Paas, F., & Sweller, J.(2014). Implications of cognitive load theory for multimedia learning. In
R.E. Mayer (Ed.), The Cambridge handbook of multimedia learning (2nd ed., pp.27–42).
NewYork: Cambridge University Press. https://doi.org/10.1017/CBO9781139547369.004.
Penney, C.G. (1989). Modality effects and the structure of short-term verbal memory. Memory &
Cognition, 17(4), 398–422. https://doi.org/10.3758/bf03202613.
Purnell, K.N., Solman, R.T., & Sweller, J.(1991). The effects of technical illustrations on cogni-
tive load. Instructional Science, 20(5), 443–462. https://doi.org/10.1007/bf00116358.
Rey, G. D. (2012). A review of research and a meta-analysis of the seductive detail effect.
Educational Research Review, 7(3), 216–237. https://doi.org/10.1016/j.edurev.2012.05.003.
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing

142
Rey, G.D., Beege, M., Nebel, S., Wirzberger, M., Schmitt, T.H., & Schneider, S. (2019). A meta-
analysis of the segmenting effect. Educational Psychology Review,31(2), 389–419. https://doi.
org/10.1007/s10648-018-9456-4.
Richter, J., Scheiter, K., & Eitel, A. (2016). Signaling text-picture relations in multimedia learn-
ing: A comprehensive meta-analysis. Educational Research Review, 17, 19–36. https://doi.
org/10.1016/j.edurev.2015.12.003.
Schmidt-Weigand, F., Kohnert, A., & Glowalla, U. (2010). A closer look at split visual attention
in system- and self-paced instruction in multimedia learning. Learning and Instruction, 20(2),
100–110. https://doi.org/10.1016/j.learninstruc.2009.02.011.
Schneider, S., Beege, M., Nebel, S., & Rey, G. D. (2018). A meta-analysis of how signaling
affects learning with media. Educational Research Review, 23, 1–24. https://doi.org/10.1016/j.
edurev.2017.11.001.
Schroeder, N.L., & Cenkci, A.T. (2018). Spatial contiguity and spatial split-attention effects in
multimedia learning environments: A meta-analysis. Educational Psychology Review, 30(3),
679–701. https://doi.org/10.1007/s10648-018-9435-9.
Seufert, T., Schütze, M., & Brünken, R. (2009). Memory characteristics and modality in multime-
dia learning: An aptitude-treatment-interaction study. Learning and Instruction, 19(1), 28–42.
https://doi.org/10.1016/j.learninstruc.2008.01.002.
Shepard, R.N. (1967). Recognition memory for words, sentences, and pictures. Journal of Verbal
Learning and Verbal Behavior, 6(1), 156–163. https://doi.org/10.1016/s0022-5371(67)80067-7.
Singh, A.-M., Marcus, N., & Ayres, P. (2012). The transient information effect: Investigating the
impact of segmentation on spoken and written text. Applied Cognitive Psychology, 26(6), 848–
853. https://doi.org/10.1002/acp.2885.
Skulmowski, A., & Rey, G. D. (2018). Realistic details in visualizations require color cues
to foster retention. Computers & Education, 122, 23–31. https://doi.org/10.1016/j.
compedu.2018.03.012.
Spanjers, I.A. E., van Gog, T., & van Merriënboer, J.J. G. (2010). A theoretical analysis of how
segmentation of dynamic visualizations optimizes students’ learning. Educational Psychology
Review, 22(4), 411–423. https://doi.org/10.1007/s10648-010-9135-6.
Stiller, K.D., Freitag, A., Zinnbauer, P., & Freitag, C. (2009). How pacing of multimedia instruc-
tions can inuence modality effects: A case of superiority of visual texts. Australasian Journal
of Educational Technology, 25(2), 184–203. https://doi.org/10.14742/ajet.1149.
Stull, A.T., Gainer, M.J., & Hegarty, M. (2018). Learning by enacting: The role of embodi-
ment in chemistry education. Learning and Instruction, 55, 80–92. https://doi.org/10.1016/j.
learninstruc.2017.09.008.
Sweller, J.(2015). In academe, what is learned, and how is it learned? Current Directions in
Psychological Science, 24(3), 190–194. https://doi.org/10.1177/0963721415569570.
Sweller, J.(2016). Cognitive load theory, evolutionary educational psychology, and instructional
design. In D.C. Geary & D.B. Berch (Eds.), Evolutionary perspectives on child development
and education (pp.291–306). Cham: Springer. https://doi.org/10.1007/978-3-319-29986-0_12.
Sweller, J., & Sweller, S. (2006). Natural information processing systems. Evolutionary
Psychology, 4, 434–458.
Sweller, J., Ayres, P., & Kalyuga, S. (2011). Cognitive load theory. NewYork: Springer.
Sweller, J., van Merriënboer, J. J. G., & Paas, F. (2019). Cognitive architecture and instruc-
tional design: 20 years later. Educational Psychology Review, 31(2), 261–292. https://doi.
org/10.1007/s10648-019-09465-5.
Tarmizi, R.A., & Sweller, J.(1988). Guidance during mathematical problem solving. Journal of
Educational Psychology, 80(4), 424–436. https://doi.org/10.1037/0022-0663.80.4.424.
Tricot, A., & Sweller, J. (2014). Domain-specic knowledge and why teaching generic skills
does not work. Educational Psychology Review, 26(2), 265–283. https://doi.org/10.1007/
s10648-013-9243-1.
J. C. Castro-Alonso et al.

143
van Gog, T. (2014). The signaling (or cueing) principle in multimedia learning. In R.E. Mayer
(Ed.), The Cambridge handbook of multimedia learning (2nd ed., pp. 263–278). NewYork:
Cambridge University Press. https://doi.org/10.1017/CBO9781139547369.014.
Vekiri, I. (2002). What is the value of graphical displays in learning? Educational Psychology
Review, 14(3), 261–312. https://doi.org/10.1023/a:1016064429161.
Wiegmann, D.A., Dansereau, D. F., McCagg, E.C., Rewey, K.L., & Pitre, U. (1992). Effects
of knowledge map characteristics on information processing. Contemporary Educational
Psychology, 17(2), 136–155. https://doi.org/10.1016/0361-476X(92)90055-4.
Wilson, T. D. (2015). Role of image and cognitive load in anatomical multimedia. In L.K. Chan
& W.Pawlina (Eds.), Teaching anatomy: A practical guide (pp.237–246). Cham: Springer.
https://doi.org/10.1007/978-3-319-08930-0_27.
Wong, A., Leahy, W., Marcus, N., & Sweller, J.(2012). Cognitive load theory, the transient informa-
tion effect and e-learning. Learning and Instruction, 22(6), 449–457. https://doi.org/10.1016/j.
learninstruc.2012.05.004.
Xie, H., Wang, F., Hao, Y., Chen, J., An, J., Wang, Y., etal. (2017). The more total cognitive load
is reduced by cues, the better retention and transfer of multimedia learning: A meta-analysis
and two meta-regression analyses. PLoS One, 12(8), e0183884. https://doi.org/10.1371/jour-
nal.pone.0183884.
Youssef, A., Ayres, P., & Sweller, J.(2012). Using general problem-solving strategies to generate
ideas in order to solve geography problems. Applied Cognitive Psychology, 26(6), 872–877.
https://doi.org/10.1002/acp.2888.
5 Instructional Visualizations, Cognitive Load Theory, andVisuospatial Processing