Content uploaded by Richard Mayer
Author content
All content in this area was uploaded by Richard Mayer on Jul 03, 2016
Content may be subject to copyright.
Journal
of
Educational Psychology
2001,
Vol. 93, No.
1,
187-198Copyright 2001
by the
American Psychological Association,
Inc.
0022-0663/01/S5.00
DOI:
10.1037//0022-0663.93.1.187
Cognitive Constraints
on
Multimedia Learning: When Presenting More
Material Results
in
Less Understanding
Richard E. Mayer, Julie Heiser, and Steve Lonn
University
of
California, Santa Barbara
In
4
experiments, college students viewed
an
animation
and
listened
to
concurrent narration explaining
the formation
of
lightning. When students also received concurrent on-screen text that summarized
(Experiment 1) or duplicated (Experiment 2) the narration, they performed worse on tests of retention and
transfer than
did
students
who
received
no
on-screen text. This redundancy effect
is
consistent with
a
dual-channel theory
of
multimedia learning
in
which adding on-screen text
can
overload
the
visual
information-processing channel, causing learners
to
split their visual attention between
2
sources. Lower
transfer performance also occurred when
the
authors added interesting
but
irrelevant details
to the
narration (Experiment
1) or
inserted interesting
but
conceptually irrelevant video clips within (Experi-
ment
3) or
before
the
presentation (Experiment 4). This coherence effect
is
consistent with
a
seductive
details hypothesis
in
which
the
inserted video
and
narration prime
the
activation
of
inappropriate prior
knowledge
as the
organizing schema
for the
lesson.
Multimedia scientific explanations—in CD-ROMs,
on the
World Wide
Web, and in
classroom demonstrations—are
po-
tentially valuable instructional tools.
For
example, Figure
1
shows selected slides from
a 140-s
multimedia explanation
intended
to
help students understand
how
lightning storms
develop.
The
multimedia explanation uses animation
to
depict
the steps
in
lightning formation along with corresponding
narration
to
describe them.
Our
research
has
documented
that well designed multimedia explanations formatted like
the
one
in
Figure
1 can be
highly effective
in
promoting students'
understanding,
as
indicated
by
their ability
to
generate accept-
able answers
to
open-ended transfer questions (Mayer,
1997,
1999a, 1999c; Mayer
&
Moreno,
1998;
Moreno
&
Mayer,
1999).
How can we improve
on
multimedia scientific explanations like
the
one
summarized
in
Figure
1? In
this study,
we
examine
two
suggestions:
(a)
make
the
explanation more accommodating
by
adding on-screen text
and
(b) make
it
more entertaining
by
adding
interesting words
and
video.
Richard
E.
Mayer, Julie Heiser,
and
Steve Lonn, Department
of Psy-
chology, University
of
California, Santa Barbara.
Julie Heiser
is now at the
Department
of
Psychology, Stanford
University.
The original multimedia program
was
developed
by
Roxana Moreno.
The video clips
(for the
video-interspersed, video-before,
and
video-after
groups
in
Experiments
3 and 4)
were added
by
Travis Apple. Experi-
ments
1 and 2 are
based
on an
undergraduate honors thesis submitted
by
Julie Heiser. Experiments
3 and 4 are
based
on an
undergraduate honors
thesis submitted
by
Steve Lonn.
Correspondence concerning this article should
be
addressed
to
Richard
E. Mayer, Department
of
Psychology, University
of
California, Santa
Barbara, California 93106. Electronic mail
may be
sent
to
mayer®
psych.ucsb.edu.
Add On-Screen Text
to
Accommodate Individual
Learning Preferences
A straightforward suggestion
is to
add redundant on-screen text
to the narrated animation (using the words from the narration).
The
rationale
is
that students will have two ways
of
learning the words
(i.e.,
from narration
and
on-screen text) rather than one (i.e., from
the narration).
In
this
way, the
presentation accommodates
stu-
dents who prefer to process words visually and students who prefer
to process words auditorily.
The
proposal that
two
modalities
are
always better than
one is
based
on
the idea that each modality
is a
delivery system
for
information
so
having
two
deliveries
of the
same information
is
better than having
one.
Kalyuga, Chandler,
and
Sweller (1998,
in
press) have used
the
term redundancy effect
to
refer
to any
multimedia situation
in
which "eliminating redundant material results
in
better perfor-
mance than when the redundant material
is
included" (Kalyuga
et
al.,
1998,
p. 2). For
example,
in
some cases, eliminating printed
text from
a
multimedia instructional presentation results
in
better
learning, presumably because
the
same information
is
already
presented
by
means
of
diagrams
or
other sources (Bobis, Sweller,
& Cooper, 1993; Chandler
&
Sweller, 1991, 1996; Kalyuga
et al.,
1998;
Sweller
&
Chandler, 1994).
In
the
present study,
we use the
term redundancy effect
in a more restricted sense to refer to multimedia learning situations
in which presenting words
as
text
and
speech
is
worse than
presenting words solely
as
speech (Kalyuga
et al., in
press;
Mann, 1997).
For
example,
in a
pioneering study, Kalyuga
et al.
(in press) provided training
in
soldering (i.e., techniques
for
joining metals) through
the use of
diagrams with accompanying
printed text, diagrams with accompanying speech (which
con-
tained
the
same words
as the
printed text),
and
diagrams with
accompanying text
and
speech.
A
redundancy effect
was ob-
tained
in
which
the
learners receiving diagrams with speech
per-
formed better
on
subsequent tests than
did the
learners receiving
diagrams with both speech
and
text.
In the
present study,
we
187
188MAYER, HEISER, AND LONN
"Cool moist air moves over a wanner surface and
becomes heated."
"As the air in this updraft cools, water vapor
condenses into water droplets and forms a cloud.'
"Eventually, the water droplets and ice crystals
become too large to be suspended by the updrafts.
"When downdrafts strike the ground, they spread
out in all directions, producing the gusts of cool
wind people feel just before the start of
the
rain."
"Wanned moist air near the earth's surface rises
rapidly."
"The cloud's top extends above the freezing level,
so the upper portion
of
the
cloud is composed of
tiny ice crystals."
"As raindrops and ice crystals fall through the
cloud, they drag some
of
the
air in the cloud
downward, producing downdrafts."
"Within the cloud, the rising and falling air currents
cause electrical charges to build."
Figure 1. Selected frames and corresponding narration from the no-text/no-seductive-details program.
extend the investigation of the redundancy effect by (a) examining
whether it can also occur in a multimedia environment in-
volving animation, on-screen text, and speech and (b) using a rich
set of dependent measures aimed at assessing both retention and
transfer.
Add Interesting Words and Video to Stimulate Learner Interest
A second suggestion for improving multimedia explanations is
to make them more interesting. For example, given the visual
appeal of video, we could insert a few short video clips to spice up
MULTIMEDIA LEARNING189
"The charge results from the collision of
the
cloud's
rising water droplets against heavier, falling pieces
of
ice."
"The negatively charged particles fall to the bottom
of
the
cloud, and most
of
the
positively charged
particles rise to the top."
"A stepped leader
of
negative charges moves
downward in a series
of
steps.
It
nears the ground.'"A positively charged leader travels up from such
objects as trees and buildings."
"The two leaders generally meet about 165-feet
above the ground.""Negatively charged particles then rush from the
cloud to the ground along the path created by the
leaders.
It is
not very bright."
"As the leader stroke nears the ground,
it
induces an
opposite charge, so positively charged particles
from the ground rush upward along the same path."
"This upward motion
of
the current is the return
stroke. It produces the bright light that people
notice as
a
flash
of
lightning."
Figure
1
(continued)
the presentation. Appendix A summarizes six video clips that are
intended to add interest to the lightning presentation. As one can
see,
each video clip provides an interesting elaboration on a point
made in the narrated animation but is not directly relevant to the
explanation of lightning formation. Thus, the video clips are anal-
ogous to seductive details in text passages, which can be defined
as interesting but conceptually irrelevant material that is added to
a passage to arouse the learner's interest (Garner, Brown, Sanders,
& Menke, 1992; Garner, Gillingham, & White, 1989). Research on
seductive details in a book-based environment has shown that
190MAYER, HEISER, AND LONN
adding interesting but conceptually irrelevant text to a text passage
can reduce the amount of relevant material that the learner remem-
bers (Garner et al, 1992, 1989; Hidi & Baird, 1988; Mohr, Glover,
& Ronning, 1984; Shirey & Reynolds, 1988; Wade & Adams,
1990).
Similarly, Harp & Mayer (1997, 1998) reported that adding
interesting but conceptually irrelevant illustrations to a text-and-
illustrations explanation results in poorer performance on tests of
retention and transfer.
Mayer (1999a, 1999c) has used the term coherence effect to
refer to situations in which adding words or pictures to a
multimedia presentation results in poorer performance on tests
of retention or transfer. The present research extends this work
on textbook-based coherence effects by examining what hap-
pens when narrated video clips are inserted in a narrated
animation.
Overall, the goal of this research is to provide a better
understanding of the role of well intentioned adjuncts in learn-
ing from scientific animation and narration. In Experiments 1
and 2, we examined the role of adjuncts intended to accommo-
date individual learning preferences, namely, the addition of
redundant on-screen text. In Experiments 3 and 4, we examined
the role of adjuncts intended to promote learner interest,
namely, the addition of interesting but conceptually irrelevant
words and video.
Experiment 1
The case for adding on-screen text is based on the information
delivery hypothesis, which states that students learn more when the
same information is delivered by means of more paths rather than
fewer paths. Students have more exposure to the material when it
is delivered in three ways—animation, narration, and on screen—
than when it is delivered in two ways—animation and narration. In
addition, presenting the same words in two modalities is better
than presenting them in one because it allows students to choose
the modality that best fits their learning style. In a multimedia
environment, when words are presented both visually (as on-
screen text) and auditorily (as narration), learners can attend to
whichever format best meshes with the way they prefer to process
verbal information. Multiple presentation modes may be particu-
larly important for students with disabilities such as hearing im-
pairments, although only regular education students were involved
in our research. The importance of accommodating individual
differences in learning style has long been recognized in educa-
tional psychology (Jonassen & Grabowski, 1993) and has been
tested in a multimedia learning environment (Plass, Chun, Mayer,
& Leutner, 1998). The information delivery hypothesis predicts
that adding on-screen text to a narrated animation will result in
better performance on tests of learning that focus on remembering
the verbal explanation (i.e., retention test) and being able to use the
explanation to solve new problems (i.e., transfer test).
On the other hand, the case against adding on-screen text fol-
lows from the cognitive theory of multimedia learning, which is
based on three research-grounded assumptions: (a) the dual-
channel assumption, in which learners possess separate visual
and verbal information-processing channels (Clark & Paivio,
1991;
Paivio, 1986); (b) the limited capacity assumption, in which
each channel is limited in processing capacity (Baddeley, 1992;
Chandler & Sweller, 1991); and (c) the generative learning as-
sumption, in which meaningful learning occurs when learners
attend to relevant portions of the incoming auditory and visual
information, organize the material into visual and verbal mental
representations, and integrate the two representations (Mayer,
1996,
1999b; Wittrock, 1990).
As shown in Figure 2, narration enters the verbal channel via the
ears.
If the learner attends to the incoming material, the learner
may select relevant words for further processing in the verbal
channel in working memory. In working memory, the learner may
mentally organize the selected words into a coherent verbal mental
model and may mentally integrate the verbal representation with a
visual representation and with prior knowledge from long-term
memory. Similarly, the animation enters the visual channel via the
eyes.
If the learner attends to the incoming material, the learner
may select relevant images for further processing in the visual
channel in working memory. In working memory, the learner may
mentally organize the selected images into a coherent pictorial
mental model and may mentally integrate the pictorial represen-
tation with the verbal representation and with prior knowledge
from long-term memory. In contrast, on-screen text enters the
visual channel via the eyes (as indicated by the arrow from words
to eyes) and later can be converted to sounds (as indicated by the
arrow from images to sounds) that are used to build a verbal
representation in the verbal channel.
A straightforward implication of the cognitive theory of multi-
media learning is the split-attention hypothesis, which states that
when words are presented visually, learners must split their visual
attention between the on-screen text and the animation, thereby
failing to adequately attend to some of the presented material. In
short, the eyes initially receive input from two sources (as indi-
cated by the arrow from words to eyes and from pictures to eyes).
This hypothesis is derived from a cognitive theory of multimedia
MULTIMEDIA
PRESENTATIONSENSORY
MEMORYWORKING MEMORYLONG-TERM
MEMORY
Words
Pictures
\
s\Ears
Eyes
selecti
wore
selecti
lmag
ig.
s ""
Sounds
!i
r
Images
organiz
wo«l
oreaniz
imag
s
ing ^
s ""
Verbal
Model
Pictorial
Model
~L
T
i;ratingPrior
Knowledge
Figure 2. The cognitive theory of multimedia learning.
MULTIMEDIA LEARNING
191
learning in which visually presented material is processed (at least
initially) in a limited-capacity visual channel, whereas auditorily
presented material is processed (at least initially) in a limited-
capacity auditory channel (Mayer, 1997, 1999a, 1999c; Mayer &
Moreno, 1998). The split-attention hypothesis predicts that adding
on-screen text to a narrated animation will result in poorer perfor-
mance on retention and transfer tests. This prediction holds only
for multimedia presentations in which the visual and verbal ma-
terial is presented rapidly and the pace of presentation cannot be
controlled by the learner.
In Experiment 1, students received a multimedia presentation
explaining the formation of lightning and answered questions on
retention and transfer tests. Some students received animation and
concurrent narration (no-text/no-seductive-details group). For
other students, we added either on-screen text that summarized the
concurrent narration with the same words as in the narration
(text/no-seductive-details group); six additional sentences, inter-
spersed in the narration, that contained entertaining but conceptu-
ally irrelevant information (no-text/seductive-details group); or
both on-screen summary text and seductive details (text/seductive-
details group).
Method
Participants
and
Design
The participants were 78 college students recruited from the Psychology
Subject Pool
at the
University
of
California, Santa Barbara.
We
used
a
2X2 between-subjects design, with
the
first factor being presence
or
absence
of
on-screen text that summarized
the
narration
in a
multimedia
presentation,
and the
second factor being whether extraneous details were
added
to the
narration
and
text. There were
22
students
in the
no-text/no-
seductive-details group, 19
in
the text/no-seductive-details group, 21
in the
no-text/no-seductive-details group,
and 16 in the
text/seductive-details
group.
The
mean combined Scholastic Aptitude Test
(SAT)
score
was
1159, the
mean
age was 18.4
years,
and
33%
of the
sample
was
comprised
of
male students.
All
participants
in the
study reported
low
levels
of
knowledge about meteorology,
as
indicated
by low
scores
on a
meteorology knowledge questionnaire (i.e.,
7 or
less
out of 11).
Four
participants were excluded because
of
high scores
on the
meteorology
knowledge questionnaire (i.e., greater than
7),
yielding
a
sample
of 78
remaining participants.
Materials and Apparatus
The paper-based materials were typed
on 8.5 X
11-in. sheets
of
paper
and consisted
of a
one-page questionnaire,
a
one-page retention test,
and a
four-page transfer test.
The
materials
in
this
and the
other three experi-
ments included
an
additional test that was given
at the end of
the session
but
was not
used
in the
data analysis.
The
materials
are
similar
to
those
used by Mayer and Moreno (1998). The questionnaire asked participants
to
indicate their age, gender,
and
SAT score;
it
also contained
a
meteorology
knowledge scale
in
which participants were asked to rate on
a
5-point scale
(1
=
very little,
5 =
very much) their level
of
knowledge
of
meteorology
and
to
place check marks next
to
each
of six
weather-related items that
applied
to
them (e.g.,
"I
know what
a low
pressure system
is" or "I can
distinguish between
a
cumulus
and
nimbus cloud").
The
retention test
contained
the
following printed instruction: "Please write down
an
expla-
nation of how lightning works." The transfer test consisted of the following
four questions, each printed
on a
separate sheet: "What could
you do to
decrease the intensity
of
lightning?"
"Suppose you see clouds
in
the sky but
no lightning,
why
not?" "What does
air
temperature have
to do
with
lightning?"
and
"What causes lightning?"
The computer-based materials consisted
of
four multimedia programs
constructed using Director
4.0
(Macromedia, 1994).
The
no-text/no-
seductive-details version included
a
140-s animation depicting
the
forma-
tion
of
lightning along with concurrent narration that was broken into
16
segments (i.e., listed
as 1
through
16 in
Appendix
B).
Figure
1
shows
selected frames from the animation along with concurrent narration (shown
as text
in
quotation marks below each frame). The animation was intended
to depict
the
major steps
in the
formation
of
lighting,
and the
text
was
intended
to
describe
the
major steps
in the
formation
of
lightning.
The
materials were intended
to be
well designed
and to
include
the
same
essential information
as
found
in
science textbooks. The text/no-seductive-
details version
was
identical
to the
no-text/no-seductive-details version
except that each
of
the 16 on-screen text summaries appeared
at
the bottom
of the screen during
the
same time that
the
corresponding narration
seg-
ment was being spoken. Corresponding portions
of
animation and text were
presented simultaneously.
The
on-screen text summaries contained
se-
lected words from the narration and
are
shown
in
brackets
in
Appendix
B.
The no-text/seductive-details version was created
by
adding
six
narration
segments—which
we
call "seductive details"—throughout
the
no-text/no-
seductive-details presentation.
The
seductive details
(and
their placement
in the series
of
statements)
are
indicated
by
SD-1 through SD-6
in
Appen-
dix
B. The
text/seductive-details version
was
identical
to
this except that
each
of the 16
on-screen text summaries
and the 6
on-screen seductive
details summaries appeared
at the
bottom
of the
screen during
the
same
time that the corresponding narration segment was being spoken. Appendix
B lists the 22 on-screen summaries, which are indicated
in
brackets.
In
both
the text/no-seductive-details version
and
the text/seductive-details version,
the text summaries were entirely redundant with
the
narration
in
that
the
words used
in the
summaries were selected directly from
the
narration
segments, with
no
additional words added.
The apparatus consisted
of
five Macintosh computers with 15-in. color
monitors
and
Koss earphones.
A
stopwatch was used
to
time
the
tests.
Procedure
Participants were randomly assigned to treatment groups and were tested
in groups
of up to 5 per
session. Each participant
was
seated
at an
individual computer station
and
tested
in an
individual cubicle. First,
the
participants completed
the
questionnaire
at
their
own
rates. Second,
par-
ticipants were instructed that they would receive
a
brief multimedia
pre-
sentation about lightning formation, that they should pay attention, and that
after the presentation they would have
to
answer some questions about the
material. Participants were instructed
to put on the
headphones
and
press
any
key on the
keyboard
to
begin
the
presentation. Third, following
the
presentation, participants were given the retention sheet and asked
to
write
an explanation of how lightning forms. The sheet was collected after 6 min.
Fourth, participants were given each
of
the transfer test sheets individually
for 2.5 min each
in
the order listed
in
the Materials and Apparatus section
of Experiment
1.
Participants were instructed
to
keep working until
the
sheet was collected,
and
each sheet was collected before the next one
was
handed
out.
After taking
the
tests,
the
participants were thanked
and
excused.
All tests were scored
as in
Mayer and Moreno (1998). The meteorology
knowledge scale was scored by giving
1
point for each item checked on
the
list
of
weather knowledge items (i.e., from
0
to 6) and
1
point
for
each level
of self-assessed knowledge
of
meteorology (i.e., from
1 to
5).
The
maxi-
mum score
was
11,
and
participants scoring below
8
were classified
as
having
low
experience
in
meteorology.
The
retention test
was
scored
by
determining
how
many
of
eight
key
idea units were included
in the
protocol. One point was given
for
each
of
the following ideas, regardless
of
wording:
(a) air rises, (b) water condenses, (c) water and crystals fall,
(d)
wind
is
dragged downward,
(e)
negative charges fall
to the
bottom
of
the
cloud,
(f)
the leaders meet, (g) negative charges rush down, and (h) positive
charges rush up. To test
for
interrater reliability,
a
second rater scored 10%
of the retention tests;
the
correlation between
the two
raters was
.99.
192MAYER, HEISER, AND LONN
The transfer test was scored by tallying the number of acceptable
solutions to each of the four transfer test questions. We consider these
answers to be creative solutions because students must invent them rather
than simply recall material directly from the presentation. For example,
acceptable answers included removing positive ions from the ground for
the first question, stating that the top of the cloud might not be above the
freezing level for the second question, stating that the air must be cooler
than the ground for the third question, and stating that there must be a
difference in electrical charges within the cloud for the fourth question.
Participants received no more than 8 points overall, although there was no
a priori upper bound. To test for interrater reliability, a second rater scored
10%
of the transfer tests; the correlation between the two raters was .94.
Results and Discussion
Table 1 presents the mean scores and standard deviations for
each group on each of the two tests. We conducted a two-way
analysis of variance, with presence or absence of on-screen text as
the first factor and presence or absence of seductive details as the
second factor.
Does Adding an On-Screen Text Summary Affect
Retention or Transfer Performance?
Students who received on-screen text summaries remembered
significantly fewer of the idea units on the retention test
(M = 4.09, SD = 1.10) than did students who received no
on-screen text (M = 4.74, SD =
1.82),
F(l, 74) = 3.94,
MSE = 2.38, p = .05. Students who received on-screen text
summaries produced fewer creative solutions on the transfer test
(M = 2.91, SD = 1.55) than did students who received no
on-screen text (M = 4.02, SD =
1.76),
F(l, 74) = 8.73,
MSE = 2.78, p <
.01.
The effect sizes were .36 on retention and
.63 on transfer. Overall, these results show that adding on-screen
text summaries to a multimedia presentation hurt student learning.
Thus,
the findings are most consistent with the split-attention
hypothesis.
Does Adding Seductive Details Affect Retention or
Transfer Performance?
Students who received seductive details remembered signifi-
cantly fewer of the idea units on the retention test {M = 3.76,
SD = 1.60) than did students who did not receive seductive details
(M = 5.07, SD =
1.32),
F(l, 74) = 14.39, MSE = 2.38,p < .001.
Students who received seductive details produced fewer creative
solutions on the transfer test (M = 3.05, SD = 1.69) than did
Table 1
Mean Scores and Standard Deviations on Retention and
Transfer
Tests
for Four Groups—Experiment 1
Group
No text/no seductive details
Text/no seductive details
No text/seductive details
Text/seductive details
Retention
M
5.41
4.68
4.05
3.38
SD
1.53
1.11
2.09
1.09
Transfer
M
4.59
3.21
3.43
2.56
SD
1.56
1.65
1.94
1.41
Note. Maximum possible retention score is 8, and maximum obtained
transfer score is 8.
students who did not receive seductive details (M = 3.95,
SD =
1.61),
F(l, 74) = 5.67, MSE = 2.78, p < .05. The effect
sizes were .56 on retention and .55 on transfer. Overall, consistent
with research on seductive details in printed materials (Harp &
Mayer, 1997, 1998), these results show that adding seductive
details to a multimedia presentation hurt student learning.
There was no significant interaction between the two factors
(i.e.,
presence or absence of on-screen text and presence or absence
of seductive details) on the retention test, F(l, 74) = 0.01,
MSE = 2.38, p = .94, or the transfer test, F(l, 74) = 0.46,
MSE = 2.78, p = .50.
Experiment 2
Experiment 1 provided evidence for a redundancy effect in
which adding redundant on-screen text to a multimedia explana-
tion resulted in poorer student learning. The on-screen text may
have created cognitive load either by competing with the anima-
tion for cognitive resources in the visual channel or by demanding
resources in the auditory channel to reconcile the auditory and
text-presented versions. Experiment 2 tested these two explana-
tions by comparing students who received no additional on-screen
text, additional on-screen text that summarized the narration (as in
Experiment 1), or on-screen text that contained all of the exact
words in the narration. Both the summary text and full text
contained all eight target idea units. If the redundancy effect
observed in Experiment 1 is caused mainly by students trying to
reconcile the on-screen text summary with the full narration, then
the summary-text group should perform more poorly than the
no-text group and the full-text group. If the redundancy effect
observed in Experiment 1 is caused mainly by an overload of the
visual channel, then both the summary-text and full-text groups
should perform more poorly than the no-text group.
Method
Participants and Design
The participants were 109 college students recruited from the Psychol-
ogy Subject Pool at the University of California, Santa Barbara, with each
student serving in one of three treatment groups. There were 36 students in
the no-text group, 37 in the summary-text group, and 36 in the full-text
group. The mean combined SAT score was 1185, the mean age was 18.8
years,
and male students comprised
31%
of the sample. All participants in
the study reported low levels of knowledge about meteorology, as indicated
by low scores on a meteorology knowledge questionnaire (i.e., 7 or less out
of 11). Six participants were excluded because of high scores on the
meteorology knowledge questionnaire (i.e., greater than 7), yielding 109
remaining participants.
Materials and Apparatus
The questionnaire, retention test, and transfer test were identical to those
in Experiment 1. Three multimedia programs were used in Experiment 2:
the no-text version was identical to the no-text/no-seductive-details version
used in Experiment 1; the summary-text version was identical to the
text/no-seductive-details version used in Experiment 1; the full-text version
was identical to the text/no-seductive-details version used in Experiment 1
except that the text shown at the bottom of the screen was a full transcript
of the corresponding narration rather than a summary. The computers used
in Experiment 2 were identical to those used in Experiment 1.
MULTIMEDIA LEARNING193
Procedure
The procedure was identical to Experiment 1 except that after complet-
ing the questionnaire, all participants took two additional 3-min tests
(which were not used in the data analysis in this study). After completing
these tests, participants received either the no-text, summary-text, or full-
text version of the multimedia presentation.
Results and Discussion
Table 2 presents the mean scores and standard deviations for
each group on each of the two tests. We conducted a one-way
analysis of variance, with type of on-screen text as the between-
subjects factor. Pairwise Tukey Honestly Significant Difference
(HSD) tests were conducted with p < .05. We used the Tukey
HSD test because it is not unnecessarily conservative and controls
the familywise error rate.
Does Adding an On-Screen Text Summary Affect
Retention or Transfer Performance?
There were significant retention score differences among groups
receiving no text (Af = 5.22, SD =
1.51),
summary text
(M =
4.51,
SD =
1.92),
and full text (M = 3.92, SD =
1.71),
F(2,
103) = 5.89, MSE =
2.80,
p <
.01.
Tukey HSD tests revealed that
students who received no on-screen text remembered significantly
more of the idea units on the retention test than did students who
received full on-screen text, whereas the summary-text and full-
text groups did not differ significantly. There were significant
differences on transfer score among groups receiving no text
(M = 4.64, SD =
1.57),
summary text (M = 2.78, SD =
1.25),
and
full text (M = 2.75, SD =
1.54),
F(2, 103) = 28.48, MSE = 1.62,
p <
.001.
Tukey HSD tests revealed that students who received no
on-screen text created significantly more solutions on the transfer
test than did students who received summary on-screen text and
students who received full on-screen text, whereas the latter two
groups did not differ significantly from one another. The effect
sizes comparing the full-text group with the no-text group were .86
for retention and 1.20 for transfer. Overall, these results show that
adding on-screen text to a multimedia presentation hurt student
learning and that summary text (as used in Experiment 1) and full
text produced indistinguishable results. Thus, the findings are most
consistent with the split-attention hypothesis.
Experiment 3
Experiments 1 and 2 showed that the well intentioned addition
of a simple adjunct—on-screen text—resulted in poorer under-
Table 2
Mean Scores and Standard Deviations on Retention and
Transfer
Tests
for Three Groups—Experiment 2
Group
No text
Summary text
Full text
M
5.22
4.51
3.92
Retention
SD
1.51
1.92
1.71
M
4.64
2.78
2.75
Transfer
SD
1.57
1.25
1.54
Note. Maximum possible retention score is 8, and maximum obtained
transfer score is 8.
standing of a multimedia explanation of how lightning works.
In Experiments 3 and 4, we explored a different kind of ad-
junct—the insertion of several short video clips depicting light-
ning storms. Like the on-screen text added in Experiments 1
and 2, video clips are adjuncts (i.e., material added to the
narrated animation) that are intended to enhance student learn-
ing. The theoretical rationale for adding video clips is that,
unlike the on-screen text used in Experiments 1 and 2, they will
make the material more interesting. According to the emotional
interest hypothesis, adding interesting material to an instruc-
tional presentation increases the learner's enjoyment of the
presentation, and this increase in enjoyment causes the reader to
pay more attention and work harder to understand the material.
In short, emotion-grabbing adjuncts improve the learner's af-
fect, which in turn improve the learner's cognitive processing
(Garner et al., 1992, 1989). In Experiment 3, if the emotional
interest hypothesis is correct, then interspersing interesting
video clips to a narrated animation about lightning formation
should result in improvements in retention and transfer
performance.
In contrast, the cognitive theory of multimedia learning is
based on the idea that learners actively seek to make sense of
the presented material by mentally selecting pieces of the
presented material, mentally organizing the selected material,
and mentally integrating it with relevant existing schemas in
their long-term memory. In the case of the lightning explana-
tion, learners must build a mental model (which we call a
"system model") of the lightning system. The system model
consists of a cause-and-effect chain in which a change in one
component causes a change in another component, and so on.
According to the seductive details hypothesis, which we de-
rived from the cognitive theory of multimedia learning (Harp &
Mayer, 1997, 1998), the act of constructing a coherent mental
model that makes sense to the learner gives the learner a sense
of satisfaction in which the material seems interesting to the
learner. The model-building process can be negatively influ-
enced by the video clips, which may cause learners to focus on
relating the incoming explanation to the events highlighted in
the video clips. In Experiment 3, if the seductive details hy-
pothesis is correct, then interspersing interesting video clips in
a narrated animation about lightning formation should result in
decrements (or no improvement) in retention and transfer
performance.
It is important to note that the video clips are relevant to the
topic of lightning because they are about lightning storms.
However, the video clips are not relevant to the cause-and-
effect explanation of how lightning storms develop. In sum-
mary, the video clips have topical relevance (or surface rele-
vance) to the content about lightning storms, but they lack
conceptual relevance (or structural relevance) to the explana-
tion of how lightning works. According to the emotional inter-
est hypothesis, the video clips will increase learners' interest in
the topic of lightning such that they will pay more attention
when they are given the multimedia explanation of lightning
formation. According to the seductive details hypothesis, the
video clips will prime inappropriate knowledge about lightning
storms that will interfere with learners' efforts to build a co-
herent cause-and-effect chain.
194MAYER, HEISER, AND LONN
Method
Participants and Design
The participants were 38 college students from the Psychology Subject
Pool at the University of California, Santa Barbara. Twenty-one students
served in the no-video group, and 17 students served in the video-
interspersed group. All participants indicated a low level of experience in
meteorology, as indicated by low scores (i.e., 7 or less out of 11) on a
meteorology experience questionnaire. Two students were excluded be-
cause of high scores (i.e., greater than 7) on the meteorology experience
questionnaire, yielding a sample of 38 remaining participants. The mean
combined SAT score of the no-video group was 1106 (SD = 156), and the
mean combined SAT score of the video-interspersed group was 1153
(SD = 113), «(32) = 0.97, ns. The mean score on the meteorology
experience questionnaire was 3.00 (SD = 2.00) for the no-video group
and 2.82 (SD = 1.74) for the video-interspersed group, f(36) = .286, ns.
The mean age was 19.1 (SD = 1.0) for the no-video group and 18.6
(SD = 1.0) for the video-interspersed group, t(36) = 1.51, ns. The
proportion of female students was .67 for the no-video group and .71 for
the video-interspersed group; a Fisher Exact Test revealed that the differ-
ence is not significant at the .05 level.
Materials
The paper-and-pencil materials consisted of the same participant ques-
tionnaire, retention test, and four-sheet transfer test that were used in
Experiments 1 and 2.
The computer-based materials consisted of two multimedia programs
explaining the formation of lightning. The no-video program was the same
as the no-text/no-seductive-details program in Experiment 1 and the no-
text program in Experiment 2. The video program presented the same
narrated animation except that six short narrated video clips were inserted
at various points in the presentation, each lasting approximately 10 s.
Appendix A provides the narration scripts and descriptions of the video
images for each of the six clips. The first clip was presented at the
beginning of the presentation, the second clip after the 2nd sentence, the
third clip after the 6th sentence, the fourth clip after the 15th sentence, the
fifth clip after the 20th sentence, and the sixth clip at the end of the
presentation. The video clips were intended to be interesting, to be related
to the topic of lightning, and to be unrelated to the explanation of lightning
formation.
The apparatus consisted of four Macintosh computer systems with 17-in.
color monitors and Koss earphones.
Procedure
The procedure was the same as for Experiments 1 and 2 except that
participants received either the no-video or video presentation.
Results and Discussion
Do Students Remember More of
the
Explanation When
Interesting But Irrelevant Video Clips Are Excluded
Rather Than Included in a Multimedia Explanation?
According to the seductive details hypothesis, adding interesting
video clips to the multimedia explanation will result in the same or
poorer retention performance, whereas the emotional interest hy-
pothesis predicts that adding video clips will result in improved
retention performance. The top-left portion of Table 3 shows that
the no-video group recalled more idea units than did the video
group, but this difference failed to reach statistical significance,
f(36) = 1.00, ns. The effect size is .37. The pattern of results is
consistent with the idea that adding interesting video clips does not
help learning, as measured by the retention test.
Table 3
Mean Retention and Transfer Scores for Two Groups in
Experiment 3 and Two Groups in Experiment 4
Group
Experiment 3
No video (n = 21)
Video interspersed (n = 17)
Experiment 4
Video after (n = 16)
Video before (n = 16)
Retention
score
M
3.22
2.77
3.81
3.37
SD
1.91
1.52
1.22
1.54
Transfer
M
4.43*
3.41
4.38*
2.94
score
SD
1.72
1.18
1.70
1.44
Note. Maximum possible retention score is 8, and maximum obtained
transfer score is 8.
*p< .05.
Do Students Understand More Deeply When Interesting
But Irrelevant Video Clips Are Excluded Rather Than
Included in a Multimedia Explanation?
According to the seductive details hypothesis, adding interesting
but irrelevant video clips to the multimedia explanation will result
in poorer transfer performance, whereas the emotional interest
hypothesis predicts that adding video clips will result in improved
transfer performance. The top-right portion of Table 3 shows that
the no-video group generated significantly more solutions on the
transfer test than did the video-interspersed group, t(36) = 2.074,
p <
.05.
The effect size is .86. These results are consistent with the
seductive details hypothesis and inconsistent with the emotional
interest hypothesis.
Experiment 4
In Experiment 4, we provide a deeper comparison of the emo-
tional interest hypothesis versus the seductive details hypothesis.
In particular, we tested the emotional interest hypothesis, which
holds that adding video clips before a multimedia presentation
should result in better learning (i.e., better retention and transfer
performance) than inserting them after a multimedia explanation.
The rationale for this prediction is that presenting video clips
before the multimedia presentation will increase the learner's
interest in and attention to the multimedia explanation. In contrast,
the level of interest (and hence, attention) for the explanation will
be lower if the interest-boosting video clips are not given until
after the multimedia explanation.
The seductive details hypothesis (based on the cognitive theory
of multimedia learning) allows the opposite prediction. The video
clips can prime inappropriate assimilative schemas in learners such
that they assimilate the information about lightning formation to
schemas that are not cause-and-effect chains. According to the
seductive details hypothesis, learners should perform more poorly
on retention and transfer when video clips are placed before rather
than after the narrated animation. The rationale is that conceptually
irrelevant video clips can prime inappropriate schemas if they are
presented before the target information but not if they are pre-
sented after.
MULTIMEDIA LEARNING195
Method
Participants and Design
The participants were 32 college students recruited from the Psychology
Subject Pool at the University of California, Santa Barbara. Sixteen stu-
dents served in the video-after group, and 16 served in the video-before
group. As in Experiment 3, all participants indicated a low level of
experience in meteorology, as indicated by low scores (i.e., 7 or less out of
11) on a meteorology experience questionnaire. Two students were ex-
cluded because of high scores (i.e., greater than 7) on the meteorology
experience questionnaire, yielding a sample of 32 remaining students. The
mean combined SAT score of the video-after group was 1188 (SD = 99),
and the mean combined SAT score of the video-before group was 1239
(SD = 60), f(25) = 1.62, ns. The mean score on a meteorology experience
questionnaire was 3.06 (SD = 1.65) for the video-after group and 2.12
(SD = 1.09) for the video-before group, r(30) = 1.89, ns. The mean age
was 18.9 (SD = 1.0) for the video-after group and 18.4 (SD = .5) for the
video-before group, t(35) = 1.61, ns. The proportion of female students
was .69 for the video-after group and .75 for the video-before group; a
Fisher Exact Test showed that the proportions are not significantly differ-
ent at the .05 level.
Materials and Apparatus
The materials and apparatus were identical to those used in Experiment 3
except that the two multimedia programs were video after, in which all of
the narrated animation was followed by all of the narrated video clips, and
video before, in which all of the narrated video clips preceded all of the
narrated animation. The group of six video clips lasted approximately 60 s.
Procedure
The procedure was identical to that in Experiment 3 except that the two
multimedia programs were video after and video before.
Results and Discussion
Do Students Remember More of the Explanation When
Interesting But Irrelevant Video Clips Are Presented After
Rather Than Before a Multimedia Explanation?
According to the emotional interest hypothesis, placing inter-
esting video clips at the beginning of a multimedia explanation
rather than at the end will result in higher retention performance,
whereas the seductive details hypothesis predicts the opposite
pattern. The bottom-left portion of Table 3 shows that the video-
after group recalled more idea units than did the video-before
group, but this difference failed to reach statistical significance,
f(30) = .89, ns. The effect size is .28. The retention results do not
allow the disconfirmation of either hypothesis.
Do Students Understand More Deeply When Interesting
But Irrelevant Video Clips Are Placed After Rather
Than Before a Multimedia Explanation?
According to the emotional interest hypothesis, placing inter-
esting video clips before rather than after a multimedia explanation
will result in poorer transfer performance, whereas the seductive
details hypothesis predicts the opposite pattern. The bottom-right
portion of Table 3 shows that the video-after group generated
significantly more solutions on the transfer test than did the video-
interspersed group, f(30) = 2.58, p < .05. The effect size is 1.00.
These results are consistent with the seductive details hypothesis
and inconsistent with the emotional interest hypothesis.
General Discussion
Redundancy Effect in Multimedia Learning
In two studies, learning a scientific explanation from a narrated
animation was hurt by the addition of on-screen text that contained
the same words as in the narration. The detrimental effects of
redundant on-screen text were found both when the on-screen text
was an exact copy of the corresponding narration (i.e., Experiment
2) and when it was a summary with the same words as the
corresponding narration (i.e., Experiments 1 and 2). We refer to
this finding as a redundancy effect: Adding redundant on-screen
text to a narrated animation detracts from multimedia learning.
The redundancy effects we discovered within our multimedia
environment involving animation complement the redundancy ef-
fects in a study involving diagrams reported by Kalyuga et al. (in
press).
In both cases, when presenting a multimedia lesson with
spoken words and pictures, adding words in the form of printed
text did not improve learning. The present research provides the
first documented extension of Kalyuga, Chandler, and Sweller's
finding to a multimedia environment involving animation.
Theoretical Implications of the Redundancy Effect
The redundancy effect is consistent with the cognitive theory of
multimedia learning and the split-attention hypothesis, which can
be derived from it (Mayer, 1997; Mayer & Moreno, 1998). The
locus of the effect seems to be at the point of visual attentional
scanning, as posited by the split-attention hypothesis. In particular,
the pattern of results is most consistent with the idea that the
on-screen text competes with the animation for visual attention,
thus reducing the chances that the learner will be able to attend to
relevant aspects of the animation and text.
We interpret the redundancy effect as a new piece of support for
the cognitive theory of multimedia learning and, in particular, the
idea that humans possess separate visual and auditory processing
channels that are each limited in capacity. When words are pre-
sented in the auditory channel and pictures are presented in the
visual channel, the load on these systems is minimized. In contrast,
when words are presented in both the auditory channel and the
visual channel, and pictures are presented in the visual channel, the
attentional load on the visual channel is increased. In this case,
learners are less likely to be able to carry out the active cognitive
processes needed for meaningful learning. Consistent with the
predictions of the cognitive theory of multimedia learning—and its
split-attention hypothesis—adding redundant text hurts student
learning of a multimedia explanation.
This result is not consistent with the information delivery view
of multimedia learning or the hypotheses we derived from it. The
results contradict the information delivery hypothesis, which posits
that two ways of presenting words are better than one, or the idea
that adding redundant printed text allows learners to choose which
mode of presentation—visual or auditory—fits their preferred
learning style. However, these results should not be taken to deny
the value of allowing learners some choice in adjusting multimedia
presentations to fit their learning preferences in all situations. In
fact, a better pedagogical approach may be to allow learners to
choose whether they would prefer to see words as on-screen text or
hear words as narration rather than presenting words in both forms
(Plass et al., 1998).
In contrast to situations in which "two sensory modalities are
better than one" (Tindall-Ford, Chandler, & Sweller, 1997, p. 257),
196MAYER, HEISER, AND LONN
these studies have pinpointed a situation in which two sensory
modalities are worse than one, creating what can be called a
redundancy effect. An important contributing factor may be that
presenting a narrated animation places a heavy cognitive load on
the visual and auditory channels; therefore, adding additional load
on the visual channel through on-screen text reduces the amount of
cognitive resources learners can apply to essential processes in
multimedia learning. In other studies, we have documented situa-
tions in which two modalities are better than one—which we call
a multimedia effect (Mayer, 1997, 1999a, 1999c)— as indicated by
improved learning when a narration is supplemented with corre-
sponding animation. In this case, load on the visual channel is not
increased because words are presented in the auditory channel.
According to a cognitive theory of multimedia learning, not all
techniques for removing redundancy are equally effective. For
example, in the case of multimedia explanations consisting of
animation, narration, and on-screen text, one effective solution is
to remove the on-screen text (as was done in the present studies),
but it does not follow that the same benefits would occur by
instead removing the narration. Although this eliminates redun-
dancy, it would create a split-attention situation in which the visual
channel is overloaded with visually presented words and visually
presented graphic material (e.g., animation or diagrams), thus
resulting in poorer learning and understanding. This prediction of
the cognitive theory of multimedia learning, which can be called a
split-attention effect, was confirmed by Kalyuga et al. (in press)
using diagrams and printed text and by Mayer and Moreno (1998)
using animation and on-screen text.
Practical Implications of
the
Redundancy Effect
The redundancy effect also has implications for the design of
multimedia instructional messages. This research allows us to add
a new design principle to our collection (Mayer, 1997) which we
call the redundancy principle: When making a multimedia presen-
tation consisting of a narrated animation, do not add on-screen text
that duplicates words that are already spoken in the narration. This
principle holds for situations in which the animated narration runs
at a fast rate without learner control of the presentation.
This study, and the resulting redundancy principle, should not
be taken as evidence that printed text and spoken text must never
be presented together. In our opinion, multimedia design principles
should not be taken as blanket commandments but rather should be
interpreted in light of theories of how people learn—such as the
cognitive theory of multimedia learning (Mayer, 1997). For exam-
ple,
PowerPoint presentations—in which a presenter both speaks
and presents printed words on screen—can be effective even
though words are presented in two modalities. In a PowerPoint
presentation, the rate of presentation may be slower than in the
lightning multimedia program. Thus, presenting words in spoken
and printed form may be harmful in some situations (as demon-
strated in these studies) but not in other situations (such as when
the rate of presentation is slow or when no pictorial material is
concurrently presented). Further research is needed to pinpoint the
conditions under which redundancy effects occur.
Coherence Effect in Multimedia Learning
This research extends previous research on seductive details by
showing that the same coherence effect obtained in a paper-based
environment (Harp & Mayer, 1997, 1998) also occurs in a
computer-based environment and by showing that the effect orig-
inally obtained using retention as a dependent measure can also be
obtained using transfer measures (Garner et al., 1989; Renninger,
Hidi, & Krapp, 1992). As in Harp and Mayer's (1997, 1998)
studies, Experiments 1, 3, and 4 demonstrated that adding seduc-
tive details to the lightning passage resulted in poorer transfer
performance.
Overall, Experiments 3 and 4 demonstrate that adding interest-
ing but conceptually irrelevant video clips to a multimedia expla-
nation can have negative effects on students' understanding of the
explanation. In particular, students who received video clips inter-
spersed within the narrated animation or placed before the narrated
animation displayed poorer problem-solving transfer performance
than students who received no video clips. Similarly, in Experi-
ment 1 adding interesting but irrelevant sentences to a multimedia
presentation also depresses students' understanding of the expla-
nation. Apparently, the video clips and added sentences caused
students to learn less deeply and therefore be less able to transfer
what they had learned to new problem situations.
Theoretical Implications of
the
Coherence Effect
These results support and extend the cognitive theory of multi-
media learning. First, the results are inconsistent with an emotional
interest hypothesis which holds that adding interesting details to a
passage increases overall arousal and therefore results in more
learning overall. In contrast, the results are consistent with a
seductive details hypothesis which holds that students try harder to
understand material when the material is presented in a way that
highlights the underlying conceptual structure. The results of Ex-
periment 4 further pinpoint the locus of the coherence effect as
attributable to priming of inappropriate assimilative schemas. Im-
portantly, our research shows that the same learning theories apply
to both paper-based and computer-based environments, adding
weight to the assertion that what constitutes good pedagogy in one
environment also constitutes good pedagogy in the other.
Practical Implications of
the
Coherence Effect
The results extend previous findings by Harp and Mayer (1997,
1998) from paper-based media to computer-based media. The
results provide the first documented evidence that adding interest-
ing but irrelevant video clips and narration segments to a coherent
multimedia explanation can hurt student understanding of the
explanation. This work allows us to offer an expanded statement of
the coherence principle which includes computer-based venues
(Mayer, 1997, 1999a, 1999c): Students understand a multimedia
explanation more deeply when interesting but conceptually irrel-
evant video and narration are excluded rather than included.
Overall, the role of interest in learning has a long and sometimes
forgotten history in instructional
design.
When designers wish to spice
up a multimedia presentation by adding interesting video segments,
they seem to be subscribing to emotional interest theory—the idea
that students learn more deeply when interesting adjuncts are added to
a lesson. Consistent with Dewey's (1913) admonitions more than 80
years ago, our results reject emotional interest theory and the procliv-
ity to add "bells and whistles" to multimedia presentations.
Limitations and Future Directions
This research is limited by the nature of the materials, the
learning measures, the learners, and the learning venue. First, the
MULTIMEDIA LEARNING
197
materials consisted of a short cause-and-effect explanation of how
a scientific system works. It is not clear that the same results would
occur for other text genres such as narratives or descriptive text.
Second, the learning measures focused on problem-solving trans-
fer because our goal was to examine student understanding. The
same results may not always occur for other kinds of learning
measures such as retention of the main ideas that were presented in
the narration. Third, the learners were college students who lacked
experience in meteorology and who did not have any hearing
impairments. It is unlikely that the same results would be obtained
for high-experience learners nor for learners with hearing impair-
ments. Fourth, students learned from a short multimedia presen-
tation within a laboratory setting. It is not clear whether these
results would transfer to learning in a classroom or training setting
involving more student interaction and longer-term learning tasks.
Additional work is needed to determine how redundant on-screen
text and seductive details affect multimedia learning situations that
involve other text genres (such as narration and description), other
dependent measures (such as interest and affect), other learners
(such as high-experience learners or learners with disabilities), and
other venues (such as in a classroom or training program).
Overall, these results are consistent with Moreno and Mayer's
(2000, p. 124) recommendation to "only include complementary
stimuli that are relevant to the content of the lesson." Whereas
Moreno and Mayer (2000) showed that adding extraneous music
or sounds hurt student understanding of a multimedia explanation,
the present studies show that adding redundant on-screen text or
conceptually irrelevant video clips also hurt student understanding
of a multimedia explanation. Together, these results show that
adding "bells and whistles" can hurt the sense-making process in
learners, but for quite different reasons. In the case of redundant
on-screen text, the visual channel can become overloaded (i.e.,
affecting the process of selecting words and pictures in Figure 2);
in the case of video-clips, the learner is encouraged to integrate the
presented material with inappropriate prior knowledge (i.e., affect-
ing the process of integrating in Figure 2). Further research is
needed to clarify these theoretical interpretations.
References
Baddeley, A. (1992, January). Working memory.
Science,
255, 556-559.
Bobis, J., Sweller, J., & Cooper, M. (1993). Cognitive load effects in a
primary school geometry task.
Learning
and
Instruction,
3, 1-21.
Chandler, P., & Sweller, J. (1991). Cognitive load theory and the format of
instruction. Cognition
and
Instruction,
8,
293-332.
Chandler,
P., &
Sweller,
J.
(1996). Cognitive load while learning
to
use
a
computer program. Applied Cognitive Psychology,
10, 1-20.
Clark,
J. M., &
Paivio,
A.
(1991). Dual coding theory
and
education.
Educational Psychology Review,
3,
149-210.
Dewey,
J.
(1913). Interest
and
effort
in
education. Cambridge,
MA:
Houghton Mifflin.
Gamer,
R.,
Brown,
R.,
Sanders,
S., &
Menke,
D.
(1992). "Seductive
details" and learning from text.
In
K. A. Renninger, S. Hidi, & A. Krapp
(Eds.),.
The role of interest in learning and development (pp. 239-254).
Hillsdale, NJ: Erlbaum.
Garner,
R.,
Gillingham,
M., &
White,
C.
(1989). Effects
of
seductive
details
on
macroprocessing and microprocessing
in
adults
and
children.
Cognition
and
Instruction,
6,
41-57.
Harp,
S. F., &
Mayer,
R. E.
(1997).
The
role
of
interest
in
learning
from scientific text and illustrations: On the distinction between emotional
interest
and
cognitive interest. Journal
of
Educational Psychology,
89,
92-102.
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,
414-434.
Hidi,
S., &
Baird,
W.
(1988). Strategies
for
increasing text-based interest
and students' recall
of
expository text. Reading Research Quarterly,
23,
465-483.
Jonassen,
D. H., &
Grabowski,
B. L.
(1993). Handbook
of
individual
differences, learning,
and
instruction. Hillsdale, NJ: Erlbaum.
Kalyuga,
S.,
Chandler,
P., &
Sweller,
J.
(1998). Levels
of
expertise
and
instructional design. Human Factors,
40, 1-17.
Kalyuga,
S.,
Chandler,
P., &
Sweller,
J. (in
press). Managing split-attention
and redundancy
in
multimedia instruction. Applied
Cognitive
Psychology.
Macromedia. (1994). Director
4.0
[Computer software].
San
Francisco:
Author.
Mann,
B.
(1997). Evaluation
of
presentation modalities
in a
hypermedia
system. Computers and Education,
28,
133-143.
Mayer,
R. E.
(1996). Learning strategies
for
making sense
out of
expository text: The SOI model
for
guiding three cognitive processes
in knowledge construction. Educational Psychology Review,
8, 357-
371.
Mayer,
R. E.
(1997). Multimedia learning:
Are we
asking
the
right ques-
tions? Educational Psychologist,
32, 1-19.
Mayer,
R. E.
(1999a). Multimedia aids
to
problem-solving transfer. Inter-
national Journal
of
Educational Research,
31,
611-624.
Mayer,
R. E.
(1999b).
The
promise
of
educational psychology. Upper
Saddle River,
NJ:
Prentice Hall.
Mayer,
R. E.
(1999c). Research-based principles
for
the design
of
instruc-
tional messages:
The
case
of
multimedia explanations. Document
De-
sign,
1, 7-20.
Mayer,
R. E., &
Moreno,
R.
(1998).
A
split-attention effect
in
multimedia
learning: Evidence
for
dual processing systems
in
working memory.
Journal
of
Educational Psychology,
90,
312-320.
Mohr,
P.,
Glover,
J., &
Ronning,
R. R.
(1984).
The
effect
of
related
and
unrelated details
on the
recall
of
major ideas
in
prose. Journal
of
Reading Behavior,
16,
97-109.
Moreno,
R., &
Mayer,
R. E.
(1999). Cognitive principles
of
multimedia
learning:
The
role
of
modality
and
contiguity. Journal
of
Educational
Psychology,
91,
358-368.
Moreno,
R., &
Mayer,
R. E.
(2000).
A
coherence effect
in
multimedia
learning:
The
case
for
minimizing irrelevant sounds
in the
design
of
multimedia messages. Journal of Educational Psychology,
92,
117-125.
Paivio,
A.
(1986). Mental representations:
A
dual coding approach.
New
York: Oxford University Press.
Plass,
J. L.,
Chun,
D. M.,
Mayer,
R. E., &
Leutner,
D.
(1998).
Sup-
porting visual
and
verbal learning preferences
in a
second-language
multimedia learning environment. Journal
of
Educational Psychol-
ogy,
90,
25-36.
Renninger, K. A., Hidi,
S., &
Krapp, A. (Eds.). (1992). The role of interest
in learning
and
development. Hillsdale, NJ: Erlbaum.
Shirey,
L., &
Reynolds,
R.
(1988). Effect
of
interest
on
attention
and
learning. Journal
of
Educational Psychology,
80,
159-166.
Sweller,
J., &
Chandler,
J.
(1994). Why some material
is
difficult
to
learn.
Cognition
and
Instruction,
12,
185—233.
Tindall-Ford,
S.,
Chandler,
P., &
Sweller,
J.
(1997). When
two
sensory
modes
are
better than
one.
Journal
of
Experimental Psychology:
Ap-
plied,
3,
257-287.
Wade, S., & Adams, R. (1990). Effects of importance and interest on recall
of biographical text. Journal of Reading Behavior,
22,
331-353.
Wittrock,
M. C.
(1990). Generative processes
of
comprehension. Educa-
tional Psychologist,
24,
345-376.
(Appendixes follow)
198MAYER, HEISER, AND LONN
Appendix A
Video Clips Used in Experiments 3 and 4
ClipNarration scriptVideo images
1 "Lightning can occur in virtually any season and can potentially strike
anywhere at any time."
2 "Scientists can simulate this process in a controlled laboratory
experiment."
3 "In trying to understand the various processes involved, many
scientists create lightning by launching rockets into overhead
clouds."
4 "Statistics show that more people are injured by lightning each year
than by tornadoes and hurricanes combined."
5 "When lightning strikes the ground fulgurites may form as the heat
from lightning fuses sand into the shape of the electricity's path."
6 "Many people thought that lightning was a form of heavenly fire until
Benjamin Franklin conducted his famous experiments with a kite
and key showing that lightning was really a form of electricity."
Several different clips of flashes of lightning in the sky and
clouds, and above a cluster of trees.
Clips of swirling wind within glass-enclosed cloud chamber.
Clips of scientists setting up rockets in an open field, buttons
being pressed on a control box, and small rockets soaring into
overhead clouds.
Clips of trees bending against strong winds, lightning striking into
the trees, ambulance arriving along path near the trees, victim
being carried in stretcher to the ambulance near crowd of
onlookers.
Clips of workers examining fulgurites in an open area, sweeping
off sand with a small brush, and probing fulgurites with metal
instruments.
Several different clips of cloud-to-ground lightning strikes in
various cities, from a panoramic skyline perspective.
Appendix B
Sixteen Narration Segments (Numbered 1 Through 16) and Seductive Details (Numbered SD-1
Through SD-6) With Corresponding On-Screen Summaries (Indicated in Brackets)
(SD-l) Each year lightning kills approximately one hundred and fifty
Americans and injures ten thousand. [Lightning kills 150 Americans,
injures 10,000.]
(1) Cool moist air moves over a warmer surface and becomes heated. [Cool
moist air becomes heated.]
(2) Warmed moist air near the earth's surface rises rapidly. [Warm moist
air rises.]
(3) As the air in this updraft cools, water vapor condenses into water
droplets and forms a cloud. [As the air cools, water vapor condenses into
a cloud.]
(SD-2) On a warm cloudy day, swimmers are sitting ducks for lightning.
[Swimmers are sitting ducks.]
(4) The cloud's top extends above the freezing level, so the upper portion
of the cloud is composed of tiny ice crystals. [The cloud's top is composed
of tiny ice crystals.]
(5) Eventually, the water droplets and ice crystals become too large to be
suspended by updrafts. [The droplets and crystals become too large.]
(6) As raindrops and ice crystals fall through the cloud, they drag some of
the air in the cloud downward, producing downdrafts. [Raindrops and ice
crystals fall, drag air downward.]
(SD-3) When flying through these downdrafts, an airplane ride can become
bumpy but metal airplanes that are struck by lightning usually sustain little
damage. [An airplane ride can become bumpy, airplanes usually sustain
little damage.]
(7) When downdrafts strike the ground, they spread out in all directions,
producing gusts of cool wind people feel just before the start of the rain.
[Downdrafts strike the ground, producing gusts of cool wind.]
(8) Within the cloud, the rising and falling air currents cause electrical
charges to build. [Air currents cause charges to build.]
(SD-4) Golfers are vulnerable targets because they hold metal clubs which
are excellent conductors of electrical charge. [Golfers are targets, metal
clubs are excellent conductors.]
(9) The charge results from the collision of the cloud's rising water droplets
against heavier, falling pieces of
ice.
[The charge results from the collision
of rising droplets against falling ice.]
(10) The negatively charged particles fall to the bottom of the cloud, and
most of the positively charged particles rise to the top. [The negatively
charged particles fall, the positively charged particles rise.]
(11) A stepped leader of negative charges moves downward in a series of
steps.
It nears the ground. [A stepped leader moves downward.]
(SD-5) If lightning strikes a beach, it fuses sand into the shape of the
electricity's path, frying plants and animals in its way. [Lightning fuses
sand into the shape of the electricity's path.]
(12) A positively charged leader travels up from such objects as trees and
buildings. [A leader travels up from objects.]
(13) The two leaders generally meet about 165 feet above the ground. [The
leaders meet 165 feet above ground.]
(14) Negatively charged particles then rush from the cloud to the ground
along the path created by the leaders. It is not very bright. [Negatively
charged particles rush to the ground along the path. It is not bright.]
(SD-6) In Burtonsville, Maryland, a bolt of lightning tore a hole in the
helmet of a football player, burned his jersey, and blew his shoes off.
[Lightning tore a hole in the helmet of a football player, burned his jersey,
blew his shoes off.]
(15) As the leader stroke nears the ground, it induces an opposite charge,
so positively charged particles from the ground rush upward along the
same path. [Positively charged particles rush upward along the same path.]
(16) This upward motion of the current is the return stroke. It produces the
bright light that people notice as a flash of lightning. [This current is the
return stroke. It produces bright light.]
Received May 15, 2000
Revision received July 10, 2000
Accepted July 25, 2000 •
A preview of this full-text is provided by American Psychological Association.
Content available from Journal of Educational Psychology
This content is subject to copyright. Terms and conditions apply.
