ArticlePDF Available

The effect of redundant text in multimedia instruction



The purpose of this study was to examine the redundancy effects obtained when spoken information was duplicated in writing during the learning of a multimedia document. Documents consisting of diagrams and spoken information on the development of memory models were presented to three groups of students. In the first group, no written text was presented. In the second, written sentences redundant with the spoken information were progressively presented on the screen while in the third group, these written sentences were presented together. The results show that whatever the type of text presentation (sequential or static), the duplication of information in the written mode led to a substantial impairment in subsequent retention and transfer tests as well as in a task in which the memorization of diagrams was evaluated. This last result supports the hypothesis that the visual channel is overloaded as the cognitive theory of multimedia learning suggests.
0361-476X/$ - see front matter © 2006 Elsevier Inc. All rights reserved.
Contemporary Educational Psychology 32 (2007) 588–598
The eVect of redundant text in multimedia
Eric Jamet ¤, Olivier Le Bohec
Laboratoire de Psychologie Expérimentale, CRPCC, Université Rennes 2 Haute Bretagne,
place du recteur Henri Le Moal, CS 24307 35043 Rennes Cedex, France
Available online 1 November 2006
The purpose of this study was to examine the redundancy eVects obtained when spoken informa-
tion was duplicated in writing during the learning of a multimedia document. Documents consisting
of diagrams and spoken information on the development of memory models were presented to three
groups of students. In the Wrst group, no written text was presented. In the second, written sentences
redundant with the spoken information were progressively presented on the screen while in the third
group, these written sentences were presented together. The results show that whatever the type of
text presentation (sequential or static), the duplication of information in the written mode led to a
substantial impairment in subsequent retention and transfer tests as well as in a task in which the
memorization of diagrams was evaluated. This last result supports the hypothesis that the visual
channel is overloaded as the cognitive theory of multimedia learning suggests.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Multimedia learning; Redundancy eVect; Cognitive load
1. Introduction
Multimedia material is being increasingly used as an aid to teaching, whether in the
form of on-line courses, CD-ROMs or projections of presentations using the appropriate
software. The studies undertaken in the Weld of cognitive psychology now permit us to gain
*Corresponding author.
E-mail address: (E. Jamet).
E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598 589
a better understanding of the mechanisms involved in the learning of such multimedia doc-
uments (i.e., documents consisting of multiple sources of information). The most inXuential
models to be proposed in this Weld are the cognitive load theory (Paas, Renkl, & Sweller,
2003; Sweller, 1999; Sweller, van Merrienboer, & Paas, 1998) and the cognitive theory of
multimedia learning (Mayer, 2001, 2005).
This cognitive theory of multimedia learning is based on the idea that there are separate
processing systems for pictorial and verbal information and that learning consists of estab-
lishing links between the verbal and pictorial representations. Each of these channels has a
limited processing capacity at any given time. This hypothesis is also central to the cogni-
tive load theory (Sweller, 1999). Finally, meaningful learning is deWned as the deep under-
standing of the material and presupposes the active processing of the document. It is
revealed in the ability to apply what has been learned to new situations, in particular
within the framework of knowledge transfer problems. It requires the construction of a
mental model of the document on the basis of three processes: selection of the important
elements in the presented material, the organization of these elements within a coherent
structure and their integration with existing knowledge.
As far as cognitive load is concerned, a distinction is made between three types of
sources. The Wrst relates to the central processes necessary for learning. The second is asso-
ciated with the incidental processes which are not necessary for learning but which are
induced by the design of the document or the learning situation. The third refers to the
processes involved in the working memory retention of a representation over a given
period of time.
One of the examples of overload presented by Mayer and Moreno (2003) relates to the
incidental processing associated with the way in towards the information is presented.
Here, the authors refer to a situation in which the learner has to manage the simultaneous
presentation of the same information in the written and spoken modes (redundant presen-
tation). According to the authors, this type of presentation may result in cognitive over-
load by unnecessarily increasing the incidental processing operations associated with the
reconciliation of two forms of verbal information. Indeed, this redundancy eVect has been
reported many times in the literature.
In general terms, redundancy occurs when one and the same item of information is pre-
sented in diVerent forms or when unnecessary additional information is proposed (Sweller,
2005). The term “redundancy eVect” is used more restrictively by Mayer (2001) to refer to
the negative eVects obtained when spoken explanations of an illustration are duplicated in
writing. It is this form of redundancy eVect that will be investigated in this article.
In an initial approach to this Weld, researchers studied redundancy eVects during the
presentation of written information which either or was not duplicated in the spoken
mode. In signal detection tasks, redundancy eVects are generally positive (Miller, 1982,
1991; Miller & Reynolds, 2003). Similarly, the retention of words presented both in speech
and in writing is generally better than when words are presented in one mode only (Lewan-
dowski & Kobus, 1993; Penney, 1989 for a review), and the same has been observed for the
retention of lists of numbers (Nordby, Raanaas, & Magnussen, 2002) or warnings about
the eVects of alcohol consumption inserted in television adverts (Barlow & Wogalter,
1993). Positive redundancy eVects have also been observed in connection with more com-
plex texts (Montali & Lewandowski, 1996). However, the opposite results (i.e., the superi-
ority of the spoken mode) have sometimes also been observed (Kalyuga, Chandler, &
Sweller, 2004, Exp. 3). Finally, a number of studies involving in driving or navigation tasks
590 E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598
have revealed better performances for the redundant and spoken modes compared with
the visual mode using an on-board vehicle navigation system (Liu, 2001). However, this
superior performance was not found in a study of the modes of presentation of informa-
tion communicated by air traYc controllers to pilots (Helleberg & Wickens, 2003).
The results relating to learning from illustrated documents are clearer. Even though an
illustrated written document may be particularly eVective (Mayer, 2001), it may constrain
readers to share their attentional resources between the various sources of visual informa-
tion if these are not intelligible on their own. This consequently increases the cognitive load
in the visual channel (Sweller et al., 1998). This “split-attention eVect” (Ayres & Sweller,
2005 for a review) can be reduced if parts of the text are moved to the corresponding places
on the illustration (Chandler & Sweller, 1991, 1992; Kester, Kirschner, & van Merriënboer,
2005; Sweller & Chandler, 1991; Sweller, Chandler, Tierney, & Cooper, 1990) or if pop-up
windows are integrated in the diagram (Erhel & Jamet, 2006). To avoid the split-attention
eVect observed for illustrated texts, it is possible to use the spoken mode to explain a dia-
gram (Ginns, 2005 for a review).
A number of studies undertaken in this research Weld have analyzed the eVects of adding
redundant information to information presented in writing. For example, the study con-
ducted by Kalyuga, Chandler, and Sweller (1999) evaluated the role of the mode in which
explanations of a diagram concerning the fusion of materials were presented. The explana-
tions of the diagram were presented in oral, visual or redundant conditions. In the two
tasks in the test phase, the spoken explanation group outperformed both the visual expla-
nation group and the redundant group (redundancy eVect). The subjective mental load
questionnaires also indicated that the spoken explanation group was at an advantage.
These redundancy eVects were identiWed again in later studies (Craig, Gholson, & Driscoll,
2002; Kalyuga, Chandler, & Sweller, 2000; Leahy, Chandler, & Sweller, 2003).
In a more complete study in which documents relating to the formation of lightening
were presented to students (Moreno & Mayer, 2002), the presence of redundant explana-
tions in the spoken and written modes were manipulated while an animation was or was
not presented simultaneously. They therefore compared the successive/simultaneous and
redundant/non-redundant versions. In this case, the authors observed an interaction
between the two factors. When presentation was simultaneous, the redundancy eVect was
negative, thus conWrming the results of other studies (Kalyuga et al., 1999, 2000). When
presentation was successive, the redundancy eVects were positive again. In such case, learn-
ing does not require resources to be shared between the written words and the illustrations.
In consequence, if the negative redundancy eVect is linked to the joint processing of the
various sources of visual information, one solution would consist of successive oral and
written presentations (Kalyuga et al., 2004). This is therefore more a case of repetition than
of redundancy. In cases where this type of solution is not possible, for example a Power-
Point course presented with a video projector, other solutions could be evaluated. One of
these could result from a less literal approach to redundancy. In all the studies presented
here, the written text strictly duplicated the spoken explanation. It is possible to reduce the
quantity of written information by simply presenting a summary on the screen. These
reduction had no eVect in a study conducted by Mayer and coworkers (Mayer, Heiser, &
Lonn, 2001). However, a study which we recently undertook involving the multimedia
teaching of accounting for economics students (Le Bohec & Jamet, in press) suggests that
the summary format seems to represent a good compromise. It returned performances
equal to those observed in the no-redundancy condition without being judged as
E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598 591
negatively as the no-redundancy condition in terms of speed of presentation and pleasant-
ness on a subjective satisfaction scale.
To summarize, when purely verbal information is presented, the redundancy eVects are
generally positive. When another source of visual information such as an illustration is pre-
sented at the same time as this verbal information then the redundancy eVects are negative.
These negative eVects may disappear if the illustration is presented after the redundant text
or if the written text is presented after the illustration and oral explanation.
Another solution which, to our knowledge, has never been evaluated is to use the
sequential presentation of written information in order to reduce the quantity of visual
information on the screen. This type of presentation is, for example, used in studies of
redundancy without, however, subjecting this to any speciWc evaluation (Kalyuga et al.,
1999). It may be of interest for redundancy studies to examine the impact of presenting
written information on screen all at once or, for example, cumulatively on a sentence-by-
sentence basis. This type of presentation, which is very easy to achieve using slide-show
software, would have the beneWt of restricting (at least at the start of each slide) the quan-
tity of visual information and should reduce the overload in the visual channel.
The Wrst aim of this study was to replicate the negative eVect of redundancy described
above, as well as to test the hypothesis according to which redundancy eVects are propor-
tionate to the quantity of written information presented on-screen. In other words, the
present of a written text which is redundant with the spoken explanation should have a
negative eVect on learning because it overloads the visual channel. In contrast, if the writ-
ten sentences are presented sequentially, the quantity of visual information will be smaller
at the start of each slide and the redundancy eVect should diminish.
The second aim of this study was to test the speciWc eVect of redundancy on the reten-
tion of the illustrated information. Indeed, if the negative eVect of redundancy is linked to
the overloading of the visual channel as the cognitive theory of multimedia learning pre-
dicts, it should be particularly strong in tasks requiring the retention of information pre-
sented on screen (i.e., texts or diagrams). While text retention tasks have frequently been
used, diagram retention tasks have never been employed in the studies described above
with the exception of those conducted by Moreno and Mayer (2002). However, it is in this
type of task that particularly strong eVects should be observed since, unlike texts which are
presented in both the oral and written modes, diagrams are only presented visually. Their
processing should therefore be particularly impaired by the presence of other visual infor-
mation. We predicted that redundancy would have a negative eVect on this type of infor-
mation since the presence of written information will cause learners to refocus their visual
attention away from the diagram and towards the text or, at least, share their attention
between these two sources of information.
To test these hypotheses, in the study presented below, we confronted students with
a multimedia document in the Weld of cognitive psychology dealing with the develop-
ment of memory models. In this document, written information redundant with that
presented orally was either absent, present in full or present in full but presented
sequentially on screen. The participants were then asked to complete text retention, dia-
gram completion and knowledge transfer tasks for the purposes of evaluation. We pre-
dicted that performances for all of these tasks will be aVected negatively by
redundancy, but also that this eVect should be particularly strong for the diagram com-
pletion task. Sequential presentation should reduce negative eVects of redundancy for
all of these tasks.
592 E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598
2. Experiment
2.1. Method
2.1.1. Participants and design
The participants consisted of 90 second-year undergraduate students recruited from the
psychology faculty at the University of Rennes (France). The Median age was 20 and the
overall percentage of women was 83%. A pre-test questionnaire revealed that none of them
had received any instruction concerning memory models. The students were randomly
assigned to three training groups in a single factor, between-subjects design. There were 30
students in each group.
2.1.2. Materials and apparatus
The computer-presented material consisted of three documents which were presented in
succession. The documents took the form of three illustrated texts dealing with the ques-
tion of memory functioning. The texts used presented (a) the Atkinson and ShiVrin model
(1968), (b) Baddeley’s model (Baddeley, 1986) and (c) Cowan’s model (Cowan, 1995). The
texts were constructed to be roughly equivalent in terms of number of words: 386 words
for the Atkinson and ShiVrin model, 363 words for Baddeley’s model and 379 words for
Cowan’s model. Each slide contained at most three or four sentences.
Each model was presented for a period of approximately 220s and consisted of six
slides. The diagram corresponding to the model was present on the screen at all times. The
spoken explanation was the same in all the groups. The students could not replay the nar-
ration and consequently, the presentation of the texts took 11min for all participants.
After each model, presentation stopped on a blank screen and students had to click on a
button to move on to the next model.
In the no text group, no written text was presented alongside the diagram. In the full
text group, the text corresponding to the voice presentation was displayed next to the dia-
gram. In the sequential text group, the text was displayed cumulatively on screen sentence-
by-sentence (see Fig. 1). The multimedia presentations were developed using Director
MX™ (Macromedia, 2002). The apparatus consisted of 6 Pentium III PC computer sys-
tems, each with a 15-inch monitor.
For each participant, the paper-and-pencil materials consisted of a subject question-
naire, a retention test, a transfer test and a diagram completion test. The subject question-
naire investigated the participants’ prior knowledge of memory models by means of four
general questions. Two more speciWc questions were used to verify whether the participants
had already been taught about memory. The participants completed this questionnaire
even if memory model courses were not to be available until later in the academic year in
order to detect any students who might have been retaking the year or subjects who had
chosen optional courses.
The retention test consisted of 12 literal open-ended questions (for example, “What is
the role of the central executive in Cowan’s model?”).
The transfer test consisted of 12 inferential open-ended questions which required the
participants to use their knowledge of the text and transfer this to new situations in order
to make their responses (for example, “Why do you think it is more diYcult to learn letters
which are phonologically similar than those which are phonologically dissimilar?”). The
diagram completion test consisted of three pages containing the diagrams presented during
E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598 593
the learning stage but without the captions. The participants had to complete the diagram
for example, in the case of Baddeley’s model – which corresponded to the simplest of the
diagrams – the task consisted of writing the names of the various components (central
executive, phonological loop, visuo-spatial sketchpad).
2.1.3. Procedure
To complete the experiment, each student worked at a computer with a connected head-
set. On arrival, each participant completed the prior knowledge test. They were then told
that they were going to see three texts concerning memory operation and that, after the
presentation they would have to answer a questionnaire relating to what they had seen and
Presentation of consigns and text tooks approximately 15min. At the end of the presen-
tation, they answered three questionnaires (with a 15-min time limit).
2.1.4. Scoring
A list of possible correct answers was produced for each question in the retention test
and transfer test. These questionnaires were scored by awarding one point for each ques-
tion with a correct response. For the diagram completion task, the participants received
one point for each correct caption at the right place in the diagram.
Fig. 1. Selected frames from the sequential-text group. The no-text group corresponds to the Wrst frame, the full-
text group to the last.
594 E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598
3. Results and discussion
Table 1 presents the mean scores and standard deviations for each group on each of the
three tests.
We conducted a one-way analysis of variance, with redundancy levels as the between-
subjects factor. Planned comparisons were used to explore further the diVerences between
the cells of the ANOVA design with p<.05.
3.1. Verbal retention task
There was a signiWcant diVerence in the retention scores obtained by the groups receiv-
ing no text (MD4.4, SD D2.22), sequential text (MD2.87, SD D2.24) and full text
(MD2.27, SD D2.02), F(2, 87) D7.84, MS E D36.3, pD.001, 2D.15. Planned comparisons
were conducted and showed that the no-text group performed better than the full-text
group, t(87) 2.52, pD.01, Cohen’s dD1.01, and the sequential text group, t(87) 2.88,
pD.005, Cohen’s dD0.69. There was no signiWcant diVerence between the sequential text
group and the full text group.
3.2. Diagram completion task
The ANOVA revealed a signiWcant eVect of the Information Presentation Format on
subjects’ retention performances, F(2, 87)D21.89, MSE D387, p< .001, 2D.34. Planned
comparisons indicated a signiWcant diVerence between the no-text group (MD14.98,
SD D3.12) and the full text group (MD8.22, SD D4.24); t(87) 7.04, p<.001, Cohen’s
dD1.81. Similarly, the performances of the sequential text group (MD9.5, SD D5.04) were
signiWcantly poorer than those of the no-text group, t(87) 5.06, p< .001, Cohen’s
dD1.3. There was no signiWcant diVerence between the sequential text group and the full
text group.
3.3. Transfer task
There was a signiWcant diVerence in retention scores between the groups receiving no
text (MD3.63, SD D2.22), sequential text (MD2.18, SD D1.82) and full text (MD2.37,
SD D1.75), F(2, 87) D4.94, MSE D18.7, p<.01, 2D.10. Planned comparisons were con-
ducted and revealed that the no-text group performed better than the full text group
t(87) 3.92, p< .001, Cohen’s dD0.63, and the sequential text group, t(87) 2.76,
Mean score and standard deviations on retention, diagram completion and transfer tests for three groups
Note. Maximum score is 12 for retention and transfer tests and 20 for diagram completion test.
Group Retention Diagram completion Transfer
No text 4.4 2.19 14.98 3.12 3.63 2.22
Sequential text 2.87 2.24 9.5 5.0 2.18 1.82
Full text 2.27 2.02 8.22 4.23 2.37 1.75
E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598 595
pD.007 (Cohen’s dD0.71). There was no signiWcant diVerence between the sequential text
group and the full text group.
To summarize, the Wrst hypothesis predicted that the negative redundancy eVects in the
full text condition would be observed in all the conditions but also that they would be par-
ticularly great in the diagram completion task. This hypothesis was conWrmed: while the
eVects were moderate for the transfer test (dD.63) and strong for the retention test
(dD1.01), they were greater for the diagram test (dD1.81).
The second hypothesis put forward in this study was that the sequential presentation of
the text would reduce the negative eVect of redundancy. Even though the size of the eVect
for the retention and diagram completion tests was clearly smaller in the case of sequential
presentation, no signiWcant diVerence was observed between the full text and sequential
text groups; a Wnding which was at odds with our hypothesis.
4. General discussion
The Wrst aim of this study was to replicate the negative redundancy eVects observed on
illustrated documents and then to test the eVects of sequential text presentation on this
redundancy eVect. First of all, the negative redundancy eVect already observed during the
learning of animated (Craig et al., 2002; Mayer et al., 2001; Moreno & Mayer, 2002) or
static illustrated documents (Kalyuga et al., 1999, 2000) was again identiWed in this study,
independently of the mode of text presentation.
As far as we know, the results relating to the diagram completion task are more innova-
tive. As we expected, performance on this task was particularly impaired when information
that was redundant with the spoken explanations of the diagram was presented on screen.
The study conducted by Moreno and Mayer (2002) made use of a similar task (matching
test). The result of this task was not signiWcantly aVected by redundancy when there was a
simultaneous presentation with the animation (condition closest to our study). However,
according to the authors, this task did not have a very high discriminatory value. Unlike
Moreno and Mayer’s task, the task used in this study employed a larger number of more
complex diagrams and a more varied response scale (0–20).
Our results clearly validate the hypothesis that the visual channel is overloaded as pro-
posed by the underlying theory of multimedia learning (Mayer, 2001; Mayer et al., 2001)
given that the learning performance that is most aVected by the redundancy is associated
with the information source that directly competes with the written text in terms of visual
processing (i.e., the diagram).
Our second hypothesis was that the sequential presentation of the text would reduce this
eVect. Even though the negative eVects were reduced by approximately 30% in this format,
no signiWcant diVerence was observed between the static and sequential formats. A number
of explanatory hypotheses can be advanced. First of all, this type of presentation only
reduces the quantity of information presented on screen by a few seconds since the presen-
tation of the sentences is cumulative (i.e., each sentence appears below the last one with a
maximum of four sentences on the screen). It will be necessary to test a presentation in
which each sentence disappears after being displayed in order to eliminate this problem. It
should also be noted that the number of sentences was limited to four per screen in each
type of presentation. This quantity of information is perhaps too small to reveal positive
eVects in sequential presentations. Finally, it is possible that the redundancy eVect is to
some extent independent of the quantity of written information present on the screen. In
596 E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598
eVect, the cognitive management of a redundant multimedia document consists of at least
three components. One of these consists of listening to sentences and selecting the impor-
tant information. The second, which is speciWc to redundant documents, consists of visu-
ally scanning the screen in order to Wnd a heard sentence in the written text and then read
it. The third component, which does not necessarily occur after the second, consists of
looking at the illustration and identifying the part that corresponds to the text in order to
establish referential links between the verbal and pictorial representations (Mayer, 2001,
As far as the second component is concerned, our hypothesis was that the stage of visual
scanning of the text would be aVected by sequential presentation. We did not observe any
such eVect. This absence could be due to the fact that, in this case, sequentiality only acted
on one of the stage with a low level of explanatory power in terms of the redundancy eVect.
In other words, if sequentiality is capable of inXuencing the written text scanning phase,
there is nevertheless no reason to believe that the process of reading the sentence can be
inXuenced by this presentation. However, this reading time may sometimes be suYcient to
impair the phase during which the illustration is processed because the learners are not
able to control the speed of the oral presentation. This hypothesis seems to be supported in
our study by the strong negative eVects of the presence of a written text – whether sequen-
tial or not – on the memorization of the diagram. Otherwise stated, it is possible that it is
the time taken to read the sentence and not the time required to locate it on the screen that
impairs the processing of the illustration.
Similarly, the sequential presentation of the text cannot aVect the phase during which
the referential links between the verbal and pictorial representations are established. How-
ever, this phase is often complex and imposes a high cognitive load, in particular when the
text and illustrations are presented separately on the screen (Ayres & Sweller, 2005). In the
future, it will be interesting to test a sequential presentation of the text in parallel with the
presentation of elements of the diagram in order to facilitate this phase during which the
referential links are constructed. In eVect, in such a case, the successive presentation of the
illustrated and textual elements could facilitate the establishment of relations between
them and, in particular, the search in the diagram since the last element to appear would be
the element referred to in the text. This type of presentation would thus resemble other
modes of signaling the illustrated elements, the positive eVects of which are already known
(Craig et al., 2002; Jeung, Chandler, & Sweller, 1997; Tabbers, Martens, & van Merriënb-
oer, 2004).
As far as the limitations of this study are concerned, we should Wrst of all point out that
all the participants found it very diYcult to understand the document as the relatively low
scores on the retention and transfer tests indicate. This relative diYculty could have
aVected the conclusions and more research is required in this regard.
Secondly, the learners did not have very many ways of controlling the process (a pause
after each model). This situation is probably suYcient for a proportion of the observed
cognitive overload to be associated with the absence of possible compensations in terms of
time. A presentation in which the introduction of each phase is controlled by the learner
would probably help complement the present results.
Finally, it is necessary to note that the three documents used (the memory models pro-
posed by Atkinson, Baddeley and Cowan) varied in complexity at both the graphical and
conceptual levels. It was not possible to counterbalance the order of the documents
because there is a theoretical (and historical) logic to the order used in our study. However,
E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598 597
the eVects are probably very diVerent for relatively simple models such as Baddeley’s and
considerably more complex models such as the one developed by Cowan. Nothing in our
study enables us to evaluate the eVects of diagram complexity on the redundancy eVect but
it is likely that they are non-negligible. More research is needed to evaluate these eVects, in
particular through the experimental manipulation of this level of complexity.
As far as the practical implications of the study are concerned, it seems clear that when
learners learn a multimedia document without having control over the presentation,
redundant written texts should not be used. The negative eVects of this redundancy are
particularly strong in the case of diagram retention. However, they are also observed in
text retention and knowledge transfer tasks. The learning situation used here was relatively
similar to lecture situations involving the use of a video projector. Apart from their genu-
ine didactic content and relatively long duration, these multimedia documents were pre-
sented in a way which the learners could not control. Under such circumstances, the results
of our study clearly indicate that the written duplication of the text leads to a considerable
performance impairment. Despite this, this type of redundant presentation remains fre-
quent both in the context of classwork and oral presentations in workshops.
Atkinson, R. C., & ShiVrin, R. M. (1968). Human memory: A proposed system and its control processes. In K. W.
Spence & J. T. Spence (Eds.), The psychology of learning and motivation (Vol. 2). New York: Academic Press.
Ayres, P., & Sweller, J. (2005). The split-attention principle in multimedia learning. In R. B. Mayer (Ed.), The
Cambridge handbook of multimedia learning (pp. 135–146). Cambridge: University Press.
Baddeley, A. D. (1986). Working memory. Oxford: Oxford University Press.
Barlow, T., & Wogalter, M. S. (1993). Alcoholic beverage warnings in magazine and television advertisements.
Journal of Consumer Research, 20(1), 147.
Chandler, P., & Sweller, J. (1991). Cognitive load theory and the format of instruction. Cognition & Instruction, 8,
Chandler, P., & Sweller, J. (1992). The split-attention eVect as a factor in the design of instruction. British Journal
of Educational Psychology, 62, 233–246.
Cowan, N. (1995). Attention and memory: An integrated framework(Vol. 26). New York: Oxford University Press.
Craig, S. D., Gholson, B., & Driscoll, D. M. (2002). Animated pedagogical agents in multimedia educational envi-
ronments: EVects of agent properties, picture features, and redundancy. Journal of Educational Psychology,
94(2), 428–434.
Erhel, S., & Jamet, E. (2006). Using pop-up windows to improve multimedia learning. Journal of Computer
Assisted Learning, 22, 137–147.
Ginns, P. (2005). Meta-analysis of the modality eVect. Learning and Instruction, 4, 313–331.
Helleberg, J. R., & Wickens, C. D. (2003). EVects of data-link modality and display redundancy on pilot perfor-
mance: An attentional perspective. International Journal of Aviation Psychology, 13(3), 189–210.
Jeung, H.-J., Chandler, P., & Sweller, J. (1997). The role of visual indicators in dual sensory mode instruction. Edu-
cational Psychology, 3, 329–343.
Kalyuga, S., Chandler, P., & Sweller, J. (1999). Managing split-attention and redundancy in multimedia instruc-
tion. Applied Cognitive Psychology, 13, 351–372.
Kalyuga, S., Chandler, P., & Sweller, J. (2000). Incorporating learner experience into the design of multimedia
instruction. Journal of Educational Psychology, 92(1), 126–136.
Kalyuga, S., Chandler, P., & Sweller, J. (2004). When redundant on-screen text in multimedia technical instruction
can interfere with learning. Human Factors, 46(3), 567–581.
Kester, L., Kirschner, P. A., & van Merriënboer, J. J. G. (2005). The management of cognitive load during complex
cognitive skill acquisition by means of computer-simulated problem solving. British Journal of Educational
Psychology, 75(1), 71–85.
Le Bohec, O., & Jamet, E. (in press). Level of redundancy, note-taking and multimedia learning. In J. F. Rouet, R.
Lowe, & W. Schnotz (Eds.), Multimedia Comprehension.
598 E. Jamet, O. Le Bohec / Contemporary Educational Psychology 32 (2007) 588–598
Leahy, W., Chandler, P., & Sweller, J. (2003). When auditory presentations should and should not be a component
of multimedia instruction. Applied Cognitive Psychology, 17, 401–418.
Lewandowski, L. J., & Kobus, D. A. (1993). The eVects of redundancy in bimodal word processing. Human Per-
formance, 6(3), 229.
Liu, Y.-C. (2001). Comparative study of the eVects of auditory, visual and multimodality displays on drivers’ per-
formance in advanced traveller information systems. Ergonomics, 44(4), 425–442.
Mayer, R. E. (2001). Multimedia learning. New York: Cambridge University press.
Mayer, R. E. (2005). The Cambridge handbook of multimedia learning. Cambridge: Cambridge University Press.
Mayer, R. E., Heiser, J., & Lonn, S. (2001). Cognitive constraints on multimedia learning: When presenting more
material results in less understanding. Journal of Educational Psychology, 93(1), 187–198.
Mayer, R. E., & Moreno, R. (2003). Nine ways to reduce cognitive load in multimedia learning. Educational Psy-
chologist, 38(1), 43–52.
Miller, J. (1982). Divided attention: Evidence for coactivation with redundant signals. Cognitive Psychology,
14(2), 247–279.
Miller, J. (1991). Channel interaction and the redundant-targets eVect in bimodal divided attention. Journal of
Experimental Psychology: Human Perception and Performance, 17(1), 160–169.
Miller, J., & Reynolds, A. (2003). The locus of redundant-targets and nontargets eVects: Evidence from the psy-
chological refractory period paradigm. Journal of Experimental Psychology: Human Perception and Perfor-
mance, 29(6), 1126–1142.
Montali, J., & Lewandowski, L. (1996). Bimodal reading: BeneWts of a talking computer for average and less
skilled readers. Journal of Learning Disabilities, 29(3), 271.
Moreno, R., & Mayer, R. E. (2002). Verbal redundancy in multimedia learning: When reading helps listening.
Journal of Educational Psychology, 94(1), 156–163.
Nordby, K., Raanaas, R. K., & Magnussen, S. (2002). The expanding telephone number part 1: Keying brieXy pre-
sented multiple-digit numbers. Behaviour & Information Technology, 21(1), 27–38.
Paas, F., Renkl, A., & Sweller, J. (2003). Cognitive load theory and instructional design: Recent developments.
Educational Psychologist, 38, 1–4.
Penney, C. G. (1989). Modality eVects and the structure of short-term verbal memory. Memory & Cognition, 17(4),
Sweller, J. (1999). Instructional design in technical areas. Melbourne: ACER Press.
Sweller, J. (2005). The redundancy principle in multimedia learning. In R. B. Mayer (Ed.), The Cambridge hand-
book of multimedia learning (pp. 159–168). Cambridge: Cambridge University Press.
Sweller, J., & Chandler, P. (1991). Evidence for cognitive load theory. Cognition & Instruction, 8, 351–352.
Sweller, J., Chandler, P., Tierney, P., & Cooper, M. (1990). Cognitive load as a factor in the structuring of technical
material. Journal of Experimental Psychology: General, 119(2), 176–192.
Sweller, J., van Merrienboer, J. J. G., & Paas, F. (1998). Cognitive architecture and instructional design. Educa-
tional Psychology Review, 10, 251–296.
Tabbers, H. K., Martens, R. L., & van Merriënboer, J. J. G. (2004). Multimedia instructions and cognitive load
theory: EVects of modality and cueing. British Journal of Educational Psychology, 74(1), 71–81.
... Both CLT and CTML hold that redundancy impedes learning, though they make different theoretical assumptions (which will be further discussed in section two). Numerous studies support this idea (e.g., Kalyuga et al., 1999;Mayer et al., 2001;Jamet and Bohec, 2007;Austin, 2009). However, many studies report empirical results suggesting that redundancy enhances learning (e.g., Adegoke, 2017) or has no effect on learning (e.g., Chu, 2006). ...
... 2];Craig et al., 2004, [exp. 1];Debuse et al., 2009; DeLeeuw and Mayer, 2008, [exp. 1 and 2];Dooley, 2015;Dousay, 2016;Farías et al., 2014;Fenesi et al., 2015;Jadin et al., 2009;Jamet and Bohec, 2007; Kalyuga et al., 1999, [exp. 1];Kalyuga et al., 2000, [exp. ...
Full-text available
Regarding the redundancy effect in multimedia learning environments, more consistency is needed in the theoretical assumptions and investigation of this effect. Current research lacks a comprehensive account of different redundant scenarios in which materials facilitate or inhibit learning and provides little conceptual guidance on how learning processes are affected by different types of redundancy. Theoretical assumptions refer to redundancy as a contentual overlap of information provided by the learning material; in this case, processing duplicated information strains the learners’ limited cognitive capacities. Other assumptions refer to the role of processing limitations in working memory channels, including separate processing for visual and verbal information. In this case, an ineffective combination of sources leads to an overload of the limited working memory capacity. This paper reviews empirical research on the redundancy effect (63 studies) and classifies two types of redundancy: (1) content redundancy, and (2) working memory channel redundancy. From an instructional psychology perspective, the analyses reveal four different implementations of redundant scenarios: (1) adding narration to visualizations, (2) adding written text to visualizations, (3) adding written text to narration, and (4) adding written text to narrated visualizations. Regarding the effects of the two redundancy types within these scenarios, analyses indicate positive effects of content redundancy (affected by learners’ prior knowledge), negative effects of working memory channel redundancy (regarding visualizations and written text), and positive effects of working memory channel redundancy (regarding narration and written text). Moreover, results point to factors that might moderate the effect of redundancy and illustrate interactions with existing multimedia effects. Overall, this review provides an overview of the state of empirical research and reveals that the consideration of both redundancy types provides further explanations in this field of research.
... Empirical studies concerning the redundancy principle focused on comparing students' learning outcomes in the animation with narration group with those in the animation, narration, and verbatim on-screen text group (Craig et al., 2002;Jamet & Le Bohec, 2007;Mayer et al., 2001;Moreno & Mayer, 2002). Immediate posttest results indicated that students in the animation with narration group outperformed those in the redundant text group, meaning that redundant text hinders learning. ...
... Like the studies that gave rise to the modality principle, studies relevant to the redundancy principle were limited by a common validity threat generated by uncontrolled multimedia lessons. For example, topics in multimedia lessons ranged from the formation of lightning (Craig et al., 2002;Mayer et al., 2001;Moreno & Mayer, 2002), human memory (Jamet & Le Bohec, 2007), C++ Computer Programming (Cheah & Leong, 2019), to stem cells (Tarchi et al., 2021). Recently, Liu et al. (2021) and Baceviciute et al. (2022) both examined the redundancy principle in virtual reality condition. ...
Full-text available
The modality and redundancy principles are two fundamental principles used to inform the design of multimedia instruction. They are based on a variety of experimental studies that utilized different types of multimedia lessons to compare input modes of graphics+audio, graphics+text, and graphics+audio+text with each other. However, a lack of control of multimedia lessons in previous studies creates a threat to validity because a single case scenario without following certain principles is not sufficient to represent a construct. Therefore, this study addressed this inherent validity threat and reinvestigated the applicability of the modality and redundancy principles when students learned during a controlled multimedia lesson. In this study the multimedia lesson was developed to follow a series of multimedia learning principles. These principles ensured that the lesson was representative of different types of multimedia lessons. Additionally, they ensured that the multimedia lesson was conducive to learning, since those that were not helpful would not be utilized for instruction in the first place. Eighty-six students in a research university in the US took a prior knowledge survey. They were then randomly assigned to the three input mode conditions and watched the multimedia lesson about the formation of lightning. Subsequent retention and transfer tests revealed that there were no statistically significant differences among the three input mode conditions. Therefore, both the redundancy and modality effects disappeared. This study provided an updated understanding of the applicability of the two important principles for multimedia instruction. Limitations and implications were discussed.
... Generally, redundancy is considered as negative, individuals learn better when the pedagogical document is constructed of pictures and spoken words rather than pictures, spoken and written text presented simultaneously (Jamet & Le Bohec, 2007;Kalyuga et al., 2004;Mayer & Fiorella, 2014). ...
... The current study aims to investigate the impact of full subtitles and keywords (i.e., partial subtitles) on learning in video. When focused on the redundant effect, previous studies appeared either to focus on the effect of subtitles (e.g., Jamet & Le Bohec, 2007) or on the effect of keywords (e.g., Mayer & Johnson, 2008). In a study conducted by de Koning, van Hooijdonk, and Lagerwerf (2017), the effects of oral speech, written subtitles and on-screen labels on procedural learning were analyzed. ...
Full-text available
Learning via videos presents many positive aspects (e.g., animation, multi-modality) but also has some constraints. For example, when subtitles are provided, a split-attention effect could occur between the oral narration, written text, and visual illustration. The presentation of only a few written keywords instead of subtitles may be a good solution in terms of how to guide learners into their information selection process. In the current study, 96 participants were distributed among four experimental conditions. They were shown a 12-minutes video with or without subtitles, and with or without highlighted information (i.e., keywords). The results showed no effect of subtitles, but keywords had a negative impact on content memorization, comprehension, and on the time allocated to learning. The results are discussed in terms of metacognition and learners’ strategies. It is then hypothesized that learners did not use keywords as relevant scaffolds. Instead of being guided into the selection process, learners may have considered that the keywords replaced it, and overestimated their learning.
... In this particular example, the user is guided by the text shown on the screen into selecting the "Manage" option from a list of options displayed on a dropdown menu. Audio explanations could also have been added to the annotated text, but the additional information provided in this manner would have been redundant and counterproductive to the learning process (Jamet & Le Bohec, 2007). The modules with annotated text also functioned as passive tutorials because they did not require the students to provide any input during a tutorial session. ...
... When sound, subtitles and video are presented at the same time, there will be a phenomenon of competing cognitive resources, which will produce cognitive load and then affect the learning effect ( [26], p 62-64). Jamet and Lebohec [16], Ritzhaupt et al. [35] confirmed this phenomenon through experiments. ...
Full-text available
Learning through video is an important learning method at present. In various teaching videos, subtitle size varies greatly. Therefore, it is essential to evaluate the influence of subtitle size on different teaching videos. In this paper, 90 college students from a Chinese University were randomly divided into three groups: small subtitle teaching video group, middle subtitle teaching video group and large subtitle teaching video group. The experiment was conducted to explore the effect of subtitle size on learners’ cognitive load, academic performance, and learning satisfaction. The results showed that: (1) compared with the large subtitle teaching video, learners had lower cognitive load when watching small and medium subtitles teaching videos; (2) compared with the large subtitle teaching video, the learners who watched the small subtitle teaching video had better performance; (3) compared with the large subtitle teaching video, the learners who watched the small and medium subtitle teaching video had higher learning satisfaction. Based on these results, this paper emphasizes the importance of subtitle size in teaching video. We suggest that the relevant departments should further refine the production standards of teaching videos; teachers should try to make teaching videos with smaller subtitles; students should choose teaching videos with smaller subtitles; researchers and technical teams should strengthen the research and development of personalized subtitle technology, so as to allow the private customization of subtitle size based on learners’ needs and preferences.
... Mayer et al. (2001) defined 'redundancy effect' as a detrimental consequences of repeating verbal remarks in writing form. Because it must be processed alongside other visual components on the slide, the redundant information offered in the form of subtitles (written representations of spoken words) may overwhelm the visual processing channel and impede learning (Jamet & Bohec, 2007). When learners are confronted with several and conflicting sources of information at the same time, they may get overwhelmed, resulting in superfluous cognitive load that can be harmful to learning (Mayer, 1997;Mayer et al., 2001). ...
Full-text available
Language assistance becomes an integral component of the teaching–learning process in non‐English– speaking multilingual societies where education takes place predominantly in local languages. Subtitles have been found to be beneficial in facilitating understanding of English language media in such situations. However, when accompanied with subtitles, information‐rich instructional materials can add to the learner's cognitive load as their attention must be split between the content and the subtitle. To evaluate the cognitive and affective impacts of subtitle language on learning, we conducted a multimodal study using eye tracking, electroencephalogram (EEG), self‐report, and pre–post test data of 51 individuals watching a 12‐minute educational video with either no subtitles, English subtitles (L2) or subtitles in their native language (L1). We discovered (a) positive learning gains for groups with subtitles, whether L1 or L2, which suggested positive impact of subtitles on learning outcome; (b) native language subtitling yielded the highest instructional efficiency and supported effective distribution of visual attention between slide content and subtitle; (c) native language subtitles elicited the lowest cognitive load, as evidenced by both EEG measurements and self‐report data; and (d) gaze data highlighted key strategies of high and low performers during interaction with subtitled media. Practitioner notes What is already known about this topic Use of subtitles for language acquisition is an extensively researched topic. It is well established that subtitles are beneficial for language acquisition. Use of subtitles for content learning is relatively less understood with no robust finding. What this paper adds Provides a rich literature review about the impact of subtitles, namely, beneficial, detrimental or neutral, for content learning. Offers empirical results demonstrating positive impact of native language subtitles for content learning. Adds to the limited literature of multimodal investigation of learning with subtitles. Implications for practice and/or policy English language e‐learning content designers may consider adding native language subtitles for global adoption of learning products.
Digital and online learning is more prevalent than ever, making multimedia learning a primary objective for many instructors. The Cambridge Handbook of Multimedia Learning examines cutting-edge research to guide creative teaching methods in online classrooms and training. Recognized as the field's major reference work, this research-based handbook helps define and shape this area of study. This third edition provides the latest progress report from the world's leading multimedia researchers, with forty-six chapters on how to help people learn from words and pictures, particularly in computer-based environments. The chapters demonstrate what works best and establishes optimized practices. It systematically examines well-researched principles of effective multimedia instruction and pinpoints exactly why certain practices succeed by isolating the boundary conditions. The volume is founded upon research findings in learning theory, giving it an informed perspective in explaining precisely how effective teaching practices achieve their goals or fail to engage.
This study examines the effect of modality and redundancy in an Augmented Reality (AR) based language learning environment for teaching vocabulary in a foreign language. The study aims to determine the effectiveness of the redundancy and modality principles, examine the relationship between cognitive load, satisfaction, anxiety, willingness, and achievement, identify predictors of achievement, and understand how students perceive AR-based vocabulary learning in three different presentation modes (animation + narration + text, animation + narration, and animation + text). The study used a pre- and post-test control group quasi-experimental design, with 97 Turkish-speaking undergraduate students from the Faculty of Education at a state university participating. The study participants were divided into three groups and practiced using the same AR-based flashcards in three different modalities (animation + narration + text, animation + narration, and animation + text). After completing a short distractive task during the practice session, the participants were given an achievement test. The analysis of the data revealed that while the group that used animation + narration outscored the other two groups, the difference was only significant between the animation + text group and the animation + narration group, which supported the modality principle. However, there was no significant difference between the animation + narration + text group and the animation + narration group, which did not support the redundancy principle. The study also included qualitative data from a focus group interview to support the quantitative findings. The conclusion of the study is that the principles of CTML (Cognitive Theory of Multimedia Learning) should be reconsidered in relation to the use of AR technology in instructional settings.
Full-text available
In this study, we examined the impact of irrelevant visual images in a PowerPoint lecture on attention and information retention performance. We found students viewing text-only PowerPoint slides retained less information than students viewing text-and-image slides, but the difference did not reach significance. However, when examining both fixation counts and fixation duration, we found students viewing text-and-image PowerPoint slides who spent more time looking at the text retained more information. We discuss the impact of varied visual attention (moving back and forth from text to image) on information integration and retention and establish empirical questions for future research.
Full-text available
Several cognitive concepts can provide hypotheses concerning appropriate structures for instructional material in mathematics, science, and technology: (a) Schema acquisition is the primary component of skilled problem-solving performance; (b) learning through schema acquisition is interfered with if instructional material misdirects attention and imposes a heavy cognitive load; (c) for these reasons, conventional problem solving can sometimes interfere with learning; and (d) instructional material that requires learners to mentally integrate disparate sources of mutually referring information (e.g., text and diagrams) also interferes with learning by misdirecting attention and imposing a heavy cognitive load. Using mathematics and engineering materials, we found evidence for both the relevant cognitive concepts and the efficacy of alternative instructional materials generated by the concepts. Radical recasting of current instructional formats in most technical areas is called for.
Two experiments investigated alternatives to split-attention instructional designs. It was assumed that because a learner has a limited working memory capacity, any increase in cognitive resources required to process split-attention materials decreases resources available for learning. Using computer-based instructional material consisting of diagrams and text, Experiment 1 attempted to ameliorate split-attention effects by increasing effective working memory size by presenting the text in auditory form. Auditory presentation of text proved superior to visual-only presentation but not when the text was presented in both auditory and visual forms. In that case, the visual form was redundant and imposed a cognitive load that interfered with learning. Experiment 2 ameliorated split-attention effects by using colour coding to reduce cognitive load inducing search for diagrammatic referents in the text. Mental load rating scales provided evidence in both experiments that alternatives to split-attention instructional designs were effective due to reductions in cognitive load. Copyright © 1999 John Wiley & Sons, Ltd.
Three studies investigated whether and under what conditions the addition of on-screen text would facilitate the learning of a narrated scientific multimedia explanation. Students were presented with an explanation about the process of lightning formation in the auditory alone (nonredundant) or auditory and visual (redundant) modalities. In Experiment 1, the effects of preceding the nonredundant or redundant explanation with a corresponding animation were examined. In Experiment 2, the effects of presenting the nonredundant or redundant explanation with a simultaneous or a preceding animation were compared. In Experiment 3, environmental sounds were added to the nonredundant or redundant explanation. Learning was measured by retention, transfer, and matching tests. Students better comprehended the explanation when the words were presented auditorily and visually rather than auditorily only, provided there was no other concurrent visual material. The overall pattern of results can be explained by a dual-processing model of working memory, which has implications for the design of multimedia instruction.
The redundancy principle (or redundancy effect) suggests that redundant material interferes with rather than facilitates learning. Redundancy occurs when the same information is presented concurrently in multiple forms or is unnecessarily elaborated. According to cognitive load theory, coordinating redundant information with essential information increases working memory load, which may interfere with learning. Eliminating such redundant information removes the requirement to coordinate multiple sources of information. Accordingly, instructional designs that eliminate redundant material can be superior to those that include redundancy. This chapter summarizes research and theory concerned with the effect of processing redundant information in multimedia learning, a history of research in instructional redundancy, the conditions of applicability of this principle, and its instructional implications.
This chapter is divided into two parts. The first describes the effect of Pat Rabbitt's influence in encouraging the first author to use the increasingly sophisticated methods of ageing research to answer questions about the fundamental characteristics of working memory, together with reflections on why so little of this work reached publication. The second part presents a brief review of the literature on working memory and ageing, followed by an account of more recent work attempting to apply the traditional method of experimental dissociation to research on normal ageing and Alzheimer's disease. The discussion suggests that even such simple methods can throw light on both the processes of ageing and the understanding of working memory.
In Experiment 1, inexperienced trade apprentices were presented with one of four alternative instructional designs: a diagram with visual text, a diagram with auditory text, a diagram with both visual and auditory text, or the diagram only. An auditory presentation of text proved superior to a visual-only presentation but not when the text was presented in both auditory and visual forms. The diagram-only format was the least intelligible to inexperienced learners. When participants became more experienced in the domain after two specifically designed training sessions, the advantage of a visual diagram-auditory text format disappeared. In Experiment 2, the diagram-only group was compared with the audio-text group after an additional training session. The results were the reverse of those of Experiment 1: The diagram-only group outperformed the audio–text group. Suggestions are made for multimedia instruction that takes learner experience into consideration. (PsycINFO Database Record (c) 2012 APA, all rights reserved)