The role of spatial descriptions in learning from multimedia
Florian Schmidt-Weigand
a,
*
,1
, Katharina Scheiter
b,1
a
Institute of Psychology (FB 4), University of Kassel, Hollaendische Strasse 36, 34109 Kassel, Germany
b
University of Tuebingen, Konrad-Adenauer-Strasse 40, 72072 Tuebingen, Germany
article info
Article history:
Available online xxxx
Keywords:
Multimedia
Animation
Spatial information
Working memory
Redundancy
Conjoint retention
abstract
In the reported experiment we investigated how spatial information conveyed in an expository text influ-
enced multimedia learning. It was based on a 2 2-design with the degree of spatial information given in
the text (high vs. low spatial text) and the presentation format (written text-only vs. written text + ani-
mation) as between-subjects factors. As dependent variables learning outcomes as well as self-reported
cognitive load were assessed. The results revealed that there was a multimedia effect with regard to
learning outcomes only for low spatial text, but not for high spatial text. Moreover, the cognitive load
measures showed an overall multimedia effect irrespective of the degree of spatial information conveyed
by the text (i.e., higher cognitive load ratings in the text-only conditions). These results can be explained
as a special instance of the redundancy effect as well as a consequence of processing interference within
visuo-spatial working memory.
Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction
There is a long-standing tradition of educational research show-
ing that combinations of text and visualizations improve learning
outcomes compared to learning from text-only (Anglin, Vaez, &
Cunningham, 2004; Fletcher & Tobias, 2005). There are two lines
of argumentation for explaining this so-called multimedia effect.
The first line of argumentation focuses on the cognitive processes
that are facilitated when studying visualizations rather than text
(i.e., process-oriented; e.g., Larkin & Simon, 1987). The second view
is more outcome-oriented in that it is argued that learning from
visualizations will yield a richer internal representation of the to-
be-learned content compared to learning from text (e.g., Baggett,
1984; Paivio, 1991).
1.1. A process-oriented view on learning from visualizations
Larkin and Simon (1987) have suggested that visualizations are
often more computationally efficient than verbal descriptions for
accomplishing tasks that require the processing of visuo-spatial
properties by reducing the need to search for multiple information
elements related to a single idea as this information is grouped in
visualizations. Moreover, visualizations may support inference and
reasoning processes grounded in perception allowing perceptual
processes to replace more demanding logical inferences (Goldstone
& Son, 2005; Zhang, 1997). Thus, according to the process-oriented
view, visualizations may aid learning in cases where learning is
dependent on extracting and reasoning with visuo-spatial informa-
tion by enabling and facilitating cognitive processes that are less
demanding than the processes that would otherwise be required
when learning from text.
1.2. An outcome-oriented view on learning from visualizations
According to this view, visualizations support learning, because
compared to text their processing yields an additional and qualita-
tively different representation in memory, thereby facilitating re-
call of the conveyed information (e.g., Baggett, 1984; Paivio, 1991).
For instance, the bushiness hypothesis by Baggett (1984) sug-
gests that knowledge acquired from visual rather than verbal rep-
resentations will be better accessible in memory because the
respective nodes in memory share more associations with other
nodes in the semantic network, that is, they are ‘‘bushier”. Because
concepts with more associations can be retrieved more easily from
memory, visualizations should facilitate recall of the content.
Similarly, according to the Dual Coding Theory (Paivio, 1991),
verbal and non-verbal external information sources are assumed
to yield different internal representations. Verbal information re-
sults in a symbolic linguistic representation (logogens), whereas
non-verbal information is encoded analogically (imagens). Process-
ing verbal and pictorial information compared to text alone is as-
sumed to result in a dual coding of this information, which in
turn is better accessible in memory and hence easier to recall.
The same assumption has been made in the context of the con-
joint-retention hypothesis that specifically deals with spatial infor-
mation representations (Verdi & Kulhavy, 2002). This hypothesis
0747-5632/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.chb.2010.05.007
* Corresponding author. Tel.: +49 561 8043576; fax: +49 561 8043586.
E-mail address: f.schmidt-weigand@uni-kassel.de (F. Schmidt-Weigand).
1
These authors contributed equally to this work.
Computers in Human Behavior xxx (2010) xxx–xxx
Contents lists available at ScienceDirect
Computers in Human Behavior
journal homepage: www.elsevier.com/locate/comphumbeh
ARTICLE IN PRESS
Please cite this article in press as: Schmidt-Weigand, F., & Scheiter, K. The role of spatial descriptions in learning from multimedia. Computers in Human
Behavior (2010), doi:10.1016/j.chb.2010.05.007