ERIC ED393435: Using Visual Media Technologies To Investigate Cognitive Representation of Technical Diagrams.

10 Pages·1993·0.27 MB·English

by  ERIC

DOCUMENT RESUME
IR 017 771
ED 393 435
Lowe, Richard
AUTHOR
Using Visual Media Technologies To Investigate
TITLE
Cognitive Representation of Technical Diagrams.
PUB DATE
93
10p.; In: Verbo-Visual Literacy: Understanding and
NOTE
Applying New Educational Communication Media
Technologies. Selected Readings from the Symposium of
the International Visual Literacy Association
(Delphi, Greece, June 25-29, 1993); see IR 017
742.
Speeches/Conference
Descriptive (141)
PUB TYPE
Reports
Papers (150)
MF01/PC01 Plus Postage.
EDRS PRICE
Cognitive Structures; *Computers; *Concept Formation;
DESCRIPTORS
*Data Collection; Diagrams; Foreign Countries;
Information Technology; Meteorology; *Research
Methodology; *Video Equipment; *Visual Aids;
Weather
ABSTRACT
This paper describes how visual media technologies
have been used to support research into the way technical diagrams
are represented in people's minds. Two groups of subjects
participated in this research, individuals with either a high or a
low level of expertise in meteorology, specifically professional
meteorologists and non-meteorologists. Two general types of visual
media technologies--computers and video-rwere used to investigate the
mental representation of diagrams. Two main types of data were
collected by the computer and video methodologies: (1) process data,
the actions subjects perform as they work through n task; and (2)
product data, the outcomes (drawings) of copying, :recalling, or
completing a weather map. Examples of the way process and product
data were handled during analysis are presented. The strength of
video technology is that it provides for the capture of highly
detailed visual data. However, it also presents problems with the
time-consuming nature of transcribing the data into a form suitable
for statistical analysis. In contrast, computer technology can
directly monitor some aspects of a subject's performance and save the
generated data in files that are suitable for statistical analysis,
although it Is not currently capable of directly providing (for
reasonable cost and effort) the type of facilities for dealing with
visual data that are available with video. Recent merging of the two
technologies with the advent of digital video is opening up new
possibilities for research. (Contains 15 references.) (AEF)
***********************************************************************
Reproductions supplied by EDRS are the best that can be made
from the original document.
***********************************************************************
TO INVESTIGATE
USING VISUAL MEDIA TECHNOLOGIES
TECHNICAL DIAGRAMS
COGNITIVE REPRESENTATION OF
U.S DEPARTMENT OF EDUCATION
Office of Educafionai Researcn and encnfireinent
by
EDUCATIONAL RESOURCES INFORMATION
CENTER (ERIC)
Richard Lowe
1:1 This document has been reproduced as
received from the person Of organization
originating it
0 Minor changes have been made to
accessible to us for the purposes of study
improve reproduction quality
This paper describes a selection of
Points of view or opinions stated in this
(unlike the outcomes of a diagram-based
technologies have
document do not necessarily represent
ways that visual media
official OERI position or policy
its
and
instructional task) so
been used to support various aspects of
characterization presents a considerable
research into the way technical diagrams
research challenge. As a result, the
people's minds. Its
are represented in
methodologies involved in such diagram
purpose is to demonstrate how
such
research of necessity rely on indirect
technologies can address some of the
approaches that aim to provide fine-
difficulties inherent in conducting research
grained data from which inferences can be
in this area. The research programme for
made about mental representation. This
which the techniques described here were
requires that these approaches are not only
goal
broad
the
has
of
devised
robust and powerful, but also that they are
understanding how learners deal with
appropriate to use with pictorial material.
instructional diagrams.
A fundamental consideration in
While illustrations in general can be
designing research methodologies for this
included in instructional materials for a
area arises from the nature of the materials
range of different purposes (Levie &
to be investigated. Diagrams and similar
Lentz, 1982), a central role of diagrams of
explanatory visual materials typically are
the type used in technical domains is to
used becausc they are considered better
explain some aspect of the subject matter
suited to the presentation of the subject
(Levin, Anglin & Carney, 1987; Lowe,
matter than verbal formats (written or
1993). To date, much investigation of
spoken text). It now seems clear that
expository illustrations such as diagrams
aspects of the information structure of
has focussed upon their instructional
diagrams can provide them with cognitive
outcomes, rather than upon the processes
processing advantages over text (Glenberg
by which these outcomes are generated.
& Langston, 1992; Larkin & Simon,
rise of information
However, the
1987; Winn & Li, 1989). However, the
processing approaches to the study of
great majority of research into the way
thinking and learning in recent years has
people mentally represent information
highlighted the need to understand ments1
involves the use of verbal materials and
trying
to
processes
improve
in
verbally-oriented methodologies. There
instructional outcomes. With this has come
an increasing interest in the mental
has been considerable development in the
collection of detailed verbal data in recent
processes involved in the comprehension
years and approaches to analysis such as
of pictorial materials such as diagrams
those discussed by Ericsson and Simon
(e.g. Winn, 1991). Fundamental to these
(1984) are well established. However, the
mental processes is the way in which the
direct adoption of such methodologies for
information upon which they operate is
the investigation of the way people deal
rewesented in our minds. Consequently,
with visual materials would require visual-
before we can fully understand how
to-verbal translations, so compromising
diagrams are processed, we need to know
tal character. In other
their intrinsic
about the mental reprrzentation that people
N
words, collecting verbal data about the
develop for the graphic elements which are
those
fabric
of
the
processing of visual material (for example,
diagrams.
think-aloud protocols generated by a
person's
mental
Unfortunately, a
subject performing a diagram processing
representation of a diagram is not directly
"PERMISSION TO REPRODUCE THIS
MATERIAL HAS BEEN GRANTED BY
Alice D. Walker
BEST COPY AVAILABLE
2
TO THE EDUCATIONAL RESOURCES
INFORMATION CENTER &RIC)."
.118
Verbo- V isual Literacy
6. Deal with static and dynamic
task) would be of dubious value on its
components of the investigation.
Own (although it may be a valuable source
of supplementary data). What is required
As will be illustrated below, a
instead are approaches that properly
suitable combination of the capacities of
acknowledge the visual nature of these
video and computer technology can
materials and deal with them without the
need for any translation. So for an
address these types of needs in a variety of
ways. In addition, the capacity of these
investigation of the way people deal with
technologies for dealing with dynamic data
diagrams, it is not sufficient to confine the
visual aspect of the procedure to the
gives them a distinct advantage over more
conventional (static) approaches to the
stimulus materials. It is important that the
reporting of results from this type of
outputs that subjects produce should also
be visual as should any intervening
investigation.
manipulations of the materials they
Diagrams and Subjects
perform. The need- to use a consistently
visual approach across the various aspects
All the methodologies described here
of investigation guided the development of
were applied to the study of how weather
the methodologies described in this paper.
map diagrams are mentally represented.
This development was informed by the
However, it should be noted that weather
approaches used by researchers in the
maps were used purely as illustrative
investigation of cognitive aspects of
examples of the type of diagram
various visually-oriented domains such as
(Chase & Simon,
chess
commonly used for instructional purposes
1973),
(there was no interest in the investigation
architectural drawings (Akin, 1986), and
electronic circuit diagrams (Egan &
of weather maps per se). A discussion of
the advantages of using weather maps for
Schwartz, 1979).
this purpose appears elsewhere (Lowe,
1990). In principle, the methodologies are
Before dealing with specific aspects
applicable to a variety of diagram types
of the methodologies that have been
developed for the current research
from a wide range of technical disciplines.
The two groups of subjects who
program, some general considerations
this research were
participated
regarding this type of research will be
in
individuals with either a high or a low
discussed. For a methodology to be useful
level of expertise in the domain of
for finc-grained investigation of the way
meteorology, specifically professional
people interact with diagrams, it needs to
meteorologists and non-meteorologists.
be able to:
Because it involved comparison of two
I. Display
the diagrammatic
groups with greatly differing levels of
material effectively to the research
domain expertise, this research approach
can loosely be described as fitting within
subjects,
the "expert-novice" paradigm (Chi,
Glaser, & Farr, 1988). However, because
2. Provide appropriate ways for the
the meteorologists were competent
subject to interact with the diagram,
practitioners rather than acknowledged
expert performers in their domain, the
3. Capture detailed information
expert-novice characterization should be
about subject-diagram interactions,
treated with some caution.
4. Characterize these
subject-
Data Collection
diagram interactions with appropriate
measures,
In the following sections, the use of
general
5. Permit generated data to be
types visual
media
two
technologies (computers and video), for the
analyzed in a meaningful way,
investigation of the mental representation
3
229
Visual Media Technologies
through the required interactions by
of diagrams will be discussed. Although
approaches were specifically
these
appropriate instructions provided in a side
developed for use with very abstract
bar on the screen. To reveal what was
under a particular square, the subject
diagrams of a technical nature, there is no
moved the computer's mouse pointer onto
reason in principle why they could not be
adapted for various other types of
that square and clicked once, upon which
graphically-simple pictorial materials that
the square turned black to signal that it had
are used as instructional illustrations.
been selected. The subject was then able to
either confirm the selection with a second
Computer Technology
click on the blackened square, or select a
different square by clicking on it instead.
The second (confirmatory) click on a
The computer-based methodology
is an adaption and
described here
square caused the hidden meteorological
extension of an earlier manual technique
markings on that part of the weather map
that has been reported in detail elsewhere
to be revealed. For any square exposed in
In general terms, the
this way the background was changed
(Lowe, 1989).
from white to grey so that the subject
methodology was used to investigate a
could easily tell that it had already been
subject's mental representation for weather
map diagrams by monitoring the way the
revealed (this was necessary since not all
of square segments comprising the map
subject explored an unfamiliar example of
such a diagram which was initially hidden
actually contained weather map markings).
from view. This monitoring relied on the
As this revelation process continued, the
instruction side bar gave the subject
diagram having been segmented so that the
subject's exploration across individual
updated information about how many of
followed. A
regions
could
be
the quota of 10 squares remained available
superimposed 6 x 5 grid divided the
for selection. Once all 10 of the allowed
weather map into 30 square pieces (the
squares had been revealed, the subject's
diagram consisted of a map outline and its
map showing the exposed meteorological
meteorological markings). When a subject
markings was printed off. The subject then
began the experimental task, the display
attempted to complete the whole weather
map by drawing
any further
was completely covered with squares that
in
were blank except for the map outline
meteorological markings that were thought
(i.e.,
no
meteorological markings).
to exist in the remaining 20 grid squares.
During the task, the subject gradually
Because
interactions just
the
revealed parts of the display by removing
described are somewhat involved, each
the covering squares from the display one
subject was trained in the necessary
at a time. This allowed determination of
the order in which the selected regions of
operational processes before moving on to
the display were inspected. However a
the experimental task. The sequence of
presentation was as follows. After an
condition of the task was that the subject
introduction to basic processes such as
was permitted to expose a total of only one
moving and clicking the mouse, the
third of the whole display (10 of the grid
squares) and told that from this limited
subject was given preliminary training in
amount of information, she/he would later
the technique of revealing information by
clicking on squares. This was done using
be required to complete the display.
a simple four-square grid (Figure 1).
In the computer implementation of
this technique, the experimental task
Next, these skills were applied to a
involved the subject interacting with a
graphic display of a familiar scene (a
simple house and its surroundings)
screen depicting a blank weather map of
Australia and divided into 30 squares by a
following the same procedures that would
required
performing
for
grid as described above (the map and grid
be
the
were drawn in
black on a white
experimental task with the weather map.
background). The subject was prompted
The subject began with a bare outline of
4
Verbo-Visual Literacy
.230
for hidden details of the animal's digestive
thc house divided by a grid of 30 squares
system studied. Similarly, in the area of
(Figure 2) thcn by using the procedure
electronics, subjects could search for
described above for the experimental task,
information about the components present
clicked on 10 squares in sequence to reveal
in a hidden section of a circuit diagram.
one third of the house's details (windows,
doors, ctc.).
Video Technology
For this practice task, the subject's
response was not printed off for
The methodology described here
developed to obtain finer details about the
completion. After this practice, the subject
mental representation of weather map
carried out the experimental (weather map)
task with weather map. During the
diagrams than could be obtained from the
technique described above. In addition, its
experimental task, the computer program
basic approach to data collection was
recorded the identity of the squares chosen
and the sequence in which they were
fundamentally different, thus providing the
opportunity for triangulation. A video
selected. Finally, a screen dump of the
recorder was used to capture the sequence
subject's partly completed map (showing
of actions performed by subjects in
the 10 revealed squares) was printed off
carrying out tasks involving a previously
for thc subject to complete (Figure 3).
unseen Australian weather map. A major
part of the process involved subjects
This computer version of the original
manual method was developed in
of
map's
drawing
copy
the
a
meteorological markings onto a blank
Hypercard so that it was able to run on
map. However, the weather map they
even low-end black and white Macintosh
were to copy was not continuously visible.
computers, thus making it suitable for use
Instead, subjects could only view the
in a wider range of situations. Although
stimulus map while they were pressing a
some aspects of the methodology (such as
button with their drawing hand. This
the training segment) were altered to make
the technique less dependent on the
meant that they could either take a look at
course of
the
experimenter,
the stimulus map or draw markings onto
the
the blank map but not continuously view
experimental task remained fundamentally
and draw at the same time. Hence,
unaltered. The computer-based version has
a number of benefits over the original
subjects were forced to alternate viewing
and drawing in a succession of discrete
procedure such as the automatic collection
of data regarding which squares were
actions. An indicator light was switched
chosen by a subject and their order of
on when the viewing button was pressed
selection. It also has the potential to be
and remained illuminated for the duration
of each viewing period (i.e., as long as the
used for displays other than weather maps
viewing button was depressed).
as will now be explained. In the weather
map version described above, the map
The subject
out
carried
the
outline constitutes one level of information
at a specially
tasks
experimental
(provided for the subject) and the set of
that depict meteorological
constructed table that had a glass window
markings
inserted into its top with a mirror set
information constitutes another (to be
beneath the window at an angle of 45
provided by the subject). However, since
degrees. A video recorder aimed at the
the computer program is essentially a shell
mirror allowed each subject's drawing
into which a particular display is fitted, the
behavior to be recorded from below
method could easily be adapted for
without obstruction (the response sheet
investigating the way subjects search other
containing the blank map was made from
diagrams that can be characterized in terms
tracing paper so that the subject's drawing
of two levels of information. For example,
was visible from below). The viewing
in thc arca of biology, subjects could be
indicator light was positioned so that it
presented with an outline of an unfamiliar
could be included in the area being
animal's body and the way they searched
231
Visual Media Technologies
recorded by the video and the camera was
fitted with a counter that marked each
frame with elapsed time. Thus in addition
to recording the subject's drawing of each
meteorological marking, the length of each
viewing period was automatically recorded
C
on videotape. As well as drawing a copy
of the weather map, subjects later drew
what they could recall of the map's
Notice how its gone black
markings. Video records of this process
to slow you chcked on it.
were also made to enable comparison of
Now click it a second time.
the intake of diagram information (during
copying) with its output (during recall).
Figure 1.
Grid used for training subjects in procedure for
Data Analysis
selecting squares (black indicates square 'A' has
been clicked once; a further click would reveal
Two main types of data are collected
what was under this square).
by the techniques that have been
MI
described. The first are process data which
come from the actions subjects perform
10 to
during an investigation as they work
go.
)1MI
through a task. Of particular interest here
are the temporal order in which the
Click
oat
meteorological markings are processed,
slum
the location of those markings, and the
to
select
durations of various stages that comprise
it.
the processing. The second type are
product data such as the outcomes
(drawings) of copying, recalling, or
completing a weather map.
Figure 2.
Analyzing each type of data presents
Practice task graphic showing outline within 30
its own challenges, some of which
square grid. Note instructions to subject in side
concern practicalities such as how to carry
column.
out particular measurements while others
concern more fundamental theoretical
laidialliffi
issues involving the choice, treatment, and
"to
interpretation of those measurements.
go.
MEE rd
Am
Some examples of the way process and
product data were handled during analysis
Click
will now be presented.
ort
EP.A.....
ra.
Fe
Process Data
seyet
wiltir
't.
:41475,7
A number of assumptions were made
it was assumed that the
here. First,
temporal order in which the meteorological
markings were processed was in some
Figure 3.
respects influenced by the nature of the
mental representation of the weather map
Screen dump showing markings on a subject's 10
information. This assumption parallels that
selected squares. Subject would next attempt to
made by Taylor and Tversky (1992) in
complete the map' markings.
of the
their
investigation
mental
organizations of environments. As these
BEST COPY AVAILABLE
6
Media Technologies
231
subject's processing sequence to be used
authors have observed, one of the research
to identify subsets within the complete set
difficulties in using drawings as data is
of graphic elements comprising the
how to score them. Taylor and Tversky
diagram. By examining the constitution of
used the order in which elements of a map
these subsets, inferences could be made
arc drawn as an index of organization on
about inter-element relations that might be
the assumption that the clustering of items
the subject's mental
involved
in
in a recall sequence occurs because those
representation of the diagram as a whole.
items are closely related in some way. In
For example, it appeared that the non-
weather
video-based
the
map
meteorologists structured the mental
investigations referred to in the present
representation of information in the
paper, this type of clustering assumption
diagram on the basis of relations involving
was made about the sequences produced
the visuospatial characteristics of the
during the copying of items as well as
markings themselves whereas
the
during their recall. In the computer-based
meteorologists' structuring was based
earlier,
investigation
mentioned
upon relations
reflected
that
the
assumptions about clustering were made
meteorological significance of these
from the order in which the 10 squares
markings.
were selected for revelation. Depending on
the particular task a subject was
performing, the structure of these
A major disadvantage of the process
aspect of the video-based investigation
sequences may provide an indication of
was the time consuming and labour-
matters such as the effects of established
intensive nature of the data transcription
habits of perceiving and drawing, the
required. The raw data consisted of a
extent of visual impact (in perceptual
videotape record of each subject's viewing
terms) of different types of markings, the
and drawing behavior in which each
tendency of subjects to group the markings
videotape frame was marked with the time
because of similarities in their visuospatial
elapsed since the start of the task. In
characteristics, the ease with which
markings of different types were
addition, the videotape of the copying task
also contained records of the periods
processed or the relative meteorological
during which the viewing indicator light
importance attached to different markings.
was switched on and switched off.
Transcription of the videotape involved
Thc second assumption was that the
locating precise frames at which (a) the
relative durations of particular aspects of
viewing indicator light was switched on or
processing (such as the lengths of viewing
was switched off and (b) the subject's
periods) were of significance. In the
pencil just began to draw a mark or just
current context, this assumption concerns
finished drawing that mark. The elapsed
video-based
the
only
investigation
time for each of these frames was
On most occasions
described above.
transferred manually to build up a table
during the copying task, subjects viewed
that set out each subject's viewing and
the stimulus diagram using a series of
relatively short glimpses. After each
drawing behavior. In addition, each of the
marks drawn on the map was labelled with
glimpse, a subject typically drew one or
position in
the overall drawing
its
more marks. However, at various stages
sequence. In contrast to the video-based
during the process of copying the stimulus
investigation,
the
computer-based
diagram, the subject would spend a much
earlier
investigation
described
longer time gazing at the diagram before
changing to drawing. These longer
automatically collected process daia about
viewing periods were interpreted as
the identity of the 10 squares that a subject
indicating thc start of a new phase of
revea!,:d and the order in which they were
selected. These data required no pre-
processing in which the subjcct's attention
treatment after collection and were ready
shifted to a different group (or chunk) of
for immediate transfer to a statistical
graphic elements within the stimulus
However,
analysis
program.
the
diagram. This interpretation permitted each
233
Visual Media Technologies
entities because pilot studies showed that
computer-based approach did not offer the
these different components were drawn by
same opportunities for collecting the subtle
subjects as a number of separate
types of data that can be captured by the
subgroups (in the case of a cold front, a
use of video. Clearly an integration of the
capacity of video technology to collect
line plus a group of triangular barbs).
detailed visual data and the automatic data
Once individual graphic entities have
collection capacity of computer technology
been defined, the next problem is in
would be the ideal combination.
making measurements of the position,
Product Data
size, shave, and orientation of each entity.
Recent developments in image analysis
Map markings drawn by subjects
computer software have greatly simplified
this measurement task. These programs
and the larger graphic structures formed by
the conjunction of drawn and existing
process an image that has been converted
to digital form and then use various
markings can be characterized by various
measures. These measures allow a range
algorithms to characterize properties of the
image. In the case of the weather map
of comparisons to be made including those
between (a) the markings on the stimulus
diagrams used for the research referred to
materials and those produced in the
here, the image was captured by means of
subjects' responses, (b) different classes
a flatbed scanner to form a disc file (in a
of subject (such as experts and novices)
format such as PICT). Some preliminary
treatment using image enhancement
and (c) products generated at different
stages of an investigation (such as copied
software was usually necessary to clean up
and recalled markings). Measures of
the scanned image and to prepare it for
position, size, shape, and orientation are
analysis. For example, regions bounded
useful to characterize the properties of
by particular sets of lines might be filled
with "paint" to isolate them so that the
individual markings both in terms of their
basis visuospatial properties and in terms
image analysis program can later identify
them as distinct features to be measured
of the meteorological significance of those
markings. For example, the sizes and
(each such region is typically described as
a "blob"). Once the prepared image is
relative positions of markings in of a set of
concentric closed isobars that together
loaded into the analysis program, the user
constitute a cyclone not only show the
can then specify the desired measurements
and select the blobs of interest. Simple
physical extent of that feature, they also
indicate its likely severity. As well as
measures of distance (such as length or
making measurements of each marking on
perimeter) and area are made by counting
its own, it is important to characterize
pixels which are then converted into more
larger composite structures made up of
familiar units. The position of a blob is
two or more markings which constitute
based upon its center of gr:.,vity while a
variety of measures such as symmetry,
meteorologically significant assemblies.
roundness, and the ratio of the axes of the
A fundamental matter to be resolved
best fitting ellipse help to characterize its
shape. For orientation, a measure of the
in developing an analysis procedure is the
definition of what constitutes a discrete
inclination of the longest axis of the best
entity for the purpose of analysis. In cases
fitting ellipse can be used. Data from these
where a marking consisted of a single
measurements are readily transferred to
graphic stroke or a set of physically
statistics packages for further processing.
connected strokes, this generally did not
Reporting Results
constitute a problem since it was usually
quite clear that a single entity was
intended. However, for other markings
A combination of the capacities of
such as a cold front symbol, physically
computer and video technologies has
connected but geometrically distinct
enabled the findings obtained from the
markings were considered as separate
types of investigations described above to
8
Media Technologies
234
The strength of video technology is that it
be displayed in a manner that is more
provides for the capture of highly detailed
appropriate than conventional methods.
visual data. However, it also presents
This is particularly well demonstrated in
problems with the time-consuming nature
the case of results obtained from analysis
of some aspects of transcribing the
of the process data where the nature of the
generated data into a form suitable for
subjects' performance over time was of
statistical analysis. In contrast, computer
special interest. For example, a computer
technology can directly monitor some
animation program was used to produce
aspects of a subject's performance and
essential
of
dynamic summaries
save the generated data in files that are
in how those
the
in
differences
suitable for analysis by a statistical
meteorologist and non-meteorologist
package without further processing.
subject groups built up the meteorological
Unfortunately, it is not currently capable
markings as they copied a weather map.
of directly providing (for reasonable cost
These summaries collapsed the individual
and effort) the type of facilities for dealing
marking sequences produced by subjects
with visual data that are available with
within these two groups into an overall
video. However, the recent merging of
sequence for each group as a whole. In
video and computer technologies with the
addition, the animation greatly compressed
advent of digital video opens up exciting
the time scale so that the broad temporal
possibilities for developing a whole range
patterns present in the sequence became
of new research tools that can be used to
much more apparent. Once prepared, the
increase our understanding of the way
animated sequence for each of the subject
people deal with visual information.
for
groups was transferred to videotape
This
casy presentation and distribution.
References
animated form of the results captures
subjects'
process aspects
the
of
Akin, 0. (1986).
Psychology of
performance far more faithfully than is
architecturaldesign. London: Pion.
possible in a static figure of the type
typically used to report the findings of an
Chase, W.G., & Simon, H. A. (1973).
investigation. Further, complex results of
chess.
Perception in
Cognitive
the type generated in this investigation
Psychology, 4, 55-81.
appear to be considerably more accessible
if presented in an animated format than if
Chi, M. T. H., Glaser, R., & Farr, M. J.
presented in a more conventional manner.
The nature of expertise.
(1988).
Hillsdale, NJ: Erlbaum.
Conclusion
Egan, D. E., & Schwartz, B. J. (1979).
This paper has dealt with several
recall of symbolic
Chunking in
illustrative examples of how video and
drawings. Memory & Cognition,
7,
computer technology can be used to
support basic research into the way
149-158.
diagrams are represented in people's
Ericsson, K. A., & Simon, H. A. (1984).
minds The approaches that have been
Cambridge, MA:
developed can address a variety of
Protocol analysis.
MIT Press.
challenges that face the researcher in this
field, ranging from those of presenting the
Glenberg, A. M., & Langston, W. E.
initial stimulus material to subjects through
(1992). Comprehension of illustrated
to reporting the research findings in an
Pictures help to build mental
text
appropriate manner. In common with other
Journal of Memory and
models.
research methodologies, in order to make
use of thcse approaches effective,
Language, 31, 129-151.
sufficient attention must be given to
Larkin, J. H., & Simon, H. A. (1987).
theoretically significant aspects of data
Why a diagram is (sometimes) worth
collection, analysis, and interpretation.
9
235
Visual Media Technologies
ten thousand words.
memory: Explaining the role of skill
Cognitive
and experience. Research in Science
Science, 11, 65-99.
Education, 20, 191-199.
Levie, W. H., & Lentz,
(1982).
R.
(1993).
Lowe,
R.
Effects of text illustrations: A review
K.
Successful
research.
o f
instructional diagrams.
London:
Edwational
Communication and Technology
Kogan Page.
Journal, 30, 195-232.
Taylor, H. A., & Tversky, B.
(1992).
Descriptions and depictions of
Levin, J. R., Anglin, G. J., & Carney, R.
N. (1987). On empirically validating
environments. Memory & Cognition,
functions of pictures in prose. In D.
20(5), 483-496.
M. Willows, & H.A. Houghton
Winn, W. (1991). Learning from maps
(Eds.), The psychology of illustration:
Vol. I. Basic research.
New York:
and diagrams. Educational Psychology
Review, 3, 211-247.
Springer-Verlag.
Lowe, R. K.
Winn, W. D., & Li, T-Z. (1989, April).
Search strategies
(1989).
and inference in the exploration of
Do diagrams permit more rapid and
scientific diagrams.
accurate problem solving than text?
Educational
Psychology, 9(1), 27-44.
Paper presented at the annual meeting
of the American Educational Research
(1990).
Lowe,
Diagram
K.
R.
Association, San Francisco, CA.
information and its organization in
10