An Empirical study of Attending and ComprehendingMultimedia Presentations

Pete Faraday & Alistair Sutcl@e

The Centre for HCI Design

City University,

Northampton Square

London, EC1 OHB, UK+44-171-477-8994

P.M, Faraday@city.ac.uk

ABSTRACT

The paper reports two studies concerning attention to and

comprehension of Multimedia presentations. The MM

sequence used was taken from a commercially produced

CD-ROM, ‘The Etiology of Cancer’. First, an eye tracking

study of the presentation is reported. A second study was

then cortductekl ml the memorisation of the materials used in

the eye tracking study. The results of the studies are used to

propose guidelines to improve design of MM presentations.

KEYWORDS

Eye Tracking, Comprehension, Presentation Design.

INTRODUCTION

One of the problems in the design of Multimedia (MM)

interfaces is knowing whether the presentation will

successfully deliver its content to its audience. For a user,

understanding a multimedia presentation requires a series of

cognitive processes composed of visual and auditory

attention, comprehension into a proposition, and integration

into a mental model. This paper outlines two studies we

performed on these cognitive processes, and attempts to use

their results to provide design advice for MM presentations.

The studies made use of a commercially produced MM

presentation, ‘The Etiology of Cancer’ ( figure 1).

Our first study examined visual attention to thepresentation. Unlike auditory attention, in which speech is

processed in a linear manner [19]; visual attention

processing is more complex. Visual attention is required to

extract still and moving visual objects [7, 14], and text label

or caption elements from the presentation [12]. It can

process only a limited amount of visual information at any

one time, since it can only take visual input from a

Permission to rneke dlgitrd/hard copies of all or patt of this material forpersonal or classroom use is granted without fee provided that the copiesare not made or distributed for profit or commercial advantsge, the. copy-right notice, the title of the publication and its date appear, and notice isgiven that cop yright is by permission of the ACM, lnc. To copy otherwise,

to republish, to post on servers or to redlatribute to tists, requires specific

permission and/or fee.ACM Multimedia 96, Boston MA USA01996 ACM ()-89791_S71_l/9tj/11 ..$3.50

relatively small area, around 2 degrees of angle from the

eye [12, 14]. Processing is serial, fixation has to be shifted

each time to process the next piece of visual information

[9]. Whilst processing, attention is locked and the eye is

held still in a fixation, toward the point being attended to

[4,12, 21].

In order to investigate visual attention we examined the

order in which visual information is processed by the

viewer using an eye tracking system. We were interested in

the attentional processing of different visual presentation

elements, such as text captions, labels and still and moving

objects; and in how the serial nature of visual attention

would cope with processing a complex and rapidly

changing MM presentation. Few studies have made use of

eye tracking equipment to look at animated materials [4,21],

and none have attempted to track subjects viewing a

complex multimedia presentation. However, visual attention

is vital if visual information is to be successfully processed

from the presentation.

Our second study examined comprehension of the

presentation. The comprehension process acts upon the

elements processed by attention, and combines them

together with the user’s existing long term memory (LTM)

knowledge [1]. The outcome of this is a proposition, or unit

of meaning, built from the presentation elements [10, 22].

For instructional presentations the proposition will be one

of the following types, based on [3] : (i) object proposition,

defining object descriptors or identifiers; (ii) action

proposition, defining the role(s) of objects in the action,

together with path or state change produced by the action;

(iii) procedure proposition : giving sequence or cause-effect

relationships between a group of actions

These proposition types may be built fkom the speech, still

or moving visual objects, label or caption text elements

attended to. The individual propositions are then combinedtogether to form a mental model [10], representing the

meaning across the presentation.

processed, the mental model will

objects, actions and procedures.

As the presentation is

contain a hierarchy of

265

Figure la Figure lb

Despite significant improvements in diagnosis and ... the major cause of death in most cases of cancer is

treatment ... metastasis ...

Figure lC Figure Id

.. the spread of tumour cells fkom the primary neoplasm ... where secondary growth occurs,to distant organs .,.

Figure 1e Figure 1f

266

In our second study, we investigated the comprehension

process by testing free recall, in which subjects wrote down

what they could remember of the presentation. We were

interested in the information types recalled from the

presentation and in relating then to the type of presentation

elements shown. We also investigated the effects of the

users existing LTM knowledge upon comprehension by

using two subject groups, one with and one without

background knowledge of the subject area,

STUDY ONE

The first study investigated the pattern of fixations and

attention upon the MM presentation,. using a Pupilometer

eye tracking system. The objectives of the study were to

attempt to match the shifting of fixations to the visual

presentation elements shown, and to investigate any

common patterns.

Study Materials

A 27 second animation was taken tlom the ‘Etiology of

Cancer’ showing the spread of cancer, termed metastasis. It

is illustrated as a series of screen shots in figure 1. The

speech track which accompanied the presentation is given

under the relevant screen shots.

In figure la,b the malignant tumour object in the centre of

the screen is shown moving down toward the bottom edge

of the screen. The scene ends as shown in figure 1b. A new

scene begins in figure 1c with the tumour breaking up into a

series of tumour cells which move in a diagonal from the

tumour across to the blood vessel. Figure 1d shows the next

scene in which turnour cells move from the top right of the

screen in a diagonal following the blood vessel to the

bottom left of the screen. The final scene is shown in figure

1e,f in which the tumour cells proliferate from the small

cluster shown in figure 1e to the secondary tumour shown in

figure 1f.

Figure 2: Single subject scan path on figure 1c

Study method & Procedure

Eight subjects were drawn flom research students within

City University. A pupilometer system was fixed upon the

subjects eye, whilst they viewed the MM presentation. Due

to equipment problems, two subjects had to be discarded.

The tracker produced an x,y location of the pupil every 50

msec and filtered out spurious data from subject blinks.

Subjects used a bite clamp avoid head movement and were

held approximately 50cm from a 17 inch VGA CO1OW

monitor.

Data AnalysisThe raw data of pupil location was time sliced into five

second parts for each subject to ease analysis. A PC based

tool was built to plot the data for each subject as a scan

path, The data for the six subjects was synchronised over

time to build up a composite trace for each part. This was

placed upon a 32 by 24 grid so that it could be mapped to

the 640x480 VGA screen resolution and compared with the

MM presentation. An example of the data analysis is shown

in figure 2.

ResultsThe fixation maps were compared with the presentation

sequence to produce a pattern for fixation clusters against

the presentation across the six subjects. Whilst it is known

that pupil location and attention may not always coincide, in

general a fixation, or pause, of the pupil at a location will

indicate attentional processing [21]. The parts of the

presentation fixated were analysed to produce an aggregate

of the shitls in fixation between each. The parts were then

ordered in directed graphs, shown on the screen shots and

graphs as a circled number (figure 3,4,5,6). The arrows in

the directed graphs show the direction of fixation, whilst the

number at the end of each arrow shows how many subjects

shifted from one part to another; if subjects shifted back and

forth between parts, the number will be larger. The entry

and exit points are shown on the graphs as bold numbers.

Figure 3a : Fixations on figure 1a

267

(1) All subjects initially produced orientation fixations,

beginning by transversing over the large central tumour

object presentation (figure 3a)

(2) Five of the subjects then shifted to fixate upon the

region of the ‘Metastasis’ title caption (figure 3a)

(5) Revealed caption ‘Cells break through basal Lamina’

concurrent with animation. Subjects tended to fixate upon

this and the motion of tumour cells (4). They also refixated

between the other labels (3a), (3b) (figure 3b)

(3a, 3b, 3c) Shift to the labels ‘malignant tumour in

bronchial epitheliums’, ‘connective tissues’. No particular

order was found across the subjects. The ‘basal lamina’ label

(3c) was not fixated upon. This may be because it was

removed shortly after the start of the presentation and

replaced with label (5) during the animation (4) (figure 3a)

Figure 4a : Fixations on figure lC

,--3

Figure 3b : Fixations on figure lb

J(.=--~@T”mO”’”biect 2 q

(

(

(2~Metast~s;label ~--.,_=@ tvldignantt.fnourl abel

, (2 ‘!)l-l;$

* ia Connective tissue label A— 5 Cells breakthrough label

,> ‘$23: y \2

4 Tumour path~>,

Figure 3C : Directed graph for figure 3a, 3b

(4) Animation of the malignant tumour cell spreading

downwards. The motion caused a general shift of fixation to

the centre of the presentation; but the shift did not happen at

the start of the motion for most subjects. This could be due

to attention still being locked on the labels (3a), (3b) at thestart of the motion. Once attention shifted, subjects did nottrack the motion, but tended to fixate upon its general area

and edges. This may be because the animation was internal,changing the shape, but not the position of the object as a

whole so did not require tracking. (figure 3b)

Figure 4b : Directed graph for figure 4a

(6) The scene was changed. Subjects produced orientation

fixations and then generally shifted to the title ‘Metastasis’

or directly to the tumour path. The tumour path showed the

tumour object partly broken up as a still, before animating it

in (8). (figure 4a)

(7) Subjects shifted to caption ‘Cells break away’ (figure 4a)

(8) Animation of the tumour cells moving in a diagonal

f?om top-middle to bottom-right. All subjects showed an

alignment upon the tumour cells, with subsequent fixations

tracking their path of motion. (figure 4a)

(9) Shift to revealed caption ‘Cells invade capillary’ at the

end of the path of(8). Several subjects then shifted fixations

back and forth between the tumour cell objects and labels

‘Cells break away’ and ‘Cells invade capillary’. (figure 4a)

268

Figure 5a : Fixations on figure 1c

Figure 5b : Directed graph for figure 5a

(lOa, 10b) The scene was changed. Subjects produced

orientation fixations and then shifted to the title ‘Metastasis’(lOa) or to the tumour object itself (lOb) (figure 5a)

(1 la) Animation of the tumour cells moving in a diagonal

from top-right to bottom-left. Only two subjects fixated

upon the start of the motion. The other subjects fixations

tended to be either still on (1 O), or on (1 lb), before picking

up the path about l/4 way through (figure 5a)

(1 lb) Static label ‘Cells travel through blood stream’.

Subjects in general shifted from (1 O) to (1 lb), and then to

the animation (11 a). (figure 5a)

(12a, 12b) Shift to secondary tumour object at the end of

the path (11 a) and then to revealed label ‘Cells adhere to

capillary wall’; subjects shifted fixations back and forth

between the tumour cells and label. (figure 5a)

Figure 6a : Fixations on figure 1e

(13a, 13b) The scene was changed. Subjects produced

orientation fixations and then generally shifted to the

secondary tumour object (13a) or cells escape label (13b).

Subjects shifted fixations back and forth between the object

and label. (figure 6a)

(14) Several subjects switched from the secondary object

or cells escape label to the metastasis label. (figure 6a)

Figure 6b : Fixations on figure lf

Figure 6C : Directed graph for figure 6a, 6b

(15a) Animation of secondary tumour cell growth; subjectstended to fixate upon its general area and edges. The

269

animation was internal, changing the object shape, and

again did not produce tracking fixations (figure 6b)

(15b) Revealed label ‘Cells proliferate to form liver

metastasis’; subjects shifted fixations back and forth

between the secondary tumour growth object and label

(figure 6b)

Discussion

The study provided some evidence for the effect of design

techniques upon fixation and attention within presentations.

Several tentative conclusions are proposed :

● Animation has a strong effect on attention. This follows

the previous results of [7,12,21]. The onset of an

animation will generally produce a shift of attention to

the object in motion. Evidence for this can be found in

animations (8) and (15a).

However, in animations (4) and (11a) the start of the

animations were generally not shifted to immediately. In

animation (4) subjects did not seem to shift to the

downward spread of the tumour cells until late in their

motion; whilst in (1 la) the path of the tumour celk was

not tracked by any subjects until the animation was

around 1/4 complete. Instead attention may have been

locked on the title ‘Metastasis’ (2), (10) and labels (3a, 3b,

3c) and (11 b) in the two frames. If the start of the motion

is important, then ensure that attention will be free, not

processing labels or captions.

. Attention will generally be locked when processing

motion, making it unavailable to shift to other elements

which change or are revealed. This is inkeeping with [7].

In (4) the animation of the tumour cells moving

downwards may have locked attention before it could

process the label ‘basal lamina’ (3c). Using animation may

be unwarranted if the object is not important; trackingmay prevent other fixations.

● Animation which shows an object moving as a whole

along a path produces fixations which will track the

motion of the object and will be directed toward the end

point of the path [8]. Tracking fixations are clearly shownin animation of turnour cells in (8) and (11a). Subsequentelements to be processed should be placed at the end of

path of the motion. Animation which shows an objectchanging state, in which the object remains as a whole but

it shape alters, produces dispersed fixations over the area

and edges of the object. In animation (4) and (15a)

fixations are spread over the tumour and its edges whilstit changed shape.

● Changing scenes causes attention to re-orientate.

Reorientation fixations were found after scene changes in

(2), (6), (10) and (14). Each time attention seemed to

process the scene from fi-esh, eg refixating on the

previously attended title ‘Metastasis’ afier each scene

change. Similar results are given by [11, 16], who suggest

that using a cut in film tends to enhance attention to the

material following the cut. Avoid revealing labels or

object motion to allow re-orientation to take place after

changing to a new scene.

. By gradually revealing labels, symbols and objects,

fixations can be shifted eg between the tumour and labels

in (5), (9). Reveal may be used to order attention fi-om

one part of the topic to the next. Revealing several

presentation elements simultaneously may obscure any

ordering effects; reveal important elements singularly.

● Static presentation elements are not attended to

uniformly. Those which are Iabelled, which are bright in

colour, or appear to be different from their surroundings

will be attended to in preference to other objects. Use

strong colours or labels to draw attention. Avoid using

multiple labels, such as in (3a, 3b, 3c) which may lead to

unpredictable shifting of attention. .

. Reading captions and labels will lock attention [12],preventing other elements being processed, Reading also

requires longer fixation time than for identifying objects.

Attention may have been locked on the title ‘Metastasis’

(10) and caption (1 lb), preventing processing of the start

of animation (11a). Allow time for text elements to be

processed : this may range from a 1/5 second for a simple

word to several seconds if the word is unfamiliar [12].

● Labelling an object will produce shifts of fixations

between the object and label eg (13a,b) and (15a,b), The

findings of [6] are similar suggesting that switches of

attention are used to integrate information from the label

with the object. Ensuring that an object and its label areboth available together may aid this process.

STUDY TWO

The results of study one gave some insight into how visual

attention would process an MM presentation. However, it

did not address what was extracted fi-om the presentation by

comprehension. Study two investigated recall of theinformation content of the original presentation. The issues

of interest here were what types of proposition would berecalled from the presentation; and how the different

presentation elements, such as speech, still and moving

objects, label and captions would influence what wasrecalled. The effect of the subject’s existing domain

knowledge upon comprehension was also tested by

comparing a group of low knowledge subjects with a

second group of subjects who had expert knowledge of

biochemistry.

270

Problem

i) Metastasis is problem despitesignificant advances in diagnosis

Cause

Ii) Metastasis is thespread of cancer

E&d

xiii) Major causeof Death

Seauence

iii) Tumour cells break awayfrom primary tumour

i ‘----- ._Obiect state Acfion

iv) Cells m v) CellsBronchial breakthroughEpithehum Basal Lamina

._.r–--—-—

vii) Tumour cells travelthrough blood stream

I~..

—-~

vi) Cells break viii) Cells Invade ix) Cells Adherethrough Connectwe Capillary to Capillarytissues

Figure 7: Graph of propositions

Study Method & Procedure

Study two used two groups of eight subjects. The first

group was drawn from research students within City

University, who had low background knowledge of the

presentation domain. The second group of subjects was

recruited from biochemistry post-graduates who had high

background knowledge, but were pretested to ensure that

they had no specific knowledge of the subject matter itself.

This provided low ‘v’ high background knowledge

conditions. Both groups of subjects were shown the

presentation once and were then instructed to write down

what they could recall of the presentation.

Materials

The same materials were used as for study one (figure 1).

Data Analysis

In order to score the accuracy of subject recall the MM

presentation was analysed into a set of thirteen propositions,

representing the main ideas was created, The propositions

were further grouped into either high level organizing

concepts of sequence, problem or cause-effect, or into

lower level actions or object states within each part of the

sequence. This is shown as a graph in figure 7, with higher

level concepts at the top and lower level actions at the base.

The recall transcripts were analysed and categorised into

these propositions.

Results

Recall was scored for each of the thirteen propositions for

the two groups of subjects. These are shown in the table 1.

DiscussionThe background knowledge of the subjects appeared to

make little difference to their recall performance. Thissuggests that performance was based on what was given in

I,Seauence

x) Secondary tumourgrowth is formed

/vdQL1 mxi) Cells escape xii) Ceilsfrom Capillary invade

another organ

the presentation, rather than what was already known by

subjects from their existing long term memory.

Recall of higher level sequence propositions iii), vii), xii),

representing a group of actions, was good for both expert

and novice groups. Most of these sequence propositions

were given as animations. Animations are known to

improve comprehension of general action information

[15, 17]. The use of extra animations may have improved

recall performance for eg the tumour breaking through

the basal lamina, tumour adhere and extravisation actions.

Recall of specific action, object state and object role

propositions v), vi), ix), x) was poor. Most of these

propositions were given in text caption or label elements.

Comprehension difficulties may be in part due to

attentional problems with the use of multiple labels and

animations, discussed earlier. Others problems may be

related to poor synchronisation of the text labels with the

speech track, see below.

Recall of propositions which were only given in the

speech track was generally poor. The high level problemproposition i) and effect proposition xiii) were both

poorly recalled. However, the speech ‘metastasis spread

of cancer’ proposition ii) was well recalled. This may

have been because it was reinforced with the title

‘Metastasis’ in each fi-ame of the presentation. Generally,

if the speech information is complex or important then it

may be best to also concurrently present it in a caption. It

may also be useful to use captions or labels which take

the exact wording given in the speech track, to ensuretight linking. It is concluded by [5] that ‘only exact

duplicates on the screen of what is being said in thenarration appear to allow it to be processed unimpeded’.

271

●

●

Proposition type

Problem

Cause

Sequence

Object state

Action

Action

Sequence

Action

Action

Action

Action

Sequence

Effect

Proposition Low Domain High Domain Total (/1 6)

knowledge (/8) knowledge (/8)

i) Metastasis is a problem despite 1 0 1

significant improvements in diagnosis

and treatment

ii) Metastasis is the spread of cancer. 6 7 13

iii) Tumour cells break away from 6 4 10

primary tumour.

iv) Malignant tumour cells are in 1 1 2

bronchial epitheliums

v) Tumour cells breakthrough basal o 2 2

Iamina

vi) Tumour cells break away through o 1 1

the basal lamina and connective tissue

vii) Tumour cells travel through the 7 5 12blood stream.

viii) Tumour cells invade the blood 4 5 9

vessel capillary

ix) Tumour cells adhere to the o 2 2

capillary wall in liver

x) Tumour cells escape from capillary 2 4 6

xi) Tumour cells invade another 6 5 11organ.

xii) A secondary tumour is formed. 8 5 13

xiii) Metastasis is major cause of 3 1 4death

Table 1: Recall results for low ‘v’ high domain knowledge

Linking of the speech track to visual elements of the

presentation was generally poor. The speech track rarelyprimed or re-inforced the propositions presented visually.

Where speech did prime the animation toward the end of

the presentation in propositions xi) and xii) describing the

secondary tumour, it improved comprehension. Mostly

however, the speech propositions i), ii), xiii) did not

directly relate to the visual elements being shown. It is

noted by [2, 5] that it is important that speech and text

captions, labels, animation or objects combine to form asingle composite proposition. Comprehension may fail ifdifferent propositions are presented by the two media. A

study by [5] found that if speech and visual informationwere unrelated, subjects would generally ignore one ofthe media, usually the speech track. [2] also notes that itis vital that speech describing object or actions should

coincide with the object or action being shown.

The structure of the mental model formed from the

propositions can be supported by the use of advanced

organisers and summaries [13, 22], The use of a the

general title ‘Metastasis’ improved recall for proposition

ii). Titles, advanced organisers and summaries may help

to structure the mental models formed by the user from

the individual propositions, keeping high levelpropositions active in the mental model.

IMPROVING DESIGN

We have used these conclusions to produce a re-authored

version of the first part of the presentation (figure 1a, b),which showed the tumour breakthrough the basal lamina

and enter the connective tissue.

We began by providing an advanced organiser for the

sequence, using a caption and speech track whichintroduced the main topic of Metastasis (figure 8). The

caption was tightly linked to the speech track to reinforce

the propositions given. The organiser will form the top level

of the mental model for this part of the presentation, below

which other propositions could then be integrated,

272

Metastasis is the spread of cancer from the primary tumour

to a secondary tumour...

Figure 8

In the first scene itself, the title ‘Metastasis’ was removed

and placed in the organiser as any unnecessary titles may

distract attention following orientation. The original

presentation also used three other labels, which were shown

concurrently. These may cause attention to shift

unpredictably between objects. Instead a single central label

was used to direct attention to the central tumour mass;

other labels were then revealed as required. The tumour

mass had a different colour to its background, this was left

as-is, since objects which stand out against their ground will

usually attract attention (see figure 9).

,,, The malignant primary tumour shown is in the patients

bronchial epitheliums.

Figure 9

Following a pause to allow reading time, a revealed captionwas then used to shift attention to the basal lamina (figure

10). The basal lamina was

basal lamina object wasnot brightly coloured, so a newdrawn which stood out more

clearly, to make it easier to attend to. Again reading time

was allowed for the caption.

In order to spread, the cells of the primary tumour must

break through the basal lamina ...

Figure 10

Next an animation was added to show the basal lamina split

apart. To improve comprehension of the action for the

proposition ‘Tumour cells break through basal ku-nina’ the

next frame (figure 10) used an additional animation of the

basal Iamina splitting. The speech track was also changed to

reinforce the text captions, Using the same phrases in

speech and captions will tightly link what is processed into

a single proposition. The start of the animation was then

synchronised to the speech track, so that the spread of the

tumour began as it was introduced by the speech. This may

again help to produce a single unified proposition.

.,. and then enter the surrounding connective tissues.

Figure 11

The animation will shift attention to the basal lamina and

bottom of the tumour. As the animation will also lock

attention, it was placed prior to the existing animation of the

273

tumour spreading down into the connective tissue which

otherwise might have competed for attention.

After the end of the new animation, the existing animation

was then used to shift attention to the tumour spreading

down into the connective tissue. With attention shifted to

the spreading tumour, this animation was then paused half

way through and a label revealed describing the animation,

figure 11. By pausing the animation, it should allow

attention to move to the revealed label. After leaving time

for the label to be read, the end of the animation was

completed.

MetastasisThe spread of cancer from Primary Tumour .

(i)TumourCells in Bronchial Epilhelium cross Basal

Lamlna and enter connective Tissues

Figure 12

The final frame (figure 12) returned to the advanced

organiser and gave a summary of the previous animations

and captions to provide a high level proposition to ensure

the mental models fi-om the captions and animations were

correctly integrated with the organiser.

EVALUATION OF RE-AUTHORED VERSION

Our current work is to evaluate the re-authored presentation

with a further group of subjects. Preliminary results indicate

that comprehension is improved by the design changessuggested. However, we are also seeking to clarifi whether

changes in the content of the verbal information provided,

or the addition of visual elements such as animations and

captions are responsible for these improvements, To

investigate these issues, we intend to produce a third

presentation, which will be a text and speech only condition

based upon the re-authored presentation. By comparing thiswith the re-authored presentation, we will be able to judge

how effective our visual against verbal changes have been.

CONCLUSIONS

Little current work on MM has addressed the evaluation

and successful design of MM presentations. It is believedthat the issues of attention and comprehension are strongly

interlinked with producing ‘good’ MM presentations : a

‘good’ presentation is one which can understood by the userin the manner intended by the designer. The two studies

reported have shown some of the dangers of attempting to

design MM presentations without concern for attention and

comprehension.

Our eye tracking results explored the potential problems

that using captions, labels, animations and scene changes

may cause, Attention can easily be overloaded, resulting in

presentation elements being missed. Attentional design

should take into account the need to order shifts of attention

according to the content being shown. It should allow for

attention being locked by animation and reading, such as

providing time to read captions and labels and to pause for

re-orientation when scenes are changed. Where the

presentation is complicated it may also be useful to guide

attention to important elements using animation, reveals,

colouring or labels.

Our second study of subject’s recall of the presentation

showed that design problems may lead to difficulties incomprehension. We suggested comprehension may be

improved by ensuring that the elements of the presentation

could be formed together into unified propositions. This

may be achieved by scripting the speech track to more

tightly reference and synchronise with the visual elements

of the presentation; and by using animations and captions to

support the propositions being formed. We also proposed

that the formation of the subject’s mental model could be

supported by applying titles, organisers and summaries.

This work provides some tentative steps toward placing

MM design and evaluation within a firmer theoretic

framework. Several problems remain for future research. As

noted, the re-authored version of the presentation is

currently being tested with firther subjects. We also feel it

is important to address how design advice can be integrated

into the life cycle of MM production, following our work in

[20]. Because of the costs of producing MM presentations,

leaving evaluation until the product is complete is not

practical. Design advice needs to be given during the

formative processes of authoring, such as when scriptingand choosing media.

ACKNOWLEDGMENTS

The Etiology of Cancer CD-ROM is produced by and

copyright of American Medical Television and Silver

Platter,

REFERENCES

1. Baddeley, A. Human Memory. LEA, 1990.

2. Baggett, P. The Role of Temporal overlap of Visual and

Auditory Material in Forming Dual Media Associations.

Journal of Educational Psychology 76,1984,408-417

3, Bieger, G.R. & Glock, M.D. The Information Content

of Picture-Text Instructions. Journal of Experimental

Education 53 (1984), 68-76.

274

4. Brandt, H. The Psychology of Seeing. New York : The

Philosophical library (1954)

5. Grimes, T. Mild Auditory-Visual dissonance in TV

News may exceed viewer attention. Human

Communication Research 18,2 (1991), 268-298

6. Hegarty, M. & Just, M.A. Constructing mental models

of machines fi-om text and diagrams. Journal of Memory

& Language, 32,717-742, 1993.

7. Hillstrom, A.P. & Yantis, S. Visual motion and

attentional capture. Perception & Psychophysics, 55(4),

(1994), 399-411.

8. Hochberg, J & Brooks, V. The Perception of Motion

Pictures. In Handbook of Perception and Human

Pe~formance X

9. Johnson, W .A. & Dark, V..J. Selective Attention.

Annual review ofpsychology, 37,43-45.

10. Johnson-Laird, P. Mental Models. Harvard Press, 1983.

11. Kraft, R. The role of cutting in the evaluation andattention of film. Journal of Experimental Psychology :

Learning, Memory & Cognition, 12, 1 (1986), 155-163.

12. Loftus, G.R. Eye Fixations on Text and Scenes. In EyeMovements in Reading, Pet-ception, and Language

Processes. Academic Press, 1983, pp.359-376.

13. Mayer, R. E. Aids to text comprehension. Educational

Psychologist, 19(1 984), 30-42

14. Nelson, W.W. & Lotlus, G.R. The functional field

during Picture Viewing. Journal of Experimental

Psycholo~ : Human Learning& Memoq, 6(4), (1980),

391-399.

15. Park, O. & Hopkins, R. Instructional conditions for

using dynamic visual displays : a review. Instructional

Science 21 (1993), 427-449

16. Reeves, B. The effects of scene changes and semantic

relatedness on attention to television, Communication

Research, 20(2), 1993, 155-175

17. Rieber, L. P. Animation in Computer Based Instruction.

Educational Technology Research & Development 38

(1990), 77-86

18, Rollins, H.A. & Thibadea, R. The effects of auditory

shadowing on recognition of information received

visually. Memory & Cognition, 1(2), (1973), 164-168

19. Salame, P. & Baddeley, A. Disruption of STM by

unattended speech : Implications for the structure of

Working Memory. Journal of Verbal Learning &

Verbal Behaviour,21, (1982), 150-164

20. Sutcliffe, A.G. & Faraday, P.M. Designing Presentation

in Multimedia Interfaces. In Proceedings CHZ’94

Human Factors & Computer systems. ACM, 1994, pp.

92-98

21, Yarbus A. Eye Movements& Vision. Plenum. 1967.

22. Van Dijk, T.A. & Kintsch, W. (1983) Strategies of

Discourse Comprehension. New York: Academic Press.

275