ATTENTION FACTORS IN
TELEVISED. MESSAGES:
EFFECTS ON LOOKING BEHAVIOR
AND RECALL

ThesisforAtheDegréev'offPh_.fIDE..;i fl 1
. MICHIGAN STATE UNIVERSITY;
> ' SHERILYN‘. KAY ZEIGLER' y ‘

' 1969

I 11 gm;

 

This is to certify that the

thesis entitled

Attention Factors in Televised Messages:
Effects on Looking Behavior and Recall

presented bg I

Sherilyn Kay Zeigler

has been accepted towards fulfillment
of the requirements for

PhoDO degree in communication

7/511 ////

Major professor

Date May 8, 1969

0—169

 

 

Accepted by the faculty of the Department of Communication,
College of Communication Arts, Michigan State University, in partial

fulfillment of the requirements for the Doctor of PhilOSOphy degree.
I}

 

Guidance Committee: , Chairman

 

 

 

 

 

 

ABSTRACT
ATTENTION FACTORS IN TELEVISED MESSAGES:
EFFECTS ON LOOKING BEHAVIOR AND RECALL
By

She rilyn Kay Zeigler

The Problem

In the field of mass communication today, a great deal of
effort is directed toward attracting the attention of consumers
amidst a multitude of competing stimuli. Yet, relatively little is
known about techniques of obtaining attention and of maintaining it
for the duration of a message. An experiment was conducted for
the purpose of examining attention factors in televised messages.
The study sought answers to the following questions:

1. Given the same visual content in televised messages,
how are looking behavior and recall affected by varying camera
techniques ?

2. Given the same visual content, how are looking behav—

ior and recall affected by relevant and irrelevant audio?

She rilyn Kay Zeigler

3. How are looking behavior, recall, and content
evaluations affected by simple and complex displays of the same

Vi sual content ?

Method

An eye-movement camera, utilizing Optical techniques
developed by Dr. Norman H. Mackworth, was equipped with a 16mm
reflex motion picture camera. Eye movements across televised
messages were thereby recorded. Three message manipulations
were employed: (1) visual hesitation (operationalized as "blank
spaces" appearing between series of items on the screen); (2) audio-
video interlock (operationalized as visual presentation of items and
simultaneous verbal mention of them); and (3) visual simplicity and
complexity (operationalized as symmetrical and nonsymmetrical
arrangements of a given set of items).

Analysis of variance was used to test independent and inter-
active effects of the above on looking behavior and retention. In
addition, the effects of visual simplicity and complexity on evalua—
tions of Visual content were examined. Subjects were forty—eight
volunteers from undergraduate classes at Michigan State Univer-

sity.

She rilyn Kay Zeigler

Findings

NO significant differences inlooking behavior or recall
were found in connection with the visual hesitation manipulation.
Likewise, no differences occurred in looking behavior, recall, or
content evaluations, in the case of the visual simplicity and com-
plexity manipulation. The audio-video interlock treatment did pro-
duce a significant effect in one instance. Respondents seeing a
series of items on the screen while the announcer simultaneously
discussed them recalled more of the items than did respondents
seeing the same Objects while the announcer gave irrelevant infor-
mation. The difference occurred, however, only when the recall
test was administered immediately after exposure to the message.

A delayed recall test (one to four weeks later) revealed no difference
between groups. Also, when the same treatment was used with a
different set of items (in a different scene), no significant effect was
observed on either an immediate or a delayed basis.

Nineteen analysis of variance tests were run; thirteen of
them produced results in the predicted direction. It seemed, there-

fore, that the phenomena studied did have some effect on the subjects.

Conclusions

 

Conclusions were that Visual hesitation, audio—Video inter-

lock, and visual simplicity and complexity had little effect on looking

Sherilyn Kay Zeigler

behavior or recall. It was recommended, however, that all three
manipulations be studied in connection with subjects' tendency to
follow the action occurring on the screen, and with their inclination
to anticipate information not yet exposed visually.

The eye camera apparatus has just begun to be used in the
area of televised messages. This study suggests that it may provide
a useful tool to communication research, as more attention factors
are identified and isolated for consideration of their effects on

learning and evaluation of the material presented.

ATTENTION FACTORS'IN TELEVISED MESSAGES:

EFFECTS ON LOOKING BEHAVIOR AND RECALL

By

She rilyn Kay Zeigler

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Communication

1969

6:757 Y0

//'/7-'67

ACKNOWLEDGMENTS

Special thanks must be given to Dr. Verling C. Troldahl,
Associate Professor of Communication, at Michigan State Univer-
sity, first, for stimulating my interest in eye-camera research,
and second, for guidingand directing this study through to its com-
pletion. The valuable assistance of Dr. Kenward L. Atkin, Pro-
fessor of Advertising and Communication, and the encouragement
of Dr. William J. E. Crissy, Professor of Marketing, and
Mr. Leo A. Martin, Professor of Television and Radio, are also
gratefully acknowledged.

Credit goes to Mr. George Robeck, for teaching me the
hows and whys of eye camera operation, and I am indebted to
Mr. James Culley, my laboratory assistant, who aligned and ran
the eye camera for the duration of the experiment. Finally, appre-
ciation is given to Mr. ElDean Bennett, who provided audio for the
commercials; to Mr. John Hofmeister, who assisted in the reliability
check; to students in the advanced course in Television and Radio
Advertising who took part in production of the test commercials;

and to students in Advertising classes who served as subjects.

ii

i TABLE OF CONTENTS

Page
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . v
LIST OF ILLUSTRATIONS . . . . . . . . . . . . . . . . vi
Chapter

I. BACKGROUND AND RELATED LITERATURE . 1

Need for Research 1

The Problem 6

Significance of Study 6

Limitations . . . . . . . . . . . . . . . 8

Hypotheses . . . . . . . . . . . . . . . . . 10

II. METHODOLOGY . . . . . . . . . . . . . . . . . 20

Recording Instrument . . . . . . . . . . . . . 20

Additional Equipment . . . . . . . . . . . . . 20

Subjects . . . . . . . . . . . . . . . . . . . 22

Pretest . . . . . . . . . . . . . . . 22

Experimental PrOCedure . . . . . . . . . . . 24

Experimental Messages . . . . . . . . . . . 26

Questionnaire . . . . . . . . . . . . . . . 43

Problems Encountered . . . . . . . . . 44
Operationalization of Looking Behavior

Variables . . . . . . . . . . . . . . . . 49

Statistical Design . . . . . . . . . . . . . . 53

III. FINDINGS . . . . . . . . . . . . . . . . . . . . 56

Reliability Check . . . . . . . . . . . . . . 57

Visual Hesitation . . . . . . . . . . . . . . 58

Interlock . . . . . . . . . . . . . . . . . . 64

iii

Chapter

Visual Complexity
Visual Complexity and Interlock
Summary

IV. CONCLUSIONS

Conclusions Based on Findings
Summary of Findings .
Discussion

Suggestions for Future Research .

LITERATURE CITED .

APPENDIX: QUESTIONNAIRE

iv

Page
68
73
77
78
78
84

85
100

105

108

Table

LIST OF TABLE S

Means for Immediate Re call Under Conditions
of Format Complexity and Interlock

Means for Delayed Recall Under Conditions
of Format Complexity and Interlock

Page

74

76

LIST OF ILLUSTRATIONS

Figure Page

1. Side View: Eye Movement Recorder,
ModelV—1164-1................. 21

vi

I- BACKGROUND AND RELATED LITERATURE

Need for Research

 

In the field of mass communication today, a great deal of
effort is directed toward attracting the attention of consumers
amidst a multitude of competing stimuli. Yet, relatively little is
known about techniques of obtaining attention and of maintaining it
for the duration of a message.

Studies have suggested that attention requires focusing of
the eyes and mind in such a way that a stimulus may later be
recalled with a high degree of accuracy. Sperling defined attention
according to the focus of a subject' s eyes at a given instant, and
his verbal reactions to changes in visual stimuli. As eye fixations
shifted, so did attention—-and so did the speed and accuracy of
reported visual changes. 1

Woodrow measured the attention value of items by noting

respondents' reaction time to a change in intensity or "brightness"

 

1George Sperling, "The Information Available in Brief Visu-
al Presentations," Psychological Monqgraphs, LXXIV, No. 11
(1960).

 

of a stimulus. Subjects reported verbally whether it was the right
side or left side of a projected image which decreased in intensity
at a given moment.

Levy was interested in the attention-gaining advantage of
various positions on outdoor billboards. 'Subjects were briefly
exposed to a viewing screen divided into eight equal sections. In
seven of the sections a black cross appeared; in the eighth there
was nothing. The attention value of each position was operation-
alized as the number of correct judgments respondents made when
asked in each case where the blank appeared.

In all of the above studies, however, attention factors were
examined in isolated situations. There has been little attempt to
investigate the conditions under which specific stimulus properties
-—across a wide variety of messages-~might elicit a given degree
of eye-movement activity and verbal recall. Until these conditions
are explored, few predictions can be made regarding attention-
drawing and attention-holding elements in audio-visual mes-

sages.

 

2Herbert Woodrow, "The Measurement of Attention, " Fire
Psychological Monogaphs, XVII, NO. 5 (December, 1914).

3.1. M. Levy, "Experiments on Attention and Memory,
with Special Reference to the Psychology of Advertising, " Univer-
sity of California Publications in Psychology, 11, No. 2 (April 12,
1916), 157-197. ‘—

 

 

In addition, most of the research which has been conducted
concerning attention to messages has utilized the printed media.
Few studies have considered television communication, though tele-
vision set penetration in the United States is now 95 per cent, 4 and
Americans spend an average of almost six hours per day watching
TV. 5 Currently, also, the country' s 100 largest advertisers invest
over 60 per cent Of their budgets in television. 6 There is a real
need today to examine what constitutes an "attention factor" in a
televised message. The message must be seen and heard (somehow
it must attract the attention Of viewers) before it can ever be
believed, be remembered, or have any influence on the acceptance
of an idea or sale of a product.

Research in looking behavior indicates that eye movements
are closely related to the presence of attention factors in visual

stimuli. 7 Again, however, the vast majority of studies in this area

 

4Television Factbook, 1968-69 edition, NO. 38 (Washing-
ton: Television Digest, Inc.), p. 81—a.

 

5"Television Home Viewing Gains for 13 of 14 Months, TVB
Analysis of A. C. Nielsen Data Reveals, " TvB Report, V-67(68),
New York.

6"Why We Use Television" (Television Bureau of Adver-

tising, 1968). (Mimeographed.)

7H. K. Nixon, "Attention Value and Interest in Advertis—
ing, " Archives of Psycholcgy, No. 72 (1924); J. P. Guilford and

 

to date have been concerned solely with still, rather than moving,
pictures. Ocular photography, with its unique potentials for
analysis, can give the television communicator a clear concept of
the course his viewers are following, and of the relative emphasis
being placed on particular content. It may reveal what catches and
holds attention, by indicating the location and duration of fixations
and the direction and sequence of eye movements.

Development of methods for studying eye movements began
in the middle of the nineteenth century. The "after image technique, "
wherein eye movements were observed following exposure to a bright
spot of light, provided varying degrees of reliability. The EOG
(electro-oculographic) method, involving the recording of eye move—
ments through amplification of an electric field (at the site of elec-
trodes around subjects' eyes) was also crude, and resulted in a fair
amount of subject discomfort.

In the corneal reflection technique, however, first reported

by Dodge, in 1899, the front surface of a subject' 8 eye reflected

 

H. B. Hackman, "A Study of the Visual Fixation Method of Measur-
ing Attention Value, " Journal of Applied Psycholqu, XX (1936),
44-59; James Spier Karslake, ”The Purdue Eye—Camera: A Prac-
tical Apparatus for Studying the Attention Value of Advertisements"
(unpublished Ph.D. dissertation, Purdue University, 1939).

 

8Herman F. Brandt, The Psycholpgy of Seeing (New York:
Philosophical Library, 1945), pp. 54, 56.

 

light from a fixed source, so that as the eye rotated, the light was
reflected through a corresponding angle. This particular system
has been greatly refined today, and yields fairly accurate informa-
tion concerning the relationship of eye movements to a visual field. 9
In addition, the corneal reflection technique may be used in the study
of moving pictures as well as stills.

Numerous laboratory experiments have indicated that only
one stimulus can occupy the focus Of attention at any given moment.
Specifically, Baldwin found that subjects could acquire information,
for the most part, through only one channel (either visual or aural)
at a time. 11 There is a need, however, to apply these findings to
the area Of televised communication, and to work toward develop-
ment of an audio-video "mix" which would bring Optimum results in
gaining and maintaining attention. Studies investigating eye move-

ments across televised messages can easily incorporate a

 

9Willavene Wolf and Others, "An Experimental System for
Eye-Movement Studies in Dynamic Bi-Dimensional Fields, " 64-157,
Bureau of Educational Research and Service, The Ohio State Univer-
sity, pp. 1—2.

10Brandt, The Psychology of Seeing, pp. 54, 56.

 

1Thomas Frederick Baldwin, "Redundancy in Simultaneously
Presented Audio-Video Message Elements as a Determinant of Re-
call" (unpublished Ph.D. dissertation, Michigan State University,
1966).

consideration of the effects of audio-video redundancy both on looking

behavior and on recall.

The Problem

 

Given the needs stated above, an experiment was conducted
at Michigan State University during the winter of 1969, for the pur-
pose of examining attention factors in televised messages--and,
specifically, in television commercials. The study sought answers
to the following questions:

1. Given the same visual content in televised messages,
how are looking behavior and recall affected by varying camera
techniques?

2. Given the same visual content, how are looking behav-
ior and recall affected by relevant and irrelevant audio?

3. How are looking behavior, recall, and content evalua-
tions affected by simple and complex displays of the same visual

content ?

Significance of Study

 

The present study was, first, an attempt to relate existing
knowledge of looking behavior to the important but largely untested
field of television messages. Eye -movement studies were conducted

in the area of reading many decades ago, and led to a revolutionary

change in the methods of teaching reading. For example, the dis-
covery that eyes do not move in a continuous sweep across a page

(but rather move in a series of quick, short movements and fixation
pauses) resulted in instruction programs geared to achieving an
increase in the perception span of children learning to read. Pupils
learned to increase their fluency and reading speed by grouping words
into phrases. Other studies revealed that the type of material read
greatly influenced eye -movement patterns.

Textbook publishers found such information valuable in mak-
ing decisions regarding length of lines of type and physical arrange-
ment Of the printed page. 12 Experimental work Of this nature in
television might well suggest new methods of presenting visual
material and of combining audio and Video elements so that viewers!
attention could more easily be Obtained.

Second, the current study may Open the door to a new
method of measuring the effectiveness of television messages.

Quite frequently, it is assessed through tests for memorability or
believability, or through viewers' stated willingness to accept the
ideas advocated. In each of these cases, the message is evaluated

as an entity--though writers and producers spend valuable time

 

12Willavene Wolf and Others, "An Eye -Movement Study of
Children Viewing Television, " 64-157, Bureau of Educational Re—
search and Service, The Ohio State University, p. 2.

planning and arguing over the ”most effective" presentation of a
single line or camera shot. It is time that analysts backed up a
few steps--to determine just what it is which catches a viewer' s
attention in the first place, before he can ever start to "remember"
or "believe" or "become interested in accepting. "

Third, a study of this nature gives advertising research a
chance to test some of the "accepted principles" passed from copy-
writer to copywriter and from textbook to textbook Without any scien-
tific support. One such maxim is "Keep it simple. " Another is:
”Say what you show and show what you say. " In addition, this kind of
study may identify and isolate creative variables not previously re-
garded as factors affecting advertising success, and may, thus, pave

the way for future research.

Limitations

 

Brandt found that spatial position could be an important
determinant of attention. Working with numerous black squares
drawn on a 10" X 10H card, he found that subjects made considerably
more eye fixations on the left side than they did on the right (and

especially on the upper left portion). 13 Objectives of the current

study, however, focused on general characteristics of looking

 

3Herman F. Brandt, "Ocular Patterns and Their Psycho-
logical Implications, " American Journal of Psychology, LIII (1940),
260-268.

 

behavior across televised messages, and on the power of various
visual and aural manipulations to elicit recall. No attempt was
made, therefore, to determine whether or not one area of the tele—
vision screen received more fixations than another.

In photographing eye movements, Mackworth, Kaplan, and
Metlay discovered that even though subjects spent about the same
amount of time looking at each area of a visual stimulus, they shifted
from one to another at different rates. There was some indication
in this case that frequent shifting was positively related to the
accuracy and amount of recall. 14 Again, however, while present
objectives called for a determination of the total length of time sub—
jects spent looking at particular items, they did not relate to a
consideration of consecutive eye—fixations in any one area.

In a television message, the visual properties on which
subjects may fixate are Often numerous and individual items may
even be subdivided for purposes of examining subjects' fixation
behavior. For example, talent, props, and sets may be considered
as entities for eye-movement activity, or they may be broken down
into talent' 3 hands and faces, specific props and their parts or con-—
tainers, and furniture and scenery in sets. Present hypotheses,

however, limited looking behavior analysis in this study to the

 

14N. H. Mackworth, I. T. Kaplan, and W. Metlay, "Eye

Movements During Vigilance, ” Perceptual and Motor Skills, No. 18
(1964). pp. 397-402.

 

10

items and parts of items which were specifically manipulated in each

 

scene.

Hypothe se 8

 

Visual Pauses: Effects on Retention
and Looking Behavior

In working with oral communications, J ersild found that
recall was aided by the speaker' 8 use of pauses. 15 Later, Ehrens-
berger read statements to groups of students and tested them for
recall; he, likewise, found that pausing before a statement had
definite retentive value.

Little, if any, research has been done regarding the use Of

pauses--to aid retention—-in television' 3 visual channel. It seemed

 

reasonable to assume, however, that if camera techniques were
employed to keep viewers temporarily "in suspense, " the attention
value of the objects finally revealed would be increased. Since it
was found that a camera pan across items in a display could permit

II

appearance Of each only after a slight "visual hesitation, it was

 

15Arthur Jersild, "Primacy, Recency, Frequency, and
Vividness, " Journal of Experimental Psgchology, XII (1929), 58—70.

16

 

Ray Ehrensberger, ”An Experimental Study of the
Relative Effectiveness of Certain Forms of Emphasis in Public
Speaking, " Speech Monographs, XII (1945), 94-111.

 

11

proposed that recall of items would be greater in that case than it

would be when the camera panned without visual hesitation. The

fact that the visual time spent on each item could be controlled by

spacing the items farther apart in the first case than in the second

(so that for every moment spent on an item, a moment would be

spent on a ”blank spot”) led to the following hypothesis:

H1: In a televised message, there will be greater

immediate recall of relevant visualized items when

the camera pans with visual hesitation than when it
pans without visual hesitation.

 

Then, since there was no available research suggesting
anything to the contrary, the following was also hypothesized:
H2: In a televised message, there will be greater
delayed recall of relevant visualized items when
the camera pans with visual hesitation than when it
pans without visual hesitation.
Thomas examined eye-fixation behavior while subjects
looked at slides of Rorschach Ink Blots, With no real cues as to
where they should look, most of the subjects tended to fixate initially

on the center of the display. 17 In the present study, when visual

hesitation occurred, subjects momentarily faced a ”blank" screen

 

17E. Llewellyn Thomas, "Eye Movements and Fixations

During Initial Viewing of Rorschach Cards, ” Journal of Prgj'fctive
Techniques and Personality7 Assessment, XXVII (1963), 345—353.

 

 

12

(with no cues as to where they should look). When visual hesitation
did not occur, however, the screen contained suchcues at all times
--items or parts of items relevant to the sales message. Hence,
the following was proposed:
H3: In a televised message, when the camera pans with
visual hesitation across relevant display items,
first eye fixations will be closer to the center of
the frame when a "blank screen" appears than they

will be when the camera pans without visual hesitation
directly onto another item.

Definitions

 

Visual Hesitation: immediately after the camera pans
across an item, it pans for the same
amount of time across a blank space
before moving to the next item

 

No Visual Hesitation: the camera pans across each display
item in succession; at no time does
a completely blank screen appear

Relevant Visualized Items: piano, records, drum, horn, banjo,
radio (SCENE #1); pillow, thermos,
flashlight, clock, case with a shoulder
strap (SCENE #4)

Immediate Recall: the number of relevant items (SCENE
#1 and SCENE #4) recalled on the
questionnaire administered imme—
diately after exposure to the message

Delayed Recall: the number of relevant items (SCENE
#1 and SCENE #4) recalled on the
questionnaire administered one to
four weeks after exposure to the
message

13

Redundancy in the Audio-Video Relationship:
Effects on Retention and Looking Behavior

Baldwin obtained maximum recall from viewers of a motion
picture film clip when there was high redundancy in the audio-video
relationship. 18 Therefore, it was proposed in the current study
that recall of video features would be greater, in the case of tele-
vision messages, when the audio and video "interlocked" (that is,
when they were redundant) than when the audio was irrelevant to the
visual presentation.

H4: In a televised message, there will be greater
immediate recall of relevant visualized items when

there _i_s_ audio-video interlock than when there is no
audio-video interlock.

 

Likewise, since there was, again, no research to suggest
otherwise:

H : In a televised message, there will be greater
delayed recall of relevant visualized items when
there _is audio—video interlock than when there is no
audio-video interlock.

In one case, it was supposed that the audio would draw

attention to the items appearing visually; in the other, they might

 

18Baldwin, "Redundancy in Simultaneously Presented

Audio—Video Message Elements as a Determinant of Recall. "

14

well go unnoticed by subjects. Looking behavior would be differ-
entially affected by the two situations.
H6: The amount of time spent looking at relevant
visualized items will be greater when there is

audio-video interlock than when there is 29 audio-
video interlock.

Definitions

 

Audio-Video Interlock: when items are shown visually, they
are simultaneously verbalized by the
announcer

No Audio-Video Interlock: when items are shown visually, the
announcer presents irrelevant mate-
rial '

Relevant Visualized Items: four batteries, cord for use with an
AC outlet, holder for the cord in the
radio case (SCENE #3); adjustable
antenna, earphone, front button to
light the dial (SCENE #5)

Immediate Recall: number of relevant items (SCENE #3
and SCENE #5) recalled on the ques-
tionnaire administered immediately
after exposure to the message

Delayed Recall: number of relevant items (SCENE #3
and SCENE #5) recalled on the ques-
tionnaire administered one to four
weeks after exposure to the message

Looking Time: number of eye fixations appearing on
any part of the relevant items (SCENE
#3 and SCENE #5)

15

Complexity of Visual Patterns: Effects on
Retention, Looking Behavior, and Evaluations

Two key tenets of Gestalt psychology are, first, that per-
ception is organized, and second, that the organization tends to be as
good as the stimulus conditions permit. An important principle of

H

perceptual organization has been called ”perceptual grouping. In

this regard it is believed that there is a perceptual tendency to group
elements into a ”good form"—-one which is "balanced" or "closed. "19

Berlyne defined "complex" patterns as those lacking sym-
metry of form, or housing structural elements in a random, rather
than a systematic, fashion. 20 The contrast was explained in terms
of specific and diversive exploration. According to Berlyne, spe-
cific exploration was prompted by incomplete perception of a stimu-
lus pattern, leaving subjects with uncertainty regarding its properties;
such exploration was encouraged when the stimulus was "complex"
and exposure to it was brief. The rating of ”interesting" here was
thought to indicate subjects' wishes to continue looking at the

stimulus, in hopes of obtaining additional information about it; in

other-words, subjects were seeking closure.

 

9Morton Deutsch and RobertM. Krauss, Theories in Social
Psychology (New York: Basic Books, Inc., 1965), pp. 16-21.

20D. E. Berlyne, "Curiosity and Exploration, " Science,
CLIII (July 1, 1966), 25-33.

 

16

Diversive exploration was motivated by a desire for
entertainment or diversion, and was encouraged when the stimulus
was "simple. ” In this case, subjects found the designs "pleasing"
to look at, since there was no uncertainty as to information con-
tained; that is, closure was easily obtained. 21

It was proposed that the same would be true for television
messages which included a complex and a simple display of objects.
Both displays contained nine items (three each of three different
sizes), but in one case they were arranged in a random, unbalanced
manner (complex), and in the other they were set up in a balanced,

pyramidal form (simple).

H7: In a televised message, subjects will rate a complex
display of items more "interesting” than a simple
one, when tested immediately after presentation.

H : In a televised message, subjects will rate a simple
display of items more "pleasing" than a complex
one, when tested immediately after presentation.

Once more, given no available research to the contrary, it

was hypothesized that the above differences would hold over time:

H9: In a televised message, subjects will rate a complex
display of items more "interesting" than a simple
one, when tested one to four weeks after presentation.

 

21D. E. Berlyne, "Complexity and Incongruity Variables

as Determinants of Exploratory Choice and Evaluative Ratings, "
Canadian Journal of Psychology,XVII (1963), 274—290.

 

k

17

H10: In a televised message, subjects will rate a simple
display of items more "pleasing" than a complex
one, when tested one to four weeks after presentation.

 

Thomas found that the length of fixation on a display was a
function of the difficulty of extracting information from it. The
longest fixations occurred when subjects were presented with
Rorschach Ink Blots; shorter fixations followed when the stimulus
was a chest x—ray, and still shorter fixations occurred in the case
of a simple prose passage. 22 Berlyne and Lawrence concurred that
the duration of eye fixations increased with complexity of the figures
presented. 23

In the present study, it was also proposed that subjects
would spend mars time looking at items in the complex display than
they would in the simple display, since the closure they sought
would come more quickly in the latter case. In other words, it was
assumed that presentation of the complex display would lead to

greater duration of each eye fixation than would be true with presen-

tation of the simple display. The random arrangement of items (in

 

2
2Thomas, "Eye Movements and Fixations During Initial

Viewing of Rorschach Cards, " pp. 345—353.
23D. E. Berlyne and George H. Lawrence, "Effects of
Complexity and Incongruity Variables on GSR, Investigatory Behav-
ior, and Verbally Expressed Preference, " Journal of General
Psycholog , LXXI (1964), 21-45.

 

,_

the complex display) would make it more difficult for subjects to
extract information; hence, they would not be able to look at as
many different items in a given time period as would subjects see-
ing the simple display.

H : In a televised message, subjects will fixate on more

11 different items given a simple display than they will
given a complex display.

Definitions

Interest Rating: the score obtained on a seven—point
rating scale

Pleasantness Rating: the score obtained on a seven-point
rating scale

Simple Display: nine radio mock—ups, three each of
three different sizes, with accom-
panying carrying cases in plain and
plaid colors—-all arranged sym-
metrically (systematically) in SCENE
#6

Complex Display: nine radio mock—ups, three each of

three different sizes, with accom-
panying carrying cases in plain and
plaid colors-—all arranged nonsym—
metrically (randomly) in SCENE #6
Finally, since it was suggested that subjects would fixate
on more different objects given exposure to a simple display than

they would given exposure to a complex display, it was proposed

that recall of features of these objects would be greater in the

former case than in the latter. In addition, it was also suggested
that there would be greater recall of items appearing visually when
there was audio—video interlock than when there was 39 audio—video
interlock. It was proposed, therefore, that given exposure to the
complex display, recall would be greater when there was audio-
video interlock than when there was fl audio-video interlock.
(Audio-video interlock was also manipulated for SCENE #6.) Thus,
the following interaction hypotheses emerged:
H : In a televised message, immediate recall of relevant
Visualized items w1ll be lower given a complex display
than it will be given a simple display, but not as much

lower when there is audio—video interlock as when there
is no audio-video interlock.

 

H : In a televised message, delayed recall of relevant
13 . . . . —. .
V1sualized items W111 be lower glven a complex dlsplay
than it will be given a simple display, but not as much
lower when there _is audio-video interlock as when there
is 29 audio—video interlock.

 

Definitions for immediate and delayed recall, audio-video
interlock and no audio—video interlock, and recall of items for

SCENE #6 remain the same as those stated previously.

11. ME THODOLOGY

Recording Instrument

The eye-movement camera used in this experiment was
equipped with a 16mm reflex motion picture camera for purposes
of recording eye movements on film. (See Figure 1.) Optical
techniques for the system were developed by Dr. Norman H. Mack-
worth. The principle of recording used corneal reflection super-
imposed on a film of the televised message Viewed by subjects. An
immediate pictorial record was obtained; subjects' eye movements
were shown as a series of light spots. In addition, a rotating line,
centered on the spots, indicated sequence and duration of fixations.

The field of View was twenty—two degrees, with an average
registration accuracy of plus—or-minus one-half degree. The film
Speed of the motion picture camera used in this study was eight

frames per second.

Additional Equipment

The televised message was Viewed through the system of

Fear screen projection. A plexiglass "TV screen, ” 9" X 12", and

20

L

Stage

Stage aperture

21

Marker light mirror
Marker light unit

Stage light

I
IL

 

Foot switch

Power switch

 

 

 

 

 

 

 

 

 

 

“\Q

Viewing ocular

— Camera shutter release cable
Motion picture camera (V-1164-1)

 

 

 

 

 

Fligure 1.-- Side View:

    
    
 

 

Stage light controls

Bri ghtne s s
Actuator switch
Conne ctor

Eye lens tube

Eye lens
Head rest
Bite board

—Coordinate controls

—-Vertica1

_ Lateral
tFocus

_Marke r light controls

_ Rotation speed
_ Connector
rightness

 

B
EA ctuator switch

Eye Movement Recorder, Model V-1164-1

22

a 4%" X 4%" speaker were mounted in a black board, fifty—two inches
from the subject.

Subjects

Subjects were students enrolled in undergraduate courses
at Michigan State University. Volunteers were selected from four

different classes, so that subject sensitization due to social inter—

action was minimized. Thirty-six men and twelve women partici—

pated; none had had any previous exposure to the eye camera
apparatus. Likewise, no subjects had any knowledge of the experi-
mental messages or variables; a five—minute standard presentation
in the classroom regarding eye camera research was the only back—
ground anyone had prior to participating in the experiment.

All participants were enthusiastic about the study—-both
before and after they arrived at the eye camera laboratory. No one
offered any complaints regarding discomfort during viewing of the
All were cautioned

test commercial or dislike of the questionnaire.

not to reveal details of the experiment to others until the study had
been completed.

Pretest

A pretest was conducted in June, 1968, using trial com—

Iner‘cials and questionnaires. Four points proved worth noting.

23

First, subjects needed to be assured of the safety of the eye camera
apparatus at the outset of the experiment; unless put completely at
ease in this regard, they were prone to move their heads nervously
during viewing of the commercials-—in spite of the fact that bite bars
were used. Second, subjects were aligned on the apparatus much
more quickly if they were continually reassured that they were "doing
just fine, " than was true if the experimenter said nothing to them

during alignment procedures. Third, it was important that subjects

were run through the experiment as rapidly as possible, in order to
prevent: (a) subject eye strain and consequent tearing, resulting in
very hard-to—interpret eye fixation recordings; (b) subject mental
strain and discomfort to the point where questionnaires were care—
lessly answered because subjects were anxious to leave; and (c) ex—
perimenter fatigue to the extent that instructions were hastily given,
and, hence, experimental conditions were not tightly controlled.
With regard to the questionnaire, a fourth factor concerned
the actual items which subjects were asked to recall from the com—
In(i‘I‘cials. Unless the objects were readily identifiable, respondents
did not take the time to describe them, and, in many cases, simply
left the spaces blank. In talking with subjects afterwards, however,

it Was determined that they 5113 have a "mental picture" of what they

had Seen, but did not know what to call the objects, and, therefore,

did not list them at all.

L

24

All four of the above problems were eliminated in the final
experiment.
Experimental Procedure

The experiment was conducted in the eye camera laboratory

of the Department of Communication, at Michigan State University,

over a three—week period: January 23 to February 13, 1969. Total

time per subject was approximately twenty to twenty—five minutes.
The delayed recall test was administered to all subjects in their
respective classrooms during the three-day period: February 19—21,
a lapse of one to four weeks from the time subjects saw the test

commercials. The experimental situation may be described as fol—

lows:
1. The subject was greeted by one male and one female

experimenter as he entered the eye camera laboratory, and was

asked to sit in front of the eye camera apparatus.

2. He was assured of the safety of the instrument and

advised of the nature of the experiment.

3. A bite bar was fastened to the equipment; its purpose

Was explained while the subject made a teeth impression in the wax

a-1’1<:l then let it harden.

4. The subject put his teeth back into the hardened impres-

siOh and looked at a magazine picture while one of the experimenters

.J'

I!

25

focused the corneal reflection beam to form an easily visible ”eye
marker." (Time: approximately one to two minutes. )

5. The subject was introduced to the sound of the 16mm

camera as a film was taken of his code number.

6. The magazine picture was removed and in its place
was a ”target card" used to check the instrument for linearity.

1 While the subject fixated on various parts of the card, the experi—
menter adjusted the focus control for optimum clarity of the eye—
marker and precise alignment. (Time: approximately one to two
minutes. )

7. A film was taken of the subject' s eye fixations on
various parts of the target card, so that a recording of the final
alignment would be available when the time came to code fixations
across the experimental message. (Time: approximately fifteen
to twenty seconds.)

8. The stage platform was removed and stage lights
1inrned off, while the subject' s head remained steady. A film was
then taken of his eye movements across the test commercial.

9. The subject was given the questionnaire and asked to
fill it out in an adjoining room.

10. The subject was given a brief explanation of what had

taken place during his viewing of the commercial, and was cautioned

L

26

not to reveal details of the experiment to anyone until results had

been released.

11. A delayed (identical) questionnaire was administered
in classrooms one to four weeks later.
During the experiment, up until the showing of the com—
mercial, subjects were continually reassured that they were "doing

a good job, " and that all was ”going well. "

Experimental Messages

Four versions of a television commercial were prepared

for experimental purposes. Since only one subject could use the eye

camera apparatus at a time, the message to which each was exposed
had to be short; otherwise, the experimenter would not have been
able to run a sample large enough for statistical analysis. It was
felt that a television commercial would best serve the purposes of
this study, because a complete message could be delivered in a short
1iin‘ie, and also because the principal investigator was experienced
in both the writing and the production of television commercials.

No attempt was made to secure actual televised commer-
Cials; instead, four versions of a commercial for a Panasonic radio
V"Ere created especially for this experiment——for two basic reasons.

F1 1"St, it was extremely important that manipulations were carefully

controlled and that they met production specifications of the study

k

27

exactly, so that findings were not confounded by extraneous vari—
ables. Second, subject sensitization had to be minimized to help
insure reliability of the testing procedure; hence, the experimenter
had to be absolutely certain that respondents had not seen the com—
mercials either before viewing them through the eye camera
apparatus, or after viewing but before taking the delayed recall
test.

Two versions of the commercial ran one minute and nine—
teen seconds; the other two ran one minute and seven seconds. The
time differential occurred in the first and fourth scenes, due to the

Vi sual manipulation employed there. The following pages contain

copies of the four scripts and detailed explanations of their contents.

SCENE #1:

 

(:21)

SCENE #2:

 

(:12)

SCENE #3:

 

(:06)

SCENE #4:

 

(:17)

28

COMMERCIAL VERSION A

 

(1 :19)

VIDEO

PAN WITH VISUAL
HESITATION ACROSS
DISPLAY OF MUSICAL
ITEMS, HOLDING
FINALLY ON ECU
PANASONIC RADIO.

CUT TO MCU GIRL
BEHIND TABLE. SHE
LIFTS RADIO, POINTS
OUT SPEAKER, AND
TURNS ENTIRE RADIO
AROUND.

DOLLY IN FOR CU
GIRL' S HAND OPENING
BACK OF RADIO AND
POINTING TO BATTER—
IES, CORD, AND CORD
HOLDER.

CUT TO CU ACCESSO-
RIES DISPLAY. PAN
WITH VISUAL HESITA—
TION.

AUDIO

(ANNCR OVER LIGHT MUSIC)

Look. These are the
makings of listening
enjoyment--a musical
experience that goes
where you go: it' s the
newly-developed, newly-
styled transistor radio
by Panasonic.

Priced at $49. 95, the
Panasonic weighs just
three pounds. . .yet, its
new speaker system gives
you the finest in sound
reproduction.

Four penlight batteries
are all you need——or use
any AC outlet. Here, the
cord stays neatly out of
sight.

And you can have your
choice of any of these fine
extras. Your dealer‘will
make them available to
you at just a slight addi-
tional cost when you buy
a Panasonic radio.

SCENE #5:
(:12)

 

SCENE #6:
(:11)

 

29

VIDEO

CUT TO MCU THREE
BOYS AND RADIO. ONE
POINTS OUT ANTENNA,
ANOTHER THE EAR-
PHONE, AND ANOTHER

THE FRONT DIAL LIGHT.

CUT TO CU SIMPLE
RADIO DISPLAY.

AUDIO

The adjustable Panasonic
antenna always brings you
perfect reception. In
addition, each radio comes
with its own earphone. . .
and, a press of that front
button lights up the dial
for easy tuning.

Be sure you ask to see

all three sizes of Pana-
sonic radios, styled in
both plain and plaid carry-
ing cases. Remember

the name: Panasonic.

SCENE #1:

 

(:21)

SCENE #2:

 

(:12)

SCENE #3:

 

(:06)

SCENE #4:

 

(:17)

30

COMMERCIAL VERSION B

 

(1:19)

VIDEO

PAN WITH VISUAL
HESITATION ACROSS
DISPLAY OF MUSICAL
ITEMS, HOLDING
FINALLY ON ECU
PANASONIC RADIO.

CUT TO MCU GIRL
BEHIND TABLE. SHE
LIFTS RADIO, POINTS
OUT SPEAKER, AND
TURNS ENTIRE RADIO
AROUND.

DOLLY IN FOR CU
GIRL' S HAND OPENING
BACK OF RADIO AND
POINTING TO BATTER—
IES, CORD, AND CORD
HOLDER.

CUT TO CU ACCESSO-
RIES DISPLAY. PAN
WITH VISUAL HESITA-
TION.

AUDIO

(ANNCR OVER LIGHT MUSIC)

Look. These are the
makings of listening
enjoyment--a musical
experience that goes
where you go: it' s the
newly-developed, newly-
styled transistor radio
by Panasonic.

Priced at $49.95, the
Panasonic 'weighs just
three pounds. . .yet, its
new speaker system gives
you the finest in sound
reproduction.

Take your Panasonic to
the beach with you--or
just about anywhere else.
It' 5 always there when
you' re in the mood for
music.

And you can have your
choice of any of these fine
extras. Your dealer‘will
make them available to
you at just a slight addi-
tional cost when you buy
a Panasonic radio.

SCENE #5:
(z 12)

 

SCENE #6:

 

(:11)

31

VIDEO

CUT TO MCU THREE
BOYS AND RADIO. ONE
POINTS OUT ANTENNA,
ANOTHER THE EAR-
PHONE, AND ANOTHER

THE FRONT DIAL LIGHT.

CUT TO CU SIMPLE
RADIO DISPLAY.

AUDIO

Don' t forget dances
either. Your new Pana-
sonic will be the envy of
the group. And why not?
It was designed to fit in
with what' 5 happening—-
whatever the occasion. . .
wherever you are.

So don' t be satisfied with
less than the best.
Always be sure you ask
for Panasonic. And don' t
delay: see your nearby
appliance dealer--today.

SCENE #1:

 

(:14)

SCENE #2:

 

(:12)

SCENE #3:

 

(:06)

SCENE #4:

 

(:12)

32

COMMERCIAL VERSION C

 

(1:07)

VIDEO

PAN WITHOUT VISUAL
HESITATION ACROSS
DISPLAY OF MUSICAL
ITEMS, HOLDING
FINALLY ON ECU
PANASONIC RADIO.

CUT TO MCU GIRL
BEHIND TABLE. SHE
LIFTS RADIO, POINTS
OUT SPEAKER, AND
TURNS ENTIRE RADIO
AROUND.

DOLLY IN FOR CU
GIRL' S HAND OPENING
BACK OF RADIO AND
POINTING TO BATTER-
IES, CORD, AND CORD
HOLDER.

CUT TO CU ACCESSO-
RIES DISPLAY. PAN
WITHOUT VISUAL
HESITATION.

AUDIO

(ANNCR OVER LIGHT MUSIC)

Look. These are the
makings of listening

enj oyment--a musical
experience that goes
where you go: it' s the
newly-developed, newly-
styled transistor radio
by Panasonic.

Priced at $49. 95, the
Panasonic weighs just
three pounds. . .yet, its
new speaker system gives
you the finest in sound
reproduction.

Four penlight batteries
are all you need--or use
any AC outlet. Here, the
cord stays neatly out of
sight.

And you can have your
choice of any of these fine
extras. Your dealer-will
make them available to
you at just a slight addi-
tional cost when you buy
a Panasonic radio.

SCENE #5:
(:12)

 

SCENE #6:
(:11)

33

VIDEO

CUT TO MCU THREE
BOYS AND RADIO. ONE
POINTS OUT ANTENNA,
ANOTHER THE EAR-
PHONE, AND ANOTHER

THE FRONT DIAL LIGHT.

CUT TO CU COMPLEX
RADIO DISPLAY.

AUDIO

The adjustable Panasonic
antenna always brings you
perfect reception. In
addition, each radio comes
with its own earphone. . .
and, a press of that front
button lights up the dial
for easy tuning.

Be sure you ask to see

all three sizes of Pana-
sonic radios, styled in
both plain and plaid carry-
ing cases. Remember

the name: Panasonic.

SCENE #1:

 

(:14)

SCENE #2 :

 

(:12)

SCENE #3:

(:06)

SCENE #4:

(:12)

34

COMMERCIAL VERSION D

 

(1:07)

VIDEO

PAN WITHOUT VISUAL
HESITATION ACROSS
DISPLAY OF MUSICAL
ITEMS, HOLDING
FINALLY ON ECU
PANASONIC RADIO.

CUT TO MCU GIRL
BEHIND TABLE. SHE
LIFTS RADIO, POINTS
OUT SPEAKER, AND
TURNS ENTIRE RADIO
AROUND.

DOLLY IN FOR CU
GIRL' S HAND OPENING
BACK OF RADIO AND
POINTING TO BATTER-
IES, CORD, AND CORD
HOLDER.

CUT TO CU ACCESSO-
RIES DISPLAY. PAN
WITHOUT VISUAL
HESITATION.

AUDIO

(ANNCR OVER LIGHT MUSIC)

Look. These are the
makings of listening
enjoyment—-a musical
experience that goes
where you go: it' s the
newly-developed, newly-
styled transistor radio
by Panasonic.

Priced at $49.95, the
Panasonic weighs just
three pounds. . .yet, its
new speaker system gives
you the finest in sound
reproduction.

Take yourPanasonic to
the beach with you--or
just about anywhere else.
It' 3 always there when
you' re in the mood for
music.

And you can have your
choice of any of these fine
extras. Your dealer will
make them available to
you at just a slight addi-
tional cost when you buy
a Panasonic radio.

SCENE #5:
(:12)

 

SCENE #6:
(:11)

 

35

VIDEO

CUT TO MCU THREE
BOYS AND RADIO. ONE
POINTS OUT ANTENNA,
ANOTHER THE EAR-
PHONE, AND ANOTHER

THE FRONT DIAL LIGHT.

CUT TO CU COMPLEX
RADIO DISPLAY.

AUDIO

Don' t forget dances
either. Your new Pana-
sonic will be the envy of
the group. And why not?
It was designed to fit in
with what' 3 happening--
whatever the occasion. . .
wherever you are.

So don' t be satisfied with
less than the best.

Always be sure you ask
for Panasonic. And don' t
delay: see your nearby
appliance dealer- —today.

36

Summary of Manipulations Across Commercials

COMMERCIAL VERSION A

 

SCENE #1:
SCENE #2:
SCENE #3:
SCENE #4:
SCENE #5:

SCENE #6:

Visual Hesitation
No Manipulation
Interlock

Visual Hesitation
Interlock
Simplicity 8:

Interlock

COMMERCIAL VERSION C

 

SCENE #1:
SCENE #2:
SCENE #3:

SCENE #4:

SCENE #5:

SCENE #6:

No Visual Hesitation
No Manipulation
Interlock

No Visual Hesitation
Interlock
Complexity &

Interlock

COMMERCIAL VERSION B

 

SCENE #1:
SCENE #2:
SCENE #3:
SCENE #4:
SCENE #5:

SCENE #6:

Visual Hesitation
No Manipulation
No Interlock
Visual Hesitation
No Interlock
Simplicity &

No Interlock

COMMERCIAL VERSION D

 

SCENE #1:
SCENE #2:
SCENE #3:
SCENE #4:
SCENE #5:

SCENE #6:

No Visual Hesitation
No Manipulation

No Interlock

No Visual Hesitation
No Interlock
Complexity 8:

No Interlock

37

SCENE #1 (Video Manipulation):
Visual Hesitation/No Visual Hesitation
The opening scene contained the first visual manipulation.
The camera panned across five musical items (a piano, some
records, a drum, a horn, and a banjo) which respondents were
asked to recall at two different times: immediately after exposure
to the message, and one to four weeks later. The first tests of H

1
and H2 were thereby possible.

In two versions of the commercial, the items were spaced
so far apart that a completely blank screen appeared briefly between
each pair of objects. Subjects were, therefore, given the oppor-
tunity to fixate on some portion of the screen during ”visual hesita-

tions,‘ and, hence, H was tested. A sixth item in the display was

3
the advertised product, a Panasonic radio. Respondents were not
asked to recall it (since it was shown throughout the commercial),
but it provided for an additional visual hesitation after the last
musical item. Initial eye fixations on a blank screen were, there-
fore, coded five different times in the first scene--immediately
preceding the records, the drum, the horn, the banjo, and the radio.
In the other two versions of the commercial, the objects

were spaced close enough together that a completely blank screen

never appeared between pairs. They were not so close to each other,

38

however, that more than one item could be seen at a time in its
entirety. In this case, there was "no visual hesitation” in panning,
and initial eye fixations on the five objects themselves were coded--
again, the records, the drum, the horn, the banjo, and the radio.
Audio in this scene was held constant across all four
versions of the commercial. The five musical items shown in the
video channel, however, were never mentioned in the audio and never
sgai_n appeared visually. Therefore, in the measurement of eye
fixations on these items and recall of them, there was no possible

way for audio or any later video to confound the findings.

SCENE #2 (No Manipulation)

Audio and video were held constant in the second scene
across all four versions. The on-camera talent and off-camera
announcer familiarized viewers with the Panasonic radio by indicat-
ing several of its distinguishing characteristics. No hypotheses were
tested in this regard, but the scene was necessary for transition
purposes between manipulation in the first and third scenes.

SCENE #3 (Audio Manipulation):
Interlock/No Interlock

The third scene contained the first aural manipulation.

While video was held constant--the talent opened the back of the radio

39

and continued pointing out special features—-in two versions of the
commercial, the audio included simultaneous mention of the items.
Specifically, when the talent pointed to batteries, the announcer
talked about batteries. When the talent showed viewers the cord and
then its holder inside the radio case, the announcer discussed them
at the same time. The audio and video thereby "interlocked, " so
that H4 and H5 could be tested.

In the remaining two versions of the commercial, the audio
contained information completely irrelevant to the visual presenta-
tion, in an attempt to determine whether a lack of audio—video inter-
lock would affect eye fixations on and recall of the three relevant
items: the batteries, the cord, and the cord holder. Hence, H6 was
also tested here.

The three relevant objects in this scene, whether visualized
3151 discussed by the announcer (the "interlock" situation), or merely
visualized while the audio was irrelevant (the ”no interlock" situation),
were neither mentioned verbally nor shown on the screen before or

after this exact point in the commercials. Again, therefore, findings

could not be confounded by subject sensitization.

SCENE #4 (Video Manipulation):
Visual Hesitation/No Visual Hesitation

The fourth scene Offered the opportunity for a second test

of the visual hesitation manipulation. This time the camera panned

40

across five accessory items (a pillow, a thermos, a flashlight, a
clock, and a case with a shoulder strap) which respondents were
again asked to recall on both an immediate and a delayed basis. As
before, two versions of the commercial contained visual hesitations
between objects, and two did not. In this case, initial eye fixations
on a blank screen (in the "Visual hesitation" situation) were mea-
sured four times: immediately preceding exposure to the thermos,
the flashlight, the clock, and the case with the shoulder strap. Like-
wise, initial eye fixations on an actual object (in the "no visual hesi—
tation" situation) were measured four times: on the thermos, the
flashlight, the clock, and the case with the shoulder strap.

Finally as was true in the first scene, audio was held con-
stant across all four versions of the commercial. The five items
shown in the video channel were never mentioned by the announcer
and never sgsis appeared visually, so there was no way for audio or
later video to confound looking behavior and recall findings.

SCENE #5 (Audio Manipulation):
Interlock/ No Interlock

The fifth scene afforded the chance to test the ”interlock"
manipulation a second time. While video was held constant--the
talent demonstrated uses of special radio features--in two versions

of the commercial, the audio included simultaneous mention of the

41

items. Specifically, when the talent showed viewers the radio
antenna, earphone, and dial light button, the announcer commented
on them. In the remaining two versions, the audio contained infor-
mation completely irrelevant to the visual presentation.

Once more, the three relevant objects in this scene,
whether visualized ins discussed by the announcer (the "interlock"
situation), or merely visualized while the audio was irrelevant (the
”no interlock” situation), were neither mentioned verbally nor
shown on the screen before or after this exact point in the commer-
cials. Findings were not, therefore, confounded by subject sensi-
tization.

SCENE #6 (Video and Audio Manipulations):
Simplicity/ Complexity and Interlock / No Interlock

The remaining hypotheses called for presentation of a
”simple" and a "complex" display of objects. Nine mock—ups of
Panasonic radios were used in the sixth scene: three small ones,
three medium—sized ones, and three large ones. All nine came with
carrying cases; two of each size were of solid color, and one of each
size was plaid. The three medium-sized ones had handles.

Terwilliger experimented with complex patterns, and used

nonsymmetrical displays to represent complexity. 24 In the current

 

2
4Robert F. Terwilliger, ”Pattern Complexity and Affective
Arousal, " Perceptual and Motor Skills, No. 17 (1963), pp. 387-395.

 

42

study, likewise, a symmetrical and nonsymmetrical display of the
radio mock-ups represented simplicity and complexity. Content
was held constant across all versions of the commercial; the same
nine items appeared in each instance. Format, however, (actual
arrangement of the mock-ups in a display), differed. Two versions
of the commercial contained the simple display, and two contained
the complex one. In the former case, the objects were arranged in
three rows, in pyramidal fashion: the three large mock—ups on the
bottom row, the three medium-sized ones on the middle row, and
the three small ones on the top row. In the latter instance, the
mock-ups were randomly arranged in three rows: four on the bottom
row (one large, one small, and two medium-sized), three on the
middle row (one small, one medium-sized, and one large), and two
on the top row (one large and one small).

On both an immediate and a delayed basis, respondents
were asked to rate the display they saw according to how "interesting"
and how "pleasing" they thought it was. Seven-point rating scales
provided the test instruments for H7, H8’ 9, and H10.

Finally, audio-video interlock was manipulated once more
in this last scene. Two versions of the commercial discussed the

three sizes of radios displayed, and the plain and plaid carrying

cases, while the other two did not. Subjects were given immediate

43

and delayed recall tests for the size, carrying case, and color

features.

Questionnaire

 

Recall questions dealt with the following items:

(a) the five musical objects in SCENE #1 (the piano, the
records, the drum, the horn, and the banjo);

(b) the three features pointed out in SCENE #3 (the batter-
ies, the cord, and the cord holder);

(c) the five accessory items in SCENE #4 (the pillow, the
thermos, the flashlight, the clock, and the case with the shoulder
strap);

(d) the three features demonstrated in SCENE #5 (the
antenna, the earphone, and the dial light button);

(e) the three characteristics of the radio mock—ups in
SCENE #6 (the three sizes, the accompanying carrying cases, and
the plain and plaid colors).

In addition to recall, the experimenter'was interested in
respondents' ratings of the display of radio mock-ups which appeared
in the sixth scene of each of the four versions of the commercial
(two contained a simple display and two contained a complex one).

Seven-point rating scales served as measuring instruments. Scores

44

obtained on "interesting/dull" and "pleasing/annoying" scales pro-

vided a comparison with Berlyne' s findings regarding "interesting”
and "pleasing" ratings given to simple and complex designs. Three
other scales were used in addition, and the position of polar adjec-

tives was reversed, in order to minimize halo effect.

Finally, two questions asked for recall of the weight and
cost of the Panasonic radio advertised. This information was given
in the audio during SCENE #2 (in all versions of the commercial),
but was never repeated ,or shown visually.

Identical questionnaires were administered to subjects
immediately following exposure to the televised message and again
one to four weeks later. Each took between five and ten minutes to

complete. A copy of the questionnaire may be found in the appendix.

Problems Encountered

 

The first problem which arose concerned the possibility of
Type G error; since only one subject could be run at a time, forty-
eight separate treatments had to be administered. To guard against
the operation of Type G error, the two experimenters painstakingly
rehearsed their instructions and actions in the eye camera labora-
tory, until observers on hand during the pretest agreed that succes-

sive presentations were roughly identical.

45

It was discovered, however, that experimenter fatigue set
in after approximately two and one-half hours; at that point, instruc-
tions were hastily given and experimental controls were weakened.
The experiment was scheduled, therefore, over a period of three
weeks, with a two and one-half hour limit set on continuous opera-
tions.

Because of the novelty of the viewing situation, subjects
were easily distracted. The frame housing the TV screen and
speaker was deliberately painted black, and a black curtain was
drawn halfway around the subject to help prevent temptations _to look
at areas away from the screen. In addition, all lights except the
marker light on the eye camera were turned off before the televised
message was presented. Finally, the audio track, though not so
loud as to be annoying, was of sufficient volume to drown out inter-
ference from sounds outside of the laboratory.

When a subject was required to remain more than a few
minutes in one position, eye strain made it impossible for him to
fixate steadily, and tear secretion interfered with the securing of a
well-defined spot of light. The short experimental message, how-
ever, and an experimenter who was well trained in eye camera
alignment procedures, permitted completion of each eye-movement

recording in a very short time. A subject was actually in an

46

"awkward position" (though no one deemed it uncomfortable),
situated on the bite bar with the marker light shining off of his left
eye, for a maximum of six minutes--and no more than four minutes
at any one time.

Another problem centered around the fact that the test
commercials were not produced with high-caliber equipment or
professional talent, and, hence, were not regarded as commercials
which might actually be seen on a home television set. Production
limitations were explained to subjects, however, prior to viewing,
and it was understood that the study did not call for evaluation of the
messages themselves. It was strongly felt, in addition, that pro-
duction quality was not an important factor in this experiment. The
appropriate stimulus materials were present, and manipulations
were carried out precisely; hence, looking behavior could be
examined as easily as if material produced by a national network
had been available.

Because of the nature of this experiment, it was not deemed
advisable to "force" unwilling subjects to participate; hence, only
volunteers were chosen. In addition, the participants had to be
scheduled according to their own class commitments and free hours.
To overcome the problem of lack of random selection of subjects,

therefore, the order in which the commercials were shown (and,

47

thus, the order in which experimental treatments were assigned) was
randomized.

Subjects had no orientation to the viewing and listening
situation prior to the showing of the experimental messages. Each
was, however, told specifically that a television commercial would
be shown, and that it would be complete with an announcer and back-
ground music. The idea of showing some irrelevant film before the
commercial in each case—-to help acclimate subjects to the picture-
plus-sound situation--had to be discarded for two reasons. First,
time was severely limited. Class schedules and experimenter
fatigue necessitated getting through each experiment in the shortest
possible time; also, subjects were prone to get very uncomfortable
if forced to remain in one position too long.

The second reason centered around finances. Costs pro-
hibited obtaining additional film from production houses, and none
produced by the university television station was recorded with
magnetic sound (as were the experimental commercials). Projector
adjustments would, therefore, have been necessary between the
showing of irrelevant film and the commercial in each instance, and
additional time and subject discomfort would have been involved. If
just one copy of a single piece of film were shown to all subjects, a

second projector would have been required (a near technical

48

impossibility when it came to positioning for proper rear screen
projection); otherwise, considerable time would have been needed
to locate the position of the randomly—selected experimental com-
mercial on the reel after the irrelevant film had been shown.

The noise of the projector was an additional sound to which
subjects were not accustomed. Leader film placed before the start
of each commercial, however, gave subjects at least a brief expo-
sure to the projector sound before the commercial actually appeared
on the screen.

It was believed that subjects would be much more concerned
with their "looking behavior” than they would be with their "listen-

' Since, however, none of them had ever even seen an

ing behavior. '
eye camera before, the novelty of the apparatus and conditions re-
.mained constant throughout. Differences which occurred in recall
of audio, therefore, could still be attributed to the experimental
manipulations .

Because the experiment had to be spread out over three
weeks, as discussed previously, it was possible for subjects to talk
to each other about the study before all of them had actually par-
ticipated. In an attempt to minimize biases resulting therefrom,
each subject was cautioned specifically against disclosing details of

the experiment. In addition, participants were selected from four

different classes, to help prevent extensive interaction.

49

Finally, again because of the duration of the experiment,
it was not possible to administer'the delayed recall test to all sub-
jects at exactly the same time interval from the date they saw the
experimental message. The tests were given over a three-day
period, however (again, to help minimize sensitization), and were
scheduled one to four weeks from the time that each subject had
viewed the commercial. For delayed recall, most retention studies
indicate that there is an initial period of rapid forgetting of material,
followed by a gradual "leveling off” after approximately seven days.
It was felt, therefore, that the time differential would not have a

significant effect on recall scores.

Operationalization of Looking Behavior Variables

 

Drift to Screen Center

The first and fourth scenes of the commercials were con-
sidered in the test of H3. Two versions contained ”visual hesitation"
between items panned by the camera, and two contained "no visual
hesitation. " In the "visual hesitation" commercials, a blank screen

appeared immediately preceding the records, the drum, the horn,

the banjo, and the radio, in the first scene, and immediately

 

25Carl 1. Hovland, Irving L. Janis, and Harold H. Kelley,

Communication and Persuasion (New Haven: Yale University Press,
1953), p. 245.

 

50

preceding the thermos, the flashlight, the clock, and the case with
the shoulder strap, in the fourth scene. In each of these situations,
the first fixation subjects made on this "blank screen" was coded.

In the "no visual hesitation" commercials, a new item
appeared immediately after one item had vanished from sight. In
this case, the fixation relevant to H3 became the first one which
occurred on the "new item" (again, the records, the drum, the horn,
the banjo, and the radio, in the first scene, and the thermos, the
flashlight, the clock, and the case with the shoulder strap, in the
fourth scene).

In all instances, concern focused on the distance each of
these first fixations was from center. The film coder used a ruler

to measure the length of the vector extending from the center of the

screen to the eye marker.

Fixation Duration

Nixon has suggested, as have Luborsky, Blinder, and
Mackworth, that given two visual situations, the one eliciting the
longer period of visual fixation may be considered to possess the great-

er attention value. 26 In the current study, H6 called for measurement

 

26Nixon, "Attention Value and Interest in Advertising" ;

Lester Luborsky, Barton Blinder, and Norman Mackworth, "Eye-
Fixation and the Contents of Recall and Images as a Function of
Heart Rate, " Perceptual and Motor Skills, XVIII (1964), 421-436.

 

51

of the amount of time subjects spent looking at relevant items-~as
an indication of the attention value these items possessed in an
”interlock” and a "no interlock" situation. Both location and dura-
tion of eye fixations, therefore, were considered. The third scene
of the commercials contained three relevant items: the batteries,
the cord, and the cord holder. Likewise, the fifth scene contained
three: the antenna, the earphone, and the dial light button. The
film coder' 5 task, therefore, became determination of the length
of time subjects fixated on each of these six objects.

Karslake and Brandt both measured fixation duration by
assessing the distance that film in the eye camera traveled during
a fixation. In other words, they counted the number of frames which
subjects spent looking at a given area. 27 Similarly, in the present
experiment, the film coder counted the number of frames containing
fixations on the six relevant items.

A question arose concerning the number of fixations relevant
to each manipulation. Audio was manipulated in the third and fifth
scenes, while video remained constant. In each case, two versions

of the commercial contained aural mention of the items in question

 

27Karslake, "The Purdue Eye-Camera: A Practical Appa—
ratus for Studying the Attention Value of Advertisements"; Brandt,
"Ocular Patterns and Their Psychological Implications, " pp. 260—
268.

52

while two contained irrelevant information. In the "interlock"
commercials, the only fixations on these items which were relevant

to H6 were the ones which occurred during aural mention of the

 

items, while the talent simultaneously pointed to them.

In the "no interlock" commercials, the relevant fixations
were those which occurred at the same point as in the ”interlock"
commercials--while the talent pointed to the items--though the
audio made no mention of them. In the third scene, therefore, when
the talent pointed to the batteries, eye fixations on the batteries were
counted. When she pointed to the cord and cord holder, however,
fixations on the batteries were Ti): counted--only those on the cord
and holder were then relevant. Likewise, in the fifth scene, when
the talent demonstrated the antenna, eye fixations on the antenna
were counted; but when the demonstration moved to the earphone and

dial light button, only fixations on those items became relevant.

Scope of Looking Behavior

In order to test H11, the film coder had to count the number
of different items on which subjects fixated during the sixth scene of
the commercials. A display of nine radio mock-ups remained on the
screen for eleven seconds; in two versions of the commercial it was

a "simple” (symmetrical) display, and in two versions it was a

"complex" (nonsymmetrical) display. Each subject had the

53

opportunity to fixate on as many as nine different items during the
time the display was shown. No count was made of the actual num-
ber of fixations spent on each item, because such information was
not relevant to the hypothesis. Subjects who looked at eight or nine
items necessarily had to spend less time, on the average, on each
item than subjects who looked at only two or three; how much less
time was spent in each individual case was not deemed important in

this study.

Statistical Design

 

The statistical analysis focused on the independent and

interactive effects of three independent variables: visual hesitation,

 

audio-video interlock, and visual complexity, on two dependent

 

 

variables: looking behavior and recall. All subjects were randomly

 

assigned to treatments, and homogeneity of variance was confirmed

in each case. One—way analysis of variance was used to test H

 

 

1
through H11. The following is a summary of the tests which were
run:

Independent Variable Dependent Variable
H1: Visual Hesitation/ Number of relevant items
No Visual Hesitation in immediate recall:

 

SCENE #1 and SCENE #4.

Independent Variable

 

H : Visual Hesitation/
No Visual Hesitation

H : Visual Hesitation/
No Visual Hesitation

H4: Interlock/No Interlock
H5: Interlock/No Interlock
H6: Interlock/NO Interlock
H7: Simpli city/ Complexity
H8: Simplicity/ Complexity
H9: Simplicity/ Complexity
H 1 0: Simpli city/ Complexity
H1 1: Simplicity/ Complexity

In all instances involving the "visual hesitation" and "inter—

54

Dependent Variable

 

Number of relevant items
in delayed recall:
SCENE #1 and SCENE #4.

Distance of first fixations
from center:
SCENE #1 and SCENE #4.

Number of relevant items
in immediate recall:
SCENE #3 and SCENE #5.

 

Number of relevant items
in delayed recall:
SCENE #3 and SCENE #5.

Number of fixations on
relevant items:

SCENE #3 and SCENE #5.

Immediate "interesting"

 

rating: SCENE #6.

Immediate "pleasing"

 

rating: SCENE #6.

Delayed "interesting"

rating: SCENE #6.

Delayed "pleasing"

rating: SCENE #6.

Number of items fixated:
SCENE #6.

lock" manipulations, separate analyses were run for each of the

55

scenes concerned. The differing nature of the items in these cases

made independent analyses appropriate.
Two-way analysis of variance-was used to test H12 and H13,

both of which were concerned with the conjunctive effects of two of

the independent variables: audio-video interlock, and visual com—

 

 

plexity, in the sixth scene of the commercials. Here, the following

2 X 2 tables were used in the analysis:

 

 

 

 

 

 

 

 

 

Visual No Visual
Hesitation Hesitation
# OF ITEMS,
Audio-Video IMMEDIATE
Interlock RECALL:
SCENE #6
No Audio-Video
Interlock
Visual No Visual
Hesitation Hesitation
# OF ITEMS,
Audio-Video DELAYED
Interlock RECALL:
SCENE #6

 

No Audio-Video
Interlock

 

 

 

 

III. FINDINGS

Given a televised message, it was predicted that visual
pauses, redundancy in the audio-video relationship, and complexity
of visual patterns would affect viewers' looking behavior and recall.
Since previous studies had found that Mhesitations increased

recall, it was proposed that visual hesitations would do likewise.

 

Also, visual hesitation was operationalized as "blank spaces" on the
TV screen; since earlier research had found that subjects made
initial fixations on the 2.8.9.193 of a display when given no ”cues" as
to where they should look, it was hypothesized that the same would
be true in the current study (blank screens held no cues). I I

Research had indicated a positive relationship between
recall and audio—video redundancy; it was predicted, therefore, that
recall of relevant visualized items in the present study would be
higher when the audio and video tracks "interlocked" than when they
did not. , Looking behavior was expected to be differentially affected
by the two situations also.

Finally, earlier investigations had suggested that the length

of fixation on a pattern, and subjects' evaluations of it, were

56

57

influenced by the difficulty of information extraction from the pattern.
It was currently hypothesized, therefore, that subjects' fixation
time would be longer given a complex display than it would be given
a simple display, and that their recall and evaluations of the two
displays would differ.

Simple, randomized analysis of variance was used to test

two-way analysis of variance was used forvH and

H1 through H 12

11;

H13.

Reliability Check

 

The film coder coded eye fixations in five scenes in each
commercial, for purposes of analysis. Across the forty—eight
subjects, therefore, a total of 240 scenes were coded. A second
coder coded forty-eight scenes (20 per cent) for purposes of deter-
mining interjudge reliability. These forty—eight scenes were ran-
domly selected, but were chosen as evenly as possible across all
scenes (one scene per subject). Hence, SCENE #1 was coded by a
second judge in ten different commercials. SCENE #3 was coded
nine times; SCENE #4, nine times; SCENE #5, ten times; and
SCENE #6, ten times. (SCENE #2 was not coded at all since it con-
tained no manipulations relevant to the hypotheses.) Results of the

reliability check were as follows:

SCENE #1:

SCENE #3:

SCENE #4:

SCENE #5:

SCENE #6:

Number of Items for Which
Fixation Totals Agreed

58

Frequency of Agreement

 

5

4

3

out

out

out

out

out

out

out

out

out

out

Of5

of5

Of5

of3

of3

of4

of4

of3

of9

of9

 

10

Based on the above, the experimenter concluded that

reliability was good.

Visual He sitati on

 

Recall Under Conditions of Visual Hesitation

In a televised message, there will be greater
immediate recall of relevant visualized items when

 

the camera pans with visual hesitation than when it

pans without visual hesitation.

 

59

Twenty-four subjects viewed a commercial which included

visual hesitation in the first and fourth scenes; twenty-four viewed

 

one which included no visual hesitation in the same two scenes.

 

Immediately after exposure to the message, each of the forty-eight
subjects was asked to recall five visualized items from the first
scene and five from the fourth scene. The above hypothesis was,

therefore, tested twice. It was expected that the visual hesitation

 

group would remember more of the items in both scenes than would

the no visual hesitation group.

 

For the first scene, the average number of items recalled

by the visual hesitation group was 2. 17. For the no visual hesita-

 

 

tips group the average was 1. 88. An analysis of variance test, how-
ever, showed no significant difference between the two means; the
difference was so small that it could easily have occurred merely
because of sampling error. 28 H1, therefore, was rejected.

Though the difference between the means was small, it was

in the predicted direction; the visual hesitation group did remember

 

a slightly higher number of items than did the no visual hesitation

 

group.

 

28F = 0.90, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

60

For the fourth scene, the average number of items recalled

by the visual hesitation group was 1. 21, and forthe no visual hesi-

 

 

tation group, 1. 17. Analysis of variance again showed no significant
difference between the two means; again, the difference was small
enough to have occurred merely because of sampling error.

Hence, H was rejected here also.

1

As before, though, the small difference between the means

was in the predicted direction; the visual hesitation subjects again

 

remembered a slightly higher number of items than did the no visual

hesitation subjects.

 

H2: In a televised message, there will be greater
delayed recall of relevant visualized items when
the camera pans with visual hesitation than when it
pans without visual hesitation.
One to four weeks after they had seen the commercials, all
forty-eight subjects were again asked to recall the five items from

the first scene and the five from the fourth scene. Hence, H2 was

also tested twice. As before, it was expected that the visual hesi-

 

tation group would remember more of the items in both scenes than

would the no visual hesitation group.

 

 

29F = 0.02, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam—
ples.

61

For the first scene, the average number of items recalled

by the visual hesitation group was 1. 92, while for the no visual hesi-

 

 

t_ati_orl group it was 1. 50. Analysis of variance showed that the dif-
ference between the means was not significant, and, hence, could
have occurred because of sampling error. 30 H2 was, therefore,
rejected.

Again the small difference between means was in the

expected direction. The visual hesitation group recalled more

 

items, on the average, than did the no visual hesitation group.

 

Finally, an analysis of variance test was also run to
determine whether or not the change in scores between the imme-
diate and delayed recall tests differed between the visual hesitation

and the no visual hesitation groups. No difference was found. 31

 

 

For the fourth scene, the average number of items recalled

by the visual hesitation group was 0. 67. For the no visual hesitation

 

 

group it was 0.79. Analysis of variance revealed only a sampling

error difference between the two means. 32 Hence, H2 was also

rejected here.

 

30F = 1. 50, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two samples.

31Means: ~0.25 and -O. 38. F = 0.31, with degrees of free-
dom 1 and 46.

32F = 0.23, with degrees of freedom 1 and 46. An F Test
also confirmed the homogeneity of variances between the two samples.

62

This time the existing difference between means was not in

the expected direction. The visual hesitation group recalled fewer

 

items, on the average, than did the no visual hesitation group.

 

Again, an analysis of variance test was run to determine

whether or not the change scoreswere significantly different. As

before, no difference was found. 33

Looking Behavior Under Conditions of
Visual Hesitation

H : In a televised message, when the camera pans m
visual hesitation across relevant display items,
first eye fixations will be closer'to the center of the
frame when a "blank screen" appears than they‘will
be when the camera pans without visual hesitation
directly onto another item.

Twenty-four subjects who Viewed a commercial which

included visual hesitation in the first and fourth scenes had the

 

opportunity to fixate on a "blank screen" five different times in each

 

scene. The remaining twenty-four subjects, in the no visual hesi-
tation group, had the chance, instead, to fixate on five different
objects in each scene. H3 was thereby tested twice. It was expected

that the visual hesitation group would make their first fixations closer

 

to the center of the frame than would the no visual hesitation group.

 

 

33Means: -0.54 and —0. 38. F = 0.28, with degrees of free-
dom 1 and 46.

63

For the first scene, the average distance from center that

eye fixations appeared for the visual hesitation group was 3. 12 units

 

(each unit measured one-quarter inch). For the no visual hesitation

 

group it was 3. 28 units. Analysis of variance revealed that the dif-
ference between these two means could be attributed merely to

sampling error. 34 Hence, H was rejected.

3

Although the difference was small, however, it was in the

direction expected; the visual hesitation group (_i_is make their initial

 

eye fixations closer to the center of the frame than did the no visual

hesitation group.

 

For the fourth scene, the average distance from center

that first eye fixations appeared for the visual hesitation group was

 

3. 34 units, and for the no visual hesitation group, 3. 70 units.

 

Analysis of variance showed only a sampling error difference be-
tween the two means. 35 H3 was, therefore, also rejected here.

Again, although the difference was small, it was in the pre—

dicted direction; the visual hesitation subjects again made their first

 

 

34F = 0.32, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

35F = 1.42, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

64

eye fixations closer to the center of the frame than did the no visual

hesitation subjects.

 

Int erlock

Recall Under Conditions of Interlock

H : In a televised message, there will be greater
immediate recall of relevant visualized items when
there _i_s_ audio-video interlock than when there is
_ng audio-video interlock.

 

Twenty-four subjects viewed a commercial which included
an audio-video interlock treatment in the third and fifth scenes;

twenty-four viewed one which includedno interlock in the same two

 

scenes. Immediately after exposure to the message, each of the
forty—eight subjects was asked to recall three visualized features
from the third scene, and three from the fifth scene. In this way,

the above hypothesis was tested twice. It was expected that the HE‘S:
-l_o_c_k group would remember more of the features in both scenes than

would the no interlock group.

 

For the third scene, the average number of features recalled

by the interlock group was 1. 62. For the no interlock group the

 

average was 1. 08. An analysis of variance test revealed a signifi-

cant difference between the two means; in Other words, the difference

65

was too large to have been attributed merely to sampling error.
H4, therefore, was confirmed under these conditions.

For the fifth scene, the average number of features recalled

by the interlock groupwas l. 00 and for the no interlock group, 1. 17.

 

Analysis of variance here showed a difference between the two means
which was small enough to have been caused by sampling error. 37

Hence, this time, H had to be rejected.

4
The difference which did occur between the means was _IlO—li

in the direction predicted. The interlock group remembered slightly

fewer of the features than did the no interlock group.

 

H5: In a televised message, there will be greater
delayed recall of relevant visualized items when
there _i_s_ audio-video interlock than when there is
£12 audio-video interlock.

One to four weeks after they had seen the commercials, all
forty-eight subjects were again asked to recall the three features
from the third scene and the three from the fifth scene. Hence, H

5

was also tested twice. As before, it was expected that the interlock

 

36F = 4.32, with degrees of freedom 1 and 46. P < .05.

An F Test confirmed the homogeneity of variances between the two
samples.

37F = 0.56, with degrees of freedom 1 and 46. An F Test
also confirmed the homogeneity of variances between the two sam-
ples.

66

group would remember more of the features in both scenes than
would the no interlock group.

For the third scene, the average number of features recalled
by the interlock group was 1. 17, while for the no interlock group it
was 0.79. Analysis of variance showed that the difference between
the means could be attributed to sampling error. 38 H5 was, there—
fore, rejected.

The small difference between the means was in the expected
direction, however. The interlock group recalled more features,
on the average, than did the no interlock group.

An analysis of variance was run to detect recall changes
from "immediate” recall to ”delayed" recall, for the interlock and
no interlock groups. There were no significant differences. 39

For the fifth scene, the average number of features recalled
by the interlock and no interlock groups was 1. 00; there was virtually

no difference between the means. 40 H5 was also rejected here.

 

38F = 2.20, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

39Means: -0. 46 and ~0. 29. F = 0.50, with degrees of free-
dom 1 and 46.

40F = 0.00, with degrees of freedom 1 and 46. An F Test
also confirmed the homogeneity of variances between the two sam-
ples.

67

As before, an analysis of variance was run between group

change scores. Again, no significant findings occurred. 41

Looking Behavior Under Conditions of Interlock

H : The amount of time spent looking at relevant
visualized items will be greater when there i_s
audio—video interlock than when there is _rg
audio-video interlock.

Twenty-four subjects who viewed a commercial which

included interlock in the third and fifth scenes had the opportunity

to fixate on three items which were simultaneously mentioned in the

audio. The remaining twenty-four subjects, in the no interlock

 

group, had the chance to fixate on the same three items (in both
scenes) while the audio carried irrelevant material. In this way,
H6 was tested twice. It was expected that the interlock group would
spend more time looking at the relevant items than would the £12
interlock group.

For the third scene, the average amount of time spent by
the interlock group in looking at the three relevant items was 11. 67

frames (as recorded by the eye camera film), or 1. 46 seconds. The

no interlock group spent an average of 8. 42 frames, or 1. 05 seconds

 

 

41Means: 0.00 and -O.17. F = 0.88, with degrees of free-
dom 1 and 46.

68

in looking at the same three items. An analysis of variance test
revealed a difference between the means which was so small that it

could be due solely to sampling error. 42 Hence, H was rejected.

6
For the fifth scene, the average amount of time spent by

the interlock group in looking at the three relevant items was 7. 17

frames, or 0. 90 seconds. The no interlock group spent an average

 

of 5. 29 frames, or 0.66 seconds, in looking at the same three items.
An analysis of variance test here again revealed a difference between
the means which was so small that it could be due solely to sampling

error. 43 H6’ therefore, was also rejected in this case.

Visual Complexity

 

"Interesting" Ratings Under Conditions of
Visual Complexity

H : In a televised message, subjects will rate a complex
display of items more "interesting" than a simple
one, when tested immediately after presentation.

 

Twenty-four subjects saw a version of the experimental

message which included the complex display of radio mock-ups, and

 

42F = 1.08, with degrees of freedom 1 and 46 An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

43F = 0.46, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

69

twenty-four saw one which included the simple display. The rating
scale for the displays ran from 7 (very interesting) to 1 (very dull)
It was expected that the complex group would rate the display more
interesting than would the simple group.

On scales administered immediately after presentation of
the message, the mean score for-the complex group was 3.21, while
that for the simple group was 3. 17. Analysis of variance showed
that the difference between the two means was small enough to have

occurred because of sampling error. 44 Hence, H was rejected.

7
The existing difference, however, though extremely small,
was in the predicted direction. The complex group did rate the
display slightly more interesting than did the simple group.
H : In a televised message, subjects will rate a complex

9 display of items more ”interesting" than a simple
one, when tested one to four‘weeks after presentation.

 

One to four weeks after exposure to the message, each sub-
ject re-rated the display he saw on the seven-point "interesting/dull"
scale. All forty-eight subjects participated in the delayed test. It
was again expected that the complex group would rate the display

more interesting than would the simple group.

 

44F = 0.01, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

70

This time the mean score for the complex group was 3. 46;

for the simple group it was 2. 88. Analysis of variance showed no

significant difference between them, so that H9 was rejected.

Again, the small existing difference between the two means
was in the predicted direction. The complex group did rate the dis-
play slightly more interesting than did the simple group.

An analysis of variance run between change scores for both

groups on immediate and delayed tests produced nonsignificant find-

ings.

"Pleasing" Ratings Under Conditions of
Visual Complexity

H8: In a televised message, subjects will rate a simple
display of items more "pleasing" than a complex
one, when tested immediately after presentation.

Again, twenty-four subjects saw a version of the test com-
mercials which included the simple display of radio mock-ups, while
twenty-four saw one which included the complex display. This time

the rating scale for the displays ran from 7 (very pleasing) to 1 (very

\

45F = 1. 80, with degrees of freedom 1 and 46. An F Test
8-180 confirmed the homogeneity of variances between the two sam-

ples.

46Means: —0. 29 and 0.25. F = 2. 82, with degrees of free-
dom 1 and 46.

 

71

annoying). It was expected that the simple group would rate the
display more pleasing than would the complex group.

On scales administered immediately after presentation of
the message, the mean score for the simple group was 3. 83, while
that for the complex group was 3. 50. Analysis of variance showed
that the difference between the two was not significant; again, it was
small enough to have occurred because of sampling error. 47 Hence,
H8 was also rejected.

As before, however, the difference was in the expected
direction. The simple group did rate the display slightly more
pleasing than did the complex group.

H : In a televised message, subjects will rate a simple

10 display of items more "pleasing" than a complex
one, when tested one to four weeks after presentation.

 

One to four weeks after exposure to the message, each
subject was also asked to re-rate the display he saw on the seven-
point "pleasing/annoying" scale. It was again expected that the

simple group would rate the display more pleasing than would the

complex group.

 

47F = 0.62, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

72

In this case, the mean score for the twenty-four subjects
in the simple group was 3. 67, and for the complex group it was
3. 83. Analysis of variance revealed no significant difference be—
tween the means. 48 H10 was rejected.

This time, the existing difference was pp; in the direction
predicted; the simple group rated the display slightly kiss pleasing
than did the complex group.

The analysis of variance test run between change scores in
this case also produced nonsignificant findings. 49

Looking Behavior Under Conditions of
Visual Complexity

H11: In a televised message, subjects will fixate on more
different items given a simple display than they will
given a complex display.

All forty-eight subjects saw a display of nine radio mock-
ups. Twenty-four saw the mock-ups arranged in a simple display

and twenty-four saw them in a complex display. In each case,

though, subjects had the opportunity to fixate on nine different items

 

48F = 0.16, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.
49 .
Means: -0. 17 and 0.33. F = 1.95, With degrees of free-
dom 1 and 46.

73

during the eleven seconds the display was exposed. It was expected

that the simple group would fixate on more different items, on the

 

average, than would the complex group.

The average number of fixated items for the simple group
was 3. 42, and for the complex group, 2.62. Analysis of variance
showed no significant difference between the two means. 50 H11 was,
therefore, rejected.

The difference which occurred, however, was in the

expected direction. The simple group did fixate on a slightly higher

number of items than did the complex group.

Visual Complexity and Interlock

 

Recall Under Conditions of Visual Complexity
and Interlock

H : In a televised message, immediate recall of relevant
12 . . . . . .
v1suallzed 1tems W111 be lower g1ven a complex display
than it will be given a simple display, but not as much
lower when there _is audio-video interlock as when
there is _rg audio-video interlock.

 

 

Twelve subjects saw a commercial which inc-luded the com-

plex display of radio mock-ups and the audio-video interlock

 

50F = 1.04, with degrees of freedom 1 and 46. An F Test

also confirmed the homogeneity of variances between the two sam-
ples.

74

treatment. Twelve saw the complex display with no interlock,

 

twelve saw the simple‘display with interlock, and twelve-saw the

simple display with no interlock. Immediately after exposure to

 

the commercial, each of the forty-eight subjects was asked to recall
three visualized features of the mock-ups (the same three appeared
in all versions of the commercial). It was expected that the highest

recall would occur in the simple/ interlock group, the second-highest

 

in the simple / no interlock group, the third—highest in the complex/

 

interlock group, and the lowest in the complex/ no interlock group.

 

The average number of features recalled by the simple/

interlock group was 1. 17. For the simple/no interlock group it was

 

0. 58. The complex/interlock group recalled an average of 1. 17

 

features, and the complex/no interlock group, 1. 08. (See Table 1.)

 

TABLE 1

MEANS FOR IMMEDIATE RECALL
UNDER CONDITIONS OF
FORMAT COMPLEXITY AND INTERLOCK

 

 

Simple Complex

 

Interlock 1. 17 1. 17

No Interlock 0.58 1.08

 

 

 

75

Analysis of variance showed that neither the main effects nor the
interaction was significant; in otherwords, the differences between
the four means were small enough to have occurred because of

. 51 .
samphng error. H12 was rejected.

The highest degree of recall occurred in the simple/inter-

 

lock and complex/ interlock groups. The next highest occurred in

 

 

the complex/ no interlock group, and the lowest in the simple/ no

 

 

interlock group. Overall, therefore, the findings were E: in the

predicted dire ctions.

H13: In a televised message, delayed recall of relevant
visualized items will be lower given a complex display
than it will be given a simple display, but not as much
lower when there i_s_ audio-video interlock as when
there is pp audio-video interlock.

 

One to four weeks after exposure to the commercials, all
forty-eight subjects were asked to recall the same three visualized
features as they had immediately after viewing. The four groups of
twelve subjects each, therefore, remained the same. It was again

expected that the highest amount of recall would occur in the simple/

 

interlock group, the second-highest in the simple/no interlock group,

 

51F = 0.94, with degrees of freedom 1 and 44. An F

Test also confirmed the homogeneity of variances between the
two samples.

76

the third-highest in the complex/ interlock group, and the lowest in

 

the complex/ no interlock group.

 

The average number of features recalled by the simple/

interlock group was 0.58. For the simple/no interlock group it was

 

0.33. The complex/interlock group recalled an average of 0.58

 

features, and the complex/no interlock group, 0.67. (See Table 2.)

 

TABLE 2

MEANS FOR DELAYED RECALL
UNDER CONDITIONS OF
FORMAT COMPLEXITY AND INTERLOCK

 

 

 

Simple Complex
Interlock 0.58 0. 58
No Interlock 0. 33 0. 67

 

 

 

Analysis of variance showed that neither the main effects nor the
interaction was significant; once more, the differences between the
four means were small enough to have occurred because of sampling

error.52 Hence, H was also rejected.

13

 

52F = 0.63, with degrees of freedom 1 and 44. An F

Test also confirmed the homogeneity of variances between the
two samples.

max

77

The highest degree of recall occurred in the complex/no

 

interlock group. The second-highest occurred in the simple/inter-

 

lock and complex/ interlock groups, and the lowest in the simple/

 

no interlock group. None of the findings, therefore, were in the

 

predicted directions.
An analysis of variance was again run between change
scores for the four groups, and no significant differences were

found. 53

Summary

 

Of the nineteen analysis of variance tests which were run,
though only one showed a significant difference between means,
thirteen of them produced results in the predicted direction. It
appears, therefore, that the manipulations in this study did have

some effect on subjects.

 

53Means: -0.58, -0.25, -0.58, and —0.42. F = 0.12, with

degrees of freedom 1 and 44.

IV. CONCLUSIONS

The current study examined various attention factors in
televised messages. Findings indicate that none of the manipulated
variables had much effect on viewers' looking behavior or recall.
Visual hesitations, operationalized as blank spaces between items
appearing on the screen, had no influence onthe number of items
viewers could remember or on the location of their eye fixations on
the screen. Likewise, audio—video interlock (redundancy in the
audio-video relationship) did not affect either recall of relevant
visualized items or eye fixations on them. Finally, visual com-
plexity (in the format of displays) had no effect on retention of dis-
play features, fixation time on the items, or'interest and pleasant-

ness evaluations of the displays.

Conclusions Based on Findirfis

 

Recall Under Conditions of Visual Hesitation

Recall was measured first in situations where the TV
camera panned across a series of objects in two different ways. It

either revealed all of the items in immediate succession, or exposed

78

79

a blank screen before each one. Hypothesis 1, suggesting that the
choice of panning technique would influence immediate recall of the
items, was tested twice and not supported. The first scene of the
commercial contained five memory items: a piano, some records,

a drum, a horn, and a banjo; the fourth scene also contained five:

a pillow, a thermos, a flashlight, a clock, and a case with a shoulder
strap. Regardless of whether the camera showed viewers every
item in succession or whether it revealed a blank space between

each pair, immediate recall of the objects involved in both scenes

 

was the same.

Hypothesis 2, claiming that delayed recall would be simi—
larly affected, also failed to receive support in the present study.
Even after a period of one to four weeks had elapsed, the difference
in retention between the two groups of viewers was negligible. Hence,

the experimenter concluded that visual hesitation had no effect on

 

recall.

Looking Behavior Under Conditions of
Visual Hesitation

One of the looking behavior variables examined in the cur-
rent study was Operationalized as "drift to screen center. " Hypoth-
esis 3, suggesting that TV camera movement across the aforemen-

tioned ”blank spaces" would have an effect on the tendency of viewers'

80

eyes to move toward the center of the screen, was also tested twice
and not supported. Again, SCENE #1 of the commercial included
blank spaces preceding exposure of the records, the drum, the horn,
the banjo, and the radio. In SCENE #4, the blank-spaces preceded
the thermos, the flashlight, the clock, and the case with the shoulder
strap. Initial fixations on blank spaces preceding items were not
significantly closer to the center of the frame in either scene than
were initial fixations on the items themselves. Here the experi-

menter concluded that visual hesitation did not affect drift to screen

 

center.

Recall Under Conditions of Interlock

Recall was also measured in situations where the off-camera
announcer either made mention of the visualized memory items or
merely presented irrelevant information. Hypothesis 4, maintain-

ing that immediate retention of the objects in question would be

 

affected by such aural conditions, was supported in one case and not
supported in a second. For the third scene of the commercial,
viewers receiving the message containing aural support for the
visualized items (the batteries, the cord, and the cord holder)
recalled more of those items on an immediate basis than did viewers

seeing the same objects while the announcer delivered irrelevant

81

material. For the fifth scene, where items were the antenna, the
earphone, and the dial light button, immediate recall for both sets
of viewers was the same.

Hypothesis 5, holding that delayed retention would, likewise,
be influenced, was not supported in either case. After a one- to
four-week time lapse, retention of items appearing in both scenes
was the same for both viewing groups. It was thus concluded that
audio-video interlock may affect recall in some cases but not in

others.

Looking Behavior Under Conditions of Interlock

A second looking behavior variable was operationalized as
”fixation duration. " Hypothesis 6, which proposed that an announc-
er' s mention of visualized items would affect the length of time
viewers looked at the relevant objects, was again tested twice and
not supported. As before, the items concerned were those in the
third scene (the batteries, the cord, and the cord holder), and those
in the fifth scene (the antenna, the earphone, and the dial light but-
ton). Fixations on the items in question lasted just as long when the
dialog was directly relevant as when it was completely irrelevant.
Hence, the experimenter concluded that audio-video interlock did

not influence the length of fixation time.

82

"Interesting" Ratings Under Conditions of
Visual Complexity

Given one of two different visual arrangements of nine radio
mock-ups, viewers were asked to rate the display according to how
interesting it was to them. Hypothesis 7, claiming that the complex

(nonsymmetrical) format would receive immediate ratings of higher

 

interest, was not supported. On a seven-point ”interesting/dull"
scale, viewers gave the same interest ratings to the display, whether
they saw the complex one or the simple (symmetrical) one.
Hypothesis 9, which maintained that the effect would hold
over time, was, likewise, not supported. Time had no effect on the
comparison of ratings across the two groups. In this case, the con-

clusion was that visual complexity had no influence on viewers'

 

interest ratings.

”Pleasing" Ratings Under Conditions of
Visual Complexity

Viewers were also asked to rate the radio mock-ups display
they saw according to how pleasing it was to them. Hypothesis 8,
which suggested that the simple (symmetrical) arrangement would
receive more pleasing ratings on an immediate basis, was not sup-
ported. On a seven-point "pleasing/annoying" scale, viewers gave
the same pleasantness ratings to the display, whether they saw the

simple one or the complex (nonsymmetrical) one.

83

Similarly, Hypothesis 10, proposing that this same effect
would hold over time, was not supported. Time had no effect on the
comparison of ratings across the two groups. Again, the experi-

menter had to conclude that visual complexity had no effect on

 

viewers' pleasantness ratings.

Looking Behavior Under Conditions of
Visual Complexity

The last variable concerned with eye movement was opera-
tionalized as "scope of looking behavior. " Hypothesis 11, contending
that a difference in format complexity of the display of radio mock-
ups would affect the number of different mock-ups on which viewers
chose to fixate, was not supported. Viewers presented with a %
ples (nonsymmetrical) arrangement looked at just as many of the
nine mock—ups as did viewers presented with a simple (symmetrical)

arrangement Of the same objects. In conclusion, visual complexity

 

did not influence the scope of looking behavior.

Recall Under Conditions of Visual Complexity
and Interlock

The last assessment of retention came, again, in the sit-
uation involving differing visual arrangements of the nine radio

mock-ups, coupled with the varied aural conditions mentioned pre-

 

viously (relevant and irrelevant information in the audio track).

84

Hypothesis 12, which suggested that immediate recall of features

 

of the display would be influenced both by the complexity of the
visual arrangement and by the relevancy of the aural presentation,
was not supported. Even though two manipulations were allowed to
interact, viewers' immediate recall of the features in question was
not affected. Retention was the same whether they saw the simple

display supported by the announcer, the simple display not supported

 

 

by the announcer, the complex display w_i’t_h aural support, or the
complex display without aural support.

Hypothesis 13, proposing that delayed recall would be
similarly affected, also failed to receive support. One to four weeks
later, the difference in recall between the four groups was, likewise,

negligible. The interaction of visual complexity and interlock did

 

not, hence, affect recall.

Summary of Findirgs

 

Nineteen separate analyses were made on the data obtained
in this study; there were thirteen different hypotheses and six of
them were tested twice. One test produced significant results; how-
ever, given a .05 level of significance, one test in twenty may be
expected to be significant simply by chance.

Thirteen of the nineteen tests did show results in the pre-

dicted direction. It seems, therefore, that the phenomena studied

85

did have some effect on the subjects.

In addition, if greater precision had been built into the
current experiment, greater differences would probably have
occurred between treatments. Two of the manipulated scenes in
the commercials yielded recall and looking behavior data on only
five items apiece. The other three manipulated scenes produced
recall data on three items apiece andlooking behavior information
on three items in two cases and nine items in the third. Replica-
tions of this experiment might contain longer messages with a larger

number of items included for the study of retention and eye-fixations.

Discussion

 

Visual Hesitation: Effects on Recall

Contrary to the hypotheses, visual hesitation had no effect
on recall. Jersild and Ehrensberger both found that pausing before
a verbal statement had definite retentive value. The ”rehearsal"
which apparently took place in subjects' minds during the delay
interval proved even more effective for recall purposes than an

immediate repetition of the statement. 54 It may well be, however,

 

4Jersild, "Primacy, Recency, Frequency, and Vividness,"
pp. 58-70; Ehrensberger, "An Experimental Study of the Relative
Effectiveness of Certain Forms of Emphasis in Public Speaking, "
pp. 94-111.

86

that verbal hesitation has an entirely different effect than does visual

 

hesitation. Perhaps a much longer pause is needed in the visual
channel before viewers can mentally rehearse what they have seen.
Also, in the above studies, subjects were forewarned by the experi—
menter that a recall test would follow presentation of the verbal
message. In the present experiment, no forewarning was given.

Subjects knew nothing about the recall test in this case until after

 

they had viewed the experimental message. Quite possibly, there-
fore, effects of the visual hesitation manipulation were not given the
same chance to operate here as were effects of the verbal hesitation
manipulation above.

It is interesting to note that in both the immediate and
delayed recall situations, subjects in both treatment groups recalled
approximately one whole item more from the first scene of the com-
mercial than they did from the fourth scene. In the first scene, the
objects included were all musical items: a piano, some records, a
drum, a horn, and a banjo. In the fourth scene, however, the
objects did not belong to any particular class: a pillow, a thermos,
a flashlight, a clock, and a case with a shoulder strap. Perhaps
the actual nature of the objects affected retention of them as well as

looking behavior.

87

Visual Hesitation: Effects on Looking Behavior

Also contrary to the situation hypothesized, visual hesita-
tion had no effect on the tendency of viewers' fixations to drift
toward the center of the screen.

Thomas' work indicated that, given no cues as to where
they should look, subjects would tend to fixate initially on the center
of the scene before them. The stimulus material in this case, how-
ever, was a set of Rorschach Ink Blots; Thomas did not present
subjects with any ”blank" cards. 55 In the current study, given a
completely blank scene in between scenes containing actual objects,
subjects' eyes tended to drift to the area of the screen in which they
expected the next item would eventually appear. . In other words,

they anticipated the appearance of a new item. While they were pre-

 

sented with no visual cues as to where they should look, they at least

 

had a mental idea of where such cues would next appear; hence, their
eyes did not tend to drift toward the center of the screen.

Subjects who saw one item after another in immediate suc-
cession had their anticipation satisfied almost instantaneously.
When a new object came into View, they merely transferred their

gaze to it--thereby also tending to fixate away from the center of the

 

55Thomas, "Eye Movements and Fixations During Initial

Viewing of Rorschach Cards, " pp. 345-353.

 

88

screen. The finding that both groups of subjects made initial scene
fixations which were between three and four units from center (each
unit measured one—quarter inch) is, therefore, understandable.

The actual items over which the camera panned may have
had some effect on fixation location. As pointed out previously, the
series of items in the first scene was somewhat homogeneous
(musical items), while the series in the fourth scene was more
heterogeneous. Given the former set of items, (all subjects (regard-
less of which treatment group they were in) tended to make initial
scene fixations closer to the center of the screen than they did given
the latter set of items (overall averages were 3. 20 units away from
center and 3. 52 units away, respectively). Perhaps presentation
of the more heterogeneous array led to a greater sense of anticipa-
tion (and curiosity) on the part of viewers. If such were true, they
might well be expected to make initial scene fixations farther from
screen center (in anticipation of new items) than would be the case

given a homogeneous array.

Interlock: Effects on Recall

Although hypotheses proposed that audio-video interlock
would increase recall of visualized items, only weak support for

them was obtained. Baldwin received maximum recall from viewers

89

of a motion picture film clipwhen there was high redundancy in the
audio-video relationship. 56 In one instance in the current study, his
findings were supported (immediate recall of features in SCENE #3:
the batteries, the cord, and the cord holder). In the others (imme-
diate recall of features in SCENE #5: the antenna, the earphone, and
the dial light button, and delayed recall of features in both SCENE #3
and SCENE #5), they were not. It is very probable, however, that
the actual subject matter involved and the context within which it
appears may have a good deal of influence on retentive value. Per-

haps viewers' familiarity with the items in question must be con-

 

sidered.

In the present experiment, it is quite possible that subjects
were more familiar with external features of the radio advertised in
the test commercials (the antenna, the earphone, and the dial light
button in SCENE #5) than they were with internal features (the bat-
teries, the cord, and the cord holder in SCENE #3). If that were
the case, in SCENE #3, aural support for the visualized internal
features may well have led to significantly more recall of the
features on the part of subjects in the interlock treatment group (as

opposed to subjects in the no interlock group). On the other hand,

 

 

56Baldwin, "Redundancy in Simultaneously Presented Audio—
Vidto Message Elements as a Determinant of Recall. "

90

if the external features presented in SCENE #5 were very familiar
to a_ll subjects, aural support for the visual presentation (or lack of
it) would not have affected recall.

Over time, the above difference did not hold. Because
there were only three memory items, however, even a slight rate
of forgetting on the part of subjects could easily have washed out the

effect obtained previously.

Interlock: Effects on Looking Behavior

It was also hypothesized that relevant audio (the audio-
video interlock treatment) in SCENE #3 and SCENE #5 would draw
attention to the items appearing visually; in the irrelevant audio

situation (the no interlock treatment), the objects could well have

 

gone unnoticed by subjects. Looking behavior was, hence, supposed
to have been differentially affected. In the current study, however,
it is reasonable to assume that subjects were geared much more
toward looking behavior than they were toward listening behavior.
Even though the experimenter explained that audio would be included
in the test commercials, it is not difficult to imagine that each sub-
ject was highly conscious of the light reflecting off his eye, and,
hence, of the experimenter' 3 interest in the _vislal aspects of the

message. The finding that all subjects were drawn to examine the

91

visual objects for roughly the same amount of time in each scene

is, therefore, quite conceivable.

The setting in which the "relevant visualized items”
appeared in the two scenes, however, could have affected the
amount of time subjects spent looking at them. The third scene
of the commercial included one female presenter (her hands only);
the fifth scene featured three male presenters (visible from the
waist up). A fair number of fixations occurred on talent' 8 hands,
faces, and clothing; it is suggested, therefore, that varying talent
situations should be explored with regard to effects on looking
behavior.

Visual Complexity: Effects on
"Interesting" Evaluations

Contrary to expectation, viewers of the complex (nonsym—
metrical) arrangement Of radio mock—ups did not rate the display
more interesting than did viewers of the simple (symmetrical)
arrangement of the same items. Berlyne found that a complex
pattern left subjects with uncertainty regarding its properties.
Interest ratings of such patterns were higher than were those of
simple patterns, because subjects wished to continue looking at

the complex designs, in hopes of obtaining additional information

92

about them. In other words, they were seeking closure. 57 In the
current study, however, the nine items comprising the simple and
complex displays were all easily identifiable as radio mock-ups--
set in the context of a commercial for a Panasonic radio. The dif-
ficulty of information extraction may be a function both of the format
and of the content of the display. Berlyne used meaningless designs,
while the current experiment used familiar items. It is quite pos-
sible, therefore, that the eleven-second duration of the displays
was long enough for subjects to obtain all of the information which
they felt necessary. Hence, interest ratings made by subjects see-
ing the complex display would not have been necessarily higher than
those made by subjects seeing the simple display.

Visual Complexity: Effects on
"Pleasing" Evaluations

Also contrary to prediction, viewers of the simple (sym-
metrical) arrangement of radio mock-ups did not rate the display
more pleasing than did viewers of the complex (nonsymmetrical)
arrangement of the same items. Berlyne found that a simple pat-
tern left subjects with a desire for entertainment or diversion.

They rated the designs ”pleasing" to look at, since there was no

 

7Berlyne, "Complexity and Incongruity Variables as
Determinants of Exploratory Choice and Evaluative Ratings, "
pp. 274-290.

93

uncertainty as to information contained; in this case, closure was
easily obtained. 58 Once more, however, if, as suggested above,
neither of the displays in the current study left viewers with uncer-
tainty, it would be expected that pleasantness ratings would be
similar across groups.

It was not believed that the discrepancy in display ratings
found between the present experiment and previous studies could
have been influenced either by the definition of a complex format or
by the test instrument employed. As was true in the current study,
both Berlyne and Terwilliger used nonsymmetri cal displays as
complex presentations, and both had subjects rate them on seven-
point scales. 59

One unfortunate condition in the final scene of the com—
mercial--the one containing the display of radio mock-ups--was
poor lighting. The set was lighted in such a way that it was only
with some difficulty that viewers were able to distinguish between
edges of the objects and the platforms on which they were resting.
It was not too easy, either, to distinguish between light and dark

colors. Overall ratings, therefore, may have been lowered for all

 

58mm.

 

593112. ; Terwilliger, "Pattern Complexity and Affective
Arousal, " pp. 387-395.

94

treatment groups on both the "interesting" and ”pleasing" scales,
so that differences between groups would have been very difficult

to obtain.
Visual Complexity: Effects on Looking Behavior

In spite of the hypothesis to the contrary, viewers of the
simple (symmetrical) display of nine radio mock-ups did not look
at more of the mock-ups during the eleven seconds the display was
exposed than did viewers of the complex (nonsymmetrical) display.

Previous research suggested that the length of fixation on
a display was a function of the difficulty of extracting information
from it. 60 In addition, complex figures were found to make infor-
mation retrieval more difficult than were simple figures. 1 In
past studies, however, stimuli consisted of materials not readily
familiar to subjects. Thomas used Rorschach Ink Blots, and

62
Berlyne and Lawrence used nonsense patterns. As pointed out

 

60
Thomas, "Eye Movements and Fixations During Initial

Viewing of Rorschach Cards, ” pp. 345-353.

1Berlyne and Lawrence, "Effects of Complexity and
Incongruity Variables on GSR, Investigatory Behavior, and
Verbally Expressed Preference, " pp. 21-45.

Thomas, ”Eye Movements and Fixations During Initial
Viewing of Rorschach Cards, " pp. 345-353; Berlyne and Lawrence,
Effects of Complexity and Incongruity Variables on GSR, Investiga-
tory Behavior, and Verbally Expressed Preference, " pp. 21-45.

95

earlier, in the current study, the nine items comprising the
displays were all easily identifiable as radio mock -ups--set in the
context of a commercial for a Panasonic radio. It is quite possible,
again, that the difficulty of information extraction is not a function
solely of the format of the display, but also of the content. Since
content was held constant across both simple and complex displays,
the finding that one group of subjects fixated on just as many dif-
ferent items as the other becomes readily acceptable. Time of
exposure of the display (eleven seconds in all cases) was not con-
sidered important in this regard, since Berlyne and Lawrence
found no differences in looking behavior between simple and com-
plex displays at exposure times ranging from ten seconds to two
minutes. 63

Visual Complexity and Interlock:
Effects on Recall

It was suggested earlier that subjects would fixate on more
different radio mock-ups given a simple arrangement of them than
they would given a complex arrangement. The rationale was based
on previous studies which found that the length of fixation on a dis-

play‘was a function of the difficulty of information extraction

 

3Berlyne and Lawrence, "Effects of Complexity and
Incongruity Variables on GSR, Investigatory Behavior, and Verbally
Expressed Preference, " pp. 21-45.

96

therefrom. It was assumed, therefore, that recall of features of
the objects would be greater in the case of the simple arrangement.
In addition, it was proposed that an aural presentation in
support of the visual scene would lead to greater recall than would
irrelevant audio. Since, however, the current study failed to sup-
port the first contention, and only weakly supported the second, it
is not surprising that the two in combination failed to produce sig-
nificant results. It must also be remembered that in this "combina-
tion" situation, the overall sample was split into _f_olr_ groups instead
of the t_vyc_> used in testing all of the other hypotheses. The total
number of subjects per group in this case was only twelve, making

significant findings extremely difficult to obtain.

Summa ry

Only one significant finding was made in the current experi-
ment, and even that was weakly supported. Thirteen out of the nine-
teen analyses made, however, did produce findings in the expected
directions. There is a good indication, therefore, that if the same
mean differences had occurred with a larger sample of subjects, more
significant findings would have been obtained. Also, as pointed out
earlier, an increase in precision would have been valuable, particu-
larly in the case of recall, where no more than five (and sometimes

fewer) memory items were considered in any one test.

97

Other possible reasons for the lack of significant looking
behavior findings must be explored, however. Most of the research
which has been conducted in the area of eye movements has been
concerned with still pictures. It may be that the variables examined
in this context cannot be lifted therefrom and studied directly in a
motion-picture framework. Movement of the camera across a scene
may give viewers an entirely different orientation than the one held
by viewers of a still picture. Eye fixations on the center of the
frame may be very rare (regardless of the content manipulation) if
the camera is moving either horizontally or vertically over the
screen. The inclination of viewers to follow this movement with
their eyes may make consideration of ”anticipation" tendencies much
more valuable than consideration of fixations on the exact center of
the screen. Then, too, different methods of moving the camera
might also be explored for their effects on eye movements.

Movement on the set might be another important considera-
tion in this area. If physical action by the performing talent attracted
viewers' attention, actual content manipulations within scenes might,
again, have little relevance to variations in looking behavior. Per—
haps ideal combinations of talent movement and camera movement
could be determined also--if situations were discovered in which a

certain amount of movement was distracting to viewers (to the point

98

where shifts in direction of eye movements became too numerous
for any "learning" to take place).

Looking at a different side of the motion—picture medium,
it may well be that different visual variables become important when
an audio track is present (regardless of its content) than when it is
absent (in the case of still pictures). Studies of the effects of audio-
video interlock on eye fixations on relevant objects may not yield
significant findings; but studies of effects of the audio-video combina-
tion on eye -movement patterns might be very beneficial. Perhaps
the number of shifts in eye fixations is influenced by the presence
of audio--or, the amount of eye regression over parts of the tele-
vised scenes might be affected.

One other consideration might be the size of the items on
the screen (regardless of what their properties are). Does size
affect the amount of time viewers spend looking at a given object-—
and if so, are there particular conditions under which use should be
made of extreme camera close-ups, as opposed to more inclusive

medium-shots? Again, the effects of movement on the screen

 

should be considered here.
The present study was, first, an attempt to relate existing
knowledge of looking behavior to the important but largely untested

field of televised messages. It was discovered that previous findings

99

in this area may not always be directly applied to the television
medium, where both audio and moving video are present. Different
considerations may be important where movement is involved--since
viewers may well have an entirely different orientation to the mes-
sage than they would have in the case of still pictures.

Second, the current study may be the door-opener to a new
method of measuring the effectiveness of television messages. It is
time for message analysts to back up a few steps-—from the evalua-
tion of television presentations as entities, to consideration of the
attention factors of each. It is the latterwhich are responsible in
some way for inducing viewers to watch the messages in the first
place--before they ever have a chance to "remember" or "believe"
or become interested in accepting them as whole messages. Three

attention factors were examined in the present experiment: visual

 

hesitation, audio-video interlock, and visual complexity, and their

 

 

effects on looking behavior and recall assessed. It is strongly rec-
ommended that the same variables be studied with a larger sample
of subjects, and also that the other suggested manipulations be
employed in future experiments.

The eye camera apparatus has just begun to be used in the
area of televised messages. It may provide a useful tool to com-

munication research as more and more factors are isolated for

100

consideration of their effects on learning and evaluation of the scenes
presented. In addition, other criterion variables may be explored.

How does lookingbehavior affect memorability, believability, and

 

willingness to accept ideas advocated (or, in the case of commercials,
products advertised)? Criterion measures which have been used with
varying degrees of success in field studies may also be applied to

laboratory investigations utilizing instruments such as the eye camera.

Third, a study like the present one gives advertising

research a chance to test some of the ”accepted principles" passed
from cOpywriter to copywriter and from textbook to textbook with-
out any scientific support. One such principle: "Keep it simple, "

was examined in the current experiment with the Visual complexity

 

manipulation. Another: ”Say what you show and show what you say, "

was explored under conditions of audio-video interlock. A number

 

of additional maxims may be worthy of future study: "Put one major
selling point in each commercial"; "Keep the product, and hands

and face of talent prominent in commercials"; "Use personal

 

language—-the 'YOU ' approach. "

Suggsstions for Future Research

 

Eye-Movement and Scene Movement

In the current study, regardless of the manipulations in a

given scene, the experimenter noticed a definite tendency on the

101

part of subjects to follow the action occurring on the screen. Eye

 

movements seemed to follow the movements of either the TV camera
across a set or the hands and faces of talent within it. Future ex-
periments might include various kinds of movement simultaneously
in a given scene to determine which attracted the attention of view-

ers.

Eye-Movement and Talent vs. Objects

Some indication was noted in the present study that the
presence of talent (male and female) and the visualization of hands,
as opposed to faces, or hands 3E9 faces, might affect the attention
viewers gave to memory objects. Other studies could well focus on
looking behavior in scenes emphasizing differing combinations of
talent and items pointed out or demonstrated by the talent. In addi-
tion, familiar and unfamiliar talent (celebrities and "unknowns"),
as well as familiar and unfamiliar objects (well-known items and
new products) might be included to test the effects of familiarity on

both looking behavior and recall.

Eye —Movement Patterns

 

Future experiments might also examine consecutive eye

fixations as opposed to shifting movements on arrangements involving

102

more than one item. If the frequency of each were studied across
selected scenes, both the attention-drawing and the attention-holding
powers of various objects could be assessed. It would also be

interesting to note whether or not repetition of certain scenes had

 

any effect on eye ~movement patterns.

Looking Behavior and Recall

Luborsky, Blinder, and Mackworth suggested that recall

 

of items is not necessarily related to the amount of time a subject
64
spends looking at them. Similarly, in the current study, corre-
lations between retention scores and looking behavior were very
65

low. Further support is needed in this area, however, especially

when consideration is given to the other variables mentioned here.

Recall from Audio or Video Only

Experimental messages in the current study included two
memory items which were mentioned only once by the announcer and
never appeared visually on the screen: the weight and the cost of

the Panasonic radio advertised in the commercials. Recall of both

 

6(J‘Luborsky, Blinder, and Mackworth, "Eye-Fixation and
the Contents of Recall and Images as a Function of Heart Rate, "
pp. 421-436.

65 Twenty-four point-biserial correlations ranged from
-. 36 to +. 32 with no appreciable differences between treatment
groups.

103

points, across all subjects, was extremely low. On an immediate
basis, only six subjects out of forty-eight remembered the weight
of the radio; on a delayed basis, seven recalled weight. Looking at
the cost feature, on an immediate basis, only three subjects out of
forty—eight could remember it; delayed recall in this case found six
who remembered cost.

In addition, the experimental messages contained one
memory item which was visualized only onceand never mentioned
by the announcer: the fact that the three medium-sized radio mock-
ups (in the display appearing in the last scene of the commercials)
had handles. Only five subjects out of forty-eight recalled the
handles on an immediate basis, and only one remembered them
when the delayed recall test was given. Much more experimentation
is needed in this regard to determine the conditions under which
”interlock" (as used in the current study) may be required for pur-
poses of achieving any substantial degree of retention. It would
also be possible to test the relative effectiveness of the aural and

visual channels when used in varying contexts.

Primacy/ Re cency

Finally, while primacy/ recency effects have been studied

both in printed and aural situations for decades, little, if any,

104

research has been conducted in this regard using television' 3 visual
channel. In the present experiment, subjects saw two different

series of items (in two different scenes of the commercial), as the

TV camera panned across them—-either in immediate succession or
with blank spaces between each pair. Interestingly enough, while
there were no differences in recall of items for the group seeing
blank spaces between items, there was a definite recency effect for
the group seeing all items in succession. Considerably more subjects
in this group recalled the lss_t item shown in both scenes than recalled
any of the other four items in either scene. Again, further research

is needed in this area before any definite conclusions can be drawn.

 

LI TE RATURE CI TED

 

 

LI TE RATURE CITED

Books

Brandt, Herman F. The Psychology of Seeing. New York: Philo-
sophical Library, 1945.

 

Deutsch, Morton, andKrauss, Robert M. Theories in Social Psy-
chology. New York: Basic Books, Inc., 1965.

 

Hovland, Carl I. , Janis, Irving L. , and Kelley, ‘Harold H. Com—
munication and Persuasion. New Haven: Yale University
Press, 1953.

 

 

Articles and Periodicals

 

Berlyne, D. E. "Complexity and Incongruity Variables as Deter—
minants of Exploratory Choice and Evaluative Ratings. "
Canadian Journal of Psychology, XVII (1963), 274—2 90.

 

Berlyne, D. E. "Curiosity and Exploration." Science, CLIII
(July 1, 1966), 25-33.

Berlyne, D. E. , and Lawrence, George H. ”Effects of Complexity
and Incongruity Variables on GSR, Investigatory Behavior,
and Verbally Expressed Preference. " Journal of General
Psychology, LXXI (1964), 21—45.

Brandt, Herman F. "Ocular Patterns and Their Psychological
Implications. " American Journal of Psychology, L111
(1940), 260—268.

 

Ehrensberger, Ray. "An Experimental Study of the Relative
Effectiveness of Certain Forms of Emphasis in Public

Speaking." Speech Monographs, XII (1945), 94-111.

105

 

106

Guilford, J. P. , and Hackman, H. B. "A Study of the Visual Fixa-
tion Method of Measuring Attention Value. " Journal of

Applied Psychology, XX (1936), 44-59.

Jersild, Arthur. "Primacy, Recency, Frequency, and Vividness.”
Journal of Experimental Psychology, XII (1929), 58-70.

 

Levy, J. M. "Experiments on Attention and Memory, with Special
Reference to the Psychology of Advertising. " University of
California Publications in Psychology, II, No. 2 (April 12,
1916), 157—197. ’

 

Mackworth, N. H., Kaplan, I. T., and Metlay, W. "Eye Movements
During Vigilance. ” Perceptual and Motor Skills, No. 18
(1964), 397-402.

 

Nixon, H. K. "Attention Value and Interest in Advertising. ” Archives
of Psychology, No. 72 (1924).

Sperling, George. "The Information Available in Brief Visual Presenta-
tions, " Psychological Monographs, LXXIV, No. 11 (1960).

 

Television Factbook. 1968—69 edition, No. 38. Washington:
Television Digest, Inc., p. 81—a.

 

Terwilliger, Robert F. "Pattern Complexity and Affective Arousal. ”
Perceptual and Motor Skills, No. 17 (1963), pp. 387—395.

 

Thomas, E. Llewellyn. "Eye Movements and Fixations During
Initial Viewing Of Rorschach Cards. " Journal of Projective
Techniques and Personality Assessment, XXVII (1963),
345-353. '

 

 

Woodrow, Herbert. "The Measurement of Attention. " The Psycho-
logical Monographs, XVII, No. 5 (December, 1914).

Reports

"Television Home Viewing Gains for 13 of 14 Months, TVB Analysis
of A. C. Nielsen Data Reveals, " TVB Report, V—67(68).
New York.

 

107

Unpublished Material

 

Baldwin, Thomas Frederick. ”Redundancy in Simultaneously Pre-
sented Audio—Video Message Elements as a Determinant of
Recall. " Unpublished Ph.D. dissertation, Michigan State
University, 1966.

Karslake, James Spier. ”The Purdue Eye—Camera: A Practical
Apparatus for Studying the Attention Value of Advertisements.
Unpublished Ph.D. dissertation, Purdue University, 1939.

Television Bureau of Advertising. "Why We Use Television. ”
Paper from TVB, 1968. (Mimeographed.)

Wolf, Willavene, and Others. ”An Experimental System for Eye-
Movement Studies in Dynamic Bi—Dimensional Fields, "
64-157. Paper from the Bureau of Educational Research
and Service, The Ohio State University.

 

Wolf, Willavene, and Others. "An Eye -Movement Study of Children
Viewing Television," 64-157. Paper from the Bureau of
Educational Research and Service, The Ohio State Univer-
sity.

APPENDIX

 

QUESTIONNAIRE

TV COMMERCIAL STUDY
January-Februaryl 1262

Now take just a minute to answer the following questions concerning
the commercial you just saw. Please read each question carefully
before responding.

1.

.At the beginning of the commercial, several musical items were
shown suggesting different types of music heard on a Panasonic
radio. List as many of these items as you can remember.

Later on, a girl opened the back of the Panasonic radio and
indicated several important aspects of radio construction and
operation. List as many of these features as you can remember.

At one point, some radio accessories were shown -- some "extras"
'which.may'be purchased g§_§.slight additional cost with the
purchase of a Panasonic, but which are not essential to radio
operation. List as many of these accessories as you can
remember.

In addition. the three boys pointed out some special features
which come with all Panasonic radios -- at pg_additional cost.
List as many of these Special features as you can remember.

 

 

 

 

,.
-\

u

- c

 

 

5.

7.

-2-

.At the end of the commercial, a display of Panasonic radios was
presented. List as many things as you can remember about the
items you saw in that display.

Now look at the scale below. It has seven intervals to give you

an opportunity to show the intensity of your judgment. For example,
if you think FORD is a very Fast car, place an "X" in the space next
to Fast. If you think FORD is Quite Fast, place your "X" in the
second space. If only'Slightly Fast, place it in the third space.
If you feel FORD is neither Fast nor Slow, or if you have no
judgment one way or the other, place your "X" in the center space.
Likewise, if you think FORD is Very Slow, put your "X" next to

Slow; Quite Slow, second space; Slightly Slow, the third space

from Slow. Fill in the scale below for practice.

FORD CAR

Got the idea? Now think again about the display of Panasonic
radios you saw at the end of the commercial, and rate that display
on each of the following scales:

PANASONIC RADIO DISPLAY

 

 

 

 

LARGE:___: 2 z : :___}___3SMALL
DULL:___3 : : : z : :INTERESTING
STRONG: : x a z z : :WEAK
ANNOYING: : x a z : :PLEASING
EELIEVAELE: 2 z 3 z s : :UNEELIEVAELE

 

.As you recall from the commercial, how much did the Panasonic
radiO‘weigh?

How much did the Panasonic radio cost?

THANK YOU VERY MUCH FCR YOUR HELP .

 

 

 

 

 

 

 

ETLB

u’mflafif

 

WTHWIWWWMTIWIHM

9 685