A MULTNARIATE STUDY OF"
REACTIONS TO TELEVISION
COMMERClALS

JAMES RANDLE RUSSELL

 

MWNHW\\W\\WWITWWIHlI 1

3 1293 10539 6836

r '\ K3

: ' 13mm"
moment «at!

    
    

 

 

 

MSU

 

 

 

. RETURNING MATERIALS:

Place in book drop to

 

LJBRARJES‘ - remove this checkout from
.J—nc ' your record. FINES will
be charged if book is
returned after the date
'stamped below.
%' ~ /

 

 

 

 

 

 

 

 

 

 

- A--'_—44—‘.u-_--',q—————

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

requirements of the Master of Arts degree.

I ~ 3: 'JA/i' v 'L/ji‘wb /"/C{’{’\
: f I ‘ or ’7
} (:Byrector of TheSis ’

 

 

ABSTRACT
A MULTIVARIATE STUDY

OF REACTIONS
TO TELEVISION COMMERCIALS

By

James R. Russell

The purpose of this research was to examine the relationship between
liking a television commercial and being persuaded by it. The composition
of that liking or disliking was also studied along with other variables
which could affect the communication situation.

Eight hypothese were formulated, the primary one proposing that the
persuasibility of a commercial can be predicted from its likeability. Per-
suasibility was measured by purchase inclination, and likeability for each
commercial was defined by an index of the respondent's attitude toward the
commercial. Other important elements of the communication of advertising
were also measured, such as media use habits, brand preference, and atti-
tudes toward advertising in general.

Students from Michigan State University were selected randomly to
participate in the laboratory study. Each student viewed thirteen commer-
cials and rated each of them, along with providing other information. Mul-
tiple regression was utilized to analyze the predictive accuracy of the
likeability measures on persuasibility.

The findings indicate that the persuasibility of a commercial can be
predicted from its likeability, but that the relative importance of all the
elements composing the likeability measure vary greatly. Impact, interest,
.and meaningfulness of the commercial appear to account for much of the var-
iance in the prediction of persuasion. Further study on these specific

elements of commercial likeability is proposed.

.6» //'/770

A MULTIVARIATE STUDY
OF REACTIONS

TO TELEVISION COMMERCIALS

BY

James Randle Russell

A THESIS

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

MASTER OF ARTS

Department of Advertising
College of Communication Arts

1974

TABLE OF CONTENTS

ACKNOWLEDGMENTS

CHAPTER 1:

CHAPTER 2:

CHAPTER 3:

CHAPTER 4:

CHAPTER 5:

CHAPTER 6:

APPENDIX A:
. APPENDIX B:
APPENDIX C:

APPENDIX D:

INTRODUCTION
REVIEW OF THE LITERATURE

INTRODUCTION

REACTIONS TO ADVERTISEMENTS IN GENERAL
TELEVISION COMMERCIAL REACTIONS
COMMUNICATION MESSAGES IN GENERAL
ADVERTISING AS A MESSAGE FORM
METHODOLOGIES AND DESIGNS

THE PROBLEM AND STATEMENT OF PURPOSE

HYPOTHESIS
ASSUMPTIONS
LIMITATIONS OF THE SCOPE OF THE PROBLEM

METHODS AND PROCEDURES

THE LABORATORY AS A DESIGN

THE CHOICE OF TELEVISION AND COMMERCIALS
THE QUESTIONNAIRE

COLLECTING THE DATA

TREATING THE DATA

DESIGN OF ANALYSIS

RESULTS

DISCUSSION AND SUMMARY

INTRODUCTION

LIKEABILITY AND PERSUASIBILITY

LIKEABILITY AND BRAND PREFERENCE

LIKEABILITY AND ATTITUDES TOWARD ADVERTISING
LIKEABILITY AND OTHER FACTORS

THE PERSUASIBILITY MEASURES: PAST AND FUTURE
SOME CONCLUDING REMARKS

QUESTIONNAIRE

CODEBOOK

INDEXES COMPUTED

COMMERCIALS UTILIZED IN THE STUDY

Page

12
15
17
23

27

28
31
31

33

34
35
36
37
38
40

41
49
50
50
55
S7
60
62
63
65
71
75

77

APPENDIX E:

APPENDIX F:

APPENDIX C:

BIBLIOGRAPHY

TABLE OF CONTENTS (CONT.)

COMMERCIAL TOTALS
VARIABLES, MEANS, AND STANDARD DEVIATIONS

CORRELATION COEFFICIENTS

91

93

95

98

ACKNOWLEDGEMENTS

I am grateful for the excellent guidance and assistance of Dr.
John Simpkins, who also served as an able advisor while attending
Michigan State. I will always admire both his knowledge and patience
which I needed so much to complete this task.

I also wish to thank my wife and friend, Jennifer, who had to suffer
the same tribulations as I in completing this thesis. Her closeness

and understanding made the going much easier.

CHAPTER 1

INTRODUCTION

The advertising research community has been plagued for decades with
the problems of measuring advertising effectiveness and relating advertising
and purchase behavior. In seventy-five years advertising has evolved from
a simplistic notion of "selling the product" to an advanced science of
communication based on behavioral research. To give a better perspective
and provide a background for this thesis, it may be helpful to retrace the
advertising and marketing doctrines that have prevailed during this century.

In the early 1900's the function of advertising was to gain the reader's
attention with cleverness, and thereby sell the product through a persuasive
message.1 The "old inspirational method," as it was called, dominated the
advertising world until the mid-1920's, when "thé research method" was coined.
Many large advertising agencies had research departments earlier than that,2
but their functions were limited indeed. A Sign of the changing times came
in 1927, when James O'Shaughnessy, an executive secretary of the American
Association of Advertising Agencies, said, "Copy is no longer a matter of
pure inspiration. It is becoming more and more a message based on tested
and proved facts."3

Despite this revelation, research of the 1930's was focused on the
product being advertised rather than the consumer of that product. The
purpose of market research for the advertiser was stated to be "... to reach
the people who are potential buyers of his (the advertiser's) product; to
deliver the message that will influence them; to secure a response that will
increase the flow of his goods."4 This economic definition of market research

' failed to be utilized by the advertising community to any great extent. A

typical advertising textbook of that period was remarkably unusual if it

contained even one or two chapters on research.5 The Hotchkiss text of

1940 went on to make a remarkable admission concerning human behavior:

"Human beings are less constant and predictable in their behavior than guinea
pigs - much less so than vegetables or minerals."6 There at least was a

new notion that consumers were somewhat different than Pavlov's dogs. This
is the first hint in advertising that repetition alone may not be sufficient
to produce desired sales.

More importance was placed on the consumer later in the 1940's, but
certainly little headway was being made. Advertising and selling functions
remained inseparably linked. Market research and advertising testing were
added in advertising texts as a second thought, stuck in the last couple
chapters, with little evidence that a connecting link with advertising
strategy was necessary or desireable.7

In 1950 Lucas and Britt published the first text on advertising research,
but it was desperately lacking in the acknowledgement that the consumer was '
anything more than a passive participant in the decision-making process.8
The two authors concerned themselves with motives but little with communication
effects. They did not examine what the consumer brought with him to an
advertising situation (predisposition), but assumed rather that there was
some similarity in the mass of consumers.

This short-sighted view changed drastically in the 1950's, when various
behavioral scientists (Lazersfeld, Hovland, and others) discovered many com-
ponents of consumer attitude that could affect behavior. Another leader in this
new endeavor, Joseph Klapper, put it this way: "The effect of communication
"is almost always a result of a host of variables, rather thgn merely the

result of a message reaching an audience through a medium." Terms

such as selective exposure, selective perception, and predisposition were

developed by these researdhers. Unfortunately, the advertising community
was slow in accepting this change.

One can easily see this evolving change in S. Watson Dunn's three
Leditions of advertising textbooks}.o In the 1961 edition, Dunn lists‘
various new developments in advertising, one of which is "the increasing
use of communication concepts in formulating creative strategy". By the
1969 edition, he develops this into a consumer behavior reference, and stresses
the importance of coordinating the marketing mix and communication. In the
1974 edition, Dunn devotes over 140 pages to communication alone, and
another 100 pages to consumer behavior and market research. Here, in an
incredibly short span of 15-20 years, advertisers have changed their philosophy
of consumers from a passive to an active agent.

Another catalyst in this recent history and development was the blooming
of a new field in 1965 and 1966 called consumer behavior. Researchers pooled
the knowledge of such diverse fields as sociology, anthropOlogy, psychology,
and communication theory with existing marketing and advertising ideas. In
just a few years this new field would move from a simplistic model of consumer
behavior (the famous AIDA model: Attention-Interest-Desire-Actionl1) to
complex consumer behavior models with nUmerous variables. These new models
reflect the added knowledge gained since the 1950's.

These developments in advertising and market research have altered the
nature of such research. In analyzing any communication situation (including
advertising, of course) all relevant variables should be examined, if possible.
This thesis is an attempt to isolate and observe certain of these variables
which may affect reactiOns to television commercials. Furthermore, the elusive

link between these reactions and purchase behavior will be examined as an

element of persuasibility. Specifically, reactions can be studied on the basis

of various categories. For instance, apart from merely liking or disliking
a commercial, an individual may react to the commercial from a production
or technical viewpoint; he may react to varying degrees depending on his interest
in the product being advertised; he may react to the overall impact of the
commercial; and so on. The possible explanations of the situation seem endless,
and, indeed, perhaps they are. But as this historical introduction has shown,
progress can be made in delineating certain variables which play important
roles in the process of communication. By examining television commercials
in the light of previous communications research, it is hoped that a clearer
definition of problems and variables will surface.

The study of these types of variables, and reactions to commercials in
general, can be of extreme importance to the advertiser and to the consumer.
An acknowledgement by the advertiser that the consumer is an active partici-
pant in the audience, rather than simply a massive blob, will surely lead
to more creative approaches. This can be achieved only when information is
known about the communication process so that better decisions can be made
for the conSumer. This means not merely a more efficient expenditure of
advertising dollars, but perhaps a more tolerable level of advertising for

the consumer. This thesis is a step in that direction.

FOOTNOTES — Chapter 1

1
Frank Presbey, The History and Development 9f Advertising, Garden City,
New York: Doubleday, Doran, and Company, Inc., 1929, page 525.
2
Ibid., page 527.
3 .‘I
Ibid., page 528.
4
George Burton Hotchkiss, An Outline pf Advertising, New York: The
MacMillan Company, 1940, revised edition, page 148.
5
In fact, the Hotchkiss text had exactly 48 of 600 pages on research,
of which 26 pages dealt with product research.
6
Ibid., page 151.
7
A good example here is Harry P. Bridge's text, Practical Advertising,
written in 1949.
8
Darrell Blaine Lucas and Steuart Henderson Britt, Advertising Psychology
and Research, New York: McGraw Hill Book Company, Inc., 1950.
9
Joseph T. Klapper, "What We Know About the Effects of Mass Communications:
The Brink of Hope", Public Opinion Quarterly, Winter, 1957-1958, page 455.
10
S. Watson Dunn, Advertising: It's Role ig_Modern Marketing, New York:
Holt, Rinehart and Winston, editions 1, 2, 3, 1961, 1969, and 1974.
11
A good summary of the more simple models can be found in an article presented
to the American Marketing Association by Conrad R. Hill in 1972, entitled,
"Delineating the Components of Attitude for Improved Pre-Test Measures of
Advertising".

 

 

 

 

 

 

 

 

 

 

 

CHAPTER 2

REVIEW OF THE LITERATURE

‘._-..—__r—— W__.

INTRODUCTION

In an attempt to organize an orderly literature review, considering
the massive amount of available studies in this field, the review is divided
into three major areas. The first, reactions to advertisements in general,
will be examined, followed by reactions to television commercials, a division
of advertising. The second area to be explored will be communication messages
in general, with the tie-in of advertisements as a message form. In the third
area of review, the various methodologies and designs that have been utilized
in other studies dealing with this theme will be summarized. This organization
structure re-emphasizes the importance and basis for this paper: advertising
research comes under behavioral and communication research, rather than an

independent body of knowledge.
REACTIONS T0 ADVERTISEMENTS IN GENERAL

Advertising research has been concerned historically with reactions

to advertisements as a measure of "effectiveness", however defined. Plummer1
identifies four levels of consumer response, which provides a useful beginning
for reviewing the literature.

(1) the unconscious level

(2) the immediate perceptual level

(3) the retention or learning level

(4) the behavior level
Thus reactions can be analyzed from any one or more of these four levels of
'COnsumer response, depending on the purpose and goals of the research, gr

perhaps, the definition one has for effectiveness. Lavidge and Steiner

attempted to develop an admittedly simple model incorporating all these levels

of response, called AIDA.' The acronym stands for attention, interest, desire,
and action; this series of activities represents the decision-making process
of the consumer, and involves each level of Plummer's. Obviously, the model
stresses the importance of advertiSing in the process, particularly at the
attention level. Since this model was introduced in the 1950's, it has been
expanded a great deal, as a host of other, more complex variables were found
to effect the consumer. Today all consumer behavior models (such as

those of Andreason and Engel) have incorporated some part of their model

with reactions to advertising.

Now that a brief theoretical background has been introduced in this area,
it is important to examine more what exactly advertising effectiveness has
historically been defined to be. Chapter 1 traced the changes in this facet
in this century, but many elements of measurement linger. A leading market

3
and advertising research company summarizes the use of the above models and the.
possible measures of advertising effectiveness:
"By far the majority of advertising research designs
in use today - both the old and the new - attempt to
evaluate ad effectiveness by measuring one or more of
the following stepwise effects which can result from
exposure to advertising.
1- The name of the product is communicated
and remembered.
2- Sales messages are communicated and rem-
embered.
3- Attitudes toward the product change, or
are reinforced.
4- Inclination to purchase changes, or con-
tinues.
5- Sales are achieved, or maintained."
Most advertisers feel that the closer the advertising test measure can come
to the bottom of the ladder, the more ideal, and useful, the test design.

Often this has frustrated researchers, since sales effects are achieved for

a host of reasons, only one of which may be advertising. Additionally,

10.

there is a problem of measuring short-range or long-range goals, which may
be quite different, both in scope and purpose.

Sales, if that in fact is the desired measure of advertising effectiveness,
can be measured by various indexes or audits (e.g., Nielsen store checks),
if one realizes that the result may be tarnished by other variables. Coupon
tests and wave studies over time are subject to similar inaccuracies. An
alternative measure of short-range goals has been the measurement of purchase
inclination or purchase intent. A well-known method in this regard is the
Scwerin test of brand preferences. Typically, people are ushered into a theater
setting and asked to fill out a questionnaire indicating their brand preferences
on several items. They are told that this is for a grand prize drawing following
the show they are about to see, and in order to make the one lucky winner the
happiest, the grocery cart the winner will receive will be filled only with
her most favored brands. The people then are exposed to a 30- or 60-minute
film, interspersed with advertisements for brands included in the grocery cart.
Afterward, the viewers are told that the questionnaires they filled out earlier
had been misplaced, and that they need to fill it out once more. Any change
in brand preference is then attributed to the commercials. This then serves
as one measure of the effectiveness of an advertisement, a measure of purchase
inclination. Obviously, the theater technique has decided shortcomings as
a measure of purchase intent, not the least of which is the vast difference
between viewing at home and viewing in a forced exposure setting. Also, the
measure of brand switching may be for different reasons, totally apart from the
advertising. For instance, a respondent may like a certain brand better as a
free prize, but is unwilling to pay a high price for it in the store. Additionally,

other research tends to cast doubts upon the possibility that a single ad

11.

can alter purchase inclination or brand preference so completely. Other

designs similar to this type are simulated shopping areas, where after viewing

a program and commercials, people are asked to select products. Again, this method
has many of the same questionable problems as with the Schwerin technique.

Many researchers have attempted to improve on the self—reporting type
of purchase inclination measurement. Thomas Juster4 of Princeton has examined
purchase intent for several years. He has found in a series of studies that
there is a large predictive accuracy of subjective purchase probabilities
over buying intentions in the cases of products with high value. In other
words, in the case of automobiles or household durables, predicting purchase
can be most successful if the respondent is given a wide range of choices of
probable purchase, and that there are small differences between alternate
choices. Juster has found in scale development that inclination to buy is
best measured when the possible choices for purchase are slight and not
definite. GruberS and Axelrod6 further explored the scaling system of Juster
with items of lesser value and could not explain as much of the variance
between probable purchase and actual purchase. It seems the products of lesser
cost are also of lesser importance when determining purchase inclination and
probability of purchase. ’

Another method concerned with purchase inclination is utilized by the
Telpex TV Check Test,7 more about which will be discussed later. This type of
question is widely used and attempts to measure immediate reactions of an
inclining nature right after the commercial. This type of inquiry might go,

"Now that you have seen this commercial, would
you say you might be (a) a little more likely

to buy the product; (b) a little less likely to
buy the product; or (c) quite definitely neither

more or less likely to buy the product?"

This form of self-reporting is meant to discover whether the consumer is

12.

"leaning" toward purchasing the product advertised, immediately following
a commercial. It remains unclear how valid the association between these
results and actual purchase is, but certainly, it would be very short-range,

if a correlation does exist.
TELEVISION COMMERCIAL REACTIONS

To focus this review somewhat, television commercials and reactions to
them can be examined as a part of the entire realm of all of advertising.
It is surprisingly difficult to find many examples of pre-test or post-test
research designs on television commercials. It seems most commercial research
companies conceal their testing procedures well, and one is left with a small
group of tests primarily by academicians. Many of these techniques are recall
measures used in a post-test situation, while many others are reaction profiles
used primarily in pre-test designs. A variety of tests attempt to measure
consumer reaction physically or physiologically, using everything from measuring
pupil dilation to galvanic skin response. Wells and Leavitt, however, recent
leaders in television commercial testing, initiated a call for a return of
the consumer opinion in advertising.8 Their paper (written in 1971) had a
stated purpose of recognizing the importance of consumer opinion about
‘ commercials, in spite of the alleged subjective bias involved. They pointed
out that consulting consumers directly.about ads can be helpful in an immediate
fashion in checking communication results.

Three examples of research studies using bipolar scaling techniques for
measuring reactions to television commerCials have been developed in the

9

past ten years. In 1964 William Wells developed an EQ (Emotional Quotient)

scale composed of twelve statements about a commercial. 'He expanded this EQ

13.

scale through several stages, using a thesaurus, and finding which words

respondents selected to describe ads. He ended up with 26 semantic differential
scales and tested various ads, having housewives use the scales to describe

the ads. He analyzed the data using factor analysis to find three clusters

of scales: attractiveness factor (typified by adjectives such as "attractive-
unattractive", "fascinating-boring"); meaningfulness factor ("important-unimp-
ortant", "honest-dishonest"); and a vitality factor ("lively-lifeless", "new,
different-common, ordinary"). In another study he used the scales to predict

recall scores and found that the "meaningfulness" scales were the best predictors.lo

A second study by Clark Leavitt in 1970 resulted in similar findings in
rating television commercials.11 Starting with a list of 525 possible words
to describe commercials, he reduced this unweidy total to 45 through the
following process. Commercials were shown to subjects who could then check
as many or as few adjectives as they wished to portray their feelings about
a commercial they had just seen. Various types and stages of testing and a
factoring of results left the 45 words. Seven factors were found, four of
which accounted for most of the variance: the energetic factor, the personal
relevance factor, the sensual factor, and the familiarity factor. The energetic
and sensual factors were similar to the attractiveness cluster in Wells' study
described earlier, and the personal relevance factor closely resembled the
meaningfulness factor.

The two researchers combined with another colleague in a third study12
involving yet more adjective pairs and commercial ratings. Similar to the
preceding studies, a series of steps were taken to distill a number of profile
scales from a large pool of words. In an attempt to predict recall, a personal
relevance factor was again prominent in the prediction. Some of the adjectives

in this relevance factor were again the same as the earlier studies - "impor-

tant, meaningful, valuable."

14.

Not all research of television commercials has been limited to bipolar
adjectives, however. In a study that has been mentioned earlier in the
discussion of purchase measures, the Telpex TV Check Test offers a different
approach. This study, by two British researchers, Caffryn and Rogers, ex-
plored the effects of successful American and British commercials on British
audiences.13 The study involved twelve commercials, six American and six
British, nearly all of which had won prizes at international festivals, and were
all accepted as reaching a high level of creative achievement. The twelve
commercials were each tested separately through what is termed the Telpex
TV Check Test. In this type of test, seven questions were used to explore
the reactions to each commercial. Four questions are standardized and set
out to evaluate the commercial in terms of "impact, acceptance/contextual,
acceptance/personal, and persuasion." Normally, the other three questions are
specifically written for each commercial tested, to probe certain areas. For
this particular study, all seven questions were identical, the other three
evaluating the Commercial in terms of "interest/welldoneness, likeability,
and repetition". The sample consisted of 100 British adults for each commercial,
giving a total sample of 1200. When results for all twelve commercials were
analyzed, there were distinct differences between the responses to American
and British commercials. Reactions to American commercials were more extreme
and more entertaining, while the British commercials were a little more
effective in terms of persuasion. The Telpex Test appears to be an excellent
tool for probing television commercial reactions. In contrast with the bipolar
scaling tests described earlier, the test itself has a more natural and more
' easily understood format, and appears to be of more interest to the respondent.

The test also has the advantage of securing a more evaluative response without

forcing the respondent to answer as an expert, as often happens asking eval-

I).

uations about specific audio or visual elements. In an abbreviated format

the test appears to evaluate all elements of a commercial.
COMMUNICATION MESSAGES IN GENERAL

As indicated several times already, communication research has added a
great deal to the mass of knowledge associated with advertising research.
Advertising can be viewed as a form of communication message, so there is
much of this sort of research which relates to the problem here. The studies
of Lazersfeld, Hovland, Schramm, and others, all have elements of mass com—
munications effects that are helpful in understanding communication.

A well-researched component of communication is that of source credibility
and attitude change. By manipulating variables of the source, message, and
medium, the reaction of the receiver can be observed; thus, many of these
studies are experimental in design. Aronson and Golden have examined source.
credibility assoCiated with Opinion change,14 particularly experimenting with
various aspects of that credibility, finding both relevant and irrelevant
conditions.15 Hovland, Janis, and Kelly discovered general relevance factors
of the source, such as intelligence, responsibility, honesty, and sincerity.16
In both of these studies, credibility of the source is defined in terms of
expertness and/or trustworthiness of the source. In the Aronson and Golden
study mentioned before, the two researchers looked at the effects of various
conditions of the source, some relevant and some irrelevant to the situation.
They used a four-celled design to test their hypothesis that a greater change

'in the opinion of the audience will be attained if attributed to a source

having high credibility. The researchers had Negro and Caucasian speakers

16.

talk to a class of sixth-graders on education, and they were portrayed as
either engineers at dishwashers. The relevant condition here was employment
or occupation, and the irrelevant condition, race. However, the result was
that both relevant and irrelevant aspects of credibility were important.
determinants of opinion change. Similarly, Aronson and others found in 1963
that once again greater opinion change results when attributed to a highly
credible source.17 Here the phenomenon of discrepancy was examined along
with credibility. They found that increasing discrepancy (the extent of
discrepancy between the opinion advocated by the communicator and the pre-
communication opinion of the recipient) resulted in greater opinion change,
but only to a point; as the discrepancy becomes extreme, opinion change declines.
This research emphasis on the source of communication messages is by no
means the only area studied. Various message forms and the medium utilized,
has been another important area of communication research. Marshall McLuhan,
probably the most widely known of communcation philosophers, introduced a
new concept in 1966, "The medium is the message".18 MUch research has shown
that credibility of a message is affected in many ways by the medium utilized;
yet, the primary focus of research recently has been in that sector so long
ignored, that of the audience. Research began to show that the audience is
anything but an inactive mass, that people remember communications in a highly
selective way. Klapper's book in 196019 brought an end to the era of taking
communication effects for granted, and the idea of controlling communication
effects by manipulating variables in the message and source is dependent
first on the audience. Klapper sees communication as operating through mediating
factors, such as group membership, selective exposure, and defense mechanisms,
such that mass communication is only a controlling agent of communication

20
change.

ADVERTISING AS A MESSAGE FORM

If advertising can now be viewed as a component of communication, then
the summary of research mentioned above has an impact when looking at tele-
vision commercials in particular as an advertising form and as a message form.

Krugman made the distinction that television advertising is viewed as non-
21
sensical and unimportant by much of the audience. This is a form of learning

without involvement where the full perceptual impact of the message is delayed.
Thus the link between low involvement advertising communication and high
involvement communication is that advertising seeks gradual shifts in percep-
tion, followed by attitude change at some time; high involvement, on the other

hand, will have a more dramatic conflict of ideas at the conscious level of
22

attitudes and opinions that precedes changes in overt behavior.

Weilbacher makes the link more complete as he describes and lists three
23

factors he believes effect a consumer's response to an advertisement.

"First is the content of the advertisement itself.

Second are the various characteristics of the individual
consumer: his sex, his age, his intelligence, his
affluence, his consumption style, his innate predis-
position to buy particular classes of goods, his feelings
about advertising, the way in which he has learned

or trained himself to defend his senses and perceptions
from mass communication assault, whether he is alert

and inquiring or dull and sickly, and so on and so on.
The third and final factor affecting the consumer's
response to an advertisement is his past history of
exposure to the same advertisement or advertisements that
are so similar as to be perceptually equivalent."

To examine the effects of advertising, then, attempts have been made to
control as many of these factors as possible. In this manner each component
of the advertising and its audience can be examined as to its importance
and relevance. Thus the remaining section of this part of the literature

review examines these separate phenomena as they have been studied. Specifically,

18.

media use habits, attitudes toward advertising, and brand preference have
been singled out for further discussion from the maze of variables Weilbacher
lists. These include elements of life style, psychographics, and behavioral
aspects of the consumer, all omnipresent in today's literature.

The use of media habits data in segmentation studies and multivariate
analysis has been a relatively new technique. It is an important component
of life style, simply because time with the media is often a significant
portion of the hours in a day. Television alone accounts for an average of
six hours a day. Even college students devote a surprising amount of time

to television; a recent study at Michigan State revealed the average was almost
24
two hours a day, and 44.1% watched more than two hours each day.

Bauer and Greyser conducted an extensive research report for the American

Association of Advertising Agencies concerning the consumer's judgement of
25
advertising. In a nationwide sample of almost 2000 households, they found

a number of interesting surprises. Of interest here are their findings
on media use and reactions to both advertising and advertisements. Later in
this section this study will be referred to again with brand preference studies.

Bauer and Greyser make two legitimate, yet conflicting, assumptions or
26
possible hypothesis concerning ad exposure and media use:

1- An individual's reaction may be a direct and immediate
function of the advertisements to which he was exposed.
Thus a person spending a high proportion of his time with a
medium known to contain many annoying advertisements would
report many annoying advertisements. If this exposure pattern
continues for a long time, he might develop a generalized
critical attitude toward advertising and become sensitive
to annoying advertisements. Therefore, his attitude toward
both advertising and specific advertisements will be highly
critical.

2- On the other hand, constant exposure with a medium might
inoculate a person so that he develops a tolerance for
annoying advertisements, and thus be less critical.

19.

The results were somewhat ambiguous, similar to the assumptions above.

In short, "high users of radio and TV are considerably more likely to report
. 27
a higher than usual proportion of annoying ads." The pattern is uneven,

however, for the print media. In Table 1 below, the relationship of total
media exposure and ad categorization is given, again from the Bauer and

Greyser study.

28
Table 1 Total Media Exposure and Ad Categorization

Total Media
Exposure Ad Category
Annoying Enjoyable Informative Offensive

Low zoz 367: aoz 42
Medium 222 362‘. 387: .42
High 252 33% ' 372 sz

There is a slight difference shown, with those who use media to a greater
degree finding ads slightly more annoying and less enjoyable.

They went further to explore the relationship of media exposure and
attitudes toward advertising in general: "The suggestion is fairly strong
that while media exposure is a factor in the type of ads one reports as having
seen or heard, this circumstance bears no patterned relationship to his long-
run overall attitude toward advertising."29

This leads into the next section considered of importance: how a consumer
perceives advertisements as it relates to his overall attitude toward advertising.
Bauer and Greyser also lOoked at this factor in the same study previously

30
mentioned, and these are their conclusions:

20.

"Attitudes toward 'advertising' are definitely

related to reactions to 'advertisements'. This is

not only generally true, but applies to reactions

to ads for individual products as well. The tend-

ency to report favorable reactions to ads is strong-

est for the group with favorable overall attitudes,

the mixed and indifferent groups are next, and the

people unfavorable to advertising report the lowest

proportion of enjoyable and informative ads, although

still over half."
Here "advertising" is distinguished from "advertisements" in that the former
includes attitudes toward the nebulous word concerning all segments of the
business that the consumer may have attitudes toward; the latter, on the other
hand, is a set of attitudes toward a single advertisement or commercial.
Obviously, there is a part-whole issue here, and it would be easy to explain
away this effect without looking at it. But it is of importance to know that
some people will rate ads considerably lower than others simply because of
the medium - that they are advertisements. This, then, is yet another
predisposition that the consumer brings to the advertising communication
situation. It appears that selective exposure works to such a degree that
meaSurement of this variable is important.

The third variable that researchers have felt is important in studying
reactions to advertisements is the effect of brand preference. Reactions
vary with respect to whether or not the brand advertised is the preferred
brand. Several studies could be referred to here as evidence of this effect,
but the Bauer and Greyser study is the most clearly illuminating of any.

(It may seem to the reader at this point that too much emphasis is being
given to one source, but Bauer and Greyser conducted by far the most complete

'study concerning advertising in this decade.)

The conclusions in this regard are, in short: "Specifically, product

I

use and brand preference emerge as correlates of considerably greater interest
' 31
in and favorability toward advertisements for such products and brands."
Table 2 from the study plainly shows the importance of brand attitude in
reacting to advertisements of that brand.
32

Table 2 Brand Attitude and Advertisement Categorization

Annoy- Enjoy— Informa- Offen- Number of

ing able ' tive sive , Ads

My favorite brand:

I prefer it over

all others 7% 49% 44% - 1371
A brand I like

better than

most others 9% 49% 41% 1% 1315
A brand I like about

as well as any

others . 19% 43% 36%" 2% 1724
A brand that will

do if nothing else

is available 47% 24% 25% 4% 498
A brand I wouldn't -

buy 62% 13% 11% 14% 363
No Opinion - , 311
All Users of Product '

In Ad . 19% 42% 37% 2% 5582
All Advertisements 23% 36% 36% 5% 9325

Bauer and Greyser summarize their findings concerning brand preference in
the following way: "There can be no doubt that people are likely to give
‘more attention to ads for brands they prefer. Moreover, when they do give
attention to ads, they are likely to evaluate more favorably the ads for

33
their preferred brands...."

James Mullen has analyzed this brand preference phenomenon from a
slightly different perspective.34 He studied the principle of congruity
proposed by Osgood and Tannenbaum35 as it would relate to television
commercials. The principle is as follows: when two or more concepts are
related by assertion, changes in attitudinal evaluation are made in the
direction of increased congruity with the existing frame of reference.

In other words, a favorable attitude toward a product would tend to move
toward a favorable attitude of the television commercial for that product.

A summary to this point would be helpful in clarifying what research
has determined to be of sufficient and measureable importance in the analysis
of advertising reactions.

1) Response to an advertisement is a function of
a) the ad itself
b) various characteristics of the consumer
c) Spast history of exposure

2) Media use habits are important components of life style,
but their exact overall relationship to advertising
reactions is not clear.

3) Attitudes toward advertising are related to attitudes
toward advertisements.

4) Brand preference is a correlate of favorability toward

advertisements for such brands.

23.

METHODOLOGIES AND DESIGNS

The nature of communication is that it is by definition a systems
concept.36 In this regard most studies have relied on multivariate analysis
to deal with the problem. Plummer has proposed that advertising research be
based less on one-way transmission models of communication and more on.
multivariate studies with the emphasis on receivers' contributions to
communication outcomes.37 He notes three unique aspects of advertising as
a form of communication that call for such an approach:

1) the repetitive nature of advertising over time
2) the highly competitive nature of the environment
where messages exist (clutter)
3) the role advertising plays in culture
Thus advertising should be viewed as "a process that incorporates both the
messages and the receivers, with major emphasis on the viewers' perspectives".38

This in fact has been the case in the last twenty years. Research has
been conducted primarily in experimental designs, lab studies, and field
studies, depending on the desired controlling effects by the researcher.
Examples of these designs cited already in the study are the Wells and Leavitt
work with factor and cluster analysis, Caffryn and Rogers with the Telpex
test, and Bauer and Greyser's mighty field study.

In addition to the development of overall designs, specific measuring
devices and scales were needed to deal with these new complexities. Juster's

work on purchase inclination and prediction has already been mentioned.

In the area of brand and product preferences, Kevin Clancy and Robert Garsen

39
examined a wide variety of scales to more-accurately measure brand preference.

Their research on the advantages and disadvantages of monadic and comparative
preference scales has been helpful in improving measurement tools. The
results of their studies were that if a brand can be shown with others, with
an additional device to measure the relative strength of that brand choice,
the consumer is more likely to respond as he feels, and lessens the chance
of yes saying and nay saying.

In summary the most popular tool for analyzing social science data has
been the multivariate design, since it allows the best isolation and measure-

ment of each variable alone and in combination with others.

25.

FOOTNOTES - Chapter 2

1
Joseph T. Plummer, "A Theoretical View of Advertising Communication",
Journal 2£_Communication, Volume 21, December 1971, page 318.
2
Robert C. Lavidge and Gary A. Steiner, "A Model for Predictive Measure-
ment of Advertising Effectiveness", Journal 2£_Marketing, October 1961, pages 59-62.
3
Burke Marketing Research, Day-After Recall Technique, unpublished.
4
Thomas F. Juster, Anticipations and Purchases, Princeton, New Jersey:
Princeton University Press, 1964.
5
Alan Gruber, "Purchase Intent and Purchase Probability", Journal 2f
Advertising Research, Volume 10, Number 1, pages 23-27.
6
Joel N. Axelrod, "Attitude Measures That Predict Purchase", Journal
'2: Advertising Research, Volume 8, Number 1, March 1968, pages 3-17.
7
John Caffyn and Nigel Rogers, "British Reactions To TV Commercials",
Journal 2£_Advertising Research, Volume 10, Number 3, June 1970, pages 21-27.
8
William D. Wells, Clark Leavitt, and Maureen McConville, "A Reaction
Profile for TV Commercials", Journal 2f_Advertising Research, Volume 11,
Number 6, December 1971, page 11.
9 .
William D. Wells, "EQ, Son of EQ, and the Reaction Profile", Journal gfi
Marketing, Volume 28, Number 4, pages 45-47.
10

 

 

 

 

 

 

Wells, op. cit., pages 47-52.
‘11 '
Clark Leavitt, "A Multi-Dimensional Set of Rating Scales for Television
Commercials", Journal 9f Applied Psychology, Volume 54, Number 5, October 1970,
pages 331-333.

12
Wells, Leavitt, and McConville, op. cit., pages 11-15.
13
Caffyn and Rogers, op. cit., pages 21-27.
14
Aronson, Turner, and Carlsmith, "Communicator Credibility and Communicator
Discrepancy As Determinants of Opinion Change", Journal gf_Abnormal Psychology,
Volume 67, pages 31-36.
15
Elliot Aronson and Burton W. Golden, "The Effect of Relevant and Irrel-
evant Aspects of Communicator Credibility on Opinion Change", Journal 2;
Personality, Volume 30, Number 2, June 1962, pages 135-246.
' l6
C.I. Hovland, I.L. Janis, and H.H. Kelley, Communication and Persuasion,
New Haven, Connecticut: Yale University Press, 1953.

 

 

 

26.

17
Aronson and Golden, 22; cit., pages 135-146.
18
Marshall McLuhan, Understandinngedia: The Extensions pf M33, New York:
McGraw-Hill, 1966. '
19
Joseph Klapper, The Effects gf_Mass Media, Glencoe, Illinois: The
Free Press, 1960.
20
Ibid., page 5.
21
Herbert E. Krugman, "The Impact of Television Advertising: Learning
Without Involvement", Public Opinion Quarterly, 1965.
22
Ibid.
23
W.M. Weilbacher, "What Happens to Advertisements When They Grow Up",
Public Opinion Quarterly, Volume 34, Number 2, Summer 1970, pages 218-219.
24
"Ownership and Use of the Mass Media By On-Campus Student Residents At
MSU", March 1974, Television/Radio Department.
25
R.A. Bauer and Stephen Greyser, Advertising ip_America; The Consumer
View, Boston: Harvard University, 1968.
26
Ibid., page 249.
27
Ibid., page 254.
28
Ibid., page 256.
29
Ibid., page 258.
30
Ibid., page 324.
31
Ibid., page 282.
32
Ibid., page 289.
33
Ibid., page 289.
34
James J. Mullen, "The Congruity Principle and Television Commercials",
Journal 2f Broadcasting, Volume 7, Number 1, Winter 1962-63, pages 35-42.
35
Charles E. Osgood, The Measurement 2f_Meaning, Urbana, Illinois: Univer-
sity of Illinois Press, 1957.
36:
Krugman, pp; cit.
3?’
Plummer, 22; cit., page 315.
38;
Ibid., page 316.
39
Kevin J. Clancy and Robert Garsen, "Why Some Scales Predict Better",
JOurna1‘2£_Advertising Research, Volume 10, Number 5, October 1970, pages 33-38.

 

 

 

 

 

 

 

 

 

 

CHAPTER 3

THE PROBLEM AND STATEMENT OF PURPOSE

28.

The primary problem Of the study can be simply stated: What relation
exists between liking a television commercial and the likelihood of future
purchase of the product advertised? Additional subquestions to be examined
involve what composes an individual's liking or disliking of a commercial.

How are media use, attitudes toward commercials in general, and brand preference
related to the liking of a commercial? It is expected that much can be learned
in exploring these relationships of reactions to commercials concerning both

persuasibility and likeability of television advertising.
HYPOTHESES

HYPOTHESIS l: The persuasibility of a television commercial can be

predicted from its likeability.

Persuasibility in this study is defined as whether a respondent is more
or less likely to purchase the product advertised after having seen the
commercial. Admittedly, this type of subjective measure of persuasion has
liabilities, but it does seem to measure an inclination or possible purchase
expectation. Also, the respondent has the option of answering that he is
neither more or less likely to purchase the product, a third possible response
that does not force an answer where there may not be one. The likeability
of a commercial is defined as the index of the respondent's overall attitude
toward the commercial. This index includes general likeability, impact,
interest, acceptance, repetition (would like to see again), and welldoneness,
or an artistic measure of the advertisement. Thus for each respondent there
will be a likeability score and a persuasibility score for every commercial
shown. Multiple regression will be the statistical technique utilized to

then test this prediction hypothesis.

HYPOTHESIS 2:- The more times a commercial has been seen, the less is
its‘likeability.

By examining the number of previous exposures to a certain commercial,

one can obtain a "seen-before" type of measure. This then can be compared

to the commercial's likeability score for each respondent, and a determination
can be made through correlation techniques whether commercial familiarity

is associated with lower likeability.

HYPOTHESIS 3: Media use is positively related to an individual's
likeability scores for selected advertisements.

This hypothesis says that as an individual uses the media to a greater
extent, he is more apt to react positively toward television commercials.
Media use is defined as an index of the total hours spent each day with
television, radio, newspapers, and magazines. This total will be compared
with a respondent's mean likeability scores across all commercials. This
is somewhat in conflict with the previous hypothesis, since a high media user
would have a greater opportunity to see more commercials, which would lower
the likeability. However, this hypothesis takes into account the effect of
selective exposure and perception; perhaps a media user develops a resistance
to commercials. The relationship of this hypothesis and the earlier one will
also be of interest.

HYPOTHESIS 4: The likeability scores of those respondents who prefer the
brand advertised will be greater than the scores of those
respondents who do not prefer the brand advertised.

For each commercial there are two distinct and separate groups of respond-

ents: those who prefer the brand advertised in the commercial, and those who

prefer an alternative brand. This hypothesis says that there will be a
significant difference between the two mutually exclusive groups, and that
those who prefer the brand advertised will also prefer the commercial.

HYPOTHESIS 5: A respondent's likeability of a specific commercial can
be predicted from his attitudes toward television com-
mercials in general.

HYPOTHESIS 6: Semantic differential scales are better predictors of
commercial likeability than Likert scales.

HYPOTHESIS 7: Of the three clusters of semantic differential scales
(attractiveness, meaningfulness, and vitality), meaning-
fulness is the best predictor of likeability.

These three hypotheses are concerned with the relationship of an indiv-

idual's likeability scores across all commercials tested and his attitudes

toward television commercials in general. This latter measurement is

obtained from responses to both Likert scaling items and semantic differential
items concerning television commercials. The hypotheses propose that not only
can likeability scores be predicted from these two measures, but also

that semantic differential scales can predict likeability of specific ads

to a greater degree than the Likert Scales. Furthermore, these semantic
differential scales include three different types of adjective pairs, and

it is expected that commercials' overall meaningfulness is more highly correlated
with likeability than either attractiveness or vitality.

HYPOTHESIS 8: There is a relationship between those respondents who report
having been persuaded by commercials in the past and those who
report being persuaded by the test commercials.

Once again the respondents can be divided into two groups initially,

those who report having been persuaded by commercials in the past and those

who have not, or at least persuaded to a lesser degree. After all the
commercials have been shown, another measure of persuasibility can be taken
according to this recent measure. one would expect that persuasibility

to television commercials will appear as some sort of trait and that a relation-

ship would exist.

ASSUMPTIONS

A basic assumption that has been made which concerns the entire study,
is the presumption that the commercial likeability scores do reflect the
subjective feeling in each respondent about the particular commercial shown.
In arriving at the scores or indexes, the assumption was made that each of
the factors measured in liking a commercial are to be weighed equally.

That is, for example, that welldoneness and impact are thought about equally
by the respondent. Another possibility is that the respondent never thinks
about any of the rating measures utilized, a danger that exists in any research.

A primary assumption was mentioned earlier - that of persuasibility
being equated with probability of purchase. Practically any other type of
measure as an alternative would require a before-after design, which was not
feasible for this study. A simulation of a brand selection process following
the commercials would seem entirely too artificial and would be open to a

host of additional problems.

LIMITATIONS OF THE SCOPE OF THE PROBLEM

At least two important areas of advertising research are omitted by

the scope of the problem. The first is recall, an area of research that has

32.

dominated much of the literature in recent years. Many of the theories
on the effects of advertising stress the importance of recall in affecting
a purchase decision. It should be pointed out that a segment of the likeability
index is associated with recall, the notion of impact. Presumably, the
greater the impact of the commercial, the greater the recall. However,
further investigation into this area is beyond the scope of this study.

A second interesting area not under examination here is the notion of
salience. A tenet of advertising has been that a commercial is liked better
by those in the "target" audience, or those who the manufacturer feels are
purchasers of his product. This group is termed highly salient, while those
not intended for the message are of low salience. An excuse for irritating
advertising has historically been that the critic was not in the target audience,
so that he was not expected to glean much from the message. To test such a
salience hypothesis for validity, it would be necessary to assess the intended
audience for each commercial and try to match respondents in this way into
groups of high and low salience. In this study college students are the subjects
and tend to be more homogeneous than the population as a whole. Their ages
are similar; life Styles are similar; and certainly advertisers speak to them
as a group. It would be difficult indeed to divide them into groups for

purposes of comparing scores in this respect.

CHAPTER 4

METHODS AND PROCEDURES

34.

THE LABORATORY STUDY AS A DESIGN

The design utilized in this study is an ex post facto lab study, where
one is concerned with describing relations and interactions among numerous
variables. The laboratory experiment has several advantages, particularly
when dealing with multivariate analyses. First of all, most variables can
be isolated or controlled. Manipulation of one or more independent variables
is more easily accomplished than in a field study. Operational definitions
can be specified and adhered to a greater degree, and the precision of
measurements can be controlled.

Despite this high internal validity of the lab experiment, it is a
design wrought with other problems, particularly the lack of external valid-
ity. Often there is an artificiality of the experimental research situation,
which prevents any generalization to other situations. Great care must be
taken in analyzing results from lab studies for this reason.

' The lab study was selected keeping these considerations and qualifications
in mind, along with the realization of restraints in time, cost, and feasibility.
Of prime importance in this study (and others involving television commercials)
is the desire to closely approximate the actual viewing exposure as much as
possible. Any overt manipulation by the experimenter would undoubtedly be met
‘with suspicion, and hopes of realism would be daShed. A major problem of
past advertising research has been the "expert" respondent: the person who
answers questions as he feels an expert would answer them. The questionaire,

’which will be discussed later, attempted to minimize that possibility.

THE CHOICE OF TELEVISION AND COMMERCIALS

Another decision that had to be made was the choice of medium in which
to test. As for as technical problems are concerned, magazines or newspapers
have a distinct advantage over television or radio. Ads are easily accessible
for testing, and a great selection would be available. However, television
combines the audio and visual elements to a much greater degree, enabling
the respondent to react in a communication situation to more stimuli. Also,
as pointed out in Chapter 2, television is an important part of the media con-
sumption habits of most people, including college students. Television commer-
cials have become a way of American life, becoming material for comedy routines,
idle chatter, and cocktail parties. So despite the added technical problems
that were encountered with the medium, television commands the most interest
and attention‘from the public.

Several conditions had to be considered in selecting the commercials
to be tested. It was felt that there should be no more than 12-15 total
commercials, since subjects are likely to tire easily, and ratings could lose
validity. Among this total number of commercials, and in order to test the
brand preference hypothesis, the commercials had to be grouped by product
categories for the total of twelve commercials. And, in order to obtain a
fair sampling of everyday commercials, they should be selected at random on
a typical viewing day.

These criteria were met, but certian problems surfaced in obtaining
the commercials. It was generally impossible to tell whether a commercial

'would be thirty or sixty seconds long initially, so a mixture of lengths

36.

was the result. Ideally, all commercials should be identical in length to
control another possible variable. .(Note: In analyzing the results, no
significant differences were found in commercial likeability scores as a
function of length.) The thirteen commercials finally selected are listed

and described in Appendix D. The four product categories selected were

beer, motorcycles, carbonated beverages, and headache remedies. Each product
category except carbonated beverages contains three different brands. Two
additional commercials were added that were apart from a category. This
served two functions: it tended to disguise the purpose of commercial selec-
tion, and the two commercials provided additional variance across the thirteen

commercials.
THE QUESTIONNAIRE

After the thirteen commercials were taped, the two-part questionnaire,
which can also be found in Appendix A, was completed. The first part of the
questionnaire deals with media habits, brand preferences, attitudes toward
advertising, and come persuasibility questions, all to be filled out prior
to showing the commercials. The second part of the questionnaire is completed
following the commercials. '

Questions 1 through 5 of Part I are questions concerning media habits,
so that a media index can be compiled from this information for each respon-

. dent. .Question 6 is a sex question, the only demographic variable measures,
in which the respondent is to select the brand of those in the product class
he prefers, and the relative strength of this preference on a 1-5 scale. This
is an attempt to separate those of strong preferenCes (or brand loyal persons)

from non-loyal, weak preferences. In analyzing results, only those questions

regarding the products of which commercials were shown were actually utilized
(Questions 7, 8, 13, and 16). All other brand preference questions are in the
questionnaire simply to disguise the purpose. Questions 18-27 are a group of

ten semantic differential scales, rating the attribute "television commercials"
on a scale of 1 through 7. The adjective pairs were randomly alternated. Ques-
tions 28-33 are similar rating scales on television commercials in general, but
utilizing Likert scaling techniques. Questions 34 and 35 are inquiries regarding
persuasion in the past as a direct consequence of advertising.

Part II of the questionnarie reasks eight questions for each of the thir-
teen commercials. Question 1 is a seen-before measure; Question 2 is a measure
of impact; Question 3 is a like-dislike measure; Question 4 is a liking ques-
tion on the repetitive aspects of the commercial; Question 5 asks two questions -
welldoneness and interest; Question 6 is an acceptance measure or how well the
commercial made its points; Question 7 is a persuasion measure from the respond-
ent's view; and, finally, Question 8 is a persuasion measure of those other
than the respondent. All eight questions comprise the basis for rating each of

the thirteen commercials. The composition of an index will be explained later.

COLLECTING THE DATE

A random sample of 67 undergraduate and graduate students at Michigan
State University were selected to participate in the study. Three separate
days were selected, in the summer of 1974, to conduct the 45-minute sessions,
in order to work out time schedule conflicts. Approximately 8-12 subjects were
'tested at a time, since 12 was considered the maximum number of respondents who
could view the television monitor easily. When each session began, both parts

of the question-

.7

38.

naire were distributed, with a cover notice attached that explained the study
(also included in Appendix A.) Subjects were instructed to complete Part I
by answering directly on the questionnaire and to await further instructions
upon completion of that section. This took an average of 8-10 minutes, after
which the respondents were told to separate the two pages of Part II. The
following instructions were given:

"In a few moments you will see 13 different

commercials on various prOducts. Some of

these commercials you may have seen before;

others, perhaps not. After each commercial

is shown, I would like for you to rate the

commercial by answering the 8 questions on

the answer sheet provided. (Example cited

and demonstrated) Please view these commer-

cials as you normally would view them; do not

try to be advertising experts. Are there any

questions?"

Following the question and answer period, the first commercial was shown
on the video tape recorder and TV monitor. Upon its completion, each respond-
ent answered the-eight questions in Part II of the questionnaire. After all
respondents were finished with Commercial #1, the second commercial was shown
and rated, and so on for all thirteen commercials. After these thirteen eval-

uations were completed, answer sheets and questionnaires were collected. Subjects

were thanked and dismissed.

TREATING THE DATA

A complete codebook can be found in Appendix B, but of more importance
here is how the data was treated in developing certain indexes, scales, and
scores. A summary can be found in Appendix C.

The media use index was calculated in the following manner. Question
1 of Part I proved undiscriminating, since 65 of 67 respondents had a tele-

vision set in their room. Questions 2 through 5, however, were useful in

39.

developing the index. The coded response to Question 2, concerning the
amount of time spent watching television was weighted by a factor of 2.
It was felt that the television viewer is exposed to far less advertisements
per hour than any of the other media. The time with the other media remained
as is, and the index was merely a summation in total hours.

The brand preference part of the study presented quite different problems.
One problem that was expected was the possibility of some brands being much
more popular than others. This tended to give a variety of different base
sizes for each brand. For example, so few respondents preferred Anacin as a
headache remedy, yet many preferred Excedrin. This large discrepancy in base
Sizes caused some difficulty in analysis. To compound the problem was the
strength of preference question. Now the cells for analysis became even smaller,
since the division had to be made between strong and weak preferences. The
small sample size made it impossible to make the preference strength dichotomy
as planned, and only use the brand preferred. A larger sample and perhaps
a smaller list of brands could have helped alleviate the situation.

The ten semantic differential scales in Questions 18-27 were divided
into three clusters following an earlier study by Wells mentioned in Chapter
2. Questions 18, 19, 20 and 25 were assigned the "attractiveness" label;
Questions 21, 22, 23 and 26 were given the "meaningfulness" label; and,‘
Questions 24 and 27 were the "vitality" scales. In the analysis these labels
are used as predictors of commercial likeability.

Commercial likeability was defined in terms of a "commercial liking score".
A score was obtained from Questions 2 through 6 of Part II for each commercial
and for each subject. Thus, each respondent had 13 scores, and each commercial
had 67 scores. Therefore, the higher the commercial liking score, the

more the commercial likeability. Both terms are equivalent.

 

 

a...
|

a
-n.

.
n...

rout

soc

p
y

 

._‘

pd
'5K‘

40.

DESIGN OF ANALYSIS

Hypotheses 1, 5, 6 and 7 all deal with the prediction of certain
variables, and consequently, multiple regression will be the statistical
technique used in analysis. Specifically, stepwise addition regression
will produce a Multiple R that must be significant at the .05 alpha level.
Hypotheses 2 and 3 are correlation hypotheses which will be evaluated
by using the Pearson test, and again, must be significant at the .05 alpha level.
Hypothesis 4, a difference hypothesis, will be evaluated by utilizing
a t-test at the .05 alpha level.
The final hypothesis is concerned with whether a relationship exists
between variables, and it is best evaluated using the chi-square technique,

once again significant at the .05 alpha level.

CHAPTER 5

RESULTS

41.

The first hypothesis, concerned with the prediction of a commercial's
persuasibility from its likeability (or commercial liking score), was tested
with two measures of persuasion. The first measure, defined as the level
of persuasion Of the respondent himself, was whether he was more or less likely
to buy the product after having seen the commercial (Question 7). The alt-
ernative measure, defined as how the respondent thought others' persuasion
might be after having seen the commercial, is Question 8. The first persuasion
measure is labeled Persuasibility (Self) in Table l; the second measure is
labeled Persuasibility (Others) in Table 2.

In both cases Hypothesis 1 is accepted using a step-wise addition re-
gression. Multiple R's of .75 for Persuasibility (Self)and .44 for Persuasibility
(Others) were calculated. Both of these values exceed the critical value (.38)
of the multiple correlation coefficient for five variables and an n=60 at
the .05 alpha level.

In both Table l and Table 2, the dependent variable persuasion is at
the top, and the five measures of commercial likeability are listed at the
side. A full explanation of exactly what these measures include can be

found in Chapter 4 and the questionaire in Appendix A.

Table l. The Prediction of Persuasibility (SELF)

By Commercial Likeability

Persuasibility (SELF)

Multiple R R2 R2 Change
Impact .703* .494 .494
Interest/

Welldonness .745* .555 .061
Repetition .747 .559 .004
Like-Dislike .751 .563 .004
Acceptance .752 .565 .002

Table 2. The Prediction of Persuasibility (OTHERS)

By Commercial Likeability

Persuasibility (OTHERS)

Multiple R R2 R2 Change
Acceptance .344* ‘ .119 .119 I
Like-Dislike .390 .152 .033
Reptition .419 .176 .024
Impact ' .435 .189 .013
Interest/
Welldoneness .435 .190 .001

* significant at .05 level

43.

Hypothesis 2, the expected inverse relationship of likeability and
the number of times a commercial has been seen, is accepted on the basis
of an r = .268, significant at the .05 level. The critical value for n = 60
is .25. (Note: The correlation is positive due to the coding procedUre'
as explained in Appendix B. Also, all correlations can be examined in
Appendix G.)

Hypothesis 3, concerning the relationship of likeability and media use,
is not confirmed. The value of the calculated "r" (-.02) was not signifi-
cant. No relationship between those variables was found in this study.

Hypothesis 4, that likeability scores of those who prefer the brand

advertised will be greater than those who do not prefer the brand advertised,

is rejected on the basis of only 5 of 11 t-tests being significantly greater
for the brand-preferred group. However, the likeability scores were higher
for the brand-preferred group over the other group, or non-preferrers, in
10 of the 11 instances. Table 3 summarizes the results of the tests com-
puted for Hypothesis 4. Each product group includes three different brands,

except the carbonated beverages with two.

47:4.

45.

Table 3. The Relationship of Brand Preference to

Commercial Likeability

Difference Tests Between Preferred and Non-Preferred Groups

 

 

 

 

 

 

 

 

BEER Old Milwaukee Pabst Schlitz

N E Std Dev N 7; Std Dev N i Std Dev
Prefer 8 2.88 .83 19 2.28 .40 9 3.22 .51
Non-Prefer 59 2.85 .78 48 2.22 ,36 58 2.48 .56
Difference - .03 .05 - .06 .04 - .74 .05

t = .099 t = .583 t = 3.70*
HEADACHE
REMEDY Alka-Seltzer Excedrin Anacin

N i Std Dev N 2 Std Dev N i Std Dev
Prefer 6 3.30 .36 20 1.84 .25 2 2.50 .01
Non-Prefer 61 2.93 .27 47 1.73 .42 65 1.79 .33
Difference - .37 .09 - .ll .17 - .71 .32

t = 3.08* t = 1.11 t = 2.96*
MOTOR-
CYCLES Yamaha Suguki Honda

N i Std Dev N x Std Dev N ‘i’ Std Dev
Prefer 14 3.69 .26 6 2.33 .44 29 2.50 .40
Non-Prefer 53 3.44 .52 61 2.48 .71 38 2.11 .43

t = 1.74* t - .487 t = 3.90*
CARB.
BEVGES. Pspsi Faygo

N x Std Dev N x Std Dev
Prefer 14 3.30 .62 4 3.55 .13
Non-Prefer 53 3.16 .53 63 3.16 .88
Difference - .14 .09 - .39 .75

t = .820 ’ t = .870

* significant at .05 level (t 1.671, one-tailed test)

46.

Hypothesis 5, that cOmmercial likeability can be predicted from attitudes
toward commercials in general, is accepted on the basis of a Multiple R
of .56, where the critical value is .38. Table 4 illustrates the results
of two types of scales in predicting commercial likeability: the semantic

differential scales and the Likert scales.

Table 4. Attitudes Toward Commercials and

The Prediction of Commercial Likeability

Commercial Liking Scores

Multiple R R2 R2 Change
Semantic Diff. .521* .272 .272
Likert .555 .309 .037

* significant at .05 level

Hypothesis 6, that the semantic differential scales are superior in
the prediction of likeability than Likert scales, is accepted on the basis
of the Multiple R comparisons in Table 5. These were computed as two simple

regressions, whereas Table 4 was a step-wise multiple regression.

Table 5. The Comparison of Scaling Techniques

In The Prediction of Commercial Likeability

Commercial Liking Scores
Multiple R R2
Semantic Diff. .521* . .272
Likert .499* .249

* significant at .05 level

47.

Hypothesis 7, that of the three clusters of the semantic differential
scales, the Meaningfulness cluster is the best predictor of commercial likeability,
is accepted on the basis of data reported in Table 6. Although the Attract-
iveness cluster is also significant in the prediction of likeability, it
does not explain as much of the variance as does the former cluster (see the
R2 change figures). Attractiveness adds only .06 and Vitality .004 to the

explained variance of the Commercial Liking Scores.

Table 6. The Three Components of the Semantic Differential

In The Prediction of Commercial Likeability

Commercial Liking Scores

Multiple R R2 R2 Change Simple R

Meaningfulness .497* .247 .247 .497
Attractiveness ‘ .551 .303 .056 .487

Vitality. .554 .307 .004 .361

* significant at .05 level

Hypothesis 8, that a relationship exists between past persuasive messages
and the persuasiveness of the tested commercials, is rejected. The results
of the chi-square tests are reported in Table 7. Only one of the question
pairs was significant, liking a commercial in the past and buying the product
advertised, and the persuasibility of others (Questions 35 and 8 on the

questionnaire in Appendix A).

Table 7. The Relationship of Past Persuasion and Future Persuasion

Dislike Commercial

 

 

 

 

 

 

 

 

Refuse To Buy 1 2 3 4 Total
Persuasibility 1 2.63 6.90 10.18 2.29 22
(Self)
2 2.39 6.27 9.25 2.09 20
3 2.98 7.83 11.57 2.62 25
Total 8 21 31 7 67
Dislike Commercial
Refuse To Buy 1 2 3 4 Total
Persuasibility l 2.39 6.27 9.25 2.09 20
(Others)
2 3.22 8.46 12.50 2.82 27
3 2.39 6.27 9.25 2.09 '20
Total 8 21 31 7 67
Like Commercial
Buy Product 1 2 3 4 Total
Persuasibility 1 1.97 7.22 11.16 1.64 22
(Self)
2 1.79 6.57 10.15 1.49 20
3 2.24 8.21 12.69 1.87 25
Total 6 22 34 5 67
Like Commercial
Buy Product 1 2 3 4 Total
Persuasibility 1 1.79 6.57 10.15 1.49 20
(Others)
2 2.42 8.86 13.70 2.02 27
3 1.79 6.57 10.15 1.49 20
Total 6 22 34 5 67

* significant at .05 level, 6 d.f.
** critical value, 6 d.f., x

= 12.59

Chi-Square = 5.71

Chi-Square = 7.54

Chi-Square = 4.27

Chi-Square - 14.75

CHAPTER 6

DISCUSSION AND SUMMARY

49.

50.

INTRODUCTION

As indicated in earlier chapters, there is a danger inherent in studies
of this nature when discussing the results. That omnipresent peril is the
acceptance of dubious assumptions as fact without reservation, and then
compounding the error by builidng on a false foundation. Results in this
case are not only faulty but misleading as well. Hopefully, throughout this
paper, there have been adequate warnings and indications when prior research
results were unavailable, and assumptions were made. With this word of cautious

reservation, the discussion can begin.
LIKEABILITY AND PERSUASIBILITY

The findings, as reported in Chapter 5, Tables 1 and 2, seem to indicate.
a relationship between likeability and persuasibility, which along with others,
supports the hypothesis that the persuasibility of a television commercial can
be predicted from its likeability. In Table 1, the five variables which
compose the likeability measure contribute to over half the total variance
in the prediction of whether an individual is more or less inclined to purchase
the product advertised. Two of these variables, impact and interest/welldoneness,
Questions 2 and 5, are significant and account for most of the explained
variance. Their importance in the prediction of persuasion allows for further
discussion.

Possible responses to the impact question are whether the commercial
was powerful, weak, Or neither. To aid in the analysis of this question,

it is helpful to review the intrOduction to this question. The phrasing seems

51.

to indicate that the measdre is an attention-procuring one: the dramatic
qualities of the commercial, and the interest it generates in the viewer.
This is an active, rather than passive, response by the viewer; interest and
impact both play a role, it would appear. The matter of interest has been
found and important variable in most studies of television commercials, as
several studies amplified in Chapter 2, particularly those of Wells and
Leavitt. It is also supported in the communication research reported in
the same chapter: the subject and/or the way it is presented must be of some
interest to the receiver to enhance communication effects. Additionally,
the interest factor surfaces again in the other question which accounts for
another significant part of the variance in predicting persuasibility. In
addition at the presence of the interest factor, clearly the dramatic quality
of an advertisement is another aspect involved here. Its importance can be
traced similarly to the studies mentioned in Chapter 2, those associated
with Wells' measure of vitality and Leavitt's energetic and sensual factors.
The other variable (alluded to in the above paragraph) which was also
shown to be a goOd predictor of persuasibility is the interest/welldoneness
factor (Question 5). Notice here again that two areas are explored by the
one question: personal interest and the quality of execution. Once more it
appears that two overriding variables associated with a persuasibility
measure are (1) a measure of interest or personal relevance; and, (2) a measure
' of dramatic or procedural qualities. Both of these variables are measured in
each of the two questions that account for most of the variance, a finding
supported by several studies mentioned in Chapter 2.

In observing that two of the five variables contribute to most of the

variance in predicting likelihood of purChase, one is faced logically with
three variables that failed (or contributed little) in the prediction.
Probably most surprising in Table l is the failure of response to Question 3,
the liking question, to be of value. If five variables as a whole are defined
to be a commercial liking score, or a likeability measure, and as a whole
predict persuasibility, why does a simple question concerning liking the
commercial do so poorly in this regard? One possible explanation may be the
set of responses offered. Most of the questions could be answered with a neu-
tral or middle qround response; Question 3 could not. Perhaps a fifth response
of "neither like or dislike" would sharpen this variable. A second feasible
explanation is that judging a commercial as weak or powerful (Question 2), or
as interesting or well done (Question 5), involve less of a commitment than
does Question 3, where a personal judgment seems required. This could be a
reluctance to admit actually liking a commercial. A third possibility is that
Question 3 is measuring exactly what Questions 2 and 5 are, and that this I
question adds little to explaining any more variance. As evidence here, a
look at Appendix G reveals a correlation of .831 between Questions 2 and 3,
and a correlation of .932 between Question 3 and the overall likeability
scores. Thus, the addition of the liking variable would not add anything
more to the step-wise regression equation, but to say it is unimportant

would be incorrect.

The discussion to this point has been centered around Table 1 only; that
is, the prediction of "self" persuasion. Table 2, on the other hand, contains
the results of predicting the persuasion of "others," as defined by each
'respondent. This technique is a generalized "other," and it is an attempt

to get the respondent to project his views to how others might react to the

53.,

stimulus. Often in this way he may reveal certain of his own characteristics
that he may hide in direct questioning of himself. In the case of television
commercials, it may be thought by many that it is impossible to be persuaded
by stupid commercials, yet there is a recognition that some other fool may
well be (after all, all that money must be spent for a reason). Question 8,
therefore, allows a way out for this type of respondent: the "other" may be
whomever the respondent wishes. This particular technique should be kept

in mind when analyzing Table 2; for example, other results may have been
possible if the question was asked of "your friends."

As Table 2 shows, the prediction of persuasibility of others from likea-
bility is of less accuracy than the persuasibility of self. These respondents
as d whole apparently felt that they were different from "others." Only one
of the five variables, acceptance (Question 6), is significant, accounting
for most of the explained variance in the prediction of the persuasion of
others. This is no particular surprise, since the question is prefaced by,
"Bearing in mind this was an advertisement, and quite apart from whether
you liked it or not...". From here the question deals with the acceptance
of contextual aspects of the commercial, or how well it made its points. The
question is more one of faCts, as opposed to the more attitudinal queries
regarding like and dislikes. For this reason, respondents' answers to this
question could be used to predict Question 8, or, the persuasion of others;
in both cases, respondents are asked to remove themselves from the actual
condition of viewing. Also, as one might expect, looking at the results in
this light, the interest factor (Question 5) is relatively low in correlation

with the persuasion of others (Appendix G). This would seem to re-emphasize

54 .

the fact that personal relevance, perhaps by definition, does not relate
to the persuasion of a generalized other. ,

In looking at the persuasion measures overall as they relate to the
five variables of likeability, certain observations need mentioning again.
First, there is a perceived difference by the respondent when asked about
persuasion of self or the persuasion of others. This would seem to support
the communication literature in stressing the importance of considering the
audience. Clearly respondents see themselves as much different from "other"
people. A pOssible explanation to this is the sample; college students Egg
probably different from their perceived "others", if not in actuality, then
at least in perception. Perhaps another study could investigate this question
in a group more closely tied to middle class, majority values. A second
observation that should be summarized here is the importance of personal
relevance and interest in both overall likeability of a commercial and its
ability to persuade the viewer.

Of course, despite these important finding, there are additional questions
that could be explored. First of all is the basic assumption that the five
questions do in fact evaluate a commercial in regard to likeability, Indications
are that they do, comparing the results of this study with the results of
studies reported in Chapter 2. If it is accepted that the measures as a
group evaluate commercials adequately (however that is measured), the added
question is whether the individual questions themselves measure what they are
alleged to measure, certainly a point for further studies. The impact measure
(Question 2) is a puzzling phenomenon, particularly important in this study,
yet little is really known about what exactly it measures. What makes a commer-

cial powerful? Is it an attention-getting, attention-holding, fast-moving, loud,

55..

or just what? The variables measured by Question 5 also need further study;
two measures are evaluated together, so perhaps one dimension of the two-part
question is buried. The coding was done in favor of the interest factor,
depressing the welldoneness factor, a questionable decision, and one for'
additional study.

More importantly, perhaps, is the question of whether the two persuasi-
bility measures do in fact measure persuasion (Questions 7 and 8). As pointed
out in earlier chapters, this is an inclination type measure, extremely I
short-term and open to problems of error and bias. Yeasaying, the desire
of the respondent to answer to please the researcher, is a clear problem.
Still another important point is the immediacy of the question; no long-term
effects can be assumed from such a short-term measure. A before-and-after
design would seem necessary to provide a better evaluation of persuasibility,
but then one is open to new and equally frustrating problems of measurement

and analysis.
LIKEABILITY AND BRAND PREFERENCE

As pointed out in earlier chapters, there are additional variables
effecting the research situation that respondents bring to the study. One
of the more important of these is brand preference. It was hypothesized
(Hypothesis 4, Chapter 3) that likeability scores should be higher for those
who prefer the brand advertised. This notion of favorable pre-disposition
resulting in more favorable responses has been demonstrated in political and
communication studies by Hovland and examined in the selective exposure

literature - in addition to Bauer and Greyser's findings, all reported in

Chapter 2. Still, despite this overwhelming bank of evidence supporting

the hypothesis, for this study the hypothesis is rejected on the basis of

the data reported in Table 3 of Chapter 5. But before this study is denounced
as clearly failing to prove what seems to be both a logical and empirical
hypothesis, the results should be examined.

Despite the lack of a majority of Significant differences among the
eleven commercials, between brand preferrers and non-preferrers in terms of
likeability, there clearly is a pattern. Although only five of the eleven t-tests
were significantly greater for brand preferrers, ten of the eleven brands had
a difference in scores that was in the hypothesized direction. Only for
Suzuki motorcycles is the pattern broken, and this result in the opposite
direction is not statistically significant either. There does not seem to
be a discernible pattern among products or product types as far as the
five significant differences are concerned. However, one can examine other
factors in explaining the failure of the study to support the hypothesis.

First of all, perhaps more decisive results might have been achieved
if the strength of preference scales for each brand could have been utilized.
The problem, however, was that the sample size simply did not permit it.

There were too many cells created for the relatively small sample, as shown

in Table 3. Some brands were woefully meager in preferers (i.e., Old Milwaukee,
Alka-Seltzer, Anacin, Suzuki, and Faygo), so that any further division by
strength of preference was impossible. Yet clearly, this would seem to be

an important variable of brand preference. Obviously, a strong, loyal

Schlitz beer drinker differs considerably from a beer abstainer who is forced

to randomly select a brand based on some nebulous (if any at all) criterion.

57.

Thus the situation is clouded by tabulating both of these individuals as
Schlitz preferers, the small sample size restricting finer distinctions.
Perhaps a more selectively drawn, stratified sample would have helped to
alleviate some of these difficulties. A similar problem to be dealt with
involves the forced response. For instance, a mark of weak preference can
mean exactly that, or that the respondent is not knowledgeable concerning that
product class. A good example here is a female respondent who may know little
about motorcycles; she marks the weak preference for Yamaha. Doesn't she
differ from a male respondent who knows about motorcycles, yet considers them
all to be almost the same, and he marks a weak preference for Yamaha? There
are problems like these that could be ironed out with further study.

A larger sample would also be a blessing regarding the notion that a
"target" audience will like a commercial to a greater degree than the "non-target"
audience. Indeed, this excuse for substandard advertising has been the
response by advertisers, saying that if a commercial is not "aimed" at an
individual, he cannot be the judge of good and bad. This test design does not
allow for much investigation of this aspect, however, with such a small sample
size. For instance, one might expect lower commercial liking scores for
females in the instances of beer and motorcycles, but the difference was

not significant Since one is dealing with such a limited sample.
LIKEABILITY AND ATTITUDES TOWARD ADVERTISING

Another variable which may influence likeability (and subsequently
persuasibility) is a predisposition to advertising in general. This is yet

another facet of predisposition and a result of the literature review into

58.

selective exposure and how initial attitudes can affect subsequent behavior
(see Chapter 2).

Tables 4 and 5 of Chapter 5 show the results of using two different
rating and measurement techniques in the prediction Of likeability for each
of the thirteen commercials by all respondents. Both the semantic differential
and Likert scaling techniques were good predictors of commercial likeability.
In comparing the two scales, it appears that the semantic differential scale
is superior in explaining more of the variance; however, the difference in
the magnitude of the R is not at a significant level. The better precision
of the semantic scales could possibly be explained by the apparent ambiguity of
certain Likert items in describing facets of television commercials. The
former scales, offering bipolar adjectives, supply the advantage of giving
the respondent a single attribute on which to judge television commercials»
in general. Likert scales, on the other hand, seek an agreement or disagree-
ment to a particular opinion statement which may have several interwoven
attributes. A look at Questions 28-33 of Part I of the questionnaire (in
Appendix A) bears this confusion out. Whereas Questions 28, 29, and 31 deal
with primarily single thoughts, the remainder are more complex. For instance,
a strong disagreement to Question 30 (Today's commercials are below the national
intelligence level) may be interpreted as a favorable attitude toward commercials;
yet, the respondent may hold a cynical belief that the national intelligence
level is so low that even commercials (as stupid as they are) are not below it.
Similarly, other of these questions have a degree of ambiguity which could
blur the prediction of likeability scores.

The codebook in Appendix B shows the weights assigned to each scaling

59.,

device. The semantic scales were assigned codes easily, the positive attribute
given a seven, the negative attribute given a one. But, as shown in the

above paragraph, assignment of codes to the Likert scales was not as simple.
Further study concerning these Likert items could result in more refined and
less ambiguous seales. This could improve the predictability of the scales
over the semantic differential, but the results of this study report the
Opposite.

Yet another interesting part of this question can be examined in looking
at various clusters within the semantic differential scales. As reported in
Chapter 2, the scales breakdown to three homogeneous clusters (meaningfulness,
attractiveness, and vitality), a finding reported by Wells and Leavitt. Of
these three clusters, as shown in Table 6 of Chapter 5, the meaningfulness
cluster has a much larger R2 change, explaining the most variance in predicting
commercial likeability from attitudes toward commercials in general. This
finding is consonant with the earlier study cited above and in Chapter 2..

Even though the earlier studies examined individual commercials with the three
clusters, results from this study indicate their importance in looking at
commercials as a group. Meaningfulness and attractiveness are both significant
in predicting commercial likeability, but vitality is not. This follows earlier
studies, but it seems in conflict with another finding of this study, where
impact was found to be important. Yet the two measures appear not that closely
associated; Appendix G reports a correlation between the measures of only

.38, whereas one might expect it to be higher. Obviously, the questions must

be measuring different aspects of commercials; this is another area that would

be interesting to explore.

UV.

Another possible explanation of the success of the scales in predicting
commercial likeability from predispositions toward television commercials
is simply a part-whole explanation. The commercial liking score (or likeability)
is an average of thirteen commercials across several questions. The attitude
toward television commercials "in general" may in fact include some of these
commercials. This would be particularly true if the commercials are typical
and were not selected for a Special purpose, apart from product class.

In summary, however, it has been demonstrated that commercial likeability
can be predicted with some accuracy from the knowledge of predisposition of
the respondent. Also, the semantic differential technique is a more precise
measure of likeability than Likert scales, and that the meaningfulness cluster

explains the most variance within the three semantic clusters.
LIKEABILITY AND OTHER FACTORS

To this point it has been demonstrated that the dimensions within the
commercial itself, brand preferences, and predispositions to commercials, all
have an effect on the likeability of commercials. There are still two other
factors that may influence the likeability measure that were looked at in
this study: media use habits and previous exposure to the commercial. These
two hypotheses should be analyzed together, since there is some obvious conflict
between the two, which was briefly examined in Chapter 3.

In one respect, media use and commercial exposure could be seen as
related variables. As one's use of the television media increases, the number
Of commercial exposures will likewise increase. Hypothesis 3 here, however, is

that it is not a simple positive relation. Perhaps it is possible for a high media

61.

user to develop a tolerance for commercials. Indeed, the use of a medium

to a high degree implies some level of satisfaction with that medium, an
implied consent. Following this logic, since commercials are definitely a
part of the medium, there must exist a certain favorable attitude toward
commercials from the high media consumer. Somewhat paradoxically, then,
Hypothesis 2 states that for an individual commercial, as the frequency of
exposure increases, the likeability decreases. The results seem to support
this latter hypothesis, but do not support Hypothesis 3, the tolerance
hypothesis. Interestingly enough, Appendix G reveals that even media use

and commercial exposures are not highly correlated (only .22). -These results
are confusing and somewhat surprising; at any rate they deserve a closer look.

One possible explanation for the somewhat conflicting results is the
failure of the self-reporting technique for both media use and commercial
exposure. Perhaps some alternative questioning method could produce different
results. Another possibility is a third now unknown variable which could
explain the confusion but remains hidden. Or, finally, there is the good
possibility that despite the logic and "common" sense, no systematic relation-
ship exists between media use, commercial exposure, and commercial likeability.
What the study does definitively show is that commercial wearout is a problem;
that increased exposure does lower likeability scores.

Since this section concludes the discussion of commercial likeability,
perhaps it should be examined in light of the test design. Of obvious import-
lance to the study is the testing unit - each commercial used in the study.
Certainly all the relationships discussed in this research are to a certain

degree a function of the thirteen commercials tested. Naturally, with a larger

62.

sample of respondents, more commercials could have been chosen and selected,
giving a larger and possibly more representative sample of commercials. With
only thirteen commercials, there could have been a disproportionate number

of new, less-liked commercials; or, older, better-liked commercials, directly
affecting at least two of the hypotheses. Perhaps all the ads were above or
below average, or even just average, creating fictional differences and
forced responses. Despite separating likeability for each commercial, the
liking score was to some extent dependent on the others tested; certainly
comparisons were made by respondents, although not solicited. Yet another
potential problem is the mixture of thirty- and sixty-second commercials.

The main question is one of validity: if thirteen other commercials were

tested, would the results be similar?
THE PERSUASIBILITY MEASURES: PAST AND FUTURE.

. Hypothesis 8 is a move away from examining the likeability of a commercial
and its corresponding persuasibility. The questionnaire gives a self-reported
measure of general commercial persuasibility in the past, positively and
negatively, and a measure of specific commercial persuasibility for the
thirteen commercials tested as a future measure. It was expected that a
person persuaded in the past because of a commerCial will likely be persuaded
by the test commercials. This could be defined as a test for the persuasibility
trait, if it does exist. The resultant chi-squares reported in Table 7 of
Chapter 5 do not reveal this trait or relationship to exist, except for

persuasion (others) and a positive persuasion in the past. This is difficult

63.

to explain, since the other three tests produce no trait of persuasibility.
Perhaps the contrasting questioning techniques could have produced this hazy
result: Question 34 and 35 are frequency scales, and Questions 7 and 8

are inclination or likelihood scales. Certainly relating past and future
action is a difficult task, particularly when a respondent is reluctant to
admit to any persuasion by a single commercial. The effect may have been
additive, across several commercials for the same brand, or it may have

been subconscious. There is no data to support either possibility in this study.
SOME CONCLUDING REMARKS

Likeability of a commercial has been empiracally demonstrated to be
an important factor in persuasion, but likeability is of itself a complex
measure. It encompasses a spectrum of reactions and factors, many of which
are predetermined before the commercial is ever seen or perceived, such as
brand preference and an individual's overall attitude toward advertising.

Once he is exposed to the commercial, other factors come into play, such

as previous exposure and numerous properties of the commercial itself (impact,
liking, welldoneness, interest, etc.). The total number of variables which
affect reactions to commercials are numerous, and this study has explored
only a few.

If any single result should be emphasized, it is the fact that the viewer
is anything but a passive participant in the communication process of viewing
a television commercial. He brings a multitude of attitudes and predispositions
to the situation, including preconceived notions and ideas about the product,

the brand, advertising, and numbers of other experiences.

V".

The implications for the advertiser, therefore, are that he know his
audience, and that in certain situations, likeability and persuasibility are
closely related. Also, the problem of commercial wear-out is a distinct
danger, as both likeability and perSuasibility are likely to decline.

Once again, also, the value of future research must be stressed. Suggestions
have been made not only in this chapter but in other that certain measure-
ment improvements, sample size changes, and design alterations could be made.
More variables must be isolated and examined before reactions to commercials
can be more fully comprehended. Specifically, better persuasibility measures
need to be discovered that more closely approximate the real world. As more
is learned of the communication process, perhaps these answers will become

more completely understood. A multivariate study is only the beginning.

APPENDIX A

QUESTIONNAIRE

65.

The purpose of this study, developed within
the Department of Advertising, is to examine
the dimensions of TV commercial evaluations.
Part I of the questionnaire is concerned with
overall media use, brand preference, and atti-
tudes toward TV commercials in general. After
you complete Part I, 13 actual commercials will
then evaluate it by answering the 8 questions
in Part II. It is important to view the com-
mercials as you would normally view them. Try
not to be advertising "experts", but rather
answer as honestly as possible. Thank you for

your cooperation.

67.
PART I

1. Do you have a television set in your room, apartment, or house?

Yes No

2. In general, how many hours each day do you watch television? (CIRCLE ONE ONLY)
0 % l 1% 2 2% 3 3% 4+

3. In general, how many hours each day do you listen to the radio? (CIRCLE ONE ONLY)
0 % l 1% 2 2% 3 3% 4+ .

4. In general, how many hours each day do you spend reading newspapers?
(CIRCLE ONE ONLY)
0 % 1 1% 2 2% 3 3% 4+

5. In general, how many hours each day do you spend reading magazines?
(CIRCLE ONE ONLY)
0 % 1 1% 2 2% 3 3% 4+

6. What is your sex?

Male Female
For the following product classes, please check the brand you actually
use or would prefer to use over the others if you had to choose. Also,
indicate how strong your preference is for that brand checked.

 

 

 

 

7. Carbonated Beverages 8. Beer
Pepsi-Cola Old Milwaukee
Coca-Cola Schlitz
7-Up Budweiser
Dr. Pepper Pabst
Faygo Stroh's
Strong Weak Strong . Weak
Preference Preference Preference ' Preference
9. Gasoline 10. Portable Televisions
Shell Sony
Mbbil Zenith
Sunoco General Electric
Standard or American RCA
Texaco Panasonic
Strong Weak Strong Weak
Preference Preference Preference Preference
11. Toothpaste 12. Diet Soft Drinks
Ultra-Brite Fresca
Crest Tab
Close-Up Diet Pepsi
Colgate Diet-Rite
Pepsodent Diet 7-up

Strong weak Strong ‘ Weak

 

.l‘

PART II 68.
1. Have you ever seen this commercial before on television?

. Yes, at least 5 times

. Yes, 1-4 times

. No, never seen it before
. Can't remember

J-‘ri—I

2. When you see commercials on TV, sometimes one makes you sit up and watch,
perhaps because it is very interesting for you or perhaps because it is dramatic.
We call this a strong or powerful commercial. Did you think this commercial was

1. A powerful commercial
2. A weak commercial.
3. Neither, just average

3. Did you like this commercial?

. Yes, it was a very good commercial. I liked it.
. Yes, it was a fairly good commercial.

. No, it was not a good commercial.

. No, it was a terrible commercial. I hated it.

mono.—

4. Sometimes you see a commercial on TV that appeals to you and you like to
see it again; at other times you see commercials that you don't really care
about or would not like to see again. Which one of the following statements
comes closest to how you feel about the commercial you have just seen?

1. I would very much rather see this commercial than most others.

2. I would prefer to see this commercial again rather than most.

3. The commercial wasn't really any different from other commercials;
I don't care whether I see this one again or not.

4. I would prefer to see most other commercials rather than this one.

5. Here are some statements that different peOple have made about this
commercial. Which one comes closest to how you feel about it?

1. It was well done, and I was very much interested in it.

2. It was well done, but it didn't mean much to me personally.

3. I didn't like the way they did it, but I was interested in
what they had to say.

4. It was rather dumb, really - I wasn't interested.

6. Bearing in mind this was an advertisement, and quite apart from whether
you liked it or not, how well do you think it made its points?

1. Very well

2. Quite well

3. Not very well
4. Poorly

7. Now that you have seen this commercial, would you say you might be
1. A little more likely to buy the product

2. A little less likely to buy the product
3. Quite definitely neither more or less likely to buy the product

8. Quite apart from how you feel, would you say others might be

1. A little less likely to buy the product
2. A little more likely to buy the product

3. Quite definitely neither more or less likely to buy the product

69.

 

 

13. Motorcycles 14. Deodorant
Kawasaki Right Guard
Honda Dial
Yamaha Soft and Dri
Suzuki Ultra Ban
Sure
Strong Weak Strong Weak
Preference Preference Preference Preference
15. Coffee 16. Headache Remedy
Folgers Bayer Aspirin
Hills Bros. Bufferin
Max Pax Alka Seltzer
Brim Anacin
Maxwell House Excedrin
Strong Weak Strong Weak
Preference Preference Preference Preference

 

 

Please rate TELEVISION COMMERCIALS in general on the following scales.

 

 

 

 

 

 

 

 

 

18. attractive unattractive-
19. unexciting exciting

20. unpleasant pleasant

21. honest dishonest

22. convincing unconvincing

23. unimportant important

24. lively lifeless

25. boring fascinating

26. strong weak

27. new,different common, ordinary

 

Please indicate whether you agree or disagree with the following statements.

28. Some commercials are so well done, at times I like these ads better
than the program.

Strongly Strongly
Agree . Disagree

29. I like commercials in which the product is compared directly to
competing products.

Strongly Strongly
Agree Disagree

30. Today's television commercials are below the national intelligence level.

Strongly Strongly
Agree Disagree

70.

31. Deceptive television commercials are in the majority, and even those
ads that are not false are exaggerated.

Strongly , Strongly
Agree Disagree

 

32. Television commercials don't have that much effect. People don't
pay much attention. ’

Strongly Strongly
Agree Disagree

 

33. Television commercials serve as entertainment mainly, unless the product
is new.

Strongly Strongly
Agree Disagree

 

Please indicate your response to the following:

34. How often have you disliked a TV commercial so much that you refused to
ever buy or use the product advertised?

Very often
Occasionally
Seldom
Never

35. How often have you like a TV commercial so much that you decided to
buy or use the product advertised?

Very often
Occasionally
Seldom
Never

APPENDIX B

CODEBOOK

Card
Column

1-3 (Part I)
4

6

7-10

11

l3

14

15

16

18-21

23-32

Field Description

 

Respondent Number
Card Number

1. Have TV?

2-5. Hours with media

6. Sex

7. Carbonated Beverages

8. Beer

13. Motorcycles

16. Headache Remedy

7,8,13,16. Preference Strengths

18—27. Semantic Differentials

72.

Codes

Actual
Actual

1=Yes
2=No

J-‘WNI-‘O

-% hours
-1% hours
-2% hours
-3% hours
+ hours

Ln-l-‘UJNH

1=Male
2=Female

l=Pepsi-Cola
2=Coca-Cola
3=7-Up
4=Dr. Pepper
5=Faygo

l=Old Milwaukee
2=Schlitz
3=Budweiser
4=Pabst
5=Stroh's

1=Kawasaki
2=Honda
3=Yamaha
4=Suzuki

1=Bayer Aspirin
2=Bufferin
3=Alka Seltzer
4=Anacin
5=Excedrin

l=Very Weak
2=Weak
3=Neutral
4=Strong
5=Very strong

l=Very Negative
2=Negative
3=Slightly Negative
4=Neutra1
5=Slightly Positive
6=Positive

7=Very Positive

Card
Column

34-35

36-39

41

42

44-45 (Part II)

46

47

48

49

50

Field Description

 

28-29. Likert Items

30-33. Likert Items

34. Dislike ad, refuse to buy?

35. Like ad, buy product?

Commercial Number

1. Senn before?

2. Weak or powerful?

3. Like commercial?

4. Like to see again?

5. How do you feel about it?

73.

Codes

l=Strong1y Disagree
2=Disagree
3=Slightly Disagree
4=Neutral
5=Slightly Agree
6=Agree

7=Strong1y Agree

l=Strongly Agree
2=Agree

3=Slightly Agree
4=Neutral
5=Slightly Disagree
6=Disagree
7=Strongly Disagree

l=Never
2=Se1dom
3=Occasionally
4=Very Often

l=Never
2=Se1dom
3=Occasionally
4=Very often

Actual

O=Can't Remember

1=Yes, at least 5 times
2=Yes, 1-4 times

3=No, have never seen before

l=A weak commercial
2=Neither, just average
3=A powerful commercial

1=No, hated it
2=No

3=Yes

4=Yes, liked it

1=No, rather see others
2=Don't care

3=Yes

4=Yes, very much

l=It was dumb, no interest
2=Well done, but no interest
3=Not well done, but interest
4=Well done, and interest

 

Card

Column Field Discription

51 6. How well made its points?

52 7. You more or less likely to
buy?

53 8. Others more or less likely?

(Repeat for other 12 commercials Columns 44-53)

74.

Codes

l=Poorly

2=Not very well
3=Quite Well
4=Very well

1=Less likely
2=Neither
3=More likely

1=Less Likely
2=Neither
=More likely

APPENDIX C

INDEXES COMPUTED

75.

MEDIA USE INDEX

76.

 

Index:

Example:

2 times (Q2) + Q3 + Q4 + Q5 (of Part I)
Coded responses: Q2s2, Q3=3, Q4=l, Q5=2

Index=2(2)+3+1+2=_1_g

SEMANTIC DIFFERENTIAL INDEXES

 

Index:

Sum of Q18 thru Q27, divided by 10

Sub-Indexes: Attractiveness Index: Q18 + Q19 + Q20 + Q25, divided by 4

Example:

LIKERT INDEX

 

Index:

Example:

Meaningfulness Index: Q21 + Q22 + Q23 + Q26, divided by 4
Vitality Index : Q24 + Q27, divided by 2

Coded responses: Q18=3, Q19=5, Q20=7, Q21=7, Q22=7, Q23=6,
Q24=6, Q25=4, Q26=7, Q27=7

Index = 3 + 5 + 7 + 7 + 7 + 6 + 6 + 4 + 7 + 7 / 10 = 59/10 = 5.9

Attractiveness Index = 3 + 5 + 7 + 4 / 4 = l9/4 = 4.75
Meaningfulness Index = 7 + 7 + 6 + 7 / 4 = 27/4 = 6.75
= 6 + 7 / 2 = 13/2 = 6.50

Vitality Index

Sum of Q28 thru Q33, divided by 6
Coded responses: Q28=2, Q29=3, Q30=5, Q31=4, Q32=l, Q33=5

Index a 2 + 3 + 5 + 4 + l + 5 / 6 = 20/6 = 3.33

COMMERCIAL LIKING SCORES

 

Index:

Example:

Sum of Q2 thru Q6, divided by 5 (of Part II)
Coded responses: Q2=4, Q3=3, Q4=4, Q5=1, Q6=4

Index - 4 + 3 + 4 + 14+ 4 / 5 = 16/5 = 3.2

APPENDIX D

COMMERCIALS UTILIZED IN THE STUDY

77.

Commercial #1

Alka-Seltzer (30 seconds)

Utilizing a nostalgic, reminiscing tone, a newspaper editor thinks back
to his first assignment many years ago. A flashback shows the young, timid
reporter facing his stern boss after the young man had written the story.
Surprisingly, the frowning boss says he liked the story and will give the repOrter
a raise. After leaving the Boss' office, the young man takes an Alka-Seltzer
to Settle his stomach. Forty years later, now editor, the man still uses

Alka-Seltzer. "It was good back then and still is today."

78.

Commercial #2

Old Milwaukee (60 seconds)

 

The scene is a baseball field in spring training. The manager wants
to see a new batter, and asks his name. The young man replies, "George
Herman Ruth, but my friends call me Babe". The coach is skeptical but gives
the kid a Chance for a laugh. Babe hits every ball for a home run to every-
one's disbelief. A voiceover states that it's hard to live up.to a good name,
and makes the connection that Old Milwaukee is from the city that means beer.
As the commercial ends, the manager puts his arm around Babe and asks him if

he knows any kids named Gehrig.

79.

Commercial #3

Anacin (30 seconds)

A man sits behind a desk holding a bottle of Anacin. He says that there
are some things about pain relievers we should know. Not all pain relievers
get in the blood stream as fast as Anacin. He uses a graph to demonstrate
the fact, which shows Anacin superior. At the end of the commercial, he

repeats that Anacin is fast, fast, fast.

80.

Commercial #4

Pepsi-Cola (30 seconds)

 

The Pepsi song dominates, "having fun, feeling free,‘

of action films shows people having fun.
black and white, and many sipping Pepsi.

leaves, bicycling, and picnicking.

' as a college

The people are young and old,

Typical scenes are couples throwing

81.

Commercial #5

Pabst (30 seconds)

The viewer watches as a huge hammer stamps a Pabst impression on a
keg of beer. As the pounding continues, a voiceover talks of the high quality
of Pabst. The voice challenges the viewer to taste and compare Pabst with

any other beer. The commercial ends with, "Since 1844, the quality has always

come through".

82.

Commercial #6

Schlitz (30 seconds)

Two men in a fishing boat head for land. But they disagree on where
they are by the map. When they dock, they go to a bar, carrying the map,
to ask the bartender which one of them is right. It turns out both are
wrong, and there is laughing, back-slapping, and Schlitz-drinking. The
Schlitz song ends the commercial by urging all to grab for all the gusto

you can.

83.

Commercial #7

Yamaha (60 seconds)

Music and photography dominate. Mark Lindsay sings new Yamaha lyrics
to his song "Silver Bird". A young man invites his father for a ride on his
new Yamaha, and reluctantly the older man goes. The mother looks concerned.
The two men ride over hills and valleys, the father enjoying the ride. Upon
returning to the anxious mother, the son offers her a helmet to ride, and

the commercial ends.

Commercial #8

Faygo (60 seconds)

The entire commercial involves a scene on a large boat, full of happy,
enthusiastic people. They are all drinking Faygo, laughing, and singing
the nostalgic Faygo song. The words to the song flash across the bottom of

the screen, inviting the viewer to sing along. The commercial ends with the

sun setting and the singing fading away.

85.

Commercial #9

Suzuki (30 seconds)

A young motorcycle rider grinds through mounds of dirt and sails over
hills. The music in the background is hard-driving and equates the man and
his machine. Speed and durability are demonstrated by the cyclist. The

commercial ends with the logo and "Suzuki-the man's machine".

Commercial #10

Honda (30 seconds)

As the voiceover lists the seven kinds of motorcycles available from

Honda, the video shows a split-screen. .On one side, a businessman in a suit

rides to work; on the other side, the same man is enjoying the bike on

a day off as a dirt bike. The audio is subdued.

87.

Commercial #11

Qantas Airlines (30 seconds)

 

An Australian koala bear takes the viewer on a tour of a Qantas 747.

As he crawls around the plane, he complains in a wobbly voice about all the.

tourists that Qantas brings to Australia. As the commercial ends, the

price of a flight from Detroit to Australia is flashed on the screen.

88.

Commercial #12

Excedrin (30 seconds)

There is a close-up of a man about 30. In a matter-of-fact voice he
tells the viewer that he thinks his pain reliever is better than the view-
er's. He cites two studies at leading independent universities that proved
Excedrin is a better pain reliever than any other. He makes a final appeal

to try Excedrin, the extra-strength pain reliever.

89.

Commercial #13

Pringles (60 seconds)

Two construction workers sit down to lunch. As one man begins eating
regular potato chips, the other begins listing reasons why Pringles are better.
He tells him how unbroken Pringles are and that you get jUst as many chips
in a can. His friend calls him a "potato chip snob", but admits they are

better. The commercial ends with both men munching Pringles happily.

90.

APPENDIX E

COMMERCIAL TOTALS

COMMERCIAL AVERAGES ACROSS 67 SUBJECTS

 

COMMERCIAL Q1 Q2-6 ., Q7 Q8 Preferred/Non-Preferred
#1 Alka Seltzer 1.67 2.96 2.22 2.73 6 / 61

#2 Old Milwaukee 2.12 2.85 2.10 2.64 8 / 59

#3 Anacin 1.15 1.81 1.93 2.19 2 / 65

#4 Pepsi 1.19 3.19 2.48 2.67 14 / 53

#5 Pabst 1.43 2.24 2.15 2.39 19 / 48

#6 Schlitz 2.00 2.58 2.18 2.48 9 / 58

#7 Yamaha 2.18 3.49 2.57 2.78 14 / 53

#8 Faygo 1.48 3.10 2.40 2.69 4 / 63

#9 Suzuki 1.94 2.47 1.97 2.30 6 / 61

#10 Honda 2.28 2.28 2.04 2.38 29 / 38

#ll Qantas 2.22 3.13 2.60 2.63 -

#12 Excedrin 1.76 1.76 1.88 2.12 20 / 47

#13 Pringles 1.64 2.46 2.27 2.60 -

Q1: Seen-Before Measure (Note: due to coding procedure, the more times

seen, the lower the average)
Q2-6: CommerCial Liking Score
Q7: Self-Persuasion Measure

Q8: Other-Persuasion Measure

APPENDIX F

VARIABLES, MEANS, AND STANDARD DEVIATIONS

 

VARIABLE MEAN STANDARD DEV . CASES
VAROOl Seen-Before 1.7975 .4295 67 '
VAR002 Impact 2.6482 .4724 67
VAR003 Like/Dislike 2.8600 .3424 67
VAROO4 Repetitive 2.4775 .2865 67
VAROOS Interest/Welldoneness 2.3866 .5050 67
VAROO6 Acceptance 2.8343 .3980 67
VAROO7 Self-Persuasion 2.2004 .2092 67
VAR008 Other-Persuasion 2.5048 .2900 67
VAROO9 Commercial Liking Score 2.6422 .3594 67
VAROlO Media Use Index 10.7164 3.0441 67
VAROll Semantic Differential 3.7254 1.2271 67
VAROlZ Sem. Diff.-Attractiveness 3.7239 1.3811 67
VAR013 Sem. Diff.-Meaningfulness 3.5485 1.3975 67
VAR014 Sem. Diff.-Vitality 4.2164 1.4853 67
VAR015 Likert Index: 4.2413 1.0199 67

94.

APPENDIX C

CORRELATION COEFFICIENTS

C

96.

VAR VAR VAR VAR VAR VAR VAR VAR VAR VAR VAR VAR VAR VAR
001 002 003 004 005 006 007 008 009 010 011 012 013 014

VAR002 .267

Impact

VAR003 .286 .831

Like/Dislike

VAR004 .159 .745 .757

Repetitive

‘VAROOS .237 .745 .780 .688

Interest/Welldoneness

‘VAR006 .149 .657 .745 .505 .620

Acceptance

‘VAROO7 .114 .703 .620 .610 .689 .513

Self-Persuasion

‘VAR008 .016 .101 .135 .170 .127 .344 .270

Other-Persuasion

VAROO9 .268 .910 .932 .823 .886 .803 .733 .203

Commercial Liking Score

VAR01O -.225 .035-.032-.139 .028-.025-.O94-.263-.020

IMedia Use Index

VAROll ' .154 .532 .467 .211 .591 .390 .507-.008 .521 .191
Semantic Differential

VAR012 .054 .493 .456 .207 .569 .322 .432-.034 .487 .210 .870
S.D. - Attractiveness

VAR013 .152 .482 .438 .218 .527 .449 .449 .078 .497 .141 .886 .597
S.D. - Meaningfulness

VAR014 .210 .380 .326 .154 .434 .188 .414-.110 .361 .150 .821 .611 .696

S.D. - Vitality

VAR015 .193 .489 .474 .357 .477 .368 .472 .123 .499 .155 .692 .532 .683 .541
Likert Index

(Note: VAR001 is Seen-Before)

BIBLIOGRAPHY

97.

98.

BOOKS

Bauer, Raymond A. and Greyser, Stephen, Advertising iE_America: The
Consumer View, (Boston: Harvard University), 1968.

 

 

Bridge, Harry P., Practical Advertising, (New York: Rinehart and Company,
Inc.), 1949.

 

Britt, Steuart Henderson, Consumer Behavior and the Behavioral Sciences,
(New York: John Wiley and Sons, Inc.), 1966.

 

Dunn, 8. Watson, Advertisipg: Its Role ip_Modern Marketing, (New York:
Holt, Rinehart and Winston), editions 1, 2, and 3, 1961, 1969, and
1974.

 

 

Feldman, Shel, ed., Cognitive Consistency: Motivational Antecedents and
Behavioral Consequents, (New York: Academic Press), 1966.

Hotchkiss, George Burton, Ag Outline 2f Advertising, (New York: The
MacMillan Company), revised edition, 1940.

 

Hovland, C.I., Janis, I.L., and Kelley, H.H., Communication and Persuasion,
(New Haven, Connecticut: Yale University Press), 1966.

 

Juster, F. Thomas, Anticipations and Purchases, (Princeton, New Jersey:
Princeton University Press), 1964.

 

Klapper, Joseph, The Effects gf_Mass Media, (Glencoe, Illinois: The Free
Press), 1960. '

 

 

Lucas, Darrell Blaine and Britt, Steuart Henderson, Advertising Psychology
and Research, (New York: McGraw Hill Book Company, Inc.), 1950.

 

McNemar, Quinn, Psychological Statistics, (New York: John Wiley and Sons,
Inc.), 4th. Edition, 1969.

 

Osgood, Charles E., et. al., The Measurement g§_Meaning, (Urbana, Illinois:
University of Illinois Press), 1957.

 

Owen, D. B., Handbook gf Statistical Tables, (Reading, Massachusetts:
Addison-Wesley Publishing Co., Inc.), 1962.

 

Presbey, Frank, The History and Development g£_Advertising, (Garden City,
New York: Doubleday, Doran, and Company, Inc.), 1929.

 

 

V Roper, Burns W., A Ten-Year View 2: Public Attitudes Toward Television

and Other Mass Media, 1959-68, (New York: Television Information Office),
1969.

 

 

Steiner, Cary, The People Look.A£_Te1evision, (New York: Knopf Publishing),
1963.

 

PERIODICALS

 

Achenbaum, Alvin A., "Is Copy Testing a Predictive Tool?", Proceedings,
10th, Annual Conference, ARF, 1964, page 66.

 

 

Armstrong, Donald B., Jr., "Measured Advertising Results: Revisited",
article in Perspectives ip_Advertising Management, Association of
National Advertisers, April, 1969, page 203.

 

 

Aronson, Elliot and Golden, Burton W., "The Effect of Relevant and Irrel-
levant Aspects of Communicator Credibility on Opinion Change", _
Journal gf_Personality, Volume 30, Number 2, June 1962, pages 135-146.

 

Aronson, Elliot; Turner, Judith A.,; Carlsmith, J. Merrill; "Communicator
Credibility and Communication Discrepancy As Determinants of Opinion
Change" , Journal of Abnormal Social Psychology, Volume 67, Number 1,
1963, pages 31- 36.

 

Axelrod, Joel N., "Attitude Measures That Predict Purchase" , Journal of
Advertising Research, Volume 8, Number 1, March 1968, pages 3- 17.

Caffyn, John and Rogers, Nigel, "British Reactions to TV Commercials",
Journal gf_Advertising Research, Volume 10, Number 3, June 1970,
pages 21-27.

 

Canter, Stanley D., "Alternative Criteria in Evaluating Television Commer-
cials", Proceedings, 13th. Annual Conference, ARF, 1967, pages 70-77.

Clancy, Kevin J. and Garsen, Robert, "Why Some Scales Predict Better" ,
Journal of AdvertisingrResearch, Volume 10, Number 5, October 1970,
pages 33- 38.

Cronen, Vernon B., "Belief, Salience, Media Exposure, and Summation Theory",
Journal pf Communication, Volume 23, March 1973, pages 86-94.

 

Cronkhite, Gary and Goetz, Emily, "Dogmatism, Persuasibility, and Attitude
Instability", Journal 2: Communication, Volume 21, December 1971,
pages 342-352.

 

Dodge, Sherwood, "Comments on the Current Scene in Copy Research", article
in PerSpectives‘ip_Advertising Management, Association of National
Advertisers, April 1969, pages 206-219.

 

 

Fishbein, Martin, "Attitude and the Prediction of Behavior", article in
Cognitive Consistency: Motivational Antecedents and Behavioral Conse-

guents, 1966.

Fulton, R. Barry, "The Measurement of Speaker Credibility", Journal 2:
Communication, Volume 20, September 1970, pages 270-279.

100 O

Gerhold, Paul E.J., "Five Suggestions on Measuring the Effectiveness
of Advertising“, article in Perspectives 13 Advertising Management,
Association of National Advertisers, April 1969, pages 221-229.

 

 

Grass, Robert C.; Bantges, David W.; Piech, Jeffrey L; "Measuring Corp-
orate Ad Effects", Journal 2f_Advertising Research, Volume 12,
Number 6, December 1972, pages 15-22.

 

Green, Paul E. and Carmone, Frank J., "Marketing Research Applications
of Non-Metric Scaling Methods", Operational Research Quarterly,
Volume 21, Special Conference Issue, pages 73-96.

 

Gruber, Alin, "Purchase Intent and Purchase Probability", Journal 2;.
Advertising Research, Volume 10, Number 1, pages 23-27.

 

Hill, Conrad R., "Delineating the Components of Attitude for Improved
Pre-Test Measures of Advertising", Combined Proceedings, 1971 Spring
and Fall Conference, Series No. 33, 1972, American Marketing Assoc-
iation, pages 332-336.

Hovland, C. I. and Weiss, W., "The Influence of Source Credibility on
Communication Effectiveness", Public Opinion Quarterly, Volume 15,
1951, pages 635-650.

 

Klapper, Joseph T., "What We Know About the Effects of Mass Communications,
The Brink of Hope", Public Opinion Quarterly, Winter 1957-1958,
pages 455-478.

 

Langhoff, Peter, "Options in Campaign Evaluation", Journal 2f Advertising
Research, Volume 7, December 1967, pages 41-47.

 

Lavidge, Robert C. and Steiner, Gary A., "A Model for Predictive Measure-
ment of Advertising Effectiveness", Journal g£_Marketing, October 1961,
pages 59-62.

Leavitt, Clark, "A.Mu1ti-Dimensional Set of Rating Scales for Television
Commercials", Journal 2£_Applied Psychology, Volume 54, Number 5,
October 1970, pages 427-429.

 

McLaughlin, G. Harry; Haskins, Jack B.; Feinbert, Barry M.: "An Economical
Technique for Pre-Testing Commercial Campaigns", Journal 2£_Applied
Psychology, VOlume 54, Number 4, August 1970, pages 331-333.

 

Mielke, Keith W., "Education Level as a Correlate of Attitudes Toward
Television", Journal 2: Broadcasting, Volume 9, Number 4, Fall 1965,
pages 313-321.

 

Mullen, James J., "The Congruity Principle and Television Commercials",
Journal 2: Broadcasting, Volume 7, Number 1, Winter 1962-63, pages 35-42.

 

Palda, Kristian 5., "The Hypothesis of a Hierarchy of Effects", Journal 2f.
Marketing Research, February 1966, pages 13-26.

101.

Plummer, Joseph TL, "A Theoretical View of Advertising Communication",
Journal gf_Communication, Volume 21, December 1971, pages 315-325.

 

Plummer, Joseph T., "Evaluating TV Commercial Tests", Journal 2f_Advertising
Research, Volume 12, Number 5, October 1972, page 21.

 

Preston, Ivan, "Logic and Illogic in the Advertising Process", Journalism
Quarterly, Summer 1967, pages 231-239.

 

Preston, Ivan and Brown, Lawrence, "Perceiving Advertisements as Emotional,
Rational, Irrational", Journalism Quarterly, Spring 1971, pages 73-84.

 

Simon, Marji Frank, "Influence of Brand Name on Attitudes", Journal 2:
Advertising Research, Volume 10, Number 3, June 1970, pages 28-30.

Stanton, Frank, "A State of the Art Appraisal of Advertising Research
Measurements", article in Perspectives in Advertising Manggment,
Association of National Advertisers, April 1969, pages 237-245.

 

 

Weilbacher, W. M., "What Happens to Advertisements When They Grow Up",
Public Qpinion Quarterly, Volume 34, Number 2, Summer 1970, pages 216-223.

Wells, William D., "EQ, Son of EQ, and the Reaction Profile", Journal 3:
Marketing, Volume 28, Number 4, pages 45-52.

Wells, William D.; Leavitt, Clark; McConville, Maureen; "A Reaction Profile
for TV Commercials", Journal gf'Advertising,Research,rVolume 11,
Number 6, December 1971, pages ll-15.

 

Zimbardo, P. G., "Involvement and Communication Discrepancy as Determinants
of Opinion Conformity", Journal 2f_Abnormal Social Psychology, 1960,
Volume 60, pages 86-94.

 

Zuckerman, "Attitudes and Behaviors", Psychology Reports, August 1972,
page 288.

 

UNPUBLISHED SOURCES

 

Burke Marketing Research, Inc., brochure, "Standard Day-After Recall
Technique".

Mauldin, Charles, Q-study of Attitudes Toward Advertising, Michigan State
University, Department of Advertising, 1974.

"Ownership and Use of the Mass Media By On-Campus Student Residents at
Michigan State University", Department of Television-Radio, MSU,
March, 1974.

387‘? 37

HICHIGQN STRTE UNIV. LIBRQR

IIHIIILIIIUIlllllllllllllllllllllll“llllllllllllllllllllllHI

293105396836