A O-STUDY IN ROLE lDENTRFiCAUON
USiNG A SAMPLE OF ADVERTESENG
PHOTOGRAPHS

Thesis for “We Deqveé of M. Ag
MECEEG‘AN STATE 'UNWERSETY
Elke Koch-Weser

1.96.1

 

 

mmuynzllyujmmmwl{11W

ENAMWA g?

 

:- '1‘ __

 

’4’; \fl

A Q-STUDY IN ROLE IDENTIFICATION USING A SAMPLE
OF ADVERTISING PHOTOGRAPHS

By

Elke Koch-Weser

AN ABSTRACT OF A THESIS

Sub'niitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF ARTS

Departinem of General Communication Arts

1961

Approved:fi)‘n,QMQ/~_— S .DVIM. L14»; f

ABSTRACT

A Q-STUDY IN ROLE IDENTIFICATION USING A SAMPLE
OF ADVERTISING PHOTOGRAPHS

by Elke Koch-Weser

This pilot study is concerned with the operation of identification
mechanisms in the message-receiver relationship in one instance of
mass communication, that is, pictorial advertising.

It has been found that pictures presenting human models in com-
bination with the advertised product attract more attention and are
better recalled than pictures showing the advertised product alone.
This phenomenon has been attributed to identification mechanisms
which occur whenever the reader, who sees the model in a picture,
imagines himself in the place of the depicted model and thus experiences
this model‘s satisfaction from the use of the advertised product.

The general hypothesis of the study was that the more closely
characteristics of the model correspond to those of the reader himself,
the deeper the identificatory reSponse will be. A distinction was made
between the reader's identification with a depicted model in reference
to his actual way of life and in reference to his ideal way of life.
Accordingly, the extent to which persons choose different models as
objects of identification when comparing them to actual self notions
and when comparing them to ideal self notions has been investigated.
The main variables studied were age correspondence and sex corres-
pondence. Since it was anticipated that colored pictures would stimu-
late identification more successfully than black and white pictures,
the color variable was also introduced into the study in order to control

for its effects. A typology of the audience segmentation with respect

Abstract Elke Koch-Weser

to identificatory reactions to advertising messages has also been
attempted.

Stephenson's Q-technique was chosen as an instrument not only
because of its particular use in assessing unobservable phenomena,
such as identification, but also because it allows one to investigate
how individual persons relate themselves to the stimuli they find in
the messages. For Q-sorting purposes a Q-sample of full-page
advertising photographs from 1:231: and Life magazines was composed
on the basis of a factorial design which included the selected picture
variables (sex of model, age of model, color of picture). Fifteen pairs
of young, middle-aged, and old husbands and wives performed Q-sorts
where they described their actual and ideal self notions by means of
the picture cards. The data were processed by means of analysis of
variance and factor analytic methods.

The findings from this Q-study show that identification in respect
to actual as well as ideal self notions tended to be facilitated if sex and
age correspondences prevailed between the receivers of the message
and the depicted model. Concerning the color variable it was found
that men and women of all age groups gave a striking advantage to
colored pictures wheh they chose their ideal objects of identification,
which was not the case when they chose their actual objects of identifi-
cation. When audience segmentations were“ assessed by grouping
persons into particular "types" according to their identification with
similar models, six ”types" of actual self notions and six "types” of
ideal self notions could be distinguished. Each ”type" was tentatively

named by referring back to the original Q-sorts.

Abstract Elke Koch—Weser

Though the generalizability of these results is restricted due to
the limited size of the subject sample, and though advertising illus-
trations were studied in isolation from headlines and copy, these
phenomena merit the attention of the art director who wishes to inte-
grate models into an illustration in order to elicit identification

mechanisms in a special group of readers within a larger audience.

A Q-STUDY IN ROLE IDENTIFICATION USING A SAMPLE
OF ADVERTISING PHOTOGRAPHS

BY

Elke Koch- Weser

A THESIS

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

MASTER OF ARTS

Department of General Communication Arts

1961

T“ ‘3 1‘ ""
~/ ; u I”!

7 / 5’ 7/» 3

ACKNOWLEDGMENTS

During the period of preparation and execution of this study aid
and encouragement were given me by various persons whom I would
like to thank at this place. I am mostly indebted to my advisor,
Professor Malcolm S. MacLean, Jr. , who showed his expert interest
through the different phases of my work. My sincere gratitude is also
due to Professors John W. Crawford, Paul J. Deutschmann, Kenward
L. Atkin, and Lionel C. Barrow, for criticism and suggestions.

I am under obligation to Professor William Stephenson who kindly gave
me access to his unpublished manuscript on A General Theory of Mass
Communication, to Mrs. Margaret Oleson who was in charge of inter-

 

 

views, and to Miss Anne Beard who carefully scrutinized the manu-
script. To my friends and colleagues go my warmest thanks for their
personal helpfulness and assistance.

ii

TABLE OF CONTENTS

. CHAPTER Page
INTRODUCTION . .................. 1

I BACKGROUND AND IMPORTANCE OF THE STUDY . 6

II. THE BASIC FRAMEWORK. . ......... . . . . . . 13

III. IDENTIFICATION AND MODELS ........... . . 24

IV. THE DESIGN OF THE o-STU DY ....... . ...... 36

V. RESULTS: PICTURE EFFECTS AND IDENTIFICATION . 45

VI. RESULTS: TYPES OF ACTUAL AND IDEAL SELF

NOTIONS ....................... . 57

VII. SUPPLEMENTARY FINDINGS ............... 83
VIII. CONCLUSIONS ........ . . ............. 89
BIBLIOGRAPHY .. . . ....................... 102
APPENDICES . . . ......................... 106
I. SUPPLEMENTARY TABLES ............... 107

II. SUPPLEMENTARY FIGURES . . ......... . . . . 120

iii

TABLE

10.

LIST OF TA BLES

. Factorial Design of Picture Sample . . . . . . . .....
. Combinations of Levels of Independencies.

. Frequency Distribution .

. Factorial Design of Subject Sample. . .

. Distribution of Significant Picture Effects for Young,

Middle-Aged, and Old Men and Women . . . . . . .

. Relationship Between Mean Picture Scores for Significant
Effects in Table 5:

a. Actual Self . . . .
b. Ideal Self . .

. Rotated Factor Loadings for Two Sets of Picture Q-Sorts

Provided by 30 Subjects .
Clusters of Persons Defining Factors in Table 7 . .

Sex and Age Characteristics of Persons Defining Different
Types of Self Descriptions . . . . .............

Number of Men and Women of Three Age Groups Most
Representative for Each Type .

. Variance Estimates and F-Ratios for Subject Groups:

1. Actual Self . .
2. Ideal Self . .

. Mean Scores Obtained from Subject Groups in Q-Sorting

Pictures of Different Color, Showing Models of Different
Age and Sex . .

. Eta Coefficient Estimates ()7) of Variable Contribution

to Total Variance in Q-Sort Scores . .

iv

Page
37
37
38

41

46

48
49

59

61

62

64

107
109

111

112

LIST OF TABLES - Continued

TABLE

D. Correlation Coefficients for Q-Sorts of 30 Subjects:

1. Ideal Self ......
2. Actual Self . . . . .

E. Factor Loadings for Two Sets Of Picture Q-Sorts Pro-
vided by 30 Subjects .....................

F. FaCtor Scores in Standard Terms:

1. 'Ideal Self ...................

Z . Actual Self .....

Page

113
114

115

116
118

LIST OF FIGURES

FIGURE Page
I. Three Ways of Identification . . . ......... 20
II. Comparison of Picture Effects in Terms Of Magnitude of
Contribution to Total Variance in Q-Sort Sets 55
III. Top and Bottom Pictures of Ideal Self Factor-Arrays. . . 67
IV. Top and Bottom Pictures of Actual Self Factor-Arrays . . 75
V. Scatterdiagram Comparing Average Ideal and Actual
Q-Scores Obtained for 72. Pictures . . . ..... 84
VI. Scatterdiagram Showing Size and Similarity of Ideal and
Actual Q-Scores for a Single Picture in a Set ........ 97
VII. Four Combinations of Identificatory Reactions in Reference
to Actual and Ideal Self Notions . . . 99
A. Reproductions Of Pictures Described in Chapter VII . . . . 120
B. Scatterdiagram Showing Size and Similarity of Ideal and
Actual Q-Scores Obtained for Picture No. 28 ...... . 121

vi

LIST OF ABBREVIATIONS APPEARING IN

FIGURES AND TABLES

C Color Variable

S Sex Variable

A Age Variable

Subject Variables

 

Sex Variable

 

 

 

 

M male subjects
F female subjects
Age Variable
Y young subjects
M middle-aged subjects
0 old subjects
Sex-Age Combinations
YF young female subjects _
MF middle-aged female subjects
OF old female subjects
YM young male subjects
MM middle-aged male subjects
OM old male subjects
Numeration of Subjects
even numbers 2 female subjects

uneven numbers = male subjects

Picture Variables

 

H1

03%

mf
of

ym

om

bw

Sex Variable

 

male models
female models

Age Variable

 

young models
middle-aged models
Old models

Sex-Age Combinations

 

young female models
middle-aged female models
old female models

young male models

middle-aged male models
old male models

Color Variable

 

colored pictures
black and white pictures

INTRODUCTION

Role identification is generally conceived of as a fundamental
process in the socialization of an individual. PeOple, it is often said,
evaluate themselves and orient their behavior either by reference to
groups to which they belong, or by reference to groups to which they
aspire to belong.

With the framework of mass media communication, the concept
of role-identification helps explain how the individual relates himself
to messages conveyed by the mass media. Most of these messages
contain verbal or pictorial descriptions of persons, which may stimu-
late identification mechanisms in the individual.

The problem with which this pilot study1 is concerned is the
operation of identification mechanisms in the message-receiver rela-
tionship in one instance of mass communication, that is, pictorial
advertising.

A study of pe0ple‘s reactions to certain models in advertising
illustrations is important in view of the fact that an increasing number

of advertisements show portraits of human beings. It has been found

 

1The present study centers upon one part of a larger research
project in role identification during the preparations for which the
author assisted Prof. Malcolm S. MacLean, Jr. in the Communication
Research Center at Michigan State University. - Together with the Q-
sample of advertising photographs a 0- sample of adjectives was used
in order to obtain qualitative self descriptions from each subject.
Furthermore, Osgood's Semantic Differential was employed to assess
connotative meanings for self notions and for advertisements eliciting
most and least identification in reference to each self notion. In addi-
tion, a projective technique deve10ped by Atkinson e_t all. was introduced
in order to estimate the subjects' motivation to fantasy the way of life
of the models shown in each picture.

that pictures presenting human models in combination with the adver-
tised product attract more attention and are better recalled than
pictures presenting the advertised product alone. One reason for this
may be that readers, who see a model in a picture, imagine themselves
in the place of the depicted model, and thus experience this model's
satisfaction from the use of the product.

Certain qualities and actions of a model are more likely to elicit
identification in the reader than others. It was expectedthat the more
closely these qualities and actions of the model correspond to qualities
and actions of the reader himself, the deeper the reader's involvement
in the identificatory response will be. The qualities and actions of the
reader may be conceived as being of two kinds: either they characterize
him in actuality at the time he finds himself confronting the model, or
they characterize him in his ideal notions of how he would like to be.

On the basis of this difference, we may distinguish between the reader's
identification with a model in reference to his actual way of life and in
reference to his ideal way of life.

Although this study will not consider whether a model appealing
to a reader‘s actual way of life has more or less persuasive power
than a model appealing to a reader‘s ideal way of life, the extent to
which persons choose different models as objects of identification when
they are comparing them to actual self notions and when comparing
them to ideal self notions will be investigated.

One aspect of the problem may be approached by asking if Similar-
flof characteristics in the models and the readers enhances identifi-
cation. Here it was examined whether identification is facilitated when
a correspondence in sex and/or in age characteristics prevails.

Sex and age are, however, only two out of a large set of
Characteristics in the model and in the reader. As these two sets

will never correspond perfectly, all identification is bound to remain

partial identification. Nevertheless, such partial identification may
vary in intensity, depending upon the importance the reader attributed
to the shared characteristics.

It is clear that while some qualities and actions of the model may

describe some aspects of the actual or the ideal way of life of one

 

reader, they may not be at all descriptive of the actual or ideal way of
life of another reader. Thus the first reader is likely to accept the
model as an object of identification, while the second reader would
reject it. When a model is to be used in an advertisement to elicit
identification mechanisms in a maximum number of people, the pre-
diction of their heterogeneous identificatory reactions becomes the
problem.

This suggests a second aspect, which has been approached in a
preliminary fashion by asking whether identificatory reactions are
heterogeneous, or whether there are clusters of persons who choose
their objects of identification in a fairly homogeneous way. Again, a

distinction was made between "types'l of persons on the basis of identi-

 

fication with certain models rather than with others in respect to the
reader's actual and ideal ways of life.

In order to prevent colored pictures from having a different
effect upon identification than black and white pictures, the color
variable was introduced into the study where its effect could be con-
trolled.

_ This (2- study is based upon a picture sample composed of full

page advertisements from Look and Life magazines which, 'in their

 

 

illustrations, showed one or more persons in relationship to the ad-
vertised products. The group of readers, whose reactions to the de-
picted models were studied, was composed of men and women of

different age-groups, but of similar socio-economic status.

Stephenson's Q—technique served as the instrument because it is
of particular use in assessing non-observable phenomena, such as
identification. Moreover, it permits the discrimination of the two
kinds of identification processes under study, and it provides the data
which allow both an analysis of the effect of similarity in characteristics
and an analysis of "types" of persons in respect to identification. And
finally, its most fundamental asset is that it helps to investigate how

individual persons relate themselves to the stimuli they find in the

 

message.1

It was anticipated that the findings from this 0- study would Show
that people identify preferably with models similar to themselves,
that pe0plei identify with some models in reference to their actual way
of life and with other models in reference to their ideal way of life,
and that there are clusters of readers in an audience who select similar
objects of identification when their actual or ideal self notions are in-
volved. These phenomena would merit the attention of the art director
who integrates models into an illustration in order to elicit identification
mechanisms in a special group of readers within alarger audience.

However, due to the fact that a very small number of people was
studied and, also, that the pictorial section of the advertisements was
considered in isolation, caution is demanded in the generalization of
the findings. 1

After some remarks concerning the background and the relevance
of the present study (Chapter I), the central concepts will be defined,
the place of personal identification in the communication process will
be pointed out, and a model,which may help to show‘some ways a

person identifies with another person, will be introduced (Chapter II).

 

l’W. Stephenson, The Study of Behavior: Q-Technique and Its
Methodology (Chicago: The University of Chicago Press,’ 1953), pp.
47-61.

 

 

In Chapter III a number of previous empirical investigations will
be mentioned, which either shed some light upon our hypotheses or
may be ofihelp in the interpretation of the findings. In this Chapter,
the hypotheses are formally introduced, experimental controls are
discussed, and the limitations of the study are pointed out. Chapter IV
explains the design of the study, Q-sorting procedures, data collection
and statistical analyses, in addition to the major aspects of Q-methodology.

Findings concerning the effect of ”sex of the model, " "age of the
model, ” and "color of the picture" upon identificatory reactions in men
and women of various age groups will be presented in Chapter IV.

A brief discussion of "types" of persons who identify with similar
models in reference to their actual self notions and "types" of persons
who identify with similar models in reference to their ideal self notions
follows (Chapter VI). Chapter VII contains some supplementary results.

The conclusions will be presented in Chapter VIII, and some
implications for practical advertising will be indicated within the
limits of generalizability. Some propositions for more rigorous investi-

gation and testing will be added.

CHAPTER I
BACKGROUND AND IMPORTANCE OF THE STUDY

Research shows that pictures of people appear in advertisements
with increasing frequency. Klapp studied this trend over a period of
40 years. In 1900 only 22% of all advertisements in four magazines
showed pictures of human beings, while the other 78% either showed
illustrations without human models or lacked pictures of any sort.

In 1915 38% of the advertisements in the same magazines incorporated
pictures of people. In 1930 the figure had grown to 55%, reaching the
considerable amount of 74% in 1940.1’

A comparable trend was found by Kitson and Allen in a somewhat
different selection of magazines. They noticed that, while in women's
magazines there was a predominance of female models in advertising
illustrations, in men‘s magazines there was a predominance of male
models. 2

In 1925, the same year in which the latter study was undertaken,
another investigation was made of the effectiveness of showing portraits
of people in advertising illustrations. For instance, Nixon undertook
a cOmparison of the attention-value of illustrations showing pictures of

people and the attention-value of pictures showing the advertised product

 

 

1O. E. Klapp, "Imitation -Value in Advertising, " Journal of
Applied Psychology, XXV (1941), pp. 243-50. *

 

2H. D. Kitson and 1. Allen, "Minor Studies in the Psychology of
Advertising from the Laboratory of Indiana University: XIII. Pictures
of People in Magazine Advertising, ” Journal of Applied Psychology, IX
(1925), pp. 367-70.

 

alone. Registering the sequence and duration of “eye—fixations" of

his subjects, he showed them a picture of each group simultaneously.
The results revealed that pictures of people more often received the
first "eye-fixation" and were looked at longer during a limited exposure
interval than pictures of the product alone. Furthermore, he found
that pictures of people were studied with increasing interest for a
considerable time and that they were better recalled at a later stage.
Thus, he concluded that advertisements showing portraits of people

1

are the more successful attention devices. A later study, concerned

with the impact of outdoor billboard advertising, supports these findings. 2
However, advertising textbooks still Show little interest in the

issue Ofwhy portraits of people in advertising illustrations increase

the ads' attention-value. Various authors have treated this matter

incidentally, particularly in reference to the problem of how the viewer

of the illustration is expected to relate himself to the depicted model.
Among the earliest comments on this issue is Scott's. He sug-

gests that the advertiser should provide the reader with a ”disposition

to action, " which will Operate whenever such action can be completed.

He explains that our actions are affected most by what is suggested to

us through the words and actions of others, particularly people we

 

1H. K. Nixon, “Two Studies of Attention to Advertisements, "
Journal of Applied Psychology, IX (1925), pp. 176-87. A similar find-
ing concerning the attention to pictures of people in comparison to
pictures of objects or animals in a study by: G. Gallup, Factors of
Reader Interest in 261 Advertisements (Published by Liberty Magazine,
1932), p. 50, reported in D. B. Lucas and S. H. Britt, Advertising
Psychology and Research (New York: McGraw-Hill Book Company,
Inc., 1950), pp. 273-74.

 

 

 

 

 

2H. E. Burtt and T. S. Crockett, "A Technique for Psycho—
logical Study of Poster Board Advertising and Some Preliminary
Results, " Journal of Applied Psychology, XII (1928), pp. 43-55.

 

know and whose judgments we reSpect. 1 This author, however, does
not point out how he expects the reader to relate himself to the model
so that his actions may become influenced by what the depicted model
does about the advertised product or idea.

In his aforementioned study, Klapp proposed that the reader re-

acts to the model through identification, which means that he implicitly

 

puts himself into the model's place and imagines having experiences
similar to the model's in connection with the product or idea. When
identification with a model occurs, imitation follows. Thus, Klapp
sees “imitation—value” as the main asset of portraits of people shown
using the advertised product. 2

Here, it seems to be expected that ”imitation” occurs on two
different levels. During the identification process, a covert imitation
is at issue, which eventually may become overt imitation when the
initial response leads to the proposed buying or adoption behavior.
In other words, our actions are affected by the actions of the model,
because once we have, in fantasy, experienced the benefits that come
from using a certain product (covert imitation), we will try to secure
this benefit for ourselves in reality (overt imitation).

The same point is discussed by Crawford, when he explains
what the art director is trying to elicit in the reader by means of the
advertising illustration:

The tendency today is towards realism in presentation--real
people doing real things with real products in real situations.
We try to create pictures of situations that the reader can

 

1W. D. Scott, The Psychology of Advertising, completely revised
edition by D. T. Howard (New York: Dodd, Mead, and Company, 1931),
pp. 183-85.

 

ZKlapp, 2p. Elf." p. 249.

project himself into, because this increases the opportunity
for the reader to understand how readily our product can fit
into his life. 1

The art director succeeds in creating an illustration into which
the reader can project himself insofar as he presents models with
which the reader may identify himself. These models may have
attributes or hold commodities which the reader shares with them;
thus the model's situation is characteristic of the actual circumstances
in which the reader finds himself. They may also have attributes or
commodities which the reader aspires to have for himself, but which
he has so far not acquired. In both cases, whether identification
occurs through comparison of the model's way of life with the reader's
actual way of life or through comparison of an idealized model with the
reader's i_cl_e_a_l_way of life, projection or identification may contribute
to the reader's understanding of how he can integrate the product into

his life. In the former case, such integration may confirm the status

 

g_l_l_g, while in the latter, change is brought about in the direction of
ideal circumstances.

Lucas and Britt, in agreement with Klapp, attribute the effective-
ness of pictures of people to an imitation-tendency in the reader. They
expect this tendency to be reinforced if the reader can also project
himself into or identify himself with one of the depicted characters.

They note that if envy were aroused in the reader, the effectiveness of

 

the advertisement could be expected to be further increased. 2 In this
case, when the model arouses envy in the reader, the identification is

likely to take place in connection with the reader's ideal way of life:

 

 

1J. W. Crawford, Advertising: Communications for Management
(Boston: Allyn and Bacon, Inc., 1960), p. 225.

 

2D. B. Lucas and S. H. Britt, Advertising Psychology and
Research (New York: McGraw-Hill Book COmpany, Inc., 1950), p. 271.

 

 

10

when an advertisement shows a picture of someone hand-
some and important-looking, you identify yourself in your mind
as that person, and you become handsome and important-looking.l

On the other hand, the identification can take place on the basis of the
actual characteristics of the reader.

In his Q-methodology and Q-technique, Stephenson offers a suit-
able approach and instrument for the investigation of how individual
persons, in the complexity of their actual and ideal characteristics,
relate themselves to various kinds of messages. Because of its interest
in individual human beings, it differs from traditional approaches
where certain characteristics or attributes are measured in isolation
for "populations" of people under the assumption that each person has

the attribute to some extent. By means of Q-technique we do not study

 

the impact an advertising illustration has upon its total audience, but

rather, how it is grasped by individual persons in this audience. 2
In a recent manuscript, in which Stephenson discusses communi-

cation theory and research in the light of Q-methodology, the phenomenon

of identification is extensively discussed.

Fundamentally, the whole domain of the mass media is one
essentially of fantasy . . . [The mass media] are engaged in a
vast game, into which audiences project, identifying with this
or that, and enjoying this or that in so far as they so identify.3

The importance of the identificatory reaction in the reader lies
in the fact that the reader, through identification, becomes "apperceptive"

for the message intended to persuade him. Persons are selective about

 

1M. S. Hattwick, How to Use Psychology for Better Advertising
(New York: Prentice Hall, Inc., 1950), p. 160.

 

2Stephenson, o__p_. ii}, p. 58.

3‘Stephenson, A General Theory of Mass Communication (unpub-
lished manuscript for a book, l960),p . 129.

 

11

who they identify with, so we need to know which models. are likely to
stimulate identification in whom. Furthermore, we need to know which
persons in the audience identify with a given model and which do not
identify with him. According to Stephenson, there is no homogeneous

l

audience which reacts to advertising messages. By means of Q-studies

of readers it is possible to find "types” of persons which lead to a

 

segmentation of the audience. Each "type” would include those persons
having similar selective criteria in choosing their objects of identifica-
tion. In order to formulate more successful advertising messages, one
should know about prevailing audience segmentations.

. with respect to the mass media, a vast domain of identifi-
cation-possibilities exist which the advertiser . . . [and others]
must come to terms with if they wish to communicate, and in
tune, with the audience of convergent selectivity.z

The present study is concerned with the identificatory relationship
between the individual reader and the photographic representations of
people in advertising illustrations. Two lines of investigation centered
upon the following questions:

1. Which characteristics of the depicted models facilitate
identification in readers of different age and sex?

2. What kind of audience segmentation is possible on the
basis of grouping persons into particular “types" accord-
ing to their identification with similar models ?

 

As shall be explained in more detail, Stephenson's Q-technique
makes the investigation of these questions possible, as it helps to
operationalize the identification process. It also allows for a distinction
to be made between identification in reference to actual and ideal self
notions. The importance of this distinction will be pointed out more

carefully later .

 

1Ibid., p. 57.

 

2Ibid., p. 118.

 

12

Although we would expect a relationship between the identification-
variable and the attention-variable to prevail, the assessment of the
extent of such a relationship has to be postponed to further investigations.

Questions may also be raised concerning the selling-effect of the
identification mechanism, which means, the more intense the identifi-
cation, the greater the persuasive effect becomes. We have limited
identification to the reader's reaction to the model in the picture,
which is only a section of the total advertisement. On the other hand,
the persuasive effect of the advertisement may be due not only to the
illustration, but also to headlines, text, and the relevance of all three
to the advertised product. Hence, a positive correlation between the
identification-variable and the persuasion-variable is difficult to obtain.

This indicates the major limitation of the present study, namely,
that from the effectiveness of certain models in stimulating identifi-
cation in certain groups of persons, we cannot infer that the advertise-

ments in which the models appear will fulfill their purpose. 1

 

1"But the illustration must also be relevant to the product and to
the idea of the advertisement, so that it can help select out of the great
mass of readers those who can be turned into customers, right now.
One of the easiest ways to get attention is to picture a pretty girl--and
the fewer clothes she is wearing the more attention the picture gets.
But if the product we are selling is spark plugs, or hammers, or house
paint, the viewers we get for the pretty girl are not necessarily con-
verted into readers for our advertising message. The illustration must
be relevant; it must increase the reader's interest in the product,
intensity his desire to own it, persuade him to want it, and contribute
to the buying decision. " Crawford, pp. 9.1.3" pp. 222-23.

CHAPTER II

THE BASIC FRAMEWORK

This chapter introduces the basic framework of the study. The
concepts used during the investigation are defined and the assumptions
are made explicit. In conclusion, the identification mechanism will be

discus sed at length.

Definition of Concepts and Assumptions

 

Basic to this investigation of the identificatory relationship be-
tween the viewer and a model in an advertising picture is the assumption
that the careful viewer of the illustration will at least momentarily
identify with the depicted model.

Like Maccoby and Wilson, we mean by identification that the
viewer, in fantasy, takes the place of the model and momentarily feels
that what is happening to the model is happening to himself. Thus, he
covertly reproduces behavior-elements and emotions he attributes to
the model with whom identification is said to occur. I

The viewer's diSposition to imitate the model in fantasy (and, on

occasion, in reality) is likely to be enhanced only if the model shares

certain "qualities” with the viewer. 2 Now, if an individual's

 

1E. E. Maccoby and W. C. Wilson, ”Identification and Observ-
ational Learning from Films, " Journal of Abnormal and Social
Psychology, LV (1957), pp. 76-87.

 

 

ZS. Freud, Group Psychology and the Analysis of the Ego (London:
The Hogarth Press and The Institute of Psycho-Analysis, 1949), p. 149.
Here a distinction is made between identification with persons

 

l3

14

organization of "qualities" defines his self, 1 then the choice of the
object with which the individual identifies himself depends upon his
self-notions. This explains why the investigation of the _s_e_1£, both
actual and ideal, is crucial to an understanding of the identification
mechanism.

In addition, the disposition for identification may be enhanced
when the model is shown performing "actions" which the viewer him-
self can occasionally perform. In this case, there is a correspondence

in roles.2

 

Each viewer plays a multitude of roles according to the positions
he occupies in everyday life. The set of behaviors characterizing each
role is dependent upon the counterparts with whom the individual inter-
acts, upon their mutual expectations in particular contexts, and upon
their attitude toward the prevailing group norms.3 For instance, one

may play different roles with his child and with his employer. Thus, in

who are objects of a sexual instinct and those who are not. Concerning
the latter case it is said that identification ”. . . may arise with every
new perception of a common quality shared with some other person who
is not an object of the sexual instinct. The most important this common
quality is, the more successful may this partial identification become,
and it may thus represent the beginning of a new tie. "

1T. R. Sarbin, "Role Theory, ” in G. Lindzey, (ed.), Handbook of

 

Social Psychology (Cambridge, Mass.: Addison-Wesley Publishing
Company, Inc., 1954), I, p. 244. ”. . . the self is what the person
‘is, ' . . . When interested in the self, we regard the person as an

 

organization of qualities. ”

ZIbid. , ". . . the role is what a person 'does' . . . When we
study roles, we regard the person as an organization of acts. Paren-
thetically, direct observation reveals only action systems resulting '
from the interplay of self and role. "

 

 

3‘N. Gross, W. S. Mason,“ and A. W. McEachern, Explorations
in Role Analysis: Studies of the School Superintendency Role (New York:
John Wiley and Sons, Inc., 1958), p. 48. See also, J. W. Thibaut,
and H. H. Kelley, The Social Psychology of Groups (London: John
Wiley and Sons, Inc., 1959), p. 143.

 

 

 

15

his role as a father, one may identify with a man in a picture who is
playing with his child, while in his role as an employee, one may
identify with a model shown in the office of his boss. This way it is
possible to assume that a person can identify with a number of different
models which are presented in advertising pictures.

Furthermore, for each person we will distinguish between an

agtual self and an ideal self. Actual self means that set of qualities

 

 

which the individual perceives as characteristic of himself at the time
of the interview, while ideal self means that set of qualities which the
individual would like to be characteristic of himself. The two self-
notions may differ in congruence, as has been pointed out by Butler
and Haigh. 1

Similarly we will discriminate between an ideal role-configuration

 

and an actual role-configuration. By the former, we mean the multitude

 

of roles the individual would ideally like to play; by the latter, we mean
the set of roles the individual actually plays in his interrelationships
with other persons. Congruence between these role-configurations
would be dependent upon how closely the individual perceives his actual
way of life as approaching his ideal way of life.

Considering this distinction, an elaboration of the previous as sump-
tion becomes necessary. Now, it is expected that a viewer can identify
with a person in a picture by either comparing the model's role to one of
his actual roles, or by comparing it to one of his ideal roles. In doing

so, he will implicitly relate his own actual and ideal qualities to those

 

1J. M. Butler and G. v. Haigh, ”Changes in the Relation Between
Self-Concepts and Ideal Concepts Consequent upon Client—Centered Coun-
seling, " in C. R. Rogers and R. F. Dymond (eds.), Psychotherapy and
Personality Change (Chicago: University of Chicago Press, 1954), pp.

55‘760
Here it was found that client-centered therapy achieves a decrease

in discrepancies among, what we are calling, actual self and ideal self

 

 

notions .

16

he perceives in the model. If, for instance, a factory worker sees a
picture of a sun-tanned, well-to-do gentleman enjoying his leisure on
a fancy beach, he may wish to play that role, although it is remote
from his actual way of life.

Thus identification is a matter of degree. The same model may
elicit intense identification in reference to the ideal self, while it
stimulates negligible identification in reference to the actual self, and

vice versa. Or, if only one self notion is considered, one model may

 

elicit more intense identification than others do.

This leads into the last assumption: that peOpIe are able to com-
pare different intensities of identification, and thus can be asked to
rank pictures showing human models, with ”intensity of identification”

as a criterion. I

Personal Identification in the Communication Process

 

As Crawford has defined it, "advertising is the art of persuading
people to do with frequency and in large numbers something you want

them to do. "1

Advertising thus provides a vehicle by means of which
the seller of the product enters into communication with the potential
buyers and promises to satisfy their needs and wants by accelerating
the circulation of the product in the market.

According to a popular formula in the advertising business, an
advertisement should request attention, arouse interest, stimulate
desire, and lead the reader to buy the product in order to achieve its
purpose as a persuasive message.

The communication process relating the seller of a product to

the potential buyer may be said to involve three different stages which

follow each other in time. First, there are those who have the product

 

lCrawford, op. cit., p. 4.

17

for sale. As they cannot make immediate contact with the large number
of potential consumers, they delegate their functions as a persuasive
"source"--to a copywriter who produces an advertising message.

The copywriter may again delegate some of his functions as a
persuasive source by incorporating into hisadvertisement a famous
personality, who is shown enjoying the benefits of smoking a new brand
of filter cigarettes. The model in the illustration is the person the
reader of the advertisement associates with the product, when his
attention, interest, desire, and actions are stimulated.

Thus, the stages of the communication process range first, from
the seller to the copywriter, second, from the COpywriter to the model
shown in the message and, third, from the model shown in the message
to the reader of the advertisement.

Our main concern lies in this third stage of the communication

process, that is, in the message-receiver relationship. It should be

 

stressed that the model depicted in the message may be thought of as
a source, to which the viewer relates himself when he is shown the
advertising photograph. This fictitious source, which has a highly
persuasive character, could, in addition, be thought by the unSOphisti—
cated receiver to be the real source of the message. As was suggested
earlier, this model is characterized by certain qualities and by the
suggestion of a role—behavior which is chosen in such a way that it
could be productive of opinion change in the viewer, and thus lead him
to the adoption of the new product.

According to Hovland (it a}. , changes in opinion follow new learn-

1

ing experiences, which have to be motivated by the message. An adver—

tisement may persuade the receiver to adOpt a favorable opinion toward

 

1C. I. Hovland, I. L. Janis, and H. H. Kelley, Communication
and Persuasion: Psychological Studies in Opinion Change (New Haven:
Yale University Press, 1953), p. 10.

 

 

18

a new product when it provides him with a motivation to learn what

the benefits of using the product are. In choosing the model in the
picture, one motive can be provided by stressing the receiver's group-
membership. This way, the model shown may remind the reader of
his actual and/or ideal reference-group from which he derives gratifi-
cations for obeying the prevailing group norms.

Thus, not only successful businessmen may switch to a new brand
of cigarettes introduced through advertising as ”The Executive's
Cigarette, ” but also the young fellow who is just starting his business
career. While the decision of the business executive has to be traced
to his achieved group-membership, the decision of the young fellow
was motivated by his aspired group-membership.

If a new learning experience is necessary to make the receiver
shift from his initial opinion to a recommended opinion, we should
suppose that identification with the depicted model provides the receiver
with at least some of the required learning experience. When the young
fellow identifies with the successful business executive in an advertising
photograph and thus reproduces in fantasy the model's behavior and
emotions, his efforts to become a business executive himself may be-
come intensified, and he may, in the meanwhile, start to acquire the
external symbols of such a role, such as, ”The Business Executive's
Cigarette. "

Intensity of identification and persuasibility depend, on the one
hand, upon the depicted model's qualities and role and, on the other
hand, upon the receiver's qualities and roles, or his ”predispositions, "1
which actually or ideally characterize him and which underlie his

responsiveness to the advertising message.

 

1Ibid., p. 12.

19

We expected that similarities in the model and in the reader en-

 

hance identification, provided that these similarities are restricted

to actions and qualities accepted by the reader's actual or ideal reference
groups. As affiliation with the aspired reference-group is likely to
satisfy more needs and wants in the reader than the already accomp-
lished affiliation with the actual reference-group, ”role-idealization"
may be a variable worth the attention of the art director and the copy-
writer when they are choosing a model. But, as an ideal role is only

an ideal role for those who aspire to it, the copywriter must know the
segmentation of his audience with respect to choice of ideal objects of

identification.

Three Different Ways of Identification

 

The identification process may take place in many different ways.
Distinctions could be made on the basis of whether the object of identifi—
Cation is an individual or a model and whether the reference is to the
' actual or the ideal self. In this study, two ways in which the identifi-
cation process can be actualized are of specific relevance, that is,

the identification with a model in reference to one's actual or ideal

 

self.
Yet, in order to outline the generality of the identification
mechanism, the basic process of identifying with an individual in

reference to one's actual self is also set forth.1

 

The first model in FIG. I shows John in his actual role-configur-
ation which includes the roles of a son, a brother, a baseball player,

and a lazy pupil. In his role of a lazy pupil, John is facing his teacher,

 

1Though in the present context only the basic process of identify-
ing with an individual in reference to one's actual self is outlined for
purposes of comparison, it is clear that also the parallel process of
identifying with an individual in reference to one's ideal self could be

 

 

 

shown in a model.

20

FIG. I: THREE WA 8 OF IDEN'I’IFICA'I'ION

1. Actual Self Identification with an Individual

 

k--~-O~—_--—

h-—--— - -‘*---—~———--—

Mun but of Ream ignition

 

3. Ideal Self Identification with a Model

   

fast—mm§--'um-n-tfl

//";.;s\\ 5 ‘E
'. ,\\ u :

uh! ‘x. " fl
*1! ‘ . . ‘. u [I I
»-——-W::. - . t
“a hr 9‘ \ .
\Mcv . / {j \ I
wwznt‘v \‘ / u 1i
\' / [Ln-33;-nnsznuzzn'c-qu

John's Mud“ 515; Link Bud “(1.5.455 W“: W a! cum: fimtion

21

Mr. Smith, whose role-configuration includes, in addition to his role
as a teacher, the roles of a husband, a father, and of a pianist. While
the teacher expects his pupil to be studious, the latter prefers not to
be bothered. These mutual expectations characterizing their role-
relationship may have been made explicit, _or they may be guessed by
each partner. In each case, these expectations affect John's perceptions
of his teacher, and Mr. Smith's perceptions of his pupil.

John, being a lazy pupil, probably will not identify with his
teacher, as illustrated by the short bold arrow in the model. He will
have difficulty imagining1 himself in the place of the teacher because
the two are very dissimilar, John being lazy and Mr. Smith being
ambitious. Furthermore, the pupil and the teacher may find their
sources of gratification in different groups to which they reSpectively
belong, as indicated by the lack of correSpondence in their mutual
expectations.

The second model again shows John in his actual role-configuration,

 

only now he takes the role of an interviewee who is asked how intensely
he identifies with an anonymous teacher portrayed in a picture. As John
is now not performing his potential role of a pupil, he has to project

himself onto a fictitious level where, in his imagination, he actualizes

 

1By the verb ”to imagine” is meant ”to form a mental image of;
to represent 'or picture to oneself; to produce by the imagination. "
Accordingly, by "imaginatiOn" is meant "the act or power of imagining;
formation of mental images of objects not present to the senses, esp.
of those never perceived in their entirety; hence, mental synthesis of
new ideas from elements experienced separately. " Definitions from:
Webster's New Collegiate Dictionary (London: C. Bell 8:. Sons, Ltd. ,
1951), p. 414.

John's act of imagination may depend upon an "inferential set, "
which, according to E. E. Jones and R. de'Charms, is assumed to have
selective as well as transformative functions in the process of forming
impressions of personality. ”The Organizing Function of Interaction
Roles in Person Perception, " Journal of Abnormal and Social Psychology,
LVII (1958), pp. 155-164.

 

 

 

 

22

this role by drawing upon his previous experiences. On this fictitious
level, John imagines not only what he would expect from this teacher,
who is not Mr. Smith, but also what this anonymous teacher would
expect from him. The teacher in the picture may look lenient and
easy-going in comparison to Mr. Smith, so that John may think that
this teacher would not demand much effort from his pupils. Thus, he
is likely to identify more intensely with the depicted teacher than with
Mr. Smith.

The first model differs from the second one in two respects.
First, in the case where persons interact, there is an immediate role-
correSpondence between the role-incumbent and his counterpart, which
is not present when a person is confronted with a model in a picture.

In this latter case, the person has to select a suitable, latent role from
his total role-repertoire, and he has to imagine himself playing this
role in a fictitious interaction with the depicted model.

Second, each individual in the first model has his own expectations
vis-a-vis his counterpart, and he actually communicates to him; in the
second model, such mutual expectations are absent, just because the
person and the depicted model have never interacted in reality.
Drawing upon his previous experiences with various teachers, the pupil,
as a). role-incumbent in a fictitious setting, has to attribute certain
expectations to his counterpart, which he thinks the latter may have
vis-a-vis him.

The third model shows John when he is asked how much he
identifies with the same portrait of the lenient and easy-going teacher

in reference to his ideal set of roles. John again is shown in his actual

 

role—configuration, presently playing the role of an interviewee. In
order to answer the question, a preliminary step of imagination is re-

quired, that is, John has to imagine his ideal set of roles. Besides the

 

roles of a baseball star, a soldier, a, father, and a brother, even the

23

role of a lenient and easy-going teacher may be included in his ideal

 

repertoire.

From this ideal set of roles, John will select the role of a
teacher, and, in another step of imagination, he will actualize this role
in a fictitious interaction situation similar to the one described in the
second model. Now that a correspondence in roles as well as in
personal attributes is achieved, John may attribute to his-fictitious
counterpart the same expectations he himself would have vis—a-vis him.
When John also thinks that he and his counterpart would find their
sources of gratification in the same reference-group, identification at
the level of his ideal self is likely to become very intense.

Thus, the third model differs from the second one in that, when-
ever a person relates a depicted model to his ideal self, an additional
step of imagination is required. Although it seems that, whenever
John identifies with the portrait of another person, more complex
identification processes are at issue, in this case a simplification is
actually involved. 1

First, the complexity of the stimulus is reduced, because John
hardly knows about the other roles performed by the depicted model
and thus has to limit his role-comparison to the role the model overtly
enacts in the picture. Second, John's problem of guessing what role-
expectations the identificand may actually have vis-a-vis him is elimi-
nated: only those expectations which he himself attributes to the model
in the fictitious interaction context will prevail. Third, Mr. Smith's
expectations vis-A-vis John may change over time and simultaneously
John's degree of identification may vary over time. However, no such

variability in identification occurs when a person's portrait is involved.

CHAPTER III

IDENTIFICATION AND MODELS

In this chapter, a few studies are discussed which have some
bearing upon the issues under investigation. The hypotheses are
introduced, experimental controls are pointed out, and the limitations

of the study are indicated.

Implications from Previous Studies

 

There are few studies in the advertising field, which go beyond
a comparison of readership scores, to askm certain advertisements
have a greater impact than others.

Among these few studies, there is a multiple factor analytic
investigation by Twedt, which justifies to some extent our intention to
consider the pictorial part of each advertisement in isolation. Twedt
defined and measured 34 variables in business magazine advertising,
and correlated these with readership scores provided by the Advertising
Research Foundation. A factor analysis, performed on thOse variables
which were found to be significantly interrelated with the readership

variable, identified six factors. One of these factors, namely, the

"Pictorial-Color" factor,’ alone explained 41% of the variance in reader—

 

ship scores. The ”Pictorial—Color" factor was composed of the cluster
of "readership, " "square inches of illustration, " ”number of pictures

showing product in use, " and “number of colors" and other variables.

24

25

While a second, "S_i_z_e-, " also had considerable loadings on the reader-
ship variable, loadings by any other factors were found negligible.l
These findings indicate the impact of pictorial variables upon
readership scores, and suggest how carefully they have to be taken
into consideration, especially in the case of advertising illustrations
showing pictures of human models.
Recent studies have some bearing upon one aspect of the problem

investigated in this study, that is, whether the similarity of character-

 

istics in the reader and the depicted model does or does not facilitate
identification.

Rudolph reports that illustration—subjects most effective in
attracting women are, in the following order, pictures of babies,
pictures of children, pictures of women with children, pictures of
woman alone. Attractive to men rather than women are pictures of
men, pictures of objects, pictures of men and women.2 The phenomenon,
that male models appeal strongly to men and female models appeal
strongly to women, is not explained by the author. However, Lucas and
Britt, as well as Hattwick, suggest that sex-correspondence facilitates
projection or identification and, hence, increases the appeal of a model

for the like-sexed audience.3

 

1D. W. Twedt, "A Multiple Factor Analysis of Advertising Reader-
ship, " Journal of Applied Psychology, XXXVI (1952), pp. 207-15.

Criterion scores from the Advertising Research Foundation's
continuing studies of business magazine readership were of the type
"Any this AD, " that is, percentages of people who remembered reading
or seeing any part of the advertisement. --Readership had negligible
loadings on factors: T (Typographic Size and Variety), In (Informational),
F (Field Factor or the Influence of Surrounding Field of Ad. ), and A
(Advertising Schedule Previously Run).

 

211. J. Rudolph, Attention and Interest Factors in Advertising:
Survey. Analysis. Interpretation. (New York: Funk & Wagnalls Com-
pany with Printers' Ink Publishing Company, Inc. , 1947), pp. 66-89.

 

 

3’Lucas and Britt, op. cit., p. 274. Also, Hattwick, op. C_i_t.,
p. 160.

26

In a study outside the advertising field, Maccoby and Wilson were
concerned with the identification variable in adolescents' reactions to
leading characters in films. They investigated if similarity in sex and
in socio-economic status in their subjects and in the leading characters
was a criterion of choosing the identificand. 1

Three major findings were published. First, in films with a
female and a male protagonist, viewers identified themselves with the
like-sexed leading character. This tends to support the aforementioned
interpretation of Rudolph's findings.

Second, in films showing a male protagonist of lower socio-
economic status and a male protagonist of upper socio-economic status,
the protagonist chosen as identificand was the one whose social class
corresponded to the subject's aSpired social class rather than his
current status. This result, which represents a break with the
similarity criterion, was unexpected by the authors. Yet, in our
opinion, this result may be due to their disregard of a distinction

_ between identification in reference to actual self notions and ideal

 

self notions. Maccoby and Wilson's index of identification was based
on which protagonist's name appeared in two out of three answers to
the following questions:

1. Which of the two main characters did you like best?
2. Which of the two main characters would you like to be like?
3. Which of the two main characters do you feel is most like you?‘2

The third question appeals to actual self notions, and here a lower
class interviewee may have given the name of the lower class protagonist.
The second question appeals to ideal self notions, and thus may have
been answered with the name of the protagonist in the aspired social

class. The first question becomes the decisive one and is probably

 

1Maccoby and Wilson, pp. SE.

ZIbid.

27

biased in favor of ideal self involvement because it is likely that the

 

interviewee is oriented toward the character who represents his
aspired class rather than the one who represents his own unsatisfactory
group affiliation.

Therefore, a break with the similarity criterion may not have
occurred if only actual self notions had been tapped for all answers.

As soon as ideal self notions enter into consideration, the qualities of

 

the identificand have to be compared with the ideal qualities of the

 

subject, and, at this level, the similarity criterion still is likely to be

operative. As the authors compared qualities of an ideal identificand

 

with the actual qualities of the subjects, the similarity criterion did not
seem to prevail. This may emphasize the importance of the distinction
between the two different kinds of identification processes in the present
context.

The third finding of Maccoby and Wilson was that peOpIe tended
to better remember actions and words of the character with whom they
identified. Although this point would have to be tested in the advertis-
ing situation, it strengthens the expectation that people are more likely
to learn the message conveyed once they havefthrough identification,
covertly reproduced actions and emotions of the model presented in an
advertising illustration. The interrelationship of the identification-
variable with the learning-variable, however, is left to further investi-
gations.

Two other studies bear upon the second aspect of the problem

investigated in this study which concerns possible kinds of audience

 

segmentation obtained when persons are grouped into "types" according

 

to their identification with similar models.
MacLean and Hazard have studied women's interest in a variety

of pictures from Time and U. S. News & World Report. They also

 

 

 

dealt with the message-receiver interrelationship, but, while they

28

approached the issue of differences in interest scores from the

message's side, we have approached the similar issue Of differences

 

in identification scores from the reader's side. In this study, "types"
of actual and ideal selves of persons were assessed on the basis of the
variance in identification scores, while the MacLean and Hazard study
sought "types" of picture appeals on the basis of variance in interest
scores.

Six major types of picture appeals were found by MacLean and
Hazard. Among these are three which may also be responsible for
interest variations (and, perhaps, also identification variations) in
advertising illustrations:

I. Idolatry: successful, glamorous, wealthy women, especially
in "happy family" situation. Adolescent girl's dream.

III. Picturesque: salon pictures, moody pictures, dynamic
pictures, pictures portraying loneliness and escape from
other pe0ple, doing things with hands.

 

VI. Spectator Sports: action pictures of sports which command

 

big audiences. 1

The three omitted types of picture appeals (II. Social Problems,

 

IV. War, V. Blood and Violence) are more likely to be found in news

 

 

photographs than in advertising photographs. The present study will
take up later the question whether there are ”types" of persons who
identify with pictures where one of the three aforementioned appeals
predominates.

Another very recent study approximates closely the problem of
how an audience can be segmented according to identification with

models in advertising pictures. Social Research, Inc. , assessed,

 

1M. S. MacLean, Jr. and W. R. Hazard, "Women's Interest In
Pictures: The Badger Village Study, " Journalism Quarterly, XXX
(1953), pp. 139-62. '

 

29

by means of depth interviews, women's reactions to certain types of
models. Here is a summary of their major findings:

1. The fun-loving, out-door kind of girl--the All-American type--
was not appreciated by most women for being "unladylike. "
However, if these models appeared with children, their appeal
was increased.

2. The college co-ed was an effective model for those to whom
college is somewhat more than a distant ideal, because it
suggests good up- bringing.

3. The fashion model appeared to represent an ideal role for
younger women, as well as for older women of higher income.

4. Only as a saleswoman for products involving self-indulgence
or any creative activity was the suburban socialite a success-
ful model.

5. The executive wife had an appeal when she was shown in
relation to decision-making problems. 1

Although a detailed report on the project is not yet available,
identification mechanisms seem to be the main issue. Again, these
partial results seem to suggest that a more rigorous distinction between
identification in reference to actual and ideal self notions should be
made. For instance, it may be found that the All-American type of
girl is not an _i_d_e_a_l object of identification, but that it describes well

the actual way of life particularly of the younger ones among the sub;

 

jects. On the other hand, the fashion model may correspond to the

ideal self notions rather than to the actual) self notions.

 

This crucial point once more should be emphasized: one may
find a very different audience segmentation when one is concerned with
the impact of certain models upon the actual self notions of the readers
and upon the ideal self notions of the readers. If the prevalence of such

audience segmentations were supported by research findings, it would

 

1R. Alden, "Advertising: Survey Looks at the Models, " The New
York Times, April 11, 1961.

 

30

mean that the copywriter who is interested in identification mechanisms,
should be aware of the two different identification processes between
which he may choose. One person may be stimulated by an exceptionally
handsome model in an “idealized role, " while another may be stimu-
lated by a not-too-handsome model in everyday circumstances.

An ideal role could be an ideal role only for certain peOple, and,
at the same time, be an actually played role for others who may or
may not find it very ideal. Therefore, once the copywriter has decided
upon which self notion he wishes to appeal to, he needs to know about
the potential buyers of the product before he can select the model which

is likely to appeal to the maximum number of self notions involved.

A Statement of Propositions

 

The first part of this pilot study is concerned with the question
of whether similarity of characteristics of the reader of an advertise-
ment and of the model shown in the illustration can facilitate identifi-
cation. In other words, this study is an attempt to replicate Maccoby
and Wilson's findings concerning similarity as a criterion of choice of
the identificand in the advertising context. I

As in their study, the impact of sex—correspondence upon identifi-
cation was investigated. But, instead of studying their second character-
istic, which was socio-economic level, the effect of age-correspondence
upon identification was investigated.

Hence, the main hypotheses are:

1. Persons tend to identify with like-sexed models rather than

 

with models of the opposite sex.

 

2. Persons tend to identify with models of their own age-group
rather than with models of different age—groups.

 

 

 

1Maccoloy and Wilson, op. cit.

31

It was expected that pictures showing a man as well as a woman
would obtain somewhat higher average identification scores in an
audience composed of men and women because identification would be
facilitated for persons of both sexes.1

Corresponding to the above hypotheses, the statistical (or null)
hypotheses were the following:

1. Sex of models in advertising pictures has no effect upon
identification in men and women.

2. Age of models in pictures has no effect upon identification
in persons of different age-groups.

In the second part of this pilot study, an exploratory procedure

 

was adopted. Because, at this stage, theory and research do not pro-
vide a satisfactory basis for the deduction of hypotheses concerning a
segmentation of the audience with respect to identification with models
in pictures, no hypotheses were stated beforehand. The intention was
to discover hypotheses for further, more rigorous testing.2

Some of the questions, which we tried to answer, were: Are
there "types" of persons who identify with different kinds of models in
pictures? Are there persons who are of the same ”type” in their
preference for certain models, both in reference to their actual selves

and to their ideal selves? Which models elicit most and least intense

 

 

 

1Rudolph, op. cit., Findings from this study indicate that, while
men are particuTsz'ly—ihterested in male models, and women in female
models, men and women are about equally interested in pictures show-
ing a male and a female model together.

ZStephenson, The Study of Behavior, p. 329. "An overweening
curiosity, we said, rather than reasoning within a formal hypothetico-
deductive framework only, is to be welcomed and encouraged. An ex-
periment, therefore, is much more than the fulfillment of predictions,
or of the empirical testing of previously asserted prOpositions. It is
an opportunity for making discoveries. ”

 

32

identification with respect to each of the self notions? Which models
appeal with similar intensity to both self notions, and which appeal
intensely to one self notion but not to the other? About which models
do men and women agree best, and about which ones do they disagree
most in their identification? Which models cause most disagreement
and most agreement in terms of identification among persons belonging

to different age-groups?

Experimental Controls

 

Research shows that an important variable in an advertising
illustration is whether it is colored or black and white. It is main-

tained that the value of color rests in making the depicted models or

 

products look more "realistic." The use of four color reproductions
is reported to affect the attention-value, therefore producing a signifi-

1 Colored advertisements were

cant increase in the size of the audience.
also found to be more successful in respect to recall as measured by
naming the advertised product. 2

From this evidence it can be expected that the color variable
may also affect identification, as realism in the presentation of a

model is likely to facilitate the reader's imagination of his interaction

with the depicted model. 3

 

1Rudolph, op. cit., p. 37.. See also, D. B. Lucas, The Con-
trolled Recogniti-OD MEhOd for Checking Magazine Advertising Reader-
ship (New York: The Crowell-Collier Publishing Company, 1942),
pp. 1-6, reported in Lucas and Britt, 32. SE." pp. 303-304.

 

2Nixon, op. cit., p. 186. See algo, H. N. Gilbert, "An Experi-
ment with Colored and Uncolored Advertisements, " Journal of Applied
Psychology, XVII (1933), pp. 49-54.

 

 

3Refer to FIG. 1, Models 2 and 3, p. 20, supra.

33

In this study the effect of color was controlled by incorporating
it into the factorial design underlying the picture sample, which will
be presented in the next chapter. Through this procedure, the effect
becomes controlled without being eliminated, thus indicating its enhanc-
ing or restricting influence upon identification in reference to actual
and ideal self notions.

The size of the total advertisement (including headlines, illus-
tration and COpy), as well as the size of the illustration itself was found

1 Because it is difficult to control both

to affect readership scores.
variables simultaneously, in this research only the size of the total
advertisement has been standardized through the selection of one-page
advertisements only. 7‘ The size of the illustration differed across
advertisements. The effect of the position variable, that is, whether
the advertisement was published on a left hand or right hand page in
the original magazine, was eliminated through the mounting of pictures
on separate cardboards.

As the subjects were presented with the pictures in their original
context, other elements of the advertisements, such as headlines,

sub-headlines, captions, logotypes, and signatures3

may have affected
identification with the depicted models. An attempt was made to
reduce such an influence by asking subjects to consider the pictorial

element alone in estimating intensities of identification. Considering

 

the Speed at which Q-sorts were performed, it is very unlikely that

 

1Twedt, op. c_:_i_t_.

2In choosing full-page advertisements the small difference in
format of Life and Look magazines was neglected. Thus a standardiza-
tion of size of the advertisements was only approximative.

 

 

3In enumerating the elements of an advertisement we follow
Crawford, op. cit. , pp. 163-164, where also the necessary explanations
are to be found.

34

copy was read, although headlines and logotypes may have been con-
sidered by the subjects.

We are aware that, in a more rigorous study, better controls
for non-pictorial elements are necessary. Possibly, the pictorial
elements Should be presented in isolation, but this is often difficult,
as COpy and headlines may run into the pictorial field. The restriction
Of the stimuli to the illustration alone also seems to indicate the
necessity Of a control for the size of the illustration rather than for

the size of the advertisement as a whole.

Limitations of the Study

 

In the interpretation and generalization of the results from this
study, one needs to be careful for several reasons:

1. When one studies people's identification with models in adver-
tising pictures and when one finds that some features in models appeal
more than Others to certain "types" of persons, one cannot predict the
effectiveness of the total advertisement in which the model appeared.
As the impact of an advertisement depends upon the successful combi-
nation of illustration, headlines, captions, logotype, and signature,
one should not infer from the effectiveness of an isolated element, such
as the illustration, the effectiveness of the advertisement as a whole.
For instance, an illustration which is a successful attention—device
with reference to a ”Thunderbird" may be a poor one with reference to

"Lux Detergents, " even though identification with the model could be

 

the same.

2. One should be careful in generalizing results concerning
certain pictures. If, for instance, one finds that the model depicted
in one illustration Obtained the highest average identification score

with reSpect to either the actual or the ideal self notions, one should

35

keep in mind that this picture is the most effective stimulus of identi-

fication only in the set of pictures studied and for the persons inter-

 

viewed.

3. In view Of the fact that this Q- study investigates only a small
number of men and women of different age—groups, who were not meant
to be representative of a defined population, a distinction must be
made between the generalizability of two kinds of results.

On the one hand, findings like "Men tend to identify with male
models and women tend to identify with female models" can be general—
ized, because such sex—distinction in the choice of objects of identifi-
cation is likely to hold whenever men and women are compared. Here
the concern is not so much with specific intensities of identification as
with two manifestations of the same process.

On the other hand, findings like "Men and women identified most

intensely with one model in a set of pictures" should not be generalized

 

to larger audiences composed of man and women, because here it
would be neceSsary for the small audience studied to be representative
of the larger audiences. However, results of the latter kind may be

of interest when they are evaluated with caution.

CHAPTER IV

THE DESIGN OF THE Q-STUDY

Before some of the characteristics of Stephenson's Q-methodology
are pointed out, it will be explained how the Q-sample was composed,
what the Q-sorting instruCLions were, and how each subject performed
the Q-sorts. I

A Q—sample can be composed of either statements, which most
frequently has been the case in Q-studles, or pictures, each of which

must have some reference to the self, so that a subject can describe

 

himself by ranking the items according to a scale ranging from "most
similar'' to "least similar. ” The present Q- sample of 72 advertising
photographs was chosen from a Q-population defined as all full-page

advertisements in recent2 issues of Look and Life magazines showing

 

 

portraits of one or more persons in the ad illustrations.
The picture sample was stratified according to three variables:

color of the advertisement, sex of the model, and age of the model.

 

 

 

As indicated in the factorial design reproduced in TABLE 1, a dis-
tinction was made between colored and black and white pictures. In

reSpect to the sex variable, one group of pictures showed a man as

 

1For a quick introduction to Q-technique and its major methodo-
logical implications, refer to W. Stephenson, "Some Observations on
Q-Technique, " Psychological Bulletin, XL (1952), pp. 183-98. For a
detailed study refer to Stephenson, The Study OfLBehavior, 2P; (_:_i_t_.
For discussion and criticism of the approach, vylere persons instead of
tests are correlated, see C. Burt, The Factors of the Mind (London:
London University Press, 1946), and R. B. Cattell, Factor Analysis
(New York: Harper 81 Bros. , 1952).

ZLook and Life magazines from April, 1958 to April, 1960.
36

 

 

 

 

 

 

37

Table l. --Factorial Design Of Picture Sample

 

 

Independencies Levels — No. D. F.
X, Color a) colored b) black 81 white 2 1
Y, Sex of Model(s) c) male d) female e) male &female 3 2
Z, Age of Model(s) f) 15—22 g) 23-25 h) 36-70 3 2

 

the only or the central model, while another group showed a woman as
the model, and still another group showed a man and a woman in the
same picture. The three age-brackets chosen for grouping models in
pictures were found to correspond approximately to the age distribution
of models in the magazines under study. TABLE 2 indicates all the

possible combinations of effects or 18 different picture groups.1

Table 2. --Combinations of levels of independencies

 

 

aaa aaa aaa bbb bbb bbb
ccc ddd eee ccc ddd eee
fgh fgh fgh fgh fgh fgh

 

Each picture was mounted separately on cardboard in its original
form and was numbered on its backside. In the interview the pile of
picture cards was handed to the subject, who was asked to Q-sort the
pictures according to a quasi-normal forced frequency scale (TABLE 3)

which was placed in front of him.

 

1The number of combinations, where everyone encompasses one
level of each independency, Should have been replicated six times in
order to provide a total of 72 pictures. However, at a later stage in
the study, some restructuring became necessary, such that the number
of replications for the different combinations became unequal. Thus it
was necessary to use Weighted Mean procedures in the Analyses of

 

Variance.

38

Table 3. --Frequency Distribution

 

 

"most" "least"

 

Score 12 11 10 9 8 7 6 5 4 3 2 l 0
Frequency 1 2 4 6 8' 9 12 9 8 6 4 2 1(N=72)

 

In order to perform the ideal self Q-sort the subject was given

 

the following instructions:

Imagine for a moment what kind of person or what kind of persons
you see in the pictures, and imagine the way of life of each of
these persons.

Now, we would like to have you distribute the picture-cards along
the scale as to "How much would I like my way of life to be like
the way of life of the person(s) in the picture ?" Place those you
would like to be like to your left, giving those you would like to

be most like the score-value 12. Place those you would not like
to be like to your right, giving those you would like to be least like

 

the score value of 0. Distribute the cards such that for each of
the given score-numbers you find the required amount of cards.

At a later stage the subject was given the following instructions in

order to perform the actual self Q-sort.

 

Again, imagine for a moment what kind of person or what kind of
persons you see in the pictures, and imagine the way of life of
each of these persons.

Now, we would like to have you distribute the picture-cards along
the Scale as to ”How similar is my present way of life to the way
of life of the person(s) in the picture?"

Assume for the moment that a jeweller had only one person,
Airs. X, in mind when he produced an extremely valuable piece, which
Only She could afford to buy, and he wishes to advertise the jewelry in
a magazine read by Mrs. X. The art director in charge may intend to
appeal to Mrs. X's ideal self with the help of a model. Under some

pretext, he may have asked Mrs. X to perform an ideal self Q-sort as

Outlined above.

39

After replacing the pictorial description provided by Mrs. X with
a statistical one provided by the coding interviewer, the data could first
be analyzed by variance methods. The question of whether or not the
pictorial variables underlying the structure of the picture sample
affected Mrs. X's identificatory reactions could be studied. For instance,
the data would Show whether Mrs. X preferred colored illustrations,
where models and Objects seem more "realistic, " or black and white
reproductions which appealed more to her sophisticated taste. Along
the same line, it may be found which sex- and age-characteristics in
the model enhance her identificatory reactions.

In order to elaborate this example, let it be assumed that Mrs.
X performed a total Of 12 Q-sorts under 12 different conditions of
instructions. Besides actual and ideal self descriptions, she may have
described herself as a young girl, as the person She hopes to be in the
eyes Of her friends and others. It is clear that, under each of the 12
different conditions, each picture may receive very different scores.

The 12 arrays Obtained could constitute 12 "variates" for corre-

lational purposes, providing a 12x12 table. Through factor analysis,

 

the correlation coefficients in this matrix would be subsumed into
perhaps three factors, which would indicate three possible "explanations"
for the identificatory reactions of Mrs. X. For an interpretation of the
clusters of self-descriptions defined by the factors, one would have to
refer back to the Q-sorts in order to see which pictures ranked highest
and lowest under the various conditions of instructions. Here the

advertiser would find to which cluster of self concepts he should appeal

 

and with which kind of model.
This example demonstrates that in Q-studies (dealing with samples

of pictures and not with samples of persons) propositions for the

40

"1 could be tested.

"single case
In Q-sorting, each subject describes himself in terms Of his own

frame of reference. No normative judgments are involved when the

importance of a characteristic is judged in relation to a reference-

group. This point, that persons' attributes and beliefs are measured

from the basis of "centrality-to-the-self, " is fundamental to Stephenson's

 

methodology. 2 If differs from the traditional approach, where attributes
or beliefs are measured on the basis of a generality of population
attributes or beliefs.

Stephenson points out that the meaning and effect of a mass media
message is always a matter of beliefs and attitudes of persons and
never a matter of beliefs and attitudes of a collective, such as an
"audience. " Persons have to be persuaded individually and the success
of persuasion depends upon how they, as human beings 32331.9 relate
themselves to the message's content.3 This is why he proposes a new
orientation in communication research based on Q—methodology and on
Q-technique. I

It is clear that Q- studies can be concerned with a number of
"single cases. " When in the present study, the intention was to see if
women identify with different models than man do, then certainly male,

as well as female, subjects had to be asked to Q-sort the pictures.

 

1For an exemplary Q-study of the "single case" see J. C. Nunnally,
"An Investigation of Some Propositions of Self-Conception: The Case of
Miss Sun, " Journal of Abnormal and Social Psychology, L (1955), pp.
87-92.

 

2Stephenson, A General Theory of Mass Communication, pp. 72-
107. ". . . our concern with Q is . . . with a complete methodology--
that of measuring from the centrality-of-self standpoint as prior to, and
in principal totally different from the century-Old methodology of measur-
ing isolated attributes for population of people. " P. 77.

 

3Ibid., p. 127.

41

In order to show that such a sex-difference in the choice of objects of

identification prevails, any woman and any man could be asked to do

 

. the Q-sorts. Similarly, one should be able to validate the proposition
that persons identify with models of their own age-group for any person.
Hence, "one chooses examples, not samples, of persons, according to
a specified factorial design which seeks to be comprehensive rather
than representative. "1

The factorial design underlying the selection of subjects studied
in this research is given in TABLE 4. Fifteen pairs of husbands and
wives in the East Lansing area were interviewed. Groups of five pairs
represented the age-groups, which were defined in correspondence with
the ones used to group models in pictures. The range of occupational
activity for nine out of the fifteen men was large, while the other six
were students. Among the wives, only a few of the younger ones were

professionally active as secretaries.

Table 4. --Factorial Design of Subject Sample

 

 

 

Independencies Levels No,
X, Sex of Subject a) male b) female 2
Y, Age of Subject c) 15-22 d) 23-35 e) 36-70 3

 

As before, in the outline Of a Q-study of a "single case, " a number
of Q-sorts describing various self-notions of Mrs. X were compared,
in this Q- study of several "Single cases, " corresponding self-descriptions

obtained from thirty subjects were analyzed.

Again, variance methods were employed to investigate if the picture

Variables underlying the structure of the picture sample affected

1Ibid., p. 9 (footnote)

42

identification. Separate analyses were performed for groups of five
subjects each, in order to find out about their relative effect upon
identification among young men and women, middle-aged men and
women, and old men and women.

In order to find "types" of persons, both the actual self Q-sorts .
and the ideal self Q-sorts Obtained from the thirty subjects were

correlated, and the two 30x30 matrixes were factor analyzed. The

 

emerging ”types" of persons had to be tentatively named by referring
back to the original Q-sorts and to additional data available on each
subject. The discovery of an order underlying the reactions of persons
to models in advertising pictures would then allow the statement of
hypotheses for further investigations of the phenomenon. 1
Q-technique has the advantage of assessing non-observable phe-
nomena, such as the identification process. Subjects do not have to
verbalize or to justify rationally their preferences for certain models.
According to instructions, they relate themselves to the content of each
item and state their reactions in terms of sorting Operations. The
assumption, however, that persons can evaluate the intensity of identifi-

cation with one model in reference to the intensity of identification with

each of the other models appearing in the set is necessary. 2

 

1Refer to footnote 1, page 31, supra.

2While in Q- studies this assumption or "transitive postulate" refers
to intra-individual differences, it refers to individual differences in
traditional studies, or R-studies, as Stephenson calls them.

In Q—technique a person finds item X more descriptive of himself
than item Y or Z. In R-technique one person A obtains a larger score
on a particular test than person B or C. In Q-studies persons are corre-
lated, i. e. the scores one person gives to all items are correlated with
those another person gives to all items (or those the first person gave
to all items under different conditions Of instruction). In R-studies tests
are correlated, i. e.. the scores obtained in one test from all persons are
correlated with those obtained on another test from all persons. For a
concise summary of postulates underlying Q in contradistinction from R,
see Stephenson, The Study of Behavior, p. 58.

 

 

 

43

Data Collection

 

Since it has already been explained how the picture sample was
composed and which subjects were interviewed, only a few additional
remarks need to be made on this point.

In order to contact persons who would cover the factorial design
in TABLE 4, an interviewer made telephone-calls in the East Lansing
area. .She inquired about the age-group of the respondent and the
respondent's partner and askedif they would agree to participate in a
study concerned with people's reactions to advertising pictures.

Interviews were held in the interviewees' homes. As each test
sequence made use of other instruments not reported in the present
study, most interviews were held in two sessions. In no case was the
husband present when his wife performed her Q-sorts, nor was the wife

present when the husband sorted the pictures.

Statistical Analyse s

v

 

In order to decide whether color of picture, sex of model, and age
of model affected identification, analysis of variance methods were
adopted. Sets of actual and ideal self descriptions were held separate.
For each set, six analyses were performed, so that groups of five
Q-sorts provided by young, middle-aged, and Old men and women
independently were treated in a three-way classification set-up.1 The
sources of variance became the picture variables of color, sex of
model, age of model, with their double and triple interactions.

Once a source of variance was found to have a significant effect

upon identification, the underlying means were compared and significant

 

1E. F. Lindquist, Design and Analysis Of Experiments in Psy-
chology and Education (Boston: Houghton Mifflin Company, 1956),
pp. 220-28.

 

 

44

differences were determined by means of Duncan's "Multiple Range
Test. ”1 Thus, if age Of the model had been found to affect picture
scores Obtained from a certain group of subjects, this test served to
show the particular age-group of models with which identification had
been predominant. I

Eta-coefficientsz were computed as estimates of the relative
contribution of each picture variable to the total variance of Q-scores.
Here it was seen which Of the picture variables had the most decisive
effect upon identification in reference to actual and ideal self notions.

Factor analytic methods were used to answer the question Of
whether there are "types" of persons who provide similar Q-sorts, and
thus give similar descriptions of themselves. Each person's Q-sort
was correlated with all other persons' Q-sorts for actual self notions
' and for ideal self notions. The resulting correlation matrixes were
submitted to a Principal Axis factor analysis involving the computation
of "Eigenvalues." After Varimax Rotations had been performed,
clusters Of persons were found. In order to define such clusters, factor
loadings of each person were compared across factors and each person's
largest loading was check-marked. Subsequently, the checkmarked
loadings on each factor were compared with each other, and the two most
sizable loadings were checkmarked a second time. The two persons
corresponding to the double checked loadings on each factor were taken

as best describing the factors. No weighting procedures were employed. 2

 

1D. B. Duncan, "Multiple Range Tests for Correlated and Hete-
roscedastic Means, " Biometrics, XIII (1957), pp. 164-176.

 

ZQ. McNemar, Psychological Statistics (New York: John Wiley
& Sons, Inc., 1955), pp. 272-75.

 

3L. L. Thurstone, Multiple-Factor Analysis (Chicago: University
of Chicago Press, 1947), Chapter 20: "The Principal Axes. " See also,
H. F. Kaiser, "The Varimax Criterion for Analytic Rotation in Factor
Analysis, " Psychometrika, XXIII (1958), pp. 187-200.

 

 

CHAPTER V

RESULTS: PICTURE EFFECTS AND IDENTIFICATION

Our major findings concerning the influence of the picture vari-
ables--"color, " "sex of models" and "age of models"--upon the
identification processes of men and women of three age groups are
presented in this Chapter. The two hypotheses are that Similarity
of characteristics in the reader of an advertisement and in the model
shown in the illustration facilitates identification:

1) Persons tend to identify with like-sexed models rather than
with models of the Opposite sex.

 

 

2) Persons tend to identify with models of their own age-group
rather than with models of a different age-group.

 

 

We further expect that pictures showing a man and a woman will
get somewhat higher average identification scores than pictures showing
a man alone or a woman alone, as identification would be facilitated
for men and for women in the audience.

TABLE 5 represents a summary of the analysis of variance re-
sults listed in more detail in Appendix I A, l and 2. It shows the dis-
tribution of significant F-ratios (.05 and . 01 level of confidence) for
subject groups and for picture effects. Results obtained for the set
of Q-sorts describing the actual self and the set of Q-sorts describing
the ideal self are shown separately.

The pattern of F-ratios found significant when identification
occurs in reference to actual self notions is different from the pattern
of F-ratios found significant when identification occurs in reference to

ideal self notions in terms of the number of recognized effects.

45

.no pun no HAmfiwH oo. .8on dem

tsoflmdon Smog msfifinopcs norm .mHopOE :pHO: .HO\pcm :pomntofippfifib. on swim monoomtcmoa oat Eonm

25.83% igneoflflcmfim was? mfiopoe :m:50>: on soc/Mm ouoom smog oafi Coat .mHopOE 29290.»: 52> >3m

thewonm pomwfiucopw upon» wept/Ohm” .mHopOE :pHO: no :powmtoHpEE: : £55.»: noflfio £23 Efimuomonm

poflficopfi moat. .o A Jon “:5an of. a“ Bohermfiwz cm ponficmooou fay; GOEOB was?» and? mofimo
12.2: OH .633 flow Egon 2.3 m0 nonnoo $2 mop oat CM 30 .VX ego “downspoumnoucm mo ofimgmxmn

Snug potomon Aofimntmv Somme
onsaoa ucmoflficwfim Loos wcfifinopc: moOoEEmo mooning .HOm .N was A < .H Xfipcoamxw on newom 5qu

 

46

 

 

 

 

- - - - :ocx 20.? 20.? mo
- :oex - - :ocx - 20.? 2o
- hoax - Hoax :9? :9? .. E2
- :ocx Amocx Amocx :9? so»? - 22
- - - - :03.” 20.? - Ea
- Amocx :ocx - :9? 29% - 2»
new :82
- Hoax - - - :oax :oax mo
- :ocx - - Amoco.” Amocx 29% 2O
- Amocx - Amocx - :ocx - E2
- - - Amocx - - - 22
- - - - - - Eocx a;
- - - - - - - 2%
new nausea.
o- .m- :o. .O- .m .O- a. .m- .< soaoo new one. 832:2,
moHQdCe> onsuofinm uoohnsm

 

 

~5an was 52
30 pan .pomxwtofippfiE .mesow no“ maoobm 23me demoflmemwm mo :oflusfifimwmtt .m 3an

47

Since each recorded "X" indicates that the presence of the correspond-
ing picture-effect (or combination of picture effects) makes a difference
in the identificatory reactions of the corresponding subject group, we
have to interpret the pattern with the larger number of "X's" to mean
that the three picture variables affect intensity of identification more
often when ideal self notions are involved than when actual self notions
are addressed.

This suggests that whether an illustration is colored or black and
white and whether a male model or a female model is shown will be of
greater impact if an appeal to the ideal instead of the actual way of
life of our subjects is intended.

So far, however, we have not yet considered if colored or black
and white pictures were more successful in stimulating identification,
or if persons tend to match their own sex and age characteristics with
those of the models. Let us refer to TABLE 6a and 6b, 'in order to
interpret each F-ratio listed in TABLE 5 by considering the underlying
mean relationships.

A. Choice of Identificand in Reference to Actual Self Notions.

 

1. Young Men: No preferences concerning the three picture
variables or their combinations.

 

2. Young Women: Preference for young models;

 

3. Middle-aged Men: Preference for a combination of age
and sex characteristics in the models. Pictures show-
ing old men and pictures showing both a young man and
woman were significantly overchosen in comparison to
pictures showing middle-aged male models, which is

 

contrary to our expectations.

4. Middle-aged Women: Preference for women as models.
Middle-aged female models elicit significantly more
identification than male models of any age-group. Black
and white portraits of women were more successful than
pictures characterized by any other combinations of the
sex and color variables.

 

48

Table 6a. --Relationships Between Mean Picture Scores for Significant
Effects in Table 5

 

 

Actual Self

 

 

 

 

 

 

 

 

Subject
Variables Picture Variables
”Agell:
YF (0 m) (y)
OF (m) ()1 _0_)
OM (m y) (3)
"Sex”:
MF (m) (r39 (3)
OF (m) ([191 f_)
0M ”1 “2.5 TB)
"Color":
OM (bW) (_C_)
”Age'l and ItSeXH:
MF (m-m y-m m-mf o-m) (O-f y-mf o-mf y-f m-f)
MM (m-m) (O-mf y-f y-m o-f m-mf m-f y-mf O-m)
"Sex" and "Color":
MF (c-m bw-m bw-mf c-f c—mf) (bw-f)
OF (bw-m c-m bw-mf) (c-f c-mf bw-f)
OM (c-f bw-mf bw-f) (bw-m c-m c-mf)

 

 

Note: Any means not appearing together with the sublined means within
the same pa—re-ntheses are Significantly different at the . 05 level
of confiderice. Any means appearing together with the sublined
means within the same parentheses are not significantly different.

 

 

49

Table 6b. --Relationships Between Mean Picture Scores for Significant

Effects in Table 5

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Ideal Self
Subject
Variables Picture Variables
”Agell:
OF (m f) (3)
0M (m y) (3)
"Sex":
YF (m) (3999
MF (m) (51le
OF (m) (gnji)
YM (f) (r3_f_ r3)
MM (f) (r3 r33)
”Color":
YF (bw) (E)
MF (bw) (_c_)
OF (bW) (3)
YM (bW) (_C_)
MM (bW) (g)
OM (bw) (5;)
“Age" and ”Sex":
MF (m-m o-m y-m) (o -f o-mf m-mf y-f -f y-mf)
MM (O-f m-f o-mf y-f) (y-m m-m y-mf o- m-mf)
"Age" and "Color":
YM (c—o c-m C-y bw-m bw-y) (bw-o)
MM (bw-O bw-y) (bw-m c-y c-m c-O)
"Sex" and "Color":
MF (bw-m c-m bw-mf bw—f) (c-f c-mf)
YM (bw-mf c-f bw-f bw—m) (c-mf c-m)
MM (bw-mf bw-f c-f bw-m) (c-m c-mf)
OM (bw-mf c-f bw-m bw-f) (c-m c-mf)
Note: Any means not appearing together with the sublined means within

the same parentheses are significantly different at the . 05 level
Of confidence. Any means appearing together with the sublined
means within the same parentheses are not significantly different.

 

50

5. Old Men: Preference for old models independent of
their sex, and preference for male models independent
of their age. Greater appeal of colored pictures.
Selection of colored pictures showing both a man and a
woman over colored, as well as black and white portraits
of women alone.

6. Old Women: Preference for Old models1 independent of
their sex, and for women as models, independent of their
age. Black and white portraits of women have a greater
appeal than either black and white or colored portraits

 

of men.

Summarizing our findings, we see that whenever our subjects
recognize an age—effect or a sex-effect in models, they identify prefer-
ably with models of their own age and sex. Also in four cases out of
five, where certain combinations of picture-effects were liked better
than others, our hypotheses tend to be supported. Let us compare
these results with those obtained for the Q-sorts describing the ideal
self notions.

B. Choice of Identificand in Reference to Ideal Self Notions.

1. Young Men: Preference for male models and for colored
pictures. 'Preference of old models in black and white
pictures over all other combinations of the age and color
variables, and of colored portraits of men alone, or of
both men and women, over all other combinations of
sex and color characteristics.

 

2. Young Women: Preference for female models in general,
and for colored pictures in general.

 

3. Middle-aged Men: Preference for male models and for
colored pictures. Pictures showing both a middle-aged
man and woman elicit significantly more identification
than pictures showing women of the three age-groups.
Colored portraits of old men are not Significantly

 

 

1Although the mean identification score for "old" models is higher
than the mean scores for "middle-aged" and "young" models, there is
no significant difference between mean scores for "old" and "young"
models. Still, both these mean scores are significantly different from
the mean score for "middle-aged" models.

51

preferred over colored or even black and white portraits
of medium-Old men, but this still runs somewhat con-
trary to our expectations. Colored pictures of both a
man and a woman, and of men alone, are preferred
over all other combinations Of the sex and color vari-
ables.

4. Middle-aged Women: Preference for female models and
for colored pictures. Pictures showing both a young man
and a woman are significantly overchosen in comparison
to pictures showing men of the three age-groups, but
they are not significantly preferred over pictures showing
middle-aged female models. This latter result does not
support our hypothesis concerning the age variable as
well as it does the hypothesis concerning the sex variable.
Furthermore, there is a preference for colored pictures.
of both a man and a woman, and of women alone.

 

5. Old Men: Preference for old models and for colored
pictures. Colored portraits of both a man and a woman,
as well as of men alone, elicit stronger identification
than pictures characterized by all other combinations of
the sex and color variables.

6. Old Women: Preference for Old models, for women as
models, and for colored pictures.

 

Summarizing once more, we see that wherever subjects have any
Consistent preference for models of a certain age, .they choose those
in which they find their own actual age-characteristic. Similarly, we
See that where sex Of the model influences identification with any sig-
nificance, persons choose like-sexed models rather than models of the
Opposite sex. Also, in four cases out of seven, where some combi-
nations of picture-effects were preferred in the objects of identification,
Our hypotheses tend to be supported.

For both sets of data we can conclude that, for those groups of
subjects where we can reject our first statistical hypothesis by saying

that sex of models in advertising pictures has an effect upon identi-

 

cation, persons identify with like-sexed models rather than with models

 

of the opposite sex. Further, we conclude that, for those groups of

 

52

subjects where we have to reject the second statistical hypothesis by

stating that the age Of models in pictures has an effect on identification,

 

persons identify with models of their own age-group.

 

After having presented the evidence which tends to support our
theoretical hypotheses, we wish to return to TABLE 5 in order to point
out in which cases the statistical hypotheses could not be rejected.

Although the results for the actual self and for the ideal self
differ in respect to the number of significant F-ratios, as noted
earlier, there is also a similarity between the two patterns insofar as
location of the "X's" is concerned. Comparing the data for young,
middle-aged, and old persons, we find the following results for the
actual self and, to a lesser degree, the ideal self:

1. Young subjects are relatively little affected by the age and
the sex of the model, and by whether a picture is colored or
not. This means that young subjects have few consistent
preferences concerning the characteristics of the identificand. .

2. Middle-aged subjects look for certain combinations in picture-
effects in their choice of objects of identification, but they do
not yet have consistent preferences for independent criteria.
Thus, although they prefer certain combinations of sex and
age in their models, we do not find a predominance of models
of a particular age-group when sex-differences are left out of

consideration.

3. Old subjects tend tO have consistent preferences for certain
picture characteristics. They evaluate each picture effect
independently and with greater consistency.

We venture an intriguing conclusion, even if it provokes argu-

ments in the light of the rather frail evidence: that discrimination of

 

gharacteristi‘cs in the identificand and their independent and consistent
Evaluation is a learned skill.

We have explained before that when the sex and age character-
istics of the models affect the identificatory reactions at all in a sub-

ject group, then there is a tendency for persons to identify with models

53

of their own sex and age. As we now find that the choice Of sex and

age characteristics more often has a significant impact upon Old sub-
jects than upon young subjects, we infer that Older persons are more
likely to identify with models of their sex and age, which is probably
due to an increased self- and role-consciousness. This suggests that
our hypotheses are more likely to hold for older people than for younger
ones.

After finding that persons identify preferably with models of
their own sex, we want to know if pictures showing both a man and a
xvoman receive somewhat higher average identification scores than
pictures Showing'a man or a woman alone, which only enhance identifi-
cation for either male or female readers. Appendix B contains the
pertinent means for each group Of pictures, for each group of subjects,
and for each self-concept.

Male models alone receive low identification scores from women,
though high ones from men, while female models alone receive low
scores from men and high ones from women. Pictures showing a man
and a woman simultaneously gain, in a male audience, identification
scores almost as high as, or sometimes even higher than, pictures
showing a man alone. But their success lies in the fact that they gain
identification scores of about the same 'size in audiences composed of
woman. The clue to their effectiveness is that they Offer to both men
and women their preferred Objects of identification.

One more aSpect of TABLE 5 deserves comment-—the "color"
variable, which was made a criterion for structuring the picture
sample in order to become experimentally controlled.

It has already been noticed (TABLE 6a and 6b) that whenever the
fact of whether a picture is black and white or colored has had a sig-
nificant influence upon the intensity Of the identificatory reaction,

colored pictures were preferred. But, while only one group of subjects

54

gave a significant advantage to colored portraits of their actual Objects
of identification, all groups, without exception, preferred colored
portraits of their ideal Objects of identification.

Thus it seems that it is much more important that an illustration

 

be colored when the depicted model is meant to appeal to ideal self

 

notions than when it is meant to appeal to actual self notions.

 

The general preference for colored illustrations may, to some
extent be due to a sample-bias, as black and white advertisements are
often less pretentious in their models, composition, or photographic
quality.

For the more Specific success of colored portraits where ideal
self notions are involved in the identification process, we suSpect
that color enhances the imagination of the subjects' ideal role-configura-
tions, prior to identification with the depicted model on the fictitious
level. Additionally, it is likely that colors symbolize an ideal role
rather than an actual one.

Finally, we turn to a consideration of the relative importance of
each of our three picture variables. Eta-coefficients in Appendix I, C,
indicate how much of the total variance in identification scores is
explained by each picture-effect. FIG. 11 gives a summary of variable
relationships, but as most differences between the Obtained Eta-
coefficients were small, we should be cautious in generalizing our
interpretation.

Both sub-sets of Q-sorts (actual self and ideal self), as well as
the total set, indicate that decisions concerning the choice of a male
or a female model are more important than decisions concerning the
choice of a colored or a black and white illustration in the composition
of advertisements to stimulate identification.

If identification in reference to actual self notions is to be elicited,

the age attributed to the model is likely to have the predominant impact,

55

FIG. II. Comparison of Picture Effects in Terms of Magnitude of
Contribution to Total Variance in Q-Sort Sets

 

 

Actual Self Ideal Self Actual and Ideal Self
\x, "r -. ,- ”\ ’\
(5’7“? (s>(/C\ (S >(Ct
,1 ¥, \g . “ .. . \,
1‘ 4 \l 1/ \\ 4
/\ r‘\\ -/-\\\
KA . A ‘

- . A
__ ,, \\, , K /

Note: S = "Sex of Model, " A 2 "Age of Model, " C = "Color of Picture"
Symbols Of relationship (< , > , = ) indicate whether one
picture effect explains less, more, or an equal amount of

the variance in identification scores for actual and/or ideal
self than another picture effect.

 

next to the sex of the model, while the color of the illustration is least

important.

However, if identification in reference to the ideal self notions

 

is to be stimulated, the choice of a male or a female model is likely
to have the predominant impact, followed by color of the illustration,
\vhile the age attributed to the model has only a minor influence.

When a simultaneous appeal to bp_t_h_ self-notions is intended, the
sex and age characteristics in the model are of equal effect, while the
color of the illustration has a smaller impact upon identification. 1

This chapter dealt with the first part of the results. It should be
emphasized that the three picture variables--"color, " "sex of model, "

and "age of model" --explain only a small part of the total variance in

m

1The relatively small impact of color upon identification, however,
does not suggest that color is superfluous, as here its influence upon
attention or recall is left out of consideration. Our findings only indi-
cate the relative importance of color among the three picture variables,
studied. They do not indicate any magnitudes of effects.

 

56

identification scores. This can be seen from the considerable residual
variances reported in Appendix I, A 1 and 2. Criteria like "socio-
economic status, " "occupational activity, " "beauty, " among others,
can be expected to have affected the subjects' choice of their Objects
of identification. To some extent, these will be considered in the next
section of results.

Once more the reader should be reminded that particularly those
findings reported so far which do not relate to the two hypotheses do

not necessarily apply to persons other than those from whom the Q-Sorts

\Vere Obtained.

CHAPTER VI

RESULTS: TYPES OF ACTUAL AND IDEAL
SELF NOTIONS

Our major findings concerning the segmentation of the "audience"
composed of our 30 subjects will subsequently be presented. Some
questions will be answered in the following sequence: Of how many
different types are our subjects when we consider their pictorial
descriptions Of their actual ways of life, and of how many types are
they, when we consider pictorial descriptions of their ideal ways of
life? Of what age and sex are the persons in each type? Of what age
and sex are those two persons most typical of each type? Which models
elicit most and least intense identification in different types of persons,
when ideal self notions are involved, and when actual self notions are

involved ?

Number of Types of Actual Selves and of Ideal Selves

 

The 30 Q-sorts describing the ideal selves of our subjects were
intercorrelated, producing the matrix shown in Appendix I, D l, which
indicates similarities of each person's ideal self notion with every other
person's ideal self notion. A similar matrix is Shown in Appendix I,

D 2, where each correlation coefficient stands for the similarity of
one person's actual self description with another person's actual self
description. 1

Although ten factors were extracted from each matrix, only six

factor solutions were adopted for both sets of data, because in each

case no more than six factors could be defined by highest loadings of

57

58

at least two persons. The adOpted factor matrixes (Appendix I, E)
summarize the Q-sort to group of Q-sort relationships, or, in other
\vords, the person to type Of person relationships. As the obtained
factors are orthogonal factors, no group to group relationships or
type with type correlations were to be expected.

Each set of factors was rotated and the results are found in
TABLE 7. The purpose of rotation was, as usual, to differentiate
more sharply between types. Here a single asterisk indicates the
factor on which the respective person has the most sizable lOading; the
position of the asterisk shows under which type of actual self and under
which type of ideal self the corresponding person will be subsumed.
Thus, subject number 2 contributes to the definition of ideal self
Type I' as well as to the definition of actual self Type IV'.

However, when we classify persons so that each person belongs
to one type only in his actual self notion and one in his ideal self notion,
we must keep in mind that each person is only Reg: characterized by
this type, while to a minor extent he is also characterized by other
types of self notions, as an examination of any person's factor loadings
acrOSS columns will show in TABLE 7.

We find those two persons indicated by two asterisks most
characteristic of each group of persons. Only these two best repre-
sentatives of each type will be considered when we later turn to "naming"
each factor, as they produce a clearer factor differentiation than will
all persons who contribute to the definition of each group. Thus, we
shall study the Q-sorts of 12 persons out of 30 with particular attention.

Although there are six types of ideal selves and six types of

 

 

actual selves among our "audience, " it is readily seen that persons

 

who together define an ideal self type, do not also define a type through
a similarity in their actual ways of life. This, certainly, is not amaz-

ing, as individuals with very Similar actual role-configurations may

59

.093: Sumo mo Orwfloocowonmon EOE mCOmnom OBS oat moOwOHOCH n3”.
.poOmHonnoo umoe me o: Sofia? £35 :OBS:

 

 

 

 

 

ofi .0 .H1 .wcatptmoH Roam“: one. was QOmnom mcfipcommouuoo oat #053 so HOoOmM of meomofipfi n. ”302
S... oo+ 2.. 3. so- .12: 3+ *5; 8+ am- mm- 3+ o.m
om: MS: can Om: +37 m¢+ no: ¢o+ QT. 00.. MN: "SEN; 0N
mm. mm: wm+ sat so: *2; mm+ 2+ Hm- +wmt 3+ wo+ mN
mm+ +21 >o+ wot HN+ moot om: Hot harem- me. Not «5+ .5
>01 mm: 2+ NH- 2.. Vwom+ ©H+ wm+ 3. NT. MN: emw+ om
mat ncnmmt no: 00: 2+ om+ wot mm+ncwmot all :t NN+ mm

.nomt wo+ om+ mo- wan sm+ mo: 2+ mo+ nanmwt wot >o+ mm

acumen no: no. wat oat Nm+ ovt .”.Nm+ cot own 07. mo: mm
No+ No+ 3+ wot mm- saga: mo+ V”.Nm+ NM- 00.. mm- 3+ MN
wm+ Tm- AN- mm- «.01 .”.om+ mm... mo+ mo. 001 pm: undm+ Hm
Hm: mo+ Ho+ Hm- ow... *ww+ mm: so+ :+ so»: mm- mm+ om
Hm- Not >o+ mm- harem: mo+ El ramm+ No+ SN- pot wo+ m;
mm- 8+ 2+ 2. sum- 3+ 2- of. 2+ ea- 2- if m:
NH+ 3. so. .20. mm- wo+ 3m: 2+ a3- RN- 21 3+ 5
Hm: mo+ 3+ ants- mm- mm+ wetfidm+ 2. mm: oat NH+ 2:
out >o+ Ho+ hem- wot mN+ sums} 3+ m3- so: pm- 3+ ma
2+ om- oH+ 2+ Awe- 00: mo- :1 mm... hmmt 00: NH: .3
nCN+ Not *vm+ mo+ mo: :1 now: so. wat mm- mm: ma. 2
mm- NH- wm+ MN- *5... 2+ Ho+ no+ mm- scumwt oo+ so- NH
out oo+ . 3+ mm- mo+ *3: :t Ho+ om- wan "new- :1 :
wan Not ”whoa: mm- no+ mo+ 2+ Nm+ can ”now. 3+ 2+ 2
ft scam; 2+ mot om: wot *nnmot ¢o+ mo+ nCo+ hm... mm+ o
¢o+ mm- 00. om- snow- mm+ mo+ 2+ OH- amnesia... 0N+ w

harem: £1 mat vo+ NH- mot No+_ «2+ wot caisson- uvm+ \t
0H1 «2+ 00. mm- emmt 2+ on: NN+ 31 new. St nm+ o
00. no. mo- «it unmet Nm+ so: wo+ Om- «to: mNt 3.0m: m
mm: 00.. :+ *3». NH- >N+ mm- mH+ mm: ism: oat wm+ as
mm. 00.. 3+ "shoot 2- No+ Manson mo+ mot 3+ Hm- mo- m
oo+ Hm- magi 00+ om: :1 pvt «5+ No+ 3.- mo+ nAm+ N
NN+ mot snail mot 3+, mot *omt NH+ mo+ No: om- :+ H
.5 _> .2 .E .2 .H .H> .> .Z .5 .2 a “neoofloam

mom H.334 mom ~33

 

 

If

muuohndm om cnn poUTVOHnH muhomtd ouduownm mo meow 0.5% MOM mwcwpmod nouomh poumuomtn .xt OBMH.

60

have very dissimilar ideal role-configurations. How congruent or in-

congruent both role-configurations are in each individual is likely to

depend upon his degree of adjustment to the status quo. 1

 

Age and Sex of Persons Representing Each Type

TABLE 8 is nothing more than an excerpt from TABLE 7; here a
person's factor loadings are shown only for the actual and ideal factors
which explain most of the variance in his identification scores. As sub-
jects here are grouped by their sex and age, it is easily seen that some
factors are predominantly or exclusively "male" or "female, ” while

others are defined by persons of one age-group rather than the others.

TABLE 9 summarizes these relationships:

1. While five out of six types of actual selves are to be found in
both men and women, only three out of six types of ideal
selves are shared by men and women. While there is only one
exclusively "male" type of actual self, there are two such
exclusively "male” types of ideal selves, and while there is
not a single exclusively ”female“ type of actual self, there is
one such type of ideal self.

2. While the actual self notions of our male subjects are of six
different types, their ideal self notions are of only five dif-
ferent types. Similarly, while our female subjects' actual
selves are of five types, their ideal selves are of four types
only.

These observations suggest, first, that the actual ways of life of

 

Sen and women are more similar than they wished they were, and
second, that men's actual and ideal ways of life are slightly more
giversified than the corresponding ones of their wives, which does not
seem amazing when we recall the wide occupational spread in our male

subjects in contrast with the ”homemaker's" role in all female subjects.

‘—

1I. D. Nahinsky, ”The Relationship Between the Self-Concept and
the Ideal as a Measure of Adjustment, ” Journal of Clinical Psychology,
XIV (1958), pp. 360-64.

 

61

Table 8. --Clusters of Persons Defining Factors in Table 7.

 

 

 

 

 

 

 

 

 

Actual Self Ideal Self
Factors: 1 11 111 1V V VI 1 11 111 1V V VI
1 +47 -39
3 -69 '—__ -67
‘Yhd 5 -63 “" +59 __—
7 — ' :31 — :19
9 :22 :92
11 +41 -44
13 +34 -46
141v1 15 -56 -42
17 -61 —51
19 .LEé :22
21 +39 +39
23 -65 +52
(DIvI 25 -52 _——' -63
27 :13 757
29 -46 :71 _——
Sub-total: 2 3 3 2 3 2 3 2 0 2 2 6
2 +42 +51
4 -71 -54
IF 6 -53 — -66
8 —46 -71
10 :99 '_—_ -49
12 -47 -83
14 -42 'T52
3N41?‘ 16 -47 +59
18 -34 +57 ""
20 +48 -76
22 +67 +52
24 _——' -50 -85
(33* 26 +59 +45 '_‘"
28 +41 -58
30 +71 +47
W ’—
~total: 5 5 2 2 o 1 3 1 8 0 3 o
75:;FSEE 7 8 5 4 3 3 6 3 8 2 5 6
\.
Pficn;

' \\e; Underlined factor loadings indicate the two persons most representa-
tive of each "type" because of their highest or second highest corre—

lation with the group.

Sub-totals indicate the number of men and the number of women found
in each cluster of persons. The total indicates the number of persons

defining each factor.

62

Table 9. --Sex and Age Characteristics of Persons Defining Different

Types of Self Descriptions

 

 

Factors: I II III IV V VI

 

Sex Characteristic s

 

Actual Self IVI+F M+F M+F M+F 1V1 M+F

Ideal Self M+F l\4+F F M M+F M

Age Characteristics

 

Actual Self I\I+O Y+M+O Y+M Y+M Y+O Y+O

Ideal Self Y+NI+O Y+I\/I Y+M+O O I\A+O Y+M

 

If we study the age characteristics of the persons included in

each type, we make the following observations:

1.

Two types of actual self notions and two types of ideal self
notions are shared by young and middle-aged subjects. One
type of actual self and also one type of ideal self is common
to middle-aged and old subjects.

While there are two kinds of actual selves shared by young
and old people, these age-groups have no ideal self in common.
While only one type of actual self is found among subjects of
all three age-groups, two types of ideal selves are defined by
young, middle-aged, and old people simultaneously. Some old
persons have a special type of ideal self, while they have no
special kind of actual self.

Our observations vaguely indicate that ideal ways of life are not

 

is often bound to specific age-groups, as actual ways of life are. This

 

is plausible, because roles actually performed at work and at home

are likely to vary more across age-groups than ideal roles, proposing

happiness, leisure, or entertainment. The type of ideal self, which

‘was only encountered among old people, will probably be concerned

with retirement expectations .

63

Age and Sex of Persons Most Typical of Each Type

Special marks symbolize those persons in TABLE 8, who, in
their actual selves or in their ideal selves, are most representative
of each type. We note, however, that the ”representative power" is
extremely variable. Thus, we note, that subject number 1 in his
actual self is a very poor representative of his type, although he is
still the second best in his group. No more than 22% of the variance
in his identification scores can be predicted by the particular factor
he himself helps to define. For contrast's sake, the ideal self of
subject number 24 has to be mentioned as a very powerful representa-
tive of its type; here as much as 72% of the variance in identification
scores can be predicted by the particular factor defined.1

Those persons singled out as being most representative of their
types are, at the same time, relatively ”pure“ in their self notions,
which is indicated by a tendency toward small loadings on all factors
but one. They probably have very set ideas about what they are at the
time of the interview, or about what they would like to be, either in
the future or if circumstances allowed.

TABLE 10 shows the sex and age of the subjects in our sample
who have the more typical actual and ideal self notions. It is interesting
to note how similar the patterns obtained for actual and ideal selves
are.

1. When we consider types of actual self notions as well as types
'of ideal self notions, we find that in two-thirds of the cases
the highest loadings are contributed by men, while only in one-
third of the cases are they contributed by women.

1The percentage of variance of a given measurement which may
be predicted by a particular factor is calculated by squaring the re-

spective factor loading, thus obtaining the necessary coefficients of
determination.

64

Table 10. --Number of Men and Women of Three Age Groups Most
Representative of Each Type

 

 

 

 

 

Actual Selves Ideal Selves
Sex ' Sex
Age M F M+F Age M F M+F
Y 5 2 7 Y 4 l 5
l O l M l 2 3
O 2 2 4 O 3 l 4
Y+M+O 8 4 12 Y+M+O 8 4 12

 

2. When we consider types of actual selves as well as types of
ideal selves, there is a tendency for young subjects to provide
the largest number of highest loadings, with old subjects pro-
viding a somewhat smaller number, and with middle-aged
subjects providing least.

3. When we consider types of actual selves, more than one-third
of the total number of highest loadings are traceable to young
men, while for types of ideal selves, exactly one-third of the
total number is traceable to young men.

To some extent our first observation is explained by the fact,
reported earlier, that our women's actual and ideal ways of life are
less diversified than those of their husbands, or, in statistical terms,
as women's self-notions are of fewer types than their husbands, their
chances are smaller to be among the best representatives of a type.

One possible interpretation of our second observation--that
Stoung people tend to have very determined ideas concerning their actual
and ideal ways of life, which is rare among middle-aged people, but
a~ppears again, to some degree, among old people--is the following:

Lacking experience, our young subjects perceive themselves in
their actual and ideal roles in a stereotypedl fashion. Our middle-aged

mm

1Lippman used the term "stereotype" to designate standardized
ELnd rigid "pictures in our head, ” which persist because they often

65

subjects may be in a stage of redefining their actual and ideal roles,
after becoming aware of limitations in some of their abilities through
accumulated experience in all kinds of interpersonal relations. Their
self-notions may have become a compromise in the face of contradictory
expectations, thus losing consistency. Old men, however, with addi-
tional experience, may have regained a consistency for their self-
notions which no longer is due to stereotyped perceptions of actual and
ideal roles.

The fact that each of the five young men included in our sample
not only has a different self concept but also represents his respective
type best or second best (with one exception, where ideal selves are
involved) would then be due to the combined phenomena we just have
discussed.

Again, we note that the basis for the foregoing interpretations is
rather frail due to the extremely small number of persons in each sub-
ject group; even if the interpretations were acceptable for the ”audience"

under study, no generalizations to larger audiences should be attempted.

Six Types of Ideal Self Notions

In this section, a tentative interpretation and ”name" for each
factor is prOposed on the basis of those three pictures which stimulated
most intense identification (top pictures) and those which stimulated
least intense identification (bottom pictures) in the representatives of

each (cluster of persons when they were asked: “How much I would like

 

Eyway of life to be like the way of life of the person(s) in the pictures?"

\.

provide us with a consistent picture of the world in which we live. Any
disturbance of these stereotypes becomes an attack upon the foundations
of our universe, particularly because they become emotionally charged.
W. Lippman, Public Opinion (New York: Harcourt, Brace 81 Co. , Inc. ,

1922), pp. 89-96. Here we mean by ”stereotype" a standardized and
I‘igid picture of ourselves.

 

66

Appendix I, F, presents in ”standard scores'' to what extent the
representatives of each cluster identified with each of the 72 pictures.
Positive algebraic signs indicate that the respective groups of persons
perceived the model‘s way of life as similar to their own ideal ways of
life, while negative algebraic signs indicate dissimilarity. Pictures
producing a very narrow range of standard scores across factors were
excluded from further consideration because of a lack in discriminatory
power.

FIG. III offers reproductions of those three pictures which ob—
tained the largest positive scores (top pictures), and of those which
obtained the largest negative scores (bottom pictures) for each of the
six types of ideal self notions.

When in the following each type of ideal self notion is introduced,
the tentative ”name" is given first. Then it will be reported how many
persons were found to possess ideal self notions of that type, and what
their age and sex had been. Occupations of the major representatives
of each cluster will be mentioned. We shall also describe briefly the
three top and bottom pictures, as this may help the reader in his own

evaluation of the rather small pictorial reproductions in FIG. III.

1

Happy Family Life versus Status Seeking

Six men and women of all age-groups aspire to a happy family
life. Among them an old engineer and a young physical education
student are the most typical representatives.

Top pictures: (a) elderly parents with their four children
among a confusion of luggage and house-
hold goods about to move into a new home
in an unfamiliar town (TIME INC.)

(b) parents with their little boy and girl re-
laxing happily on their lawn (GOLDEN
VIGORO lawn food).

(c) two boys in letter sports shirts playing
word games with ALPHA-BITS at the
breakfast table.

67

FIG. III. TOP AND BOTTOM PICTURES OF IDEAL
SELFIUMTMNKARRAYS

I' II' III'

 

68

6.0 the hon.” nu.
“- lU(t(Y UID‘ILO

 

69

Bottom pictures: (x) women in riding habit jumping a horse
over a hurdle, watched by friends in the
background (UNITED STATES BREWERS
FOUNDATION).

(y) driver of a racing car surrounded by a
swarm of mechanics after winning the race
(CHAMPION Spark plugs).

(z) self-confident man with glass (SCHENLEY'S
whisky on the rocks) in front of heavily
framed picture and friends in an art
museum.

2. Home and Good Job versus Financial Worries

 

The ideal selves of a young air force sergant and his wife are
most representative of this factor, which otherwise is also
defined by the ideal self of a middle-aged gas station owner.

Top pictures: (a) elderly couple looking at the foundations
of their new home piled high with FLINT-
KOTE products,

(b) parents with their little boy and girl re-
laxing happily on their lawn (same as in
factor 1).

(c) two men in shirts and ties discussing de-
sign problems with a woman in an archi-
tect's office, she holding her glasses, they
smoking LUCKY STRIKE.

Bottom pictures: (x) woman, classically draped in black, hit by
' a limelight in a lonely atmosphere with a
SMIRNOFF VODKA martini.

(y) young fellow in painter‘s garb, sitting down
among buckets of paint and ladder, being
brought a ROYAL CROWN COLA.

(z) middle-aged man sitting hunched over in
a hOSpital setting, smoking nervously a
cigarette, and wondering how to pay the
hospital bill and household expenses at
the same time (MONEY).

70

3. Women's Luxury and Romance versus Rugged Men

 

Among fifteen women interviewed, eight had ideal self notions of
this type and were of all three age—groups. Most typical, how-
ever, were the middle-aged wife of the gas station owner and the
old wife of a high school teacher.

Top pictures: (a) woman in evening dress being waited on at
beach table for SMIRNOFF cocktail. Color-
ful fruit on surrounding trees gleam in the
light of the sinking sun.

(b) a couple traveling in a BOEING 707, the
man looking out of the window, the woman
talking to stewardess while knitting.

(c) "A Diamond is Forever. " A girl on the t0p
of a hill sitting against the trunk of a round
tree and a man lying down with his head in
her lap. Among the red fruit in the tree
they see a Sparkling diamond.

Bottom pictures: (x) two men in band uniforms. The older one,
with a tattooed hand, offers the younger a
light for his cigarette (MARLBORO).

(y) a man washing fire truck while another
man stands by'to talk for the length of a
Cigarette (LUCKY STRIKE).

(2) football player being carried by an en-
thusiastic crowd (LIFE).

4 . Old Men's Luxury versus Youth and Dating

 

Only two old men--an executive director and a sales manager,
both close to retirement--form this cluster of persons.

Top pictures: (a) elderly couple sitting on a white carpet in
front of a large MAGNAVOX dreaming of
symphony orchestras, jazz bands, and
folk music in foreign countries.

(b) pilot adventuring with airplane in the rugged
arctic, camping with his crew around fire
at night (CAMELS).

(c) elderly grey-haired man smiling behind
the wheel of his Open-topped CHRYSLER.

71

Bottom pictures: (x) young man in Sports shirt leaning against a
wall close to a young blonde, both smiling
(LISTERINE).

(y) boy and girl kissing over an open Lane
Sweetheart Chest crowded with things
marked ”his" and “hers. ”

(z) a man washing a fire truck while another
man stands by to talk for the length of a
cigarette (LUCKY STRIKE)--same as in
factor 3.

V\
5. Home, Family, and Possessions versus Wliskey and "Odd"
Situations

 

Two men and three women, all middle-aged or old would like
their ways of life to be of this type. Most prominent among
these are a middle-aged engineering teacher and the wife of a
secretary.

Top pictures: (a) mother surrounded by five children who
play with a variety of trading stamp re-
wards in a comfortable, rustic living—room
(S 81 H GREEN STAMPS).

(b) elderly couple looking at the foundations
of their new home piled high with FLINT-
KOTE products (same as in factor 2).

(c) boy surrounded by 30 volumes of the
Encyclopedia Americana dreaming of his
future as a successful scientist or captain
of industry.

Bottom pictures: (x) two men in cowboy attire loading a case of
IMPERIAL whiskey onto the back of a mule.

(y) man riding a motor scooter with a case of
IMPERIAL whiskey on the back and waving
to a couple on another scooter coming
towards him.

(2) old man and woman, he in undershirt and
suspenders washing a sweater in WOOLITE
Cold Water Soap with old-fashioned spectacles
on the forehead, she reading a racing form
to him.

72

6. Sports versus Dieting and Old Age

 

Out of ten young and middle-aged Inen interviewed, six have
ideal roles of being admired sportsmen. Most typical among
this group of men are the ideal self notions of a young student
and of a young policeman.

Top pictures: (a) Tommy Bolt in a McGREGOR DRIZZLER
finishing a golf swing with a gallery of fans
in the background.

(b) driver of a racing car surrounded by a
swarm of mechanics after winning the race
(CHAMPION spark plugs). This picture
was second lowest in factor 1.

(c) football player being carried by an enthusi—
astic crowd (LIFE). Same picture was
lowest in factor 3.

Bottom pictures: (x) old man and woman, each with a grocery
cart looking at box of Kellog's Special K
in a context of other material on diet and
exercise.

(y) ”Fitness is for Everyone. " Stout woman in
leotard energetically doing her setting-up
exercises.

(2) old man and woman, he in undershirt and
suspenders washing a sweater in WOOLITE
Cold Water Soap with old-fashioned spectacles
on the forehead, she reading a racing form
to him (same picture was lowest in factor 5).

Summarizing these findings it is noteworthy that certain pictures
elicit identification mechanisms only among women, while others only
among men. Also, while some pictures are successful only among
Vollnger subjects, others appeal predominantly to older subjects. This

highlights the importance of model selection when identification in

refeI‘ence to ideal self notions is intended for a specific group of potential

 

buY'Eirs in an audience where not everybody is expected to be interested

1n the product.

73

When ideal self notions were involved, more than half of the
married women of all age-groups identified themselves with pictures
showing ”Women's Luxury, " where the most appealing motives were
”holidays, ” ”travels, ” and ”love. " Here, no pictures showing men
and women in association with children ranged high in identification.

Two groups of pictures elicited identification in men only, when
their ideal self notions were involved. ‘ More than half of the young
and middle-aged married men identified themselves with pictures
.showing “Sports" scenes. Old men preferred models suggesting
"Men‘s Luxury. " Again, these two groups of men did not rank pictures
showing adults in association with children very high.

Among the three other factors, which were defined by men as
\vell as women, one was composed of men and women of all age-groups,
another of younger men and women, and a last one of older men and
\vomen. Here identification occurred very often with pictures showing
parent-children relationships and the motives were “home" and "family. ”
When these three groups of persons are fairly similar in their acceptance
of "Happy Family" type pictures, they are more dissimilar in what
they reject as objects of identification. In the first case, ”Status

Seeking" is rejected, in the second ”Financial Worries, " and in the

third "Whiskey and ’Odd' Situations. "

Six Types of Actual Self Notions

In this section, the tentative interpretations and "names" for

the Second set of six factors are proposed, on the basis of those

pictures which stimulated most and least identification in the repre-

 

 

 

sentatives of each cluster of persons when they were asked: "How

W5 my present way of life to the way of life of the person(s) in

W? ”

74

Appendix I, F 2, presents in "standard scores" to what extent

each picture was identified with, across subject groups. Again,

pictures producing a very narrow range of standard scores across

factors were eliminated from further consideration.

FIG. IV shows the pictures which obtained most and least intense

 

 

identification in reference to actual self notions.

1 . Popularity versus Wild West and High Society

 

Four out of five old women interviewed define this type, in
addition to an old man and a middle-aged man and woman. Most
typical in their actual ways of life are the wife of a retired State
Employment Commissioner and the wife of an old engineer. In
their ideal self- notions, both were of the type "Home, Family and
Possessions versus Whiskey and 'Odd' Situations" (Factor 5).

Top pictures:

Bottom pictures:

(a) mother surrounded by five children who play

with a variety of trading stamp rewards in a
comfortable, rustic living—room (S & H
GREEN STAMPS. Same as top picture in
ideal self Factor 5).

(b) man next to his FORD station wagon sur-

(C)

rounded by a crowd of friends and goods he
has transported in it during one year's time.
elderly parents with their four children
among a confusion of luggage and household
goods about to move into a new home in an
unfamiliar town (TIME INC. Same as top
picture in ideal self Factor 1).

portrait of Walt Disney‘s TEXAS JOHN
SLAUGHTER against a background fighting
scene in a Western village.

woman, classically draped in black, hit by
a limelight in a lonely atmosphere with a
SMIRNOFF VODKA martini.(Same picture
among bottom pictures in ideal self Factor 2).

(z) elegant couple being served BOURBON on

the rocks in a fashionable restaurant in front
of a framed mirror reflecting a chandelier.

75

FIG. IV. TOP AND BOTTOM PICTURES or ACTUAL

SELF FACTOR ARRAYS

I' II'

 

   

26

C

7 _ = ‘ 39
£1.11 :1 .. ‘ C Jun/(1H bW

     
  

f...

1.
'- ..
a.”

THE ONLY
LAWN FOOD!

 

 

 

76

IV'

INNSIHL h "- {FIST '

 

4 .
I~ -vI~‘ 1W 1 -;- up... ~4L.
‘A‘ \--->ku-(‘nd' .-
1 C
L ' Luu — bw

.l m 1.74 an

O

 

"nondwcnmn c

77

2. Children versus Romance, Wealth, and ”Odd'l Old Age

 

Predominantly young and middle-aged women define this type,
along with one man of each of the three age-groups. Most typical,
however, are a young student of physical education and a middle-
aged teacher of engineering. In his ideal self, the former was

of the "Happy Family Life versus Status Seeking” type, while the
latter was of the ”Home, Family and Possessions versus Whiskey
and 'Odd' Situations” type (Factors 1 and 5).

Top pictures: (a) parents with their little boy and girl re—
laxing happily on their lawn (GOLDEN
VIGORO lawn food. Same picture among
top pictures in ideal self factors 1 and 2).

(b) boy surrounded by 30 volumes of the Ency-
clopedia Americana dreaming of his future
as a successful scientist or captain of
industry (Same picture among top pictures
in ideal self Factor 5). _

(c) two boys in letter sports shirts playing word
games with ALPHA-BITS at the breakfast
table (Same picture among top pictures in
ideal self Factor 1).

Bottom pictures: (x) ”Evening in Paris. ” Couple in evening robes
leaning out of a window at Manguin's into
a snowy Paris night lit by antique lantern.

(y) man next to his FORD station wagon sur-
rounded by a crowd of friends and goods
he has transported in it during one year's
time (Same picture among top pictures in
actual self Factor 1).

(2) old man and woman, he in undershirt and
suspenders washing a sweater in WOOLITE
Cold Water Soap with old fashioned Spectacles
on the forehead, she reading a racing form
to him (Same picture was lowest in ideal
self Factors 5 and 6).

3. Married Life versus "Odd" Old Age and Lonesome
Women with Drinks

 

 

Members of this group were five young and middle-aged men
and women, who were best represented in their actual ways of
life by a young student and his wife. The husband's ideal self

78

was of the type ”Sports versus Dieting and Old Age, " while his
wife's ideal self was of the type "Women's Luxury versus Rugged
Men" (Factors 3 and 6).

Top pictures: (a) parents at breakfast table spelling "LETS
GO" with ALPHA-BITS to their son who
plays with a new toy train instead of hurry-
ing to school with the books his mother holds
for him.

(b) parents with their little boy and girl relax-
ing happily on their lawn (GOLDEN VIGORO
lawn food. Same picture was among top
pictures in ideal self Factors 1 and 2, and
was tOp picture in actual self Factor 2).

(c) two women preparing a picnic on the beach
for two men who are waiting at a distance

(COCA COLA).

Bottom pictures: (x) old man and woman, he in undershirt and
suspenders washing a sweater in WOOLITE
Cold Water Soap with old-fashioned spec-
tacles on the forehead, she reading a racing
form to him (Same picture was lowest in
ideal self Factors 5 and 6, and in actual
self Factor 2).

(y) woman, classically draped in black, hit by
a limelight in a lonely atmOSphere with a
SMIRNOFF VODKA martini (Same Picture
among bottom pictures in ideal self Factor
2 and in actual self Factor 1).

(z) women in evening dress being waited on at
beach table for SMIRNOFF cocktail. Color-
ful fruit on surrounding trees gleam in the
light of the sinking sun (Same picture was
top picture in ideal self Factor 3).

4. Courtship versus Sporty Adventures

 

Young and middle-aged men and women define this factor in
their actual ways of life. A young student of police adminis-
tration and the wife of another policeman are most repre—
sentative in their group. In his ideal self the man was of the
"Sports versus Dieting and Old Age” type, while the woman
was of the "Women‘s Luxury versus Rugged Man" type
(Factors 6 and 3).

79

Top pictures: (a) two women preparing a picnic on the beach
for two men who are waiting at a distance
(COCA COLA. Same among top pictures
in actual self Factor 3).

(b) woman behind wheel of a CHRYSLER driv-
ing her husband to his office, he in a hat
with his face half hidden by the morning
newspaper he is reading.

(c) boy and girl kissing over an open Lane
Sweetheart Chest crowded with things
marked ”his" and "hers" (Same picture
among bottom pictures in ideal self Factor 4).

Bottom pictures: (x) woman in riding habit jumping a horse over
a hurdle, watched by friends in the back—
ground (UNITED STATES BREWERS
FOUNDATION. Same among bottom pic-
tures in ideal self Factor 1).

(y) driver of a racing car surrounded by a
swarm of mechanics after winning the race
(CHAMPION Spark plugs. Same among
bottom pictures in ideal self Factor 1, and
among top pictures in ideal self Factor 6).

(2) pilot adventuring with airplane in rugged
arctic, camping with his crew around fire
at night (CAMELS. Same among top
pictures in ideal self Factor 4).

5. Leisure versus Sporty and Elegant Entertainments

 

The actual self notions of two old me11--an executive director

and a sales manager both close to retirement--and of a policeman
are of this type. One of the old men and the policeman are most
representative. The latter had an ideal self of the ”Sports versus
Dieting and Old Age" type, while both businessmen had ideal selves
of the ”Old Men's Luxury versus Youth and Dating" kind (Factors

6 and 4).

Top pictures: (a) young soldier on leave in his room, which
is decorated with trophies, stretching out
on a BEAUTYREST by SIMMONS mattress
with his dog climbing upon him.

(b) elderly couple sitting on a white carpet in
front of a large MAGNAVOX dreaming of
symphony orchestras, jazz bands, and folk
music in foreign countries (Same as top
picture in ideal self Factor 4).

80

(c) boy surrounded by 30 volumes of the Ency-
clopedia Americana dreaming of his future
as a successful scientist or captain of
industry (Same picture among top pictures
among ideal self Factor 5 and actual self
Factor 2). ’

Bottom pictures: (x) woman in riding habit jumping a horse over
a hurdle, watched by friends in the back—
ground (UNITED STATES BREWERS
FOUNDATION. Same among bottom pic-
tures in ideal self Factor 1 and in actual
self Factor 4).

(y) woman in evening dress being waited on at
beach table for SMIRNOFF cocktail. Color-
ful fruit on surrounding trees gleam in the
sinking sun (Same as top picture in ideal
self Factor 3 and as lowest picture in
actual self Factor 3).

(z) driver of a racing car surrounded by a
swarm of mechanics after winning the race
(CHAMPION spark plugs. Same among
bottom pictures in ideal self Factor 1,
among top pictures in ideal self Factor 6,
and among bottom pictures in actual self
Factor 4).

6. Work versus Sports

 

A young man, an old man, and an old woman are of this type in
their actual ways of life. Most representative are the two men.
The young one, an Air Force sergeant, had an ideal self of the
”Home and Good Job versus Financial Worries" type, while the
other, a high school teacher, had an ideal self of the ”Home,
Family and Possessions versus Whiskey and 'Odd’ Situations”
type (Factors 2 and 5).

Top pictures: (a) ”Gerhard Baecker teaches Volkswagen. ”
Man with his arms crossed in front of his
chest standing in front of blackboard showing
schemes of motor parts.

 

 

 

81

(b) elderly parents with their four children
among a confusion of luggage and household
goods about to move into a new home in an
unfamiliar town (TIME INC. Same as top
picture in ideal self Factor 1 and among top
pictures in actual self Factor 1).

(c) young fellow in painter's garb, sitting down
among buckets of paint and ladder, being
brought a ROYAL CROWN COLA (Same
among bottom pictures in ideal self Factor 2).

Bottom pictures: (x) baseball-star swinging the bat. BANTRON

"If you want to stop smoking. ”

(y) woman behind wheel of a CHRYSLER driving
her husband to his office, he in a hat, with
his face half hidden by the morning news—
paper he is reading (Same among top pic-
tures in actual self Factor 4).

(z) football player being carried by an enthusi-
astic crowd (LIFE. Same picture lowest in
ideal self Factor 3, and among t0p pictures
in ideal self Factor 6).

Summarizing once more, it seems remarkable that there is no
type of actual self notion to be found exclusively among women, similar
to the type of ideal self strongly prevailing among the women inter-
viewed. Still, two types of actual selves were found among two-thirds
of the female subjects. The first, ”Popularity versus Wild West and
High Society, " was found particularly among old ladies, while the
second, "Children versus Romance, Wealth, and 'Odd' Old Age, ” was
represented mostly by young and middle-aged married women, who
are likely to have children at home in their actual ways of life.

Only one type of actual self was exclusively found in men, where
two out of three were old, and it was ”Leisure versus Sporty and Elegant
Entertainments. ” It is interesting that the actual selves of men were
of a large variety of types; that is, they were much more diversified
in their actual self notions than in their ideal self notions where the

”Sports versus Dieting and Old Age" type united six self notions into a

group.

82

It may be more difficult to choose a model for actual self identifi-
cation with the intention to appeal to a specific group of potential buyers
in an audience where not everybody is expected to be interested in the
product. This is suggested not only by the finding that actual self
factors more often cut across sex and age differentiations, but also by

the greater similarity of the picture groups which elicited most and

 

least identification when actual selves are involved than when ideal
selves are involved.

Among the pictures which stimulated most intense identification
in reference to the actual selves of persons in the various groups, there
are; in all cases, pictures showing "happy family life" and/or pictures
showing "parent—children relationships. " Because all the interviewed
persons were married, it is not amazing that these pictures were more
often identified with in reference to actual self notions than to ideal self

notions. This shows that everyday situations are most indicative to the

 

subjects of their actual ways of life, and to a minor extent to their ideal

ways of life .

CHAPTER VII

SUPPLEMENTARY FINDINGS

Some supplementary findings will be reported in this chapter.

To what extent each of the 72 pictures included in the Q-sample stimu-
lated identification in respect to actual and ideal self notions will be
discussed, and a few outstanding pictures are pointed out.

FIG. V indicates that there is a substantial correlation between
the average Q-scores each picture obtained when the pictures were
sorted according to the instruction, "How much would I like my way of
life to be like the way of life of the person(s) in the picture?” and the
average Q-scores it obtained when sorted on the basis of the instruction,
"How similar is my present way of life to the way of life of the person(s)

in the picture? " This suggests that models who appealed to actual self

 

notions in our 30 subjects also tended to appeal to their ideal self

 

notions .

Despite this general tendency, however, it can be seen from
FIG. V that some models were more successful objects of identication
in reference to ideal self notions than in reference to actual self notions
(e. g. picture no. 31). Others appealed to actual self notions more than
to ideal self notions (e. g. picture no. 42). Still others aroused similar
intensities of identification in respect to both self notions (e. g. pictures
no. 39, 27, 11, 29). We shall briefly describe the exemplary pictures,

which can be found reproduced in Appendix I, Fig. l.

83

 

 

 

84

 

FIG. V. Scatterdiagram Comparing Average Ideal and Actual
Q-Scores Obtained for 72 Pictures
Ideal 101!‘
Self
26‘.“
91. . ,,
:‘31
o o
8 .. ' ‘
O
O
o .- .
7 «r . o (’29 O
.. . o co .
“117;; °
C
o 0 o . . . .
6+ 256‘ o
0 f . o
' o
' 0
5i; . . . . o
:’ - . ‘
E . .
9 o
4‘)" . o
(Oj39 . . /“
“" @142
O
3‘b
”Q68
2 4 +4 .L 4 4. +2 4
2 3 4 5 6 7 8 9

Actual Self

85

1. High "ideal self" identification but low "actual self”

 

identification.

 

Picture No.

31

Elderly couple looking at the foundations of
their new home piled high with FLINTKOTE

products.
Average "ideal" Q-score: 8. 9
Average ”actual" Q-score: 6. 3

. Low "ideal self" identification but high ”actual self"

 

identification.

 

Picture No. 42

Middle-aged man sitting hunched over in a
hospital setting, nervously smoking a
cigarette and wondering how to pay the hospital
bill and household expenses at the same time
(MONY).

Average "ideal” Q-score: 3.
Average ”actual" Q-score: 6.

6
4

3. Similar intensities of ”ideal self” and ”actual self”

identification.

 

Picture No.

Picture No.

Picture No.

39

27

ll

Portrait of Walt Disney's TEXAS JOHN
SLAUGHTER against a background fighting
scene in a Western village.

Average "ideal" Q-score: 3. 5
Average ”actual” Q-score: 3. 5

Young couple admiring balloons with two little
boys--one is riding on daddy's shoulders.
Accompanied by their dog they stand close to
a shiny blue RENAULT Dauphine.

Average "ideal" Q-score: 6. 0

Average ”actual” Q-score: 6. 1

Girl typing while stretched out on a shadowy
lawn in front of her CHEVROLET, which sits
in the background close to the woods.
Average ”ideal" Q-score: 6.4

Average ”actual" Q-score: 6. 3

86

Picture No. 29 Happily smiling grey-haired couple, he in
shirtsleeves and tie standing close to his wife,
who is engaged in a long distance call, while
an album with photographs of children lies
Open in front of her. (BELL TELEPHONE

SYSTEM).
Average "ideal" Q-score: 7.1
Average "actual" Q-score: 7. 2.

FIG. V also shows which pictures stimulated most and least
identification in reSpect to the ideal ways of life of the men and women
interviewed. Similarly, the most and least appealing pictures in
reference to actual ways Of life can be found. It happened that in the
present picture set the same picture has been most successful in
eliciting identification in connection with both self notions, and the same

picture has been least successful in connection with both self notions.

4. Highest "ideal self" as well as ”actual self” identification.

 

Picture NO. 26 Parents with their little boy and girl relaxing
happily on their lawn (GOLDEN VIGORO lawn
food.

Average ”ideal" Q—score: 9.4
Average "actual" Q-score: 8. 5

5. Lowest "ideal self" as well as "actual self" identification.

 

Picture No. 68 Old man and woman, he in undershirt and sus-
penders washing a sweater in WOOLITE Cold
Water Soap with old fashioned spectacles on
the forehead, she reading a racing form to him.
Average "ideal” Q-score: 2. 5
Average "actual“ Q-score: 2. 7

It should be noted that pictures 26 and 68 obtained respectively
highest and lowest average identification scores in relation to ideal
selves of young and middle aged subjects. In relation to their actual
selves, picture 26 again was most successful in eliciting identification

among young and middle-aged persons, while picture 68 was least

87

successful only for middle-aged subjects, since young subjects rejected
another picture more intensely.

Neither in respect to their ideal selves nor in respect to their
actual selves did old people provide highest or lowest average scores
for pictures 26 and 68 reSpectively. This suggests that the “happy
parents" role describes their actual and ideal ways of life to a more
minor extent than it does the actual and ideal ways of life of younger
people, which is plausible in view of the fact that all subjects were
married. Similarly, the role of the ”Odd" old couple is less vehemently
rejected by old peOple than by young ones, even if they are not at all
fond of it.

It may be asked which pictures, as Objects Of identification,
discriminated most between men's and women's actual and ideal selves,
even if FIG. V fails to provide a clue to this matter. Here, a distinc-
tion has to be made between whether a model facilitated identification
in men, while it restricted it in women, or whether it facilitated
identification in women and restricted it in men. Such pictures, with
differential appeal to male and female audiences, may be particularly

useful in winning a desired audience for a certain product.

6. High "male" identification but low "female" identification
in respect to ideal self notions.

Picture NO. 6 Pilot adventuring with airplane in the rugged
arctic, camping with his crew around fire at
night (CAMELS). '
Average ”male" Q-score: 8. 5
Average "female" Q-score: 3. 8

7. High “female" identification but low ”male" identification

in respect to ideal self notions.

 

Picture NO. 15 Portrait of a bare-shouldered blonde bending
over a golden embroidered cushion against a
golden background, holding a flask of perfume
on a thin golden chain, hypnotizing the reader

88

with her staring eyes: ”he'll concentrate .
concentrate . . . concentrate on you--when
you wear HYPNOTIQUE. "

Average "male" Q-score: 3. 7
Average ”female" Q-score: 5. 9

8. High ”male" identification but low "female” identification
in reSpect to actual self notions.

 

Picture NO. 9

Elderly retired man in sportshirt sitting in a
basement on a working bench well equipped
with tools. He is building and painting an
elaborate, small sailing boat.

Average ”male" Q-score: 7. 3
Average “female" Q-score: 4. 5

9. High ”female" identification but low "male" identification
in respect to actual self notions.

Picture NO.

50

Portrait of a young girl with closed eyes and

appreciative smile, her face and neck sprinkled
with drops of water and partly covered with the
foam Of DIAL soap, her hair tied up on her head.
Average ”male" Q-score: 5. 7
Average "female" Q-score: 8. 3

It appears that by means of Q-sorts an evaluation of the stimulat-

ing potential of single pictures in reference to other pictures included

in the Q-sample becomes possible. Q-sorting procedures provide

estimates of average scores of actual and/or ideal self involvement in

the identificatory reactions of individual subjects or subject groups.

Thus Q-technique may be a useful tool for pretesting pictures for their

effectiveness in respect to identification.

CHA PT ER VIII

CONCLUSIONS

In this last chapter, the main conclusions from the reported
findings will be drawn, and sonie of their implications for practical
advertising will be pointed out. In addition, a few prOpositions for
further investigation will be discussed.

Through an empirical investigation Of identification mechanisms,
a contribution has been made to the understanding of how the receiver
of a pictorial message may relate himself to human models depicted in
such messages.

It has been found that at least two different ways of identification
with a pictorial role-stimulus must be distinguished. That is, the
viewer may either identify himself with the model by comparing it in
its qualities and actions to himself in his actual way of life, or he may
identify himself with the model by comparing it in its qualities and
actions to himself in his ideal way of life.

Identification in respect to actual as well as ideal self notions was
found to be facilitated if sex and age correspondences prevailed between
the receivers of the message and the depicted models. Thus, women
tended to identify with female models, while men preferred to identify
with male models. Young subjects chose young models as their objects
of identification, middle-aged subjects chose middle-aged models, and

old subjects chose old models. Hence, similarity appears to be a

 

criterion Of choice of the identificand insofar as sex and age character-

istics are concerned.

89

90

However, the results suggest that the presence of correspondences
in sex and age characteristics among viewers and models increasingly
enhances identification when persons grow Older.

Concerning the relative effect of the three picture variables--
”sex of model, " “age of model, " and "color Of the picture“--upon the
intensity Of identification, it was found that whether the model is a man
or a woman is more influential than whether the picture is colored or in
black and white. This seemed to hold for identification in reference to
actual self notions as well as ideal self notions.

The age of the depicted model affected identification in particular
where actual self notions were involved. Actually, the age of the model
here became more influential than the sex of the model or the color of
the picture. This, however, was not the case when ideal self notions
were involved. Here, the age of the model became the least important
determinant of identification.

Men and women of all age-groups gave a striking advantage to
colored pictures when they chose their ideal Objects of identification,
which was not the case when they chose their actual Objects of identifi-
cation.

Since it was expected that persons are selective in respect to the
role-stimuli in pictures which they accept or reject as their objects of
identification, audience segmentations were assessed by grouping per-
sons into particular "types" according to their identification with similar
models.

In the hypothetical audience studies, six) "types" of persons could

be distinguished when identification on the basis of their ideal self

 

notions was at issue. These "types" were given the following, tentative

names:

91

Happy Family Life versus Status Seeking.

. Home and Good Job versus Financial Worries.

Women's Luxury and Romance versus Rugged Men.

. Old Men's Luxury versus Youth and Dating.

. Home, Family, and Possessions versus Whiskey and
”Odd" Situations.

6. Sports versus Dieting and Old Age.

UTr-PLMNt—d

Three of these "types" were defined by both men and ‘NUHH'H,
while the other three were either exclusively male or exclusively ft'lrlétlu.
The ”Happy Family Life" Factor was comprised of men and women ”I
all age-groups, while the "Home and Good Job" Factor was comprised
of younger men and women, and the "Home, Family, and Possessiunh”
Factor, of older men and women.

The "Women's Luxury and Romance” Factor included em lusivvly
women Of all age-groups. Actually more than half of the women inter-
viewed were Of this "type. "

A few Old men were included in the "Old Men's Luxury" Fetter,
while half of the young and middle-aged men interviewed belonged to
the ”Sports" Factor.

The three Factors, "Happy Farriily Life, " ""I ornen's Luz-it'l/ and

“A- ll .— '=' ' , . ‘ 3 . ' 1 L. ' ’ ; .
Roniance, and SportS, alO...‘ accounted for t«.0-’.r..r',-, 0: 1m: intral

’ O: x ,— 1 ' ‘1 ‘If. a 1 1+ 1. ‘ -. . : G - . I , ‘ n
581:. no--0.".b reDresenteO GITIOTIZ tfi‘: J) person: .'..’,’J /.‘,I": ,f. -‘f!’ /,"f I.‘,’
O .1

In

- . ,,
' .r 3 r r :. r"~~ .r .r‘ r r .
a 4. 1‘ A A- - J- 1,440 4 J 1‘: A a... ‘.o (.3. l ['3‘

f I
‘l . ' . 9" - -,. , o..; l , g I . l . O ..
,,,fi ” , v‘ -.. , o- - ,- ,-= r r o ( .1 {. rt r r p 4 ,. , {-
---f:-. -F“ ...... Ca--0.'l 0.. ..-.‘3 2:13.; 0. ’3.- - is... 1 .. 1 . 1 . r r: .6 ,, . . , . . 1
‘ - o -0 q z . - 0-, r yr! 7 r, " ( r f : '. V I C' I v, ‘ ‘ ‘
" -o’f“"h f f’v‘v’ 9' f f , l " . " , ' ,fl ,’ ,
a;—--J.A;_-a ---ere --$.J.J‘:.o‘: A v0 -/{ I'D-‘4 :6...- 4 .nA. on I a I: " .J I. to a /
.,. b a .
v C ’ h o
"‘ .fi " o . ¢ 0- .- - I, , ,
,- a -, 1"- ’7' rp I ‘Vr (-7 f f- Y .u...
a 4‘3: ere :- U‘:.a ,,-, o J‘-J a g, — Ar. ' r a a a - a
a . J 4
'I ~— - . 1" < ’ " . r v I. .
, .n’ro—v " ’l-- d 4” v!’ f' ‘, fr ,. ~/ I I.
-. - J.)..-'.;-.- a‘.-343 H-... H 3:- ¢ - ...,._;.. , - ,
a a
- o -u 0 I f‘ ‘ r I ' r I I
) , ,.,. ,--,.. P"""' ,. r , ‘_
”C C AAAAAA 8-- l‘:. 3‘: -J-l ’- { .0 C.,_-’ a a a J! J. I It, 0
- ' ’ ' I 7
4' ";rr z’,. I Cor; c‘,’f’- {- r/‘I ( r ”1". ;,/I , a/t‘.;rrr (: [‘1’]: r
J. -0-..- no -a H-. 4 0 v- ‘4 4 J4 .l J. 4 pop 4 av o H
,
us
- '0'- ,v f-
0-.- one- #-
’ O ' O
' .—"' "‘ov’- ’ ‘V ' " 00’ 5 " "vn
1. 1‘.- - 4---.) I ,A .14»! al.11- ' on .4 - ‘ r a.
a .
I' '— 0“
r l . O .r r‘ t II I‘
a- e v o v r fivr ’. ’7’ ’ :’ ,/’ ’ u ', f
.2. L‘:‘:A‘ ‘1 :‘1. 3A: ”1.1) r A- .4. “a ’3’ r -4 VI. - -,
.
r -. I‘ -
’ r "’ f’ ’ §rr ’ '.
C. n {. .-:.-.--, ”.13--..
p

92

Here, only the ”Leisure” Factor was exclusively composed by
men, while all others were comprised of men as well as women.

Predominantly older persons had loadings on the "Popularity"
Factor, while predominantly younger persons had loadings on the
"Married Life" and the "Courtship" Factors. People of all age-groups
made up the ”Children, " "Leisure" and ”work" Factors.

The most prominent Factors were "Popularity, " ”Children" and
"Married Life, ” because they accounted together for two-thirds of the
actual self notions represented among all the subjects interviewed.

On the whole, the six "types" of Mself notions were more
different from each other than were the six ”types” Of actual self

notions.

A Discussion Of Implications for Practical Advertising

 

Two major Obstacles make it difficult to point out implications
our findings may have for practical advertising.

The first Obstacle is a lack in evidence on the question of whether
the advertiser should try to elicit identification in the reader at all, in
order to gain attention, Obtain interest, stimulate desire, enhance
recall, or persuade to buying actions. In addition, more specific
evidence is needed on whether he should appeal to actual self notions,
ideal self notions or to both simultaneously in his attempts to fulfill
his purposes. This lack in evidence requires that any implications
drawn from the results of this study be introduced by the conditional
phrase: "If the advertiser intends to elicit identification in respect to
actual and/or ideal self notions, then. . . ."

The second obstacle is to be found in the limitations Of this study,
in regard to the Specific set of pictures employed, the small number of

persons interviewed, and the consideration Of advertising illustrations

93

in isolation from headlines, copy, and, most important, the advertised
product.

Still, a few suggestions may be made where relatively little risk
is involved.

The finding that people identify with models Of their own sex
hardly will be of interest to the advertiser who knows from previous

1 . . .
research that women are more interested in female models and men in

 

male models. He may, therefore, already have selected women as
models when he wished to sell products to women, and men as models
when he intended to sell products to men.

From the theoretical standpoint, however, the finding is important,
because it supports especially Hattwick's interpretation of this phenomenon,
when he explained it as due to identification mechanisms. 2

Also, the finding that people identify with models of their own
age-group has more theoretical than practical importance, as the art
director may already have selected young models in order to appeal to
young potential consumers, or old models for older potential consumers.

A frequently discussed issue is whether or not the increase in
audience size attributable to colored pictures, as opposed to black and
white pictures, justifies the augmentation in printing expenditures.

The present findings suggest that the art director, who intends
to enhance identification in reference to ideal self notions, should be
aware that colored pictures are much more likely to be of help in
achieving his purpose than black and white pictures are.

If, however, he plans to elicit identification in reference to actual

self notions, color appears to be of minor importance, and thus the

 

lRudolph, 313. 935., pp. 72—73.

ZHattwick, 2p. c_it., p. 160.

94

increase of costs is less likely to be defensible in this case. Certainly,
this only concerns the impact of color upon identification and does not
take into account the impact upon attention, recall, and other variables
which apply to the illustration.

Also, the noted tendency that with growing age it becomes more
important that the pictures be colored and that the models be of the age
and sex of the readers in order to enhance identification may merit the
attention of the art director, even when a more rigorous investigation
of this phenomenon is necessary.

In regard to a segmentation of the audience on the basis of the
selective reactions Of people to models in advertising pictures, the art
director should be aware that persons are of one ”type” when they
identify with models on the basis of their ideal selves, and of another
when they identify with models on the basis Of their actual selves.
Thus, a certain model may elicit identification in some Of the persons
in a larger audience when ideal self notions are at issue, while it may
elicit intense identification in other persons when actual self notions
are involved.

How many different "types" of actual and ideal self notions are
to be dealt with in audiences other than the hypothetical one studied
here cannot be estimated from our results. Still, the evidence Obtained
suggests that at least the six "types” of actual and ideal self notions
assessed and provisionally named in this study will have tO be counted
with among men and women of Iniddle socio-economic level living in an
urban environment.

In general, pictures concerning "home, " "happy family, ” and
"parent-child" relationships enhanced identification, though more
persons chose them in reference to their actual self notions than to
their ideal self notions. Thus, we find only one actual self factor which

does not exhibit a single picture Showing a child among the three top

95

pictures. But there are three ”types" of ideal selves where no children
are depicted in the three pictures best describing each of the prevailing
ideal ways Of life. In order to make these results plausible, it has to
be stressed again, that all subjects were married and most of them
probably had children.

Thus evidence from previous research suggesting that pictures
showing adults ,with children rank high in interest, may be explainable,
to some extent, by the fact that such picture subjects are successful in
eliciting identification. 1

Among the three "types” of ideal selves which were found to be
less concerned with "home, " "happy family, " and, especially, the
"parent-child" motives, two deserve the special interest of the adver—
tiser because they discriminate sharply between men and women, and
because they may be rather widely represented in advertising audiences.
The ideal self notions of younger men seemed to center upon "Sports, "
so that pictures showing sports scenes stimulated their identification
successfully. A large number Of women Of all ages described their
ideal ways of life with pictures showing "Women's Luxury and Romance. "

Hence, the ”Spectator Sports” appeal, which, according to
MacLean and Hazard, accounted for some of women's interest in pic-
tures, also appears to enhance identification in young men, while the

reported ”Idolatry" picture appeal enhances identification among women. 2

 

1Rudolph, op. cit., p. 74. See also, G. Gallup, “Let 4,979,855
Readers Tell You-What. They Read on Sunday. ” Survey made by the
Gallup Research Bureau for Kimberly-Clark Corporation, reported in
Hattwick, op. cit. , p. 157. (Pictures of children were highest in atten-
tion, follow—ad Syr—those of groups of adults, sports scenes, animals,
natural scenery.

2‘MacLean and Hazard, Op. cit. The present Q-data could be
factor-analyzed on the basis Of—piOtT-ire with picture correlations. First,
it would be interesting to compare kinds of picture-appeals in reSpect
to actual and ideal self notions. Second, it could be seen whether or not

 

 

96

The most general result is that the ”everyday situation" tends to
facilitate identification, while uncommon, extravagant, or odd situ-
ations or roles inhibit identification. This result agrees with those

reported from an investigation under taken by True Story magazine.1

 

Because persons seemed to be more different in their ”types" of
ideal selves than in their ”types” of actual selves, role-idealization
may turn out to be the more predictive attribute of the depicted model,
when identification in a particular audience group is sought.

Further research is necessary to suggest whether the art director
should aim at actual and/or ideal self involvement, when he attempts to
elicit identification in order to better achieve persuasion of the potential
consumer.

Assuming the art director wished to select the model which most
enhances actual and ideal self identification in younger men out of a set
of pictures created to advertise the same product, actual and ideal
Q-Sort scores may be obtained for each picture from a number of sub-
jects who are young men. Scatterdiagrams should be drawn for each
picture, on which the scores for actual and ideal Q-sorts would be
compared. The scatterplot most resembling the one shown in FIG. VI
would then indicate the picture which would fulfill his purpose best.

Appendix 1, FIG 2, shows, as an example, the scatterdiagram for
one of the pictures included in the present Q-sample. From it, one

can see that not one of our subjects rejected the depicted "parent-child"

 

certainkinds of picture-appeals elicit identification in certain kinds of

self notions, which were reported earlier. Third, kinds Of picture-
appeals in respect to identification could be compared with kinds of

 

picture-appeals in reSpect to interest, which were found by MacLean
and Hazard.

1How to Get People Excited. Especially prepared by the editors
of the best selling magazine on America‘s newsstands for others whose
occupation is Arousing People to Buy. (Published by True Story, 1937).

 

 

97

FIG. VI. Scatterdiagram Showing Size and Similarity of Ideal and
Actual Q-Sort Scores for a Single Picture in a Set.

m

Ideal 12

. Self

 

10 x x x

--—=- - . .
I

I.

.4 -sr.._- — .1— -~ ..

 

 

.-._._—+-4-—— ._» _..

"T‘ ‘ "'“T‘”

’ ' ......'...- 1...--.1

10 "'11 12’
Actual Self

O
.—
L»
.p.
m
0‘
\1
CD
\0

 

98

role in respect to both self notions. Some accepted it in respect to
both self notions. Others rejected it in reference to ideal but not to
actual self notions, while still others rejected it in reference to actual
but not to ideal self notions. In the diagram, it can easily be checked
if men or women or even persons of the same age-group reacted
similarly to the depicted role or not.

This little test may be Of help in the selection of models and roles,
after the art director has decided to which self notions he wishes to

address the stimulus .

 

Propositions for Further Research

On the basis of a discrimination between identification mechanisms
in reference to actual and ideal self notions, further research should
investigate if a positive correlation between interest in advertising

illustrations and identification with the depicted models exists.

 

FIG. VII prOposes how illustrations could be grouped according
to the identificatory reactions which are stimulated by the models they

ShOWT

l. A model may elicit identification neither in reference to actual
self notions nor in reference to ideal self notions.

2. A model may elicit identification in reference to actual self
notions but not in reference to ideal self notions.

3. A model may elicit identification in reference to actual and
ideal self notions simultaneously.

4. A model may elicit identification in reference to ideal self
notions but not in reference to actual self notions.

When we interpret this scheme according to Festinger's theory

of ”cognitive dissonance, "1 it appears that neither consonance nor

 

dissonance is at issue when no identification with the model occurs

 

 

1L. Festinger, A Theory of Cognitive Dissonance (Evanston,
111.: Row, Peterson and Company, 1957).

 

99

FIG. VII. Four Combinations Of Identificatory Reactions in vavrvnrv
to Actual and Ideal Self Notions.

Ideal
Self

 

 

model appeals model appeals
to ideal self to ac'tual self
but not to and to ideal
high actual self: self:
"dissonance" "(:()ns(n'1ilr1('¢'”

model appeals model Appeals
neither to to a(_t11al self
actual self nor but not to

low to ideal self: ideal self:
neither "conso- "(lissonanus“

nance I'lOI~

"dissonanc e"

low high AI mal "Jt-lf

——-.- v-i- _—..

 

(I), because the vie',:,-er Will not be led 2,0 corn-part: ar. .' of I". e 21 1.; ’ ’7
played roles with any of his ideal roles,

U" ' l ,o.' .10.,~33,03.’l +0; 5 ,5, , ,0 ,4,(
"men a rnocel s.1!’."1hla~.‘-:: 1r.‘-r1,...r_a,.or. .r. f‘..‘.f‘..’.’ I. -r, oar, 81,,

 

 

. r .. . . . .
- . 0 O~ . 9 l n -
hf" I" v’ '9 r- 9 " ~ r. a 9 fr 0 I r .r’ I r’vr (o I , ,I I -1 I . r f r ’4
A.U--O... ~J.A- ong' :n r‘:“:r‘:aa(_ {' “(J rpAI— I-p-‘rl ’ I. 1F... .. D a. I I /.,’o. ‘ j 1., a
- . .. . . - - I I ~ : ,
P ' ~~ -’r ‘ cv- -' -’ r .1( ‘rr' off / 9 ' 't" to ' r / I r r. I ‘, 4 “I r ‘, . ,
UECQ‘AS‘: 6--..61' -r.(: 'n‘: "' (:r ‘5 I ‘4‘1‘lfil 1‘. J I! 4’ O ’1 a "to! ID 04': 30” ’1 . I ’I” I. ‘.r
I
. q r g I . g . I I
r-.. ’ -..—‘ I . r o 1 v r. f , f r , r r ,‘ , I /‘, z r .1 / r ’- , ‘, .’
06-: a:..‘:‘ 4 D J- Lab‘:’ ~J4- a a n n ‘ I l‘- I: bit I .1- I ’l -1, In a a
A D
vevrargrr- r: 7" ." -’( :’;(-' 9' l-’ " 7' 'r' .’ ’ ‘ ’-’ 4" ’l' ’/ ’ "
a o
o C--.-..":... J- ’4‘» a J v a p I .d‘p I p I A I :1. 4 ’ a I I. :11 I a I o i:
,
-p-P rv‘v- I v. --y—‘ .t f I 9/ I
.4 .
a--v JJJJ4 -_.. - J A J l I I‘
. O I
. I I
4 — - v 0 .. r I .r it I I r ’ r r I r r p l 1 ,
J’l-r-Ellv a: f ‘. I 1 c v j .1 ‘1 Al: I t ‘ ‘ I I c ./.I’ lit a I I t "r I I ,
I I
- ‘ t
-- - - - — o o C a Ir ’ t I r I 1,, I a r r , z I a r -
or v ff / ’ tr r,r g r ’ f I .’ , ' I, ' I r I III '
- a- ........ r I. ’ I t l'
r w a ' o , ' , r ,1, frz ’ ,1 I a r t I g, /1 I 51 a I I / ’,, x
r r ‘—
,-J_-_ ‘-.::'.: f": z’ ’4‘. (I: .I . 1 . ’Z 0 I. . . . . . ~t ./ ., ... , ,

100

When the two kinds of models who, through identification, produce
cognitive dissonance in the viewer, i. e. (Z) and (4), were compared
with each other in respect to their efficiency in arousing interest in the
viewer, the following could be predicted:

Persons will show less interest in pictures showing models with
whom they identify in reSpect to their actual self notions but not
to their ideal self notions (2) than they will in pictures showing
models with whom they identify in respect to their ideal self
notions but not to their actual self notions (4).

1 a rationale would

Drawing upon Festinger's basic hypotheses,
be that, in the first case (2), persons tend to avoid information producing
the psychologically uncomfortable dissonance, which cannot be sub-
stituted for consonance, because they may be forced by external circum-
stances to play the unsatisfactory role.

In the second case, information may not be avoided, since here
some Opportunity is given to reduce dissonance (4). Here the model
exhibits a role which is aSpired to by the viewer, who, when he takes
the place of the model in fantasy, may at least momentarily reduce
dissonance. Furthermore, these models often bear the promise that
the viewer can reduce dissonance if he acquires the external attributes
of his aspired role, that is, if he buys the advertised product.

Also, an empirical investigation of the relative interest in models
appealing to both self notions simultaneously (3) and models appealing
to ideal self notions rather than to actual self notions (4) needs to be
undertaken. Here the concern is with the comparative interest in
models producing cognitive consonance and in one kind of models pro-
ducing cognitive dissonance. The following could be hypothesized:

Persons will exhibit more interest in pictures showing models
with whom they identify in respect to both self notions simul-
taneously (3) than they will in pictures showing models with
whom they identify in reSpect to their ideal self notions but not
to their actual self notions (4).

 

11bid., p. 3.

101

Since persons are thought to be motivated to maintain consonance
through the avoidance of all information which threatens dissonance,
and to reduce dissonance wherever it occurs, models of simultaneous
appeal to both self notions should be particularly interesting to the

reader, as they reinforce his satisfaction with the status quo.

 

Pictures stimulating identification in respect to neither actual
nor ideal self notions (1) would provide a suitable control-group, where
it could be seen what the interest is in pictures which do not stimulate
identification.

It is clear that most pictures will belong to more than one of the
four groups, depending upon the persons who judge their identification
with these pictures. Thus, the same picture may appeal to both self
notions in one person and, at the same time, to none in another person.

Also further studies are needed to find out if a positive corre-
lation between recall of advertising illustrations and previous identifi-
cation with the depicted models exists. Here, again, models who appeal
to each self notion with different intensity, thus producing dissonance,
may be forgotten sooner then models who appeal to both self notions
simultaneously and to a similar extent, thus producing consonance.

But even if identification were found to be correlated with interest
and recall, the advertiser who intended to incorporate a stimulus of
identification into his message should always keep in mind that he has
to match the illustration with the headlines and the c0py and that all

three need to be relevant to the product.

BIBLIOGRAPHY

Books

 

Atkinson, J. W. , et a_l. Motives in Fantasy, Action, and Society:
A Method of Assessment and Study. Princeton: D. van Nostrand
Company, Inc. , 1958.

 

 

Burt, C. The Factors of the lxlind. London: London University Press,
1946.

 

Cattell, R. B. Factor Analysis. New York: Harper & Bros., 1952..

 

Crawford, J. W. Advertising: Communications for Management.
Boston: Allyn and Bacon, Inc., 1960.

 

Festinger, L. A Theory of Cognitive Dissonance. Evanston, 111.:
Row, Peterson and Company, 1957.

 

Freud, S. Group Psychology and the Analysis of the Ego. London:
The Hogarth Press and The Institute of Psycho-Analysis, 1949.

 

Gross, N., Mason, W. S., and McEachern, A. W. Explorations in
Role Analysis: Studies of the School Superintendency Role.
New York: John Wiley & Sons, 1958.

 

 

Hattwick, M. S. How to Use Psychology for Better Advertising.
New York: Prentice Hall, 1950.

 

Hovland, C. 1., Janis, I. L., and Kelley, H. H. Communication and
Persuasion: Psychological Studies in Opinion Change. New Haven:
Yale University Press, 1953.

 

 

Lindquist, E. F. Design and Analysis of Experiments in Psychology
and Education. Boston: Houghton Mifflin Company, 1956.

 

 

Lippmann, W. Public Opinion. New York: Harcourt, Brace & Co.,
Inc., 1922.

 

103

Lucas, D. B. The Controlled Recognition Method for Checking Maga-
zine Advertising Readership. New York: The Crowell-Collier
Publishing Company, 1942.

 

 

Lucas, D. B. and Britt, S. H. Advertising Psychology and Research.
New York: McGraw-Hill Book Company, Inc., 1950.

 

MacNemar, Q. Psychological Statistics. New York: John Wiley 3;
Sons, 1949.

 

OSgood, C., Suci, G. S., and Tannenbaum, P. H. The Measurement
of L-leaning. Urbana: University of Illinois Press, 1957.

 

 

Rudolph, H. J. Attention and Interest Factors in Advertising. New York:
Funk & Wagnalls Company with Printers' Ink Publishing Co. ,
Inc., 1947.

 

Stephenson, W. The Study of Behavior: Q-Technique and Its Metho-
dology. Chicago: The University of Chicago Press, 1953.

 

Thibaut, J. W. and Kelley, H. H. The Social Psychology of Groups.
London: John Wiley 8: Sons, 1959.

 

Thurstone, L. L. Multiple-Factor Analysis. Chicago: University of
Chicago Press, 1947.

 

True Story' editors. How to Get PeOple Excited. True Story: 1937.

 

Webster's New Collegiate Dictionary. London: C. Bell 3; Sons, Ltd.,
1951.

 

Articles and Periodicals

 

Alden, R. ”Advertising: Survey Looks at the Models, " The New York
Times, April 11, 1961.

 

Burtt, H. E. and Crockett, T. S. ”A Technique for Psychological
Study of Poster Board Advertising and Some Preliminary Results, "
Journal of Applied Psychology, XII (1928), 43-55.

 

Butler, J. M. and Haigh, G. Y. "Changes in the Relation Between Self—
Concepts and Ideal Concepts Consequent upon Client-Centered
Counseling, ” in C. R. Rogers, and R. F. Dymond, (eds.).
Psychotherapy anciPersonalit-x Change. Chicago: Univer

1
J

U}

u, I
Av” (J1

Chicago Press, 1951', pp. 55-76.

104

Duncan, D. B. "Multiple Range Tests for Correlated and Heteroscedas-
tic Means, ” Biometrics, XIII (1957), 164-176.

 

Duncan, D. B. ”Multiple Range and Multiple I: Tests, " Biometrics, XI
(1955), 1-42.

 

Gallup, G. ”Factors of Reader Interest in 261 Advertisements, "
Liberty Magazine, 1932.

 

Gilbert, H. N. ”An Experiment with Colored and Uncolored Advertise-
ments, " Journal of Applied Psychology, XVII (1933), 49-54.

 

Jones, E. E. and deCharms, R. "The Organizing Function of Inter-
action Roles in Person Perception, " Journal of Abnormal and
Social Psychology, LVII (1958), 155-164.

 

 

Kaiser, H. F. "The Varimax Criterion for Analytic Rotation in Factor
Analysis, ” Psychometrika, XXIII (1958), 187-200.

 

' Kitson, H. D. and Allen, 1. "Minor Studies in the. psychology of Adver-
tising from the Laboratory of Indiana University: XIII. Pictures

of PeOple in Magazine Advertising, " Journal of Applied Psychology,
IX (1925), 367-370.

 

Klapp, O. E. "Immitation-Value in Advertising, " Journal of Applied
Psychology, XXV (1941), pp. 243-50.

 

 

“Maccoby, E. E. and Wilson, W. C. ”Identification and Observational
Learning from Films, " Journal of Abnormal and Social Psy-
chology. LV (1957), 76-87.

 

MacLean, M. S. Jr. and Hazard, W. R. ”Women's Interest in Pictures:
The Badger Village Study, ” Journalism Quarterly, XXX (1953),
139-162.

 

Nahinsky, I. D. ”The Relationship Between the Self-Concept and the
Ideal as a Measure of Adjustment, " Journal of Clinical Psy-
chology, XIV (1958), 360-364.

 

Nixon, H. K. "Two Studies of Attention to Advertisements, " Journal
of Applied Psychology, XII (1928), 43-55.

 

Nunnally, J. C. "An Investigation of Some Propositions of Self-
Conception: The Case of Miss Sun, " Journal of Abnormal and
Social Psychology, L (1955), 87-92.

 

 

l05

Sarbin, T. R. ”Role Theory, " in G. Lindzey (ed.). Handbook of
Social Psychology. Cambridge, Mass.: Addison-Wesley Publish-
ing Company, Inc., 1954, I, pp. 223-258.

 

 

Stephenson, W. ”Some Observations on Q-Technique, ” Psychological
Bulletin, IL (1952), 183-98.

 

Twedt, D. W. "A Multiple Factor Analysis of Advertising Readership, "
Journal of Applied Psychology, XXXVI (1952), 207-215.

 

Unpublished Material

 

Gallup, G. "Let 4, 979, 855 Readers Tell You What They Read on
Sunday, " Survey made by the Gallup Research Bureau for
Kimberley-Clark Corporation, reported in M. S. Hattwick,
How to Use Psychology for Better Advertising. New York:
Prentice Hall, 1950, p. 1.57.

 

Stephenson, W. A General Theory of Mass Communication. 1960.
(Mimeographed.)

 

APPENDICES

106

Table A, 1. --Variance Estimates and E-ratios for Subject Groups

 

 

Source

df SS

V

F P

Five Young Male Subjects on the Actual Self Q-Sort

 

Rows (color) 1 1.11 1.11 .16 --
Columns (sex) 2 4.86 2.43 . 35 --
Slices (age) 2 27.67 13.84 2.00 --
C.xS. 2 29.29 14.65 2.12 --
C.xA. 2 9.11 4.56 .66 --
S.xA. 4 13.70 3.43 .49 --
C.xS.xA. 4 6.44 1.61 .23 --
Within cells 342 2357.82 6.89

Total 359 2450.00

 

Five Middle-Aged Male Subjects on the Actual Self Q-Sort

 

 

 

Rows (color) 1 24.89 24.89 3.70 --
Columns (sex) 2 2. 26 1.13 .16 --
Slices (age) 2 3.30 1.65 24 --
C.xS. 2 7.50 3.75 55 --
C.xA. 2 10. 39 5.20 77 --
S.xA. 4 79.88 19. 97 2 97 < 05
C.xS.xA. 4 22.92 5.73 85 --
Within cells 342 2298.82 6.72

Total 359 2449. 96

 

Five Old Male Subjects on

the Actual Self Q-Sort

 

 

 

 

Rows (color) 1 19.73 19.73 5.93 <.05
Columns (sex) 2 56. 30 28.15 4. 62 <. 05
Slices (age) 2 147.01 73.51 2 07 <.Ol
C.xS. 2 72.87 36.44 5.98 <.Ol
C.xA. 2 10.74 5.37 88 --
S.xA. 4 22. 18 5.55 .91 --
C.xS.xA. 4 37.64 9.41 1.54 --
Within cells 342 2083.53 6.09

Total 359 2450.00

Continued

107

108

Table A, 1. --Continued

A—.__

 

Sourc e (if SS V F P

 

Five Young Female Subjects on the Actual Self Q-Sort

 

 

Rows (color 1 3.23 3.23 .46 --
Columns (sex) 2 3.20 1.60 .23 --
Slices (age) 2 60.06 30.03 4.48 <.05
C.xS. 2 36.13 18.07 2.70 --
C.xA. 2 4.54 2.27 . 33 ,--
S.xA. 4 8.19 2.05 .30 --
C.X5.XA. 4 47.35 11.84 1.76 --
Within cells 342 2287. 30 6.69

Total 359 2450.00

 

Five Middle-Aged Female Subjects on the Actual Self Q-Sort

 

 

 

Rows (color) 1 6.35 6.35 1.01 --
Columns (sex) 2 155.19 77.59 12.39 <.01
Slices (age) 2 8.46 4.23 .67 --
C.xS. 2 50.52 25.26 4.03 <.05
C.xA. 2 3.50 1.75 .27 --
S.xA. 4 66.51 16.63 2.65 <.05
C.xS.xA. 4 24.74 6.19 .98 --
Within cells 342 2142.69 6.26

Total 359 2457.96

 

Five Old Female Subjects on the Actual Self Q-Sort

 

Rows (color) 1 2.31 2.31 .37 --
Columns (sex) 2 109.02 54.51 8.79 <.Ol
Slices (age) 2 122.41 61.20 9.87 (.01
C.xS. 2 51.54 25.77 4.15 <.05
C.xA. 2 3.15 1.58 .25 --
S.xA. 4 14.60 3.65 .58 --
C.xS.XA. 4 52.44 13.11 2.11 --
Within cells 342 2121.51 6.20

 

Total 359 2476. 98

109

I

Table A, 2. --Variance Estimates and E—Ratios for Subject Groups

 

Sourc e df SS V F P

 

Five Young Male Subjects on the Ideal Self Q-Sort

 

 

Rows (color) 1 54.81 54.81 8.71 <.01
Columns (sex) 2 78.51 39.26 6.24 <.01
Slices (age) 2 2.73 1.37 21 --
C.xS. 2 40.47 20.24 3.21 <.05
C.xA. 2 60.50 30.25 4.80 <.Ol
S.xA. 4 23.52 5.88 0 3 --
C.xS.x.A. 4 35.62 8.90 1.41 --
Within cells 342 2153.84 6.29

Total 359 2450.00

 

Five Middle-Aged Male Subjects on the Ideal Self Q-Sort

 

 

 

 

 

 

 

 

Rows (color) 1 139. 93 139. 93 23.16 <. 01
Columns (sex) 2 41.49 20.75 3.43 <.05
Slices (age) 2 6.87 3.44 . 56 --
C.xS. 2 61.26 30.63 5.02 <.01
C.xA. 2 38.12 19.06 3.15 <.05
S.xA. 4 68.33 17.08 2.82 <.05
C.xS.xA. 4 30.34 7.58 1.25 --
Within cells 342 2067.62 6.04

Total 359 2453.96

Five Old Male Subjects on the Ideal Self Q-Sort

Rows (color) 1 57.86 57.86 9.51 <.01
Columns (sex) 2 19.19 9.59 1,57 __
Slices (age) 2 81.44 40.72 6.69 <.01
C.xS. 2 118.21 59.11 9.72 <,'01
C.xA. 2 30.79 15.39 2.53 --
S.xA. 4 43.57 10.89 1.79 -.-
C.xS.XA. 4 11.18 2.79 .45 --
Within cells 342 2079.75 6.08

Total 359 2441. 99

 

Continued

110

Table A, 2- -Continued

 

 

Sourc e (if SS V F P

 

Five Young Female Subjects on the Ideal Self Q-Sort

 

 

 

 

 

 

 

 

Rows (color) 1 132.92 132.92 22.3 <.01
Columns (sex) 2 169.83 84.92 14.24 <.01
Slices (age) 2 19.29 9.65 1.61 --
C.xS. 2 34.83 17.42 2.92 --
C.xA. 2 7.33 3.67 .61 --
S.xA. 4 33.50 8. 37 1.40 --
C.xS.xA. 4 11.82 2.95 0.49 --
Within cells 342 2040.48 5. 96
Total 359 2450. 00

Five Middle-Aged Female Subjects on the Ideal Self Q-Sort
Rows (color) 1 122.47 122.47 22.22 <.01
Columns (sex) 2 336.25 168.13 30.51 <.01
Slices (age) 2 6.62 3. 31 .60 --
C.xS. 2 39.09 19.55 3.54 --
C.xA. 2 13.09 6.55 1.18 <.05
S.xA. 4 53.66 13.41 2.43 --
C.xS.xA. 4 7.91 1.97 . 35 <.05
Within cells 342 1885.89 5.51
Total 359 2464. 98

 

Five Old Female Subjects on the Ideal Self Q-Sort

 

 

ROWS (color) 1 51.73 51.73 9.09 <.01
Columns (sex) 2 299.51 149.76 28.75? <.01
Slices (age) 2 56.57 28.29 4.97 ' <.01
C.xS. 2 21.50 10.75 1.88 --
C.XA. 2 7.35 3.68 .64 --
S.XA. 4 25.89 6.47 1.13 --
C.xS.xA. 4 38.65 9.66 1.69 --
Within cells 342 1948.80 5.69

 

Total 359 2450. 00

 

111

 

 

 

 

00.0 mm.m N00 0m.0 3.0 +0.0 m0.m 24.0 20
0nd mH.0 No.0 0m.0 M00 00.0 N40 >00 22
0H0 No.0 00.0 40.0 24.0 mm.0 $0.0 N40 2.».
mm.0 «0.0 00.0 mm.0 00.0 45.4 m0.m :20 .mO
00.0 40.0 0H0 N00 :.0 40.4 :43... 40.0 MB
HH.0 00.m NN.0 N0.0 «0.0 +0.0 231.0 00.0 rmw
30m 16.03 ”cofisnmfimwfl 0.86m 2:32
m0.0 mm.m 000 mN.0 30.0 3.0 3.0. mN.0 20
00.0 00.0 no.0 NH.0 000 000 p.010 0m..0 $3.4
M00 00.m mm.0 No.0 mw.m m~.0 m0.m 00.0 2w.
00.0 MN.m 4N0 00.0 :10 mN.m N00 00.0 .mO
N00 N00 0H0 00.0 N00 0N0 m~.0 >00 {m2
000 :t.m 00.0 40.0 m~.0 mod 00.0 000 h.»
Sow H9304 ”comusnwfimmm “zoom ado;
mozmfium> 0.353% 0 E s» we 0 E .50 o msoemssm

mo man—0.20

 

 

Xom 0cm om< 0:98:00 00 30002

meg/00m £300 Empowfifl mo mousuofim mcflpomsd c0 mQSOQO 26032.5 Scum 0028000 9:90 £802.... :m 93.95

112

Table C. --Eta Coefficient Estimates (7)) of Variable Contribution To
Total Variance in Q-Sort Scores

 

 

Subject Picture Variables '
Variables Color Sex Age C x S C x A S x A

 

Actual Self

 

 

YM .022 .045 .106 - + - - + - +
MM .101 .030 .036 + - + - —+
OM .090 .152 .245 —+ -+ -+
YF .0363 .0360 .157 + - - + - +
MF .051 .251 .058 -+ -+ +-
OF .030 .210 .222 -+ -+ —+
Variable Relationships: C < S. C < A S < A

 

Ideal Self

 

YM .150 ., .179- .033 —+ +— +-
MM .239' .130 .053 +— +- +_
OM .154 .089 .182 + - -+ —+
YF .233 .263 .089 - + + — + -
MF .223 .370 .052 - + + - + ..
OF .145 .350 .152 —+ -+ + ..
Variable Relationships: C < S C > A S > A

 

Note: Algebraic signs designate directions of relationship:
(+) designates the larger effect and (-) the smaller effect in a
pair of effects.
The symbols (>,<,:) summarize directions of relationship.

 

113

Table D, 1"‘——Correlation Coefficients for Ideal Self Q-Sorts of 30 Subjects
(1:: 72 Pit 111;: s)

 

y—d
[v
L2.)
.12
U1
0\
J
00
\O
1,—4
O
1““
p—J

12 13 14 15 16 17 18 19 20 21

N
[‘6'

23 24 25 26 27 28 29 30

 

1 0 +20 -03 +38 —07 +33 +19 +32 +19 +28 +22 +27 +11+32 +14 +07 +08 +15 -10 +18 +15
2 7 -07 +32 +26 +12 +30 +32 +42 +42 +34 +51 +42 +25 +33 +38 +09 +35 +22 +15 +38 +27
3 354 —22 +22 —O€> +39 —00 +39 —02 +42 +04 +33 +10 +22 +04 +31-05 +13 -15 +16 -26 +08 +06
4 _ +34 +58+32 +42 +23+42 +19 +48 +23+33 +28 +39 +41 +55 +29 +61 +26 +34 +47 +54 +44 +57 +23 +25 +49 +49
5 - +33 +33 +32 +191-17 +14 +02 +09 +08 +28 +02-1‘25 +31 +22 +24+g7 423.11.142.10 +41 +7171+19+10+48 +32
6 - +26 +48 +23+35 +24 +57 +37 +43+38+47+49+51+47 +65 +36 +33+61+58+30+51+20+40+31+39
7 — +62 +19 +12 +28 +06 +2 +07 +39 +21+29 +24 +16 +43 +20 +29 +20 +13 +25 +38 —01 +01 +33 +36
- +18 +10 +35 +25 +34 +26 +32 +36 +26 +39 +16 +51 +40 +42 +33 +25 +20 +48 +09 +23 +40 +40

— —06 +19 -08 +43 +01+39 +13 +39 +30 +31+21+38 +16 +27 -03 +07 +05 +13 -19 +29 +13

8

9
10 -— +00 +42 +03 +25 +11 +40 +19 +30 +19 +33 -01 +21 +36 +41 +29+39 +06 +50 +05 +33
11 +22 +29 +26 +17 +14 +16 +02 +17 +24 +13 +14 +34 +23 +22 +22 +11 +04 —01+05
12 — +29 +53+10 +33 +31+32 +24 +59 +03 +13 +39 +71 +27 +50 +20 +53 -09 +28
13 — +29 +38 +17 +35 +12 +18 +39 +06 +09 +41 +27 +16 +12 +32 +08 +O1+OO
14 — +18 +18 +28 +09 +12 +34 +01 +09 +21 +41+28 +28-+32 +43 +03 —06
15 — +44 +35+18 +39 +34+34+37 +37 +18 +22+27 +29 —07 +22 +33
16 - +20 +30 +44 +35 +26 +49 +48 +37 +26 +40 +20 +38 +13 +39
17 - +21+41+35 +23 +20 +42 +20 +23 +23 +14 +10 +24 +24
18 — +38 +60 +29 +35 +35 +49 +21+43 +01 +34 +51 +40
19 - +40 +15 +32 +61+31+21+26 +09 +12 +11+25
2.0 — +36 +24 +48 +76 +13+46 +02 +32 +21 +42
3 - +29 +18 +15 +17 +24 +19 —03 +30 +18
2 — +29 +20 +33 +60 +13 +21 +45 +52

- +51 +31 +30 +15 +19 +06 +37
24 — +18 +47 +08 +45 +09 +31

25 — +51 +43 +22 +26 +24
26 — +29 +32+39 +44
27 _ +06 +07 —06
28 - +06 +17
29 — '—36
30

 

 

Note: Subject Numbers 1, , . . . 19 are men.

3, 5
2, 4, 6, 30 are women.

1 - 10 are young.

11 — 20 are middle—aged.

21 — 30 are old.

Decimal points are omitted.

 

 

114

(n 7‘ 72 Pictures)

iable D, 2 ——Correlation Coefficients for Actual Self Q—Sorts of 30 Subjects

 

 

 

 

 

 

 

 

 

 

1 2 3 4 5 6 7 8 9 10 11 1.2 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
1 _ +24 +02 +07 -15 _10 —32 ~06 _07~42-~05+04+17+01-03—05 —01+05 -01 -14+04_11-19+02+12+07+01+02 -08 -07
3 - +08-00*10*04-OZ +09+39+17 -06 +29+11+30 +03+19+12+10 +10 -01 -08+15+05+11+06+O9+15+23+11 -01
3' - +67 +16 +29 +07 +28 +16 +28 +28 +41 +06 —07 +45 +42 +34+2O +36 +34 +06 +21 +29 +22 +05 +14 +01+18 +20 +24
4 — +29 +40 +02 +38+18+30+36 +44 -12 -11+45+42+42+35 +38+33+24+21+32+34+17+39 —02+33 +35 +43
: - +43 +09 +51+23 -14 +20 +37 +03 +28 +16 +33 +27 +36 +39+34+18 +41 1264374131425 -06 +22 +46 +28
9 - +09 +34 +06 -07 H0 +59 -09+11+i4+31+14+19 +51+22-+07.+26+32+27+01+18 -'33+08 +47 +24
7 _ +10 +20 —09 +08 +27 —18 —02-+18»+08 -05 +19 +28 +17 —12 —02+39+15+20 +07—15+18+18+05
8 - +34+13+11+31 —1&+07+22+10+35+34 +31+31+34+37+14+16 +20 +39—08 +21+56 +22
‘9 - +13 +03 +50 +04 +49 -07 +12 +14 +09 +23 +16 +19 +03+17 +23 +32 +31+15 +22 +29 —01
13‘ — +19++1+1Z-07+18+14+07 +42 +19+12_13+03+13+37_02 +23+02+38 +05—02
11 - +26 +09 +02 +29 +23 +20 +16 +20 +34 +10 +23 +46 +34 +21 +27 —02 +29 +33 +41
13 - +08 +£34.13 +42+16+44 +36 +45 ~11+29+26+34+O8+45 —08 +43 +42+26
,3 - +30 -06 +03 +01-12 -00 -11-03 +07 —18 —10 —12 —10 +18 +11—20 —11
14 - ~02 +00 +01+02 +24 -04 +05 +09 —04 —06 +00 +13 +06 -05 +09 +02
+5) - +50 +42 +24 +33 +40 +06 +26 +24 +11+08 +20 -00 +37 +40 +37
”3 - +39 +37 +23 +42 +17 +24 +30 +28 +09 +22 _ 16 +35 +33 +32
17 - +31+29 +32 +40 +10 +10 +03 —05 +12 +24 +1 +40 +13
+8 - +40 +40 +08 +28 +25 +42 +03 +32 —27 +38 +37 +18
0 - +30+11+26 +42+26 +04 +24—24+22 +47+11
2: - +18 +49 +48 +27 +11 +37 —16 +34 +38 +41
21 — +32 +09-11+34 +31+15 +14 +23 +20
22 - +26 +26 +13 +47 —10 +42 +34 +52
23 _ +50 +25 +18 —06 +32 +34 +41
24 _ +05 +19 —09 +32 +28 +37
25 _ +44 +12 +27 +24+25
26 - +07 +45 +47 +42
2“, _ +08 —13 -13
28 — +28 +29
29 _ +50
3C1 _

Note: Subject Numbers 1, 3, 5, . . .29 are men.
\ 2, 4, 6,. . . 30 are women.
1 — 10 are young.
11 — 20 are middle—aged.
21 — 30 are old.
'Deciniel points are omitted.
. 1

 

 

 

115

 

 

 

 

4.3+ 3+ 3+ 8+ 3+ 8+ 3+ 8+ 2+ 3+ 8+ 8+ 8
8- 8- 8+ 3+ 2+ 8+ 43- 8+ 8+ 8+ 43- 44+ 3
48+ 8+ 8- 8+ 3- 8+ 8- 8+ 8- 8+ 3+ 8+ 3
43- 8+ 2+ 3+ 3- 2- 8- 3+ 44- 42- 8- 3+ 3
8+ 3+ 22+ 48+ 2- 88+ 8- 8+ 43- 48+ 3+ 8+ 3
43- 8+ 2+ 48+ 8- 3+ 8- 43+ 84- 42+ 48+ 3+ 3
8+ 3+ 43- 3- .8... 8+ 8.. 42.. 8+ 3.. 84+ 88+ 43
3- 8+ 8- 2- 3+ 88+ 48+ :4 8+ 3- 48- 8+ 3
24+ 2+ 2+ 3+ 8+ 8+ 3+ 2+ :- 24+ 8- 8+ 3
4o- 8- 34+ 44+ 8- 3+ 2- 8- 8+ 2+ 8- 24+ 3
8+ 2+ 2+ 8- 2+ 8+ 8- 48- 8+ 2- 42+ 83 8.
2- 3- 8- 8- 8+ 8+ 8+ 3+ 2+ 3- 2- 48+ 2
42+ 8- 2- 2- 8- 3+ 2- 8+ 3+ 3+ 8+ 8+ 2
2- 84.. 3+ 8+ 42- 84+ 8- 2+ 8- 3- 3- 48+ 2
8+ 2- 8+ 2- 8- 88+ 8+ 2+ 8- 8+ 2+ 8+ 2
8- 42- 3+ 3- 8+ 48+ 2+ 8+ :- 8- 88- 3+ 2
8+ 2- 3- 3+ 2- 8+ 3- 8- 48- 8- 8+ 84+ 42
2+ 2- 2- 8+ 3- 8- 8- 2- 2- 84- 48- 84+ 2
8+ 8- 3- 4o- 2- 48+ 8- 8- 8- 3- 88+ 88+ 2
48+ 3+ 2+ 42- 8+ 84+ 2+ 3- 3- 8- 2- 8+ 2
8+ 2+ 8- 38- 8- 3+ 2+ 2+ 8+ 88+ 84+ 44+ 2
3- 8+ 3- 48+ 3- 38+ 2- 8+ 2+ 2- 8- 3+ 8
8+ 3- 8+ 8+ 8+ 8+ 8+ 3.. 2- 3+ 2- 88+ 8
88- 3+ 3- 8- 3+ 43+ 2+ 44- 2- 43... 3- 84+ 8
2+ 3- 2- 8- 3+ 8+ 8- 8+ 8+ 2- 2+ 8+ 8
42+ 8- 8- 3+ 2+ 8+ 2.. 8+ 2- 3+ 2- 44+ 8
42- 2- 3+ 3- 2- 8+ 8- 8+ 8+ 2+ 8+ 3+ 4
3- 3- 2+ 8- 2- 8+ 8- :4 8- 3- 88.- 3+ 8
8- 8- 48- 2+ 84-. 2+ 3- 3+ 3+ 8- 2- 8+ 3
2+ 8+ 8+ 8.. 8- 8- 8+ 2- 8+ 2- 34- 3+ 2
S > >2 2 m H 5 > >2 5 : H 88638
288 8364. :68 282

 

 

 

800.346 om >6. vofigounm muuomna 6.25me mo 80m 03-H- noH mwnflomod Houumrm-i. .M 3an

116

Table F, 1. --Factor Scores in Standard Terms for Ideal Self Q-Sorts

 

 

 

 

Picture Factors
Number I II III IV V VI

1 +1.34 —2.39 -1.04 - .23 - .68 - .25
2 +1.79 +1.09 + .41 + .91 + .68 4 .00
3 -+ .45 + .65 - .62 4 .00 + .45 + .74
4 - .22 4 .00 -1.45 - .68 - .23 - .50
5 + .22 - .22 —1.66 -1.59 -1.58 +1.24
6 + .89 + .87 -1.04 +1.81 -1.13 +2.23
7 - .22 + .65 - .21 — .68 + .23 + .50
8 + .22 —1.09 + .21 +1.81 +1.58 +2.48
9 +1.12 + .43 -1.24 +1.59 - .45 +1.49
10 -1.34 -1.52 +1.04 -1.13 4 .00 - "74
11* + .45 4 .00 + .83 + .91 4 .00 + .50
12* 4 .00 + .22 + .41 + .23 - .45 - .25
13 +1.12 +1.52 +1.04 + .23 + .45 +1.49
14 -1.56 i .00 + .21 i .00 i .00 — .99
15 - .89 -1.09 + .62 - .91 -1.35 - .99
16 +1.34 + .43 +1.04 -1.36 - .23 — .25
17 -1.56 —1.95 4 .00 - .45 -1.80 -1.74
18 + .89 + .87 +1.04 - .68 +2.25 + .25
19* 4 .00 + .43 + .21 - .45 + .23 - .25
20 —1.56 - .65 +1.66 + .23 + .45 + .99
21 -1.12 4 .00 +1.24 -1.59 + .90 + .50
22 - .67 +1.30 + .83 - .91 + .23 +1.24
23 -1.56 - .65 +2.28 - .23 + .23 - .74
24 -1.12 - .43 +1.04 +1.13 + .90 + .50
25 +1.56 +1.09 4 .00 + .91 - .90 - .50
26 +2.01 +2.17 +1.24 + .91 +1.13 +1.74
27 + .89 4 .00 4 .00 - .23 4 .00 -1.24
28 + .89 +1.30 - .21 + .68 —1.35 +1.49
29 +1.79 +1.30 - .62 +1.13 - .23 - .25
30 + .89 4 .00 +2.07 +1.81 4 .00 +1.49
31 - .67 +2.17 +1.45 +1.36 +2.03 - .25
32 -1.34 +1.52 4 .00 4 .00 +1.35 -1.49
33 4 .00 -+1.09 +1.24 +2.72 + .90 4 .00
34 + .89 + .87 + .41 + .23 +1.80 + .25
35 - .67 - .65 -1.66 -1.13 - .45 + .25
36* 4 .00 - .22 - .41 4 .00 4 .00 -+ .50
37 - .22 - .43 - .41 i .00 + .90 - .50
38 + .22 - .43 ~1845 4 .00 + .45 +2.23
39 - .89 - .43 -1.45 -1.36 -1.58 - .50
40 -2.68 —1.74 - «41 i .00 - .45 -..50
41 .67 +1.30 -1.87 -1.13 +1.58 +2.23

 

(continued)

117
Table F, 1. --Continued

 

 

 

 

Picture Factors

Number I II III IV V VI
42 + .45 -2.39 -1.24 - .23 - .90 - .50
43 -1.79 -1.30 -1.66 -1.36 + .90 +2.23
44 4 .00 + .87 -l.66 -1.13 -2.03 + .99
45 + .22 - .22 - .83 +1.59 + .23 - .50
46 - .67 -1.09 - 383 -1.36 + .68 +1.49
47 - .89 -1.30 - .83 + .91 -1.13 - .50
48 4 .00 4 .00 - .83 +1.81 + .23 + .74
49 + .45 +-.22 + .62 + .91 +1.13 - .25
50 + .22 + .87 +1.04 4 .00 - .90 + .25
51 - .45 - .65 + .21 + .68 + .45 i .00
52 4 .00 - .43 + .41 - .23 — .68 - .25
53 - .22 - .22 + .21 -1.36 - .68 -1.74
54 - .22 + .65 + .62 — .45 + .23 i .00
55 +1.34 - .43 +1.04 + .45 + .45 - .50
56 +1.34 + .65 + .62 + .45 + .90 .- .50
57 4 .00 - .65 + .62 + .45 +1.35 - .99
58 + .22 + .87 - .62 -1.36 — .68 -2.23
59 + .45 - .43 + .21 + .23 - .68 + .25
60 + .45 + .65 - .21 + .68 - .23 - .74
61 + .67 +1.52 4 .00 + .45 +1.13 -1.99
62 - .89 -1.30 - .41 + .23 + .23 - .50
63 + .67 + .43 - .21 + .23 - .45 -1.49
64 - .67 - .22 - .62 -1.59 i .00 — .74
65 + .45 +1.30 +1.45 -1.36 - .45 + .25
66 - .89 - .65 - .83 -1.36 -2.03 + .99
67 - .89 -1.30 +1.66 + .91 + .68 4 .00
68 - .89 -1.30 -1.66 + .23 -2.25 -Z.98
69 +2.23 -1.09 + .41 + .23 +1.13 + .25
70 i .00 + .43 - .62 -1.13 i .00 - .25
71 - .67 4 .00 4 .00 - .68 -1.80 -1.99
72* -1.34 - .87 - .41 -1.13 -1.13 -1.24

Standard

IDevkuions: 2.24 2.30 2.41 2.21 2.22 2.01

 

Note: Positive algebraic signs in connection with standard scores indi-
cate that the reSpective picture was accepted as object of identifi-
cation by the corresponding type of persons. Negative algebraic
signs indicate rejection of the picture as object of identification.
Sizes of standard scores represent intensities of identification.

 

 

 

Pictures lacking in discriminatory power across types.

118

Table F, 2. --Factor Scores in Standard Terms for Actual Self Q-Sorts

 

 

 

 

Picture. Factors

Number I II III IV V ' VI
1 + .88 +1.38 + .21 - .46 — .47 +1.61
2 + .66 +2.07 +1.05 - .46 - .24 +1.15
3 - .22 - .46 i .00 +1.37 +2.12 + .23
4 - .22 i .00 - .42 - .91 4 .00 -l.61
5 -l.32 - .32 4 .00 + .23 4 .00 -1.38
6 + .22 - .69 -1.47 -2.28 +1.18 +1.38
7 +1.76 -1.84 + .42 + .91 + .47 + .23
8 - .44 + .69 +1.47 -1.37 +1.89 - .92
9 - .66 + .23 i .00 - .68 + .71 - .46
10 -1.10 -1.15 i .00 - .46 - .47 - .46
11 + .88 + .46 4 .42 + .23 -1.42 + .23
12 + .22 + .46 +1.26 +1.14 -1.42 - .23
13 + .66 + .92 +1.89 +2.05 +1.18 42.00
14 i .00 -l.15 -1.47 -2.05 —1.89 - .92
15 — .66 - .69 -1.47 - .46 -1.42 4 .00
16 - .44 +1.61 +1.05 +1.14 + .24 +1.15
17 -1.97 - .92 -1.68 + .23 -1.42 - .46
18 +2.63 + .46 +1.05 - .23 - .24 + .23
19 + .22 4 .00 - .21 + .91 + .24 -1.61
20 + .22 - .69 + .42 + .68 - .94 4 .00
21 4 .00 -1.38 + .84 +1.82 - .71 - .92
22 - .44 - .23 +1.68 +1.59 + .47 - .69
23 — .66 — .92 -2.10 + .91 -2.12 - .69
24 - .44 -1.84 4 .00 - .46 -1.18 - .69
25 - .44 - .92 - .42 - .23 - .24 + .46
26 +1.76 +2.30 +2.10 -1.14 + .94 +1.61
27 - .44 - .69 +1.05 - .68 - .24 - .23
28 + .88 +1.84 +2.51 - .68 - .24 +1.38
29 + .66 - .46 - .42 + .23 +1.65 + .23
30 - .88 - .23 + .42 -1.59 +1.18 + .69
31 + .88 - .69 - .42 + .68 + .24 4 .00
32 -1.54 - .92 i .00 + .68 +1.89 +1.38
33 + .88 - .69 — .42 + .23 +2.12 4 .00
34 + .88 +2.07 - .42 - .23 +2.12 + .46
35 -1.10 - .69 -l.47 -1.l4 - .47 - .23
36 +1.54 + .23 +1.05 + .46 + .47 - .69
37 +1.10 +1.61 + .42 +1.37 + .24 + .92
38 - .88 + .92 ' - .84 + .23 4 .00 -l.61
39 -1.76 - .69 -1.05 +1.37 -1.18 ~1.61
40 -1.32 - .69 + .63 +1.37 + .94 4 .00
41 + .88 -+ .23 - .21 - .91 - .71 -2.53

 

(continued).

119

Table F, 2. --Continued

 

 

 

 

Picture Factors

Number I II III IV V VI
42 -1.54 +1.84 - .42 + .23 + .24 +1.15
43 4 .00 - .23 -l.47 -2.28 -2.36 -1.61
44 —1.32 -1.15 - .84 -1.14 + .24 -1.15
45 + .44 4 .00 - .21 + .91 + .71 - .69
46 - .88 + .92 - .63 - .23 +1.42 +2.07
47 -1.32 + .69 - .42 - .23 + .24 4 .00
48 + .66 + .69 4 .00 - .91 4 .00 + .23
49 + .66 + .46 + .84 + .91 - .24 - .69
50 + .88 +1.15' - .42 + .23 + .24 + .46
51 + .66 -1.15 +1.05 -1.37 - .24 +1.15
52 + .22 4.00 - .21 - .23 4 .00 +1.15
53 + .22 - .92 - .21 - .23 +1.18 + .46
54 + .22 + .46 - .42 +2.05 - .47 -1.84
55 + .88 + .92 +1.68 - .23 -1.42 + .92
56 + .‘2 +1.38 +1.68 -1.37 - .47 + .92
57 4 .00 - .23 -l.05 - .23 -1.18 +1.38
58 — .44 - .23 - .63 - .68 - .24 -1.38
59 - .22 i .00 +1.05 + .68 - .24 - .46
60* 4 .00 + .23 + .21 + .46 ' - .47 + .23
61 +1.32 i .00 i .00 + .23 + .24 - .23
62 +1.32 i .00 + .63 i .00 - .24 - .46
63 + .88 + .69 — .84 - .23 i .00 +1.15
64 i .00 - .69 + .21 i .00 -1.18 + .23
65 + .44 — .23 + .63 +1.59 + .24 4 .oo
66 --1.32 - .69 i .00 - .68 - .47 i .00
67 -2.41 - .69 + .63 +1.59 - .47 + .23
68 -1.10 -2.53 -1.68 + .46 - .47 -1.61
69 +1.76 + .92 -1.26 -1.37 +1.18 +2.07
70 + .66 i .00 + .42 - .46 + .71 +1.15
71 - .22 + .69 -1.47 - .68 + .94 - .23
72 -l.54 -l.38 -1.47 - .23 - .71. i .00

Standard

Deviationszl 2.28 2.17 2.39 2.20 2.12 2.18

 

Note: Positive algebraic signs in connection with standard scores
indicate that the respective picture was accepted as object of
identification by the corresponding type of persons. Negative
algebraic signs indicate rejection of the picture as object of
identification. Sizes of standard scores represent intensities of

 

 

identification.

* Picture lacking in discriminatory power across types.

120

FIG. A. REPRODUCTION OF PICTURES DESCRIBED
IN CHAPTER VII

 

39
bw

   

     
 

‘ mmemUEI) 15
- . C

121

FIG. B, Scatterdiagram Showing Size and Similarity of Ideal and
Actual Q-Scores Obtained for Picture No. 28.

 

 

 

 

 

 

Ideal g~++++~~+~~~ - _..- -..--+.+++++++-.-..- +7.... -. , - -- ---.. .__.- “14”“. ___,- 4..
Self 12 i a
11“ Y
3 ' '
t -
10 1 LE! : C1
3* -._ 1
9 . 131} I ’ 0
7+ I
8 3 a 0 15 @O 331
1 . M .
. l
7 E - {(91 © :- 193
+ ‘30
6:1..-~——--~4 -—- .. — .,-_©...______.../(;, 0..
1 181 Q”
5?
g @x‘ Q
t
45 @ 3 1:3
t
3! (r = -. 18)
T .
2 ' . 1'91 1
1 1 5
1 - ;
+ 1 ‘
O 1 g
I .'
4 4 +------+~-—--++—~~~~~+~—»'~+~- +—- ----+~~—«—--+~ + 4.. +4 '
O 1 2 3 4 5 6 7 8 9 10 1 1 12

Actual Self

 

Picture Description: Parents at breakfast table spelling "LETS GO”
with ALPHA-BITS to their son who plays with a new toy train
instead of hurrying to school with the books his mother holds
for him. (Same as top picture in Factor 3, FIG. IV).

 

 

Symbols for Subjects: Men Women
Young 9 o
lVIiddle-Aged i L91 '6

01 (1 f '5};

I?)

 

  

       

u—‘-- ‘.- ‘-

 

‘
\p

 

 

-4 &-~8.1.

   

.—‘

 

a“. '-

 

ICHIGRN STQTE UNIV. LIBRQRIES

II III “Hill” I 111111 1| lflflll

3129310280192