THESIS FOR THE DEGREE orjpnn. f i {'1
MICHIGAN STATE UNIVERSE! - 7 V _ f i

 

LIFT? "‘ I

THENfi

MAJ , [e
UlllVLIJH)’

 

This is to certify that the

thesis entitled

THE EFFECTS OF
OBSERVAB] F. AIHfl—ENCE RESPONSE
ON ATTITUDE CHANGE AND SOURCE CREDIBILITY ‘

presented by

Carroll G. Hylton

has been accepted towards fulfillment
of the requirements for

Ph. D. degree in Department of Speech

[XX/£142; 1M 19' GEM/14:14

Major professor
Date November 22, 1968

0-169

 

ABSTRACT

THE EFFECTS OF
OBSERVABLE AUDIENCE RESPONSE
ON ATTITUDE CHANGE AND SOURCE CREDIBILITY

by Carroll G. Hylton

This study sought to identify, manipulate, and
measure the effects of a new variable in the one—t0-
many communication event. It was postulated that the
perception by audience members of the feedback of
others could influence audience attitude.

Feedback has usually been Viewed as a receiver—
source 100p or interaction. A process View of communi—
cation, when applied to the one-to-many communication
event, would call for a consideration of intra-audience
interaction. Such a consideration resulted in the
conceptulization of a receiver-receiver 100p that was

designated observable audience response. The study

 

presented a heuristic model demonstrating how observable
audience response could be incorporated into the one-to-
many communication event. The literature on feedback,
research in other areas such as social facilitation,
conformity, and imitation, from which support for the
concept can be extrapolated were discussed.
Considerations of the concept and the literature

resulted in the formation of a set of experimental

hypotheses and the design for their test. It was
hypothesized that observable audience response would
change attitude toward a message topic and toward a
speaker. Three treatments were employed. Two consisted
of conditions in which coached confederates supplied
either positive or negative feedback during the
presentation of a speech advocating federal control

of education. The third treatment was one in which no
effort was made to manipulate audience response. A
control group received pre and post-tests of attitude
toward message. Ss attitudes toward both message and
speaker were measured.

Data were analyzed by means of analysis of variance
and t-tests for both within—group and between-group
differences. Although some individual hypotheses were
not supported at the .05 criterion set, the overall
results supplied evidence supporting the major predictions
of the study.

In addition to attitude measures bearing on the
a priori hypotheses of the study, two secondary analyses
were performed, supporting the assumptions that (l) the
control for the experiment was adequate, and (2) the
independent variable was operating in the experiment.

The study was discussed in terms of the implications
of its design, the major findings, and the possible
future research growing out of both the secondary and

major analyses. The failure to support some of the

hypotheses was interpreted as (1) an indication, in
some instances, that control was adequate, and (2) that
positive feedback may be a stronger intra-audience
stimulus than negative feedback.

The study was concluded by discussing the reserva-
tions that should be observed in generalizing from the
results; and consideration of the implications of
observable audience response for the communication

process.

 

THE EFFECTS OF
OBSERVABLE AUDIENCE RESPONSE

ON ATTITUDE CHANGE AND SOURCE CREDIBILITY

by

Carroll G. Hylton

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Speech

1968

lt‘

.1.‘3'.U'1:‘-.'§‘1 .'l."‘:'..'3.1I1\1.-l

\"l'l.l 1 :lltl.l‘l" .1"“".H‘:T

unl|\'ll

5:!

tlll.’\

HHH’I‘

.‘1.

31.151 :l\' 7.1

VHAH"

lint In‘l

F.

r) i Elf; 1 J i HJ'lIJ71

”(jia’mvirm on. M

ajnomouiupoi 0d} In

”Lulu

{I')HIl I i 1 lil?
’zlll 'IC?

,l'z r)’}-Ili').’l

I-rfhfqr , ,»
I. A. I 4

~. .3 ,.,_"
. u 4 I .

115:1: ‘

.'11111‘l'1'l"l'.'l HM

'Ill’l
I J; I I I 11'; If I

ACKNOWLEDGEMENTS

Many people have contributed to the completion
of this study. My sincere thanks go to Dr. William
B. Lashbrook and Dr. James C. McCroskey, the co—direc-
tors of this thesis. Their advise and encouragement
have gone far above and beyond the call of duty.

My gratitude goes, as well, to Dr. Kenneth G. Hance,
Dr. Gordon Thomas, and Dr. Hal Walsh who first contributed
to my education in the classroom and later provided
valuable help as members of my guidance committee.

I am indebted to colleagues and students at San
Jose State College for generous contributions of time
and effort at various stages in the preparation of this
paper. Thanks are due to the staff of the Department
of Speech—Communication for their assistance in securing
experimental subjects. My appreciation goes to
colleagues in other departments, especially Bob Clarke
and Joan Bailey; and to the many students who have
helped, especially Linda Squires, Jennifer Benedict,
Enid Layes, and Ken Yules.

To my good friend and colleague, Ted Benedict, I
owe much more than I can say. His efforts on my behalf
have made a major contribution to this study.

To Ruby, Mike, and Laura I am grateful for their

faith and patience.

ii

TABLE OF CONTENTS

ACKNOWLEDGEMENTS . . . . . . . . . .
LIST OF TABLES . . . . . . . . . . .
LIST OF APPENDICES . . . . . . . . . .

I. ‘INTRODUCTION

General Statement of the Problem . .
The Feedback Concept . . . . . .

The Feedback Literature . . .
Observable Audience Response Model .
OAR and Other Variables . . . . .

Hypotheses Tested . . . . . . .

II. METHOD

III. RESU

IV. DISC

REFERENCES

Procedure . . . . . . . . .
Experimental Design . . . . .

LTS
Elimination of Ss and Data . .
Secondary Analysis: The Independent

Variable . . . . . . .
Secondary Analysis: Control . . .
Major Analysis . . . . . . . .
Summary of Results . . . . . .

USSION AND CONCLUSIONS

Design Implications . . . . . .
The Hypotheses . . .
Future Research on the Effects of OAR
Some Necessary Reservations . . .
Significance of the Research . . .

iii

Page

ii

iv

44

45
54
56
65

68
70
81
88
92

95

LIST OF TABLES

Table

Page

1 Three Treatment Experimental Design . . . . 42
2 Factor Scores for Sp and ARs Rating the

Concept "This Audience" for 3 Experi-

mental Conditions . . . . . . . . . 48
3 Analysis of Variance Summary of Pre-test

Scores for Experimental and Control 88 . . 55
4 Analysis of Variance Summary of Post—test

of Attitude Toward Message for PR, FR,

NR, and CG . . . . . . . . . . . 56
5 Pre and Post-test Mean Scores for Attitude

Toward Message Topic and Speaker for

3 Experimental Groups and Control Group. . 64

iv

LI ST OF APPENDICES

Appendix Page
A Instructions to Ss Preceding Pre-test . . . 100
B Pre-test Questionnaire . . . . . . . . 101
C Speaker's Introduction . . . . . . . . 104
D Post-test Questionnaire . . . . . . . 105
E Experimental Speech . . . . . . . . . 112
F Comparison of Speaker's and ARs' Audience
Ratings . . . . . . . . . . . 127

CHAPTER I

INTRODUCTION

General Statement of the Problem

During the past two decades an increasing number of
communication scholars have come to ad0pt the "process"
view of communication. This View holds that all elements
of the phenomenon are interdependent and interacting;
that to View various elements in isolation is to falsify
the event.

Among others, Berlo (1960) argues for the necessity
and desirability of taking this process view. One of
the crucial elements in the ad0ption of this position
is the recognition that communication is not a one-way
process; i.e., messages do not move just from sources
through channels to receivers in a one-way flow. Messages
move from receivers to sources as well, and this two-way
process goes on simultaneously in the face-to-face
communication event. Such a receiver-to—source flow
has been labeled "feedback."

Communication scholars have recognized the usefulness
of the concept of feedback in explaining and predicting
communication behavior. In the past few years an
increasing number of "feedback studies" have been
conducted. Investigators, for the most part, have been

interested in observing the effects on sources of varied

l

 

2

receiver or audience response. Such observable response,
or feedback, also needs to be examined with a view toward
determining its effect on members of the audience providing
the feedback.

We need to designate this phenomonon by a name other
than "feedback" because we are no longer talking about
a receiver-source 100p, but rather a receiver—receiver

loop. Let us call this receiver—receiver loop observable

 

audience response. We will discuss it in detail later

 

in this chapter when we present a model for one-to-many
communication.

The purpose of this study is to test hypotheses
which predict the effects of observable audience response
on attitude change and source credibility.

The following sections of this chapter will (1)
examine in more detail the feedback concept, (2) discuss
some of the pertinent research in the area, (3) present
a model illustrating the interaction of observable
audience response with other elements in the communication
process, (4) consider various classes of variables with
which observable audience response might interact, and (5)

present the hypotheses tested in this study.

The Feedback Concept

There has been considerable speculation, examination
and experimentation dealing with the notion of feedback.

Some writers credit Wiener (1948), others Ashby (1952),

 

3

with the development of the concept in terms of human
communication. Certainly either or both of them were
familiar with the work of such electrical engineers
as Nyquist (1924), who began to use the concept in
theoretical formulations involving mechanical communication.
Cherry (1957) writes:
. . . the first mechanical treatment of the
stabilization of a dynamic system by feeding
information signals back from the output or
"receiver" end to the input or "transmitter"
end was made by H. S. Black in a study of
electrical feedback amplifiers in 1934, and
later deve10ped largely by the efforts of
Nyquist and of Bode, into an exact mathe-
matical method and system of design. (p. 56—57)
Johnson and Klare (1962) use the term "feedback
analogy" in describing the uses to which the term has been
put and some of the "contradictions" between disciplines
and writers in their different references to feedback.
For those interested in interpersonal communication the
concept is indeed used analogically. Engineers become
disturbed when they hear some sociologists, social—
psychologists, et a1 talking about feedback and reinforce-
ment synonymously, and using the terms positive and
negative in ways quite different from the engineers' usual
use.
Therefore, we will talk about feedback as a receiver-
source 100p which is operating simultaneously with the
"normal" source-receiver message. In a face-to-face

communication event, the source presents a message to a

receiver (or receivers), the receivers may react to that

 

4
message both overtly and covertly, i.e., in addition
to non—observable responses they may smile, frown, nod,
laugh, groan, sit quietly, stare intently — at the
source, each other, or out the window - etc. Such
responses, observed by the source, are feedback. When
a source responds to these receiver behaviors, we have
an interaction that affects the message.

Behavioral scientists (like Berlo, Il960I) talk
about positive and negative feedback in a way that is
synonymous with reward and punishment. We will use this
latter meaning. In a face-to—face communication event,
when receivers engage in behaviors which the source
interprets as punishing (e.g., frowns, head shaking, lack
of attention, etc.) we have negative feedback. When
receivers engage in behaviors which the source finds
rewarding (e.g., smiles, head nodding, enthusiastic

applause, etc.) we have positive feedback.
The Feedback Literature

Earlier we talked about feedback as a receiver-
source 100p. We also pointed out that it is sometimes
talked about as reinforcement. The behavioral science
literature is abundant with studies which examine the
effects of manipulating various kinds of feedback
and/or reinforcement, under varying conditions, as

well as various characteristics of those generating such

feedback.l

Before examining some of these studies we should
understand that feedback has been treated almost exclu—
sively as an independent variable. Investigators have
manipulated feedback and measured changes in communicators.
Such changes are then attributed to the type of feedback,
the amount of feedback, or the source of the feedback.
Feedback type might be positive or negative; amount could
be varied from no feedback to much feedback. The source
of feedback could vary in terms of prestige, i.e., it
might come from a liked person or a disliked person, etc.

What are the possible effects of feedback in terms
of these three categories (type, amount, and source)?

We might, for instance, measure (take as our dependent
variable) oral delivery. As we manipulate type, amount,
or source of feedback we are interested in such things
as fluency. Miller, et_al (1961), manipulated type of
feedback, specifically reward and non-reward, during

the presentation of a speech. They measured fluency and
speaking rate. Their manipulation of feedback resulted

in significant differences in speakers' fluency but not

 

1See for example such studies as: Leavitt and
Mueller (1951) who varied amount of feedback and measured
such dependent variables as accuracy and time; Miller,
‘et al (1961) who were interested in varying feedback and
measuring the effects of such variance onspeech patterns;
Stoltz and Tannenbaum (1963) who varied feedback in order
to measure effects on oral encoding; Amato and Ostermier
(1967) who varied feedback in order to note effect on eye
contact, nervousness, bodily movement and fluency.

 

 

6
in speaking rate. Later, Miller (1964) manipulated
type of feedback using three types: approving, disapprov—
ing, and neutral. Once again, fluency and rate were
the dependent variables. There were nine treatment
conditions, and the data indicated that speakers' fluency
and rate were not significantly influenced by the responses
to his speech.

Vlandis (1964) conducted a similar experiment.
Whereas Miller, et_al (1961) and Miller (1964) had
utilized confederates in order to measure the effects
on their 85 of feedback which differed or agreed with
that which the 83 received, Vlandis, on the other hand,
simply varied type of feedback between 88 during their
presentation of speeches. The two conditions (reward
and punishment) consisted of 33 receiving the word "yes"
or the word "no" throughout the presentation of the speech.
Again, the dependent variables were fluency and rate.

The findings indicated significant differences for both
variables between the two treatments.

Amato and Ostermier (1967) provided 83 with two
types of feedback, neutral and unfavorable, in order to
analyze effects on delivery of both oral and visual
kinds. Specifically, the dependent variables were: eye
contact, nervousness, bodily movement, and fluency. Their
data supported the hypothesis that unfavorable feedback
prompts a deterioration in speaker delivery. Their study

differed from the previously mentioned ones in several

 

 

7
ways; however, for our purposes, an important difference
is that the speakers received feedback from more than
one person, i.e., 83 presented speeches to an audience.
In addition, 85 were receiving what was probably perceived
as "peer feedback" as opposed to "authority-figure feed-
back." The latter label is used to describe the Miller
§E_El1 Miller, and Vlandis studies.

Numerous studies have been conducted in "verbal
conditioning." In the majority of them reinforced behavior
has been changed significantly in the hypothesized direction.
Experimenters have used verbal reinforcers such as "mmm
hmmm," "good," "right," "fine;" non-verbal reinforcers
such as lights, buzzers, head nods, smiles. Ss have been
conditioned to increase (or decrease) the use of selected
terms, engage in selected movement responses, and vary
rate.2

The studies we have considered to this point have
varied type of feedback. We have suggested that amount of
feedback could be manipulated as well. Considerable work
has been done in small group research, specifically in
ascertaining the effects of various communication networks.
In these studies restricted networks have been created
and such dependent variablesas accuracy, hostility-
friendliness, time, etc. have been affected by the with—

holding or limiting of feedback. Early work in restricted

 

2For a summary of results of "verbal conditioning"
studies see Krasner (1958); Berelson and Steiner, (1964).

 

networks was done by Leavitt and Meuller (1951) and
Thibaut and Coules (1952).

The familiar circle, chain, Y, and wheel communication
networks have been explored by many, perhaps the best
known work being that of Bavelas (1948) (1950) (1952).
Herefeedback has been restricted or directly manipulated
in order to determine effects on task speed, role
perception, etc.

Before many questions concerned with amount of feed—
back can be answered we will have to quantify feedback in
some more rigorous manner. Information theory may suggest
some possibilities, as may extrapolations from Asch's
work in conformity (especially those studies in which
number of confederates was varied in attempts to arrive
at an optimum number which would ensure maximum conformity).3

Source of feedback, as a variable, has been investi—
gated in various communication situations. The majority
of studies which bear on this manipulation have likewise
been conducted in small—group research.

Collins and Guetzkow (1964), summarizing the results
of studies examining communication content, where status

has been manipulated, conclude:

 

3These studies followed the more well—known ones
popularizing conformity. See S. E. Asch, "Effects of Group
Pressure Upon the Modification and Distortion of Judgments,"
in E. Maccoby, T. Newcomb and E. Hartley, Readings in Social

 

Psychology 3rd ed. (New York: Holt, 1958), 181.

 

 

 

The content of communication from low to

high power-status persons will depend on

what the low status person has learned is

more likely to obtain reinforcement. (p. 176)
Such investigators as Kelley (1951), Campbell (1958), and
Cohen (1958) lend support to the above proposition.

Returning to experiments in verbal conditioning, we
find additional data that have implications for source
variation in feedback. Sopolsky (1960) manipulated
experimenter variables such as age, sex, social status,
and prestige in order to observe their effect on Ss'
receptivity to conditioning and found significant
differences between 85 who perceived their experimenter
as compatible and those who did not. Maccoby, et_al
(1961), presented 85 with information regarding child
training and then measured 85 preferences for interaction.
85 chose to interact with those who reinforced their
(Ss') Opinions.

Recalling our original question: "What are the
possible effects of feedback in terms of type, amount,
and source?" we have seen that these effects have been
examined as they change a 8'5 delivery, his speed in
completing a given task, his perception of role, his
message content, his attitude toward another, and his
desire to communicate. The investigations noted, as
well as the areas exemplified by them, have (1) been

concerned with feedback as cause and a variety of other

phenomena as effects, and (2) rest on the single-source

 

 

lO

single-receiver model.

SOURCE —— Message ——->- RECEIVER

\ Feedback/

Now let us consider the public speaking situation.
Here we have a "one-to-many" communication event. One
source presents a message to many receivers. The Amato
and Ostermier study, mentioned previously, consisted of
this kind of communication situation, which could be

pictorialized as follows:'

 

 

‘?,.w*”"—-'feedbaCk RECEIVERl

SOURCE Message >: RECEIVERZ
III“““‘--feedback RECEIVER3
RECEIVERn

Their concern was the manipulation of feedback (neutral
and unfavorable) in order to observe its effect on

source's eye contact, body movement, etc.

Observable Audience Response Model

Let us now consider the effects of feedback not on
the source but rather on the receivers only. The question
is this: Are receivers as they observe the responses of
other receivers affected by them? We can now construct

a model which attempts to pictorialize such a concern:

 

11

A MODEL ILLUSTRATING OBSERVABLE AUDIENCE RESPONSE
IN THE ONE-TO-MANY COMMUNICATION SITUATION

4

The Source presents his Receivers with at least
three kinds of stimuli (l) Verbal (his message content),
(2) Physical ("the speaker they see") and (3) Vocal ("The

speaker they hear").
but other Receivers with similar stimuli.

Source,

Each Receiver presents not only the

Each

Receiver reacts to not only the stimuli presented by the
Source, but stimuli presented by other Receivers as they

react to the Source.

The broken arrows running from

"Other Receivers" to "Receiver" represent these stimuli
(observable audience response) acting on the Receiver
along with the stimuli presented by the Source.

 

 

 

 

 

 

 

 

SOURCE

 

 

VERBAL \ ‘TERBAL
STIMULI STIMULI
Message Content (Little or
none detect-
able by S)
PHYSICAL PHYSICAL
STIMULI STIMULI
Gesture, Facial exp.,
Movement, ‘IBody Set,

Facial exp.,
etc.

Body Mvmt.,
lApplause,
Boos, hisses,
Laughter,
<12-

 

 

VOCAL STIMULI
Rate,
Loudness,
Inflection,

 

52- }

 

 

VOCAL STIMULI
(Relegated to
degree & kind
of above)

 

 

\

 

 

"-

OTHER
RECEIVERS
Reacting
Verbally,
Physically,
Vocally

 

 

 

\

\
4

 

*— RECEIVER

 

 

%
/

 

 

OTHER
RECEIVERS
Reacting
Verbally,
Physically,
Vocally ‘

 

 

 

 

4Elements of this model (verbal, physical and vocal
stimuli) are found in Miller's (1966) model for nonverbal

communication.

providing the basis for his model.

Miller credits Randall P. Harrison for

 

 

 

12
Here we have a model which includes not only a feedback
loop running from receiver(s) to source but an intra-
receiver 100p which takes into account the perception
of the responses of the receiver by others who are
likewise receivers. For example, if we are listening
to a third person's message and if I am attending to
your reaction to that message (i.e., I am perceiving,
in addition to the third person's message, your responses
to that message), then we have the phenomenon suggested

in the model. We will call this phenomenon observable

 

audience response (OAR). Such OAR can take various

 

forms: it can be positive (approving), negative
(disapproving), or mixed.

We have already seen that both positive.and negative
responses act as reinforcers on the source. But what are
the effects of OAR on members of an audience? Are their
attitudes affected? Does their behavior change as the
response they perceive varies? Could members of an
audience with neutral attitudes toward source and/or
message be persuaded to become partisans after having
perceived other audience members reacting favorably to
source and message? None of the feedback studies to
date has treated these questions nor their various
implications.

What we have just been discussing is the effect of
a group on a member(s). Although we have no current

or past work in the literature which examine such effects

 

13
in conjunction with feedback, there are numerous studies

in the general area of social influence.
OAR and Other Variables

Festinger (1955), writing more than ten years ago,
pointed out that social influence might exist as a
"wild variable" in some of the studies then being
conducted. He was concerned with the "lag" between
immediate and delayed testing, speculating that
experimental subjects might be receiving additional
treatment as they interacted in the time period between
tests. Cohen (1964) has suggested that this lack of
control might account for the so-called "sleeper effect."
This concept evolved out of the work done at Yale
investigating persistence of retention and acceptance
of Opinion change. If experimental subjects engage in
the interaction mentioned by Festinger then it might
result in (l) a decrease in retention of factual material
coupled with an increase in opinion change and/or (2)
a disappearance of source credibility effects on message
(both findings being instances of "sleeper effects").
Allport (1924), in his pioneer studies in intra-
group effects, noted that 85 performing various tasks
in the presence of others were seemingly stimulated by
such a condition. He called this phenomenon "social
facilitation," and defined it as the enhancement of a

response by contributing social stimuli, such as the

  

 

l4

sight and sound of other people engaged in the same
activity.

Callaghan (1940), writing in the Quarterly Journal

 

of Speech, presented the essence of Allport's investi-

 

gations and suggested some implications for persuasion.
Arguing that students of Speech need to recognize social
facilitation, Callaghan made a variety of extrapolations
from Allport's work to that of other investigators on
audience effects. It would seem that such a beginning
would have produced considerable speculation, if not
research, in order to integrate the notion into Speech
principles; however, the literature reveals little
subsequent consideration of this concept by writers in
Speech.

Oliver (1950), discussing individual-audience
relationships, mentions social facilitation as "the
tendency for the group to increase its like-mindedness."
Minnick (1957), describing crowd behavior, cites
Allport's work and offers social facilitation as an
explanation of individual responses which are more
vigorous than would be apparent in "non-audience"
contexts. Eisenson, et_al (1963), considering inter-
action within the audience, mentions "social facilitation,’
along with "polarization" and "circular response," as
three significant concepts developed by social
psychologists.

The above writers, while drawing attention to the

phenomenon of social facilitation, seem to take its

 

 

15

existence in the public speaking context forgranted,
and are content to make this extrapolation from the
non—speech work of Allport.

Most modern investigators have come to refer to
the original Allport concept as "social influence"
and have been primarily concerned with three major
constructs: conformity, identification, and imitation.
Let us now examine each of these to see if they offer
some insight into the possible effects of OAR.

The studies of Sherif (1935) and Asch (1956),
demonstrating conformity, established in a variety of
contexts, with numerous replications, that individuals
can be pressured into altering original judgments
when faced with a group judgment that has been manipulated
by the experimenter.

Reitan and Shaw (1964) point out that the many
studies of conformity behavior have shown conformity
to be influenced by at least four classes of variables:
(a) the personality characteristics of the individual,

(b) the kind of stimuli evoking the response which
reflects the conformity behavior, (c) situational
factors, and (d) intra-group relations.

Studies of the effects of OAR could take into account
the personality characteristics of the persons involved.
Houston and Mednick (19) have demonstrated that high
creative personalities have a strong need for associative

novelty. What would be the effect on such persons

 

 

“war-“ w.— -1 , we

16
(compared with low creative personalities) of a
communication situation in which the OAR they perceived
was positive? Type of OAR perceived, as well as
various personality characteristics, could be manipulated
in order to examine such combinations in the communication
context.

Consideration of the kind of stimuli evoking
conformity behavior is perhaps the most limited of the
‘classes of variables out of which we might generate
studies_of the effects of OAR. We have previously
talked about the various types of feedback which have
been used as reinforcement (head nodding, lights,
buzzers, etc.). In most public speaking situations the
range of behaviors which might be viewed as OAR is
relatively narrow. Audiences are "allowed" to indicate
approval with more than hand-clapping and disapproval
with more than booing; nevertheless, most of us expect
little variance from a narrow range of norms. Within
that range we might explore the differences between
auditory approval/disapproval, visual approval/disapproval,
and mixed forms. Or we might vary degree of approval/dis-
approval, and here we might find some "boomerang effects."
Thus, the kind_of audience response perceived could be
examined as to its effects.

Situational factors provide a rich area in which
OAR can be investigated. Is OAR more effective in the

large lecture hall or in the crowd on the street corner?

 

 

—.-.m fi-_fi.

17

Is OAR more effective when directed to the after

dinner speaker than to the professor in the departmental
seminar room? Needless to say, explorations of OAR

in varying situations (as well as all other classes

of variables) requires rigorous control in order to
assure that differences can be attributed to the
variable being manipulated and not to some other
variable.

Intra—group relations is perhaps the most exciting
area for studies of OAR. Common sense and research
literature tell us that individuals who perceive the
approval/disapproval of their reference group are
affected by it. Numerous studies could be devised in
which group affiliation and observable response were
manipulated in order to examine their interaction.

Most individuals "belong" to various groups with varying
degrees of affiliation. What would be the effect on

the individual who perceived positive responses from
members of one such group and negative responses from
members of another? We might think it would depend

on the message and/or the speaker, or on the strength

of the individual's affiliation. It might; but we need
to design experiments in order to find out.

The preceding discussion of just four classes of
variables which influence conformity and the possibilities
they hold for an examination of OAR is indicative of

the many areas into which investigation might lead.

 

~1-‘gn‘

 

l8

Recalling that we suggested three major constructs
(conformity, identification, and imitation), let us now
turn to identification and imitation to see whether
these two constructs might provide some insight into
the effects of OAR.

The notion of "identification" is one which has
received extensive treatment by psychologists and
social psychologists in their examination of learning
and socialization.5 Many, if not most of these
investigators, have been concerned with reinforcement,
both secondary and vicarious. The notion of identification
as it applies to OAR is most closely linked to the view
taken by Stotland (1961).

Stotland's investigations have led him to view
identification as a perceptual-cognitive process in
which an individual observing a trait in another then
goes on to identify with that individual in terms of
another trait or traits. If Stotland's similarity
theory of identification is a tenable one, then members
of an audience who perceive other members as similar to
themselves should, when presented with the additional

perception of approval/disapproval for a given message,

 

5See such works as: O. H. Mowrer, Learning Theory
and Personality Dynamics (New York: Ronald Press, 1950),
531-561; A. Bandura, “Social Learning Through Imitation,"
Nebraska Symposium on Motivation (Lincoln: University
of Nebraska Press, 1962), 211—269; R. R. Sears, "Identi—
fication as a Form of Behavioral Development," in D. B.
Harris (ed.) The Concept of DevelOpment (Minneapolis:
University of Minnesota Press, 1957), 147-161.

 

 

 

 

 

 

19

be impelled to follow suit.

Some writers use the term "identification" as
a synonym for the term "imitation," while others
tend to separate the two, reserving identification
for the process by which an individual assumes the
meaning of another.6 Lambert and Lambert (1964)
suggest a third category, viz. "vicarious socialization."
They cite the work of Berger (1961) and exemplify the
notion by pointing to the child who, while watching
another being taught, learns along with him. They
hypothesize that the imitative habits of the watcher
may be covert and that later, when the model is no
longer present, the watcher then acts overtly in line
with these covertly learned responses. Lambert and
Lambert go on to assert: "When such possibilities
(imitation and vicarious learning) have been studied
we shall have forced into the Open two important
psychological processes . . ." (p. 20). The examination
of the effects of OAR may well contribute to the under—
standing of processes such as these.

To this point, we have been concerned with an
exploration of OAR, we have suggested an OAR effect —
not on the source of a message — but rather on receivers
who perceive the responses of others in the one-to-many

communication situation.

 

6For example, see Lazowick (1955).

 

20

It remains to explore the various ramifications
of cognitive and behavioral effects of OAR in the
communication process. The purpose of the following

hypotheses is to explore one such possibility.

Hypotheses Tested

If OAR is a significant variable in those
communication events where it can operate, then a
study is called for which would "establish" it as an
affector and examine some of its effects.

To explore the problem discussed above, an experiment
consisting of three treatments was conducted. All
treatments were concerned with testing the hypotheses
presented below. Treatment I consisted of a positive
response condition (PR). Treatment II was a "free"
response condition (FR). Treatment III was a negative
response condition (NR).

The questions raised by the line of thinking just
discussed, and by a review of the literature, led to the
following hypotheses:

I. Ss in Treatments I and II will shift attitude

in the direction advocated by the message.

1. Ss stimulated by positive OAR will express
a more favorable postcommunication than
precommunication attitude.

2. 88 where the OAR stimulus is "free" will

express a more favorable postcommunication

 

II.

III.

21

than precommunication attitude.

Ss stimulated by positive OAR will express

more favorable attitudes than 58 where the

OAR stimulus is "free."

1.

Ss stimulated by positive OAR will express

more favorable attitudes toward message

topic than 85 where the OAR stimulus is

"free."

Ss stimulated by positive OAR will express

more favorable attitudes toward a speaker

than 85 where the OAR stimulus is "free."

a.

Ss stimulated by positive OAR will
express more favorable attitudes toward
speaker's authoritativeness than 85
where the OAR stimulus is "free."

85 stimulated by positive OAR will
express more favorable attitudes toward
speaker's character than Ss where the
OAR stimulus is "free."

85 stimulated by positive OAR will
express more favorable attitudes toward
speaker's dynamism than 85 where the

OAR stimulus iS "free."

55 stimulated by positive OAR will express more

favorable attitudes than 85 stimulated by

negative OAR.

1.

Ss stimulated by positive OAR will express

 

22

more favorable attitudes toward message

than 83 stimulated by negative OAR.

2. Ss stimulated by positive OAR will express

more favorable attitudes toward speaker than

85 stimulated by OAR feedback.

a .

Ss when

express

Ss stimulated by positive OAR will
express more favorable attitudes toward
speaker's authoritativeness than 85
stimulated by negative OAR.

Ss stimulated by positive OAR will
express more favorable attitudes toward
speaker's character than 85 stimulated
by negative OAR.

Ss stimulated by positive OAR will
express more favorable attitudes toward
speaker's dynamism than $5 stimulated
by negative OAR.

the OAR stimulus is "free" will

more favorable attitudes than Ss stimulated

by negative OAR.

l. 85 where the OAR stimulus is "free" will

express more favorable attitudes toward

message than 85 stimulated by negative OAR.

2. 85 where the OAR stimulus is "free"

23
will express more favorable attitudes
toward speaker's authoritativeness than
Ss stimulated by negative OAR.
85 where the OAR stimulus iS "free"
will express more fav0rable attitudes
toward speaker's character than Ss
stimulated by negative OAR.
55 where the OAR stimulus is "free" will
express more favorable attitudes toward
speaker's dynamism than Ss stimulated by

negative OAR.

 

CHAPTER II

METHOD

Procedure

The procedure consisted of five phases: (1)
selection of stimulus, (2) selection of Ss and
confederates (Cs), (3) pre-test, (4) experimental
treatments, (5) post-test.

Selection of Stimulus

 

Since the object of the study was to examine the
effects of OAR in the public speaking situation it was
necessary to obtain a speech that was capable of modifying
attitudes, and an instrument capable of measuring
attitude change. McCroskey (1966a), in his initial
research on ethos and evidence, developed sets of
experimental speeches as well as scales for the
(measurement of speech topics and ethos. One of the
experimental speeches has subsequently been used by
McCroskey and others (Frates, 1968) in studies
involving various experimental problems.

Selecting this speech as the message stimulus had
several advantages, (1) the t0pic was such that
experimental intent was not difficult to hide, (2) the
message content was "established" as appropriate for
college student 85, (3) it had successfully changed

opinion in previous experiments.

24

25

Although the ostensible stimulus in the treatments
was the presentation of a speech, the real stimulus was
the manipulation of type of feedback present. The
preparations for generating the positive and negative
responses will be discussed below under the section on
selection of Ss and Cs.

Methods of Measurement

 

The hypotheses tested in this study required the
measurement of Ss' attitudes regarding both message and
source. As mentioned above, previous research by
McCroskey included research on ethos and resulted in
the development of both semantic differential (SD) and
Likert type (LT) scales for the measurement of (1)
attitude regarding Federal Control of Education and
(2) ethos or source credibility. Since their
publication in 1966, McCroskey's scales have received
continuing validation in over a dozen experimental
studiesemploying them.

Because of their established validity, the SD
and LT scales for attitude toward message were employed.
However, additional SD scales, designed to measure
source credibility, also were used in this study. The
explanation for this decision follows.

Students of rhetoric have been long concerned
with the Aristotelian concept of ethos. One of the
early experimental studies in Speech was undertaken by

Haiman (1948), in which he systematically varied ethos

26

or source credibility in an attempt to measure its
effect on attitude change. In the past few years

such investigators as Walter (1948), Berlo and Lemert
(1961), and McCroskey (1966b) have designed instruments
to measure this variable.

The McCroskey Ethos Scales, deve10ped in a series
of experiments designed to test their validity, consist
of two dimensions (authoritativeness and character)
that can be measured using the bi-polar construct of the
SD, or LT scales. The scales were developed using
factor and item analysis techniques.

McCroskey's research began with a survey designed
to derive terms which would describe ethos or source
credibility. Having located thirty such terms "most
frequently used" he submitted them to factor analysis
in order to derive his scales.

McCroskey's factor analyses, unlike that of Berlo
and Lemert, did not produce a "dynamism" dimension. He
is quick to point out that he did not include scales
appr0priate for this dimension, and that ". . . this
should not be interpreted as an indication that it
does not exist." (McCroskey, 1966b, p. 66) He goes
on to raise an interesting question regarding the
dynamism factor. Pursuing research in the area of
congruity, Berlo and Gulley (1957) used SD scales from
the evaluative dimension as an attitude measure. Their

study should have been confounded had dynamism been a

27

factor accounting for significant amounts of variance.
Since this was not the case, is dynamism a significant
factor in persuasive communication?

The answer to this question has been partially
answered in subsequent research conducted by McCroskey
(1968) and by Berlo, Lemert and Mertz (1966). McCroskey's
investigations indicated that dynamism could not be
manipulated without affecting other dimensions at the
same time. Berlo, Lemert, and Mertz found that
dynamism was unstable and did not account for a large
pr0portion of variance in source credibility. McCroskey,
in his discussion of his ethos construct, concludes
that the research indicates that the dynamism of a
source is important but it is questionable whether it
is a dimension of source credibility "or merely one of
the many elements that affect the other dimensions."

(p. 61)

This study presented an opportunity, at very little
additional effort, to utilize both McCroskey's scales
and Berlo and Lemert's dynamism scales in this study.
(See Appendix D) Both scale sets were used with the
h0pe that analysis would provide additional evidence
as to the nature of the dimensions of source credibility.

In summary, methods of measurement employed were:
(1) LT and SD type scales measuring Ss attitudes toward
the message, administered as both pre-tests and post—

tests; (2) LT and SD type scales measuring Ss' attitudes

28

toward the source, administered as a post-test only.

Selection of Ss and Cs

 

Subjects. 85 for the three experimental conditions

 

and the control group were randomly selected from among
Speech 2A, Public Speaking, classes at San Jose State
College. Speech 2A is a general education requirement
at San Jose State College, which means that all students
must take the course sometime during their undergraduate
years at the institution. Surveys of the composition
of the course, undertaken by the Department of Speech-
Communication, indicate that the typical section is
fairly representative of the college p0pulation.
Although the course is a lower division one, freshmen,
sophomores, juniors, and seniors are approximately
equally distributed in any given section, as are the
various fields of study and proportion of men to women.
With the exception of early morning (7:30 A.M.) and
late afternoon and evening sections (3:30 P.M. to
8:15 P.M.), the age range of the typical section is
what would be expected by the classes represented
within each section.

In order to maximize homogeniety in Ss' age, and
to control for time of day, 85 were selected from among
classes offered between 8:30 A.M. and 11:20 A.M. These
are the most "popular" hours for the course, and
selecting from this time bloc carried the further advan-

tage of providing a sufficient number of students

29

per section.

During the Spring 1967 term, 49 sections of
Speech 2A were offered at San Jose State College.
Each section contained 20 to 25 students. Sixteen
sections were offered during the time bloc 8:30 to
11:20 and represented a S pool of approximately 400
students. In order to meet limitations of space,
and to create a public speaking situation where no
voice amplification would be necessary, audience size
was limited to approximately 100 for each treatment.

83 were assigned randomly, by class, to each of
the three experimental treatments and to the control
group. Number of $5 per section and treatment will
be discussed below.

Confederates. Cs were obtained from Speech 2A

 

sections and from other Speech—Communication classes
meeting at the same hours as were Ss' sections. Cs
made up approximately one half of the audience for
each of the three experimental treatments.

The decision to use a S to C ratio of one—to—one
was made after the following considerations. First,
a review of the literature indicated no studies in a
public speaking situation in which Cs (in any proportion)
were used in order to influence Ss. Asch's (1958) work
in conformity included varying the ratio of Cs to Ss
in order to see whether any ideal ratio existed for

conformity behavior. His findings indicated that a

30

three-to—one ratio (three Cs to one S) resulted in

maximum conformity behavior; the addition of more

Cs did not significantly increase conformity.
Although Asch's work might be taken as analogous,

the problem of C to S ratio in the public speaking

 

situation remains unknown. There is some tentative

 

evidence suggesting that in the public speaking
situation a one—to-one ratio is as effective as a
three-to—one ratio. Layes (1967) undertook a research
project which incorporated some of the design features
of this study. Using the same speech and measures

for attitudes toward message and source, Layes
manipulated responses (positive and "free" conditions)
as well as pr0portion of Cs to 58. Although her
results indicated no differences between a ratio of
one C to one 8 and a ratio of three Cs to one S, the
number of $5 per group was so small2 that the evidence
is tentative at best.

Even though the evidence from the study discussed
above suggests a weak basis for the ratio selected for
this study, it is the only study available and at least
did not deny the suspicion that the public speaking
situation differed enough from the situation in Asch's
experiments to call for a different ratio.

With the number of Cs determined, the next step was

to coach them in providing the experimental stimulus.

 

2Layes' four groups contained N's of 11, 12, 13 and 13.

31

One class meeting prior to that on which the experiment
would be conducted Cs were told by the experimenter (E)
that he would like their help in an experiment; that

they were to attend a speech and that they were to present
positive/negative feedback to the speaker. ("Cs,"
students who had received no pre-test, for the FR
treatment received no orientation whatsoever.) Cs

were then asked what they thought positive/negative
feedback consisted of. Cs responded by suggesting

various means of showing approval/disapproval, both

oral and visual. E did not offer specific directions

 

for feedback but rather encouraged Cs to verbalize their
own forms, to E and to each other. E then congratulated
the Cs on their creativity and pointed out that the
important thing was that the responses should appear to
be "natural" and that they knew, better than E, what
would seem "phoney" to their peers.

Cs were not told the nature of the experiment, who
the speaker was, nor any information regarding the
message. Spot checks of various Cs following the
experiments revealed that none questioned was aware that
measurements of audience members was the purpose of the
experiment. Most believed E was interested in effects
on the speaker, the remainder had no idea what the
purpose of the experiment was.

Pre-test

 

Three weeks prior to the experiments, during their

32

regular class periods, all Ss in experimental and control
groups completed attitude measures on the speech topic.
Ss were told that the Speech Activities class at San Jose
State College was interested in gathering Opinions on
various topics and asked to c00perate in this project.
(See Appendix A for complete instructions presented to 38
immediately preceding pre-test.)

As has been noted, the general education classes at
San Jose State College are fairly representative of the
general college pOpulation. Therefore, they are frequently
used by investigators sampling Opinion in order to make
inferences about the college population. It is not
unusual for students in Speech 2A (those making up the
S pool) to be questioned in class regarding a variety
of topics by a variety of individuals and organizations.

Normally, Speech 2A sections are told about the
Speech Activities Class, both by their instructor and by
a student representative of that class, sometime during
the first half of their term. Because the students
"know" that the class is real, and because they come to
expect various surveys and opinion measures, it was
assumed that they were not overly sensitized by the
pre-test measure.

Experimental Treatments

 

The three experimental treatments consisted of
presenting the same speech by the same speaker to three

groups of Ss paired with three groups of Cs. Time and

33

treatment conditions were determined by chance, with
one exception. 83 drawn for each treatment had to be
paired with Cs meeting at the same hour. It would
obviously not have been possible to have 83 drawn for
the positive condition at one hour and Cs providing
that condition at another. The first treatment (PR)
was the one in which positive response was generated;
the second (FR) consisted of the "free" response con-
dition, i.e., "Cs" received no instruction from E,

and no attempt was made to manipulate audience response;
the third treatment was the negative response condition
(NR) .

Treatment 1. Three days prior to the experiments,

 

E contacted a student in one of his classes and asked
him to present a speech before three live audiences on
the day scheduled for the experimental treatments.

The student selected was, in E's judgment, a
typical one. He was not a Speech-Communication major;
and although he had completed a course in basic public
speaking and was then enrolled in the Theory of Oral
Communication course, he was not in any sense a "trained
speaker."

McCroskey's (1966a) and McCroskey and Dunham's
(1966) research, as well as that of Holtzman (1966), has
raised the question that source credibility may have
been a masking factor in some of the studies in which

"weaker" variables have been manipulated with seemingly

34

no effect. Therefore, the speaker was selected with a
View toward holding ethos down to prevent it from acting
as a confounding factor in the experiment.

The speaker was given a copy of the Federal Control
of Education speech (See Appendix E) so that he might
familiarize himself with it prior to presentation. The
speaker was told only that his audiences were composed
of students.

At this point it might be well to consider the
rationale for deciding to sacrifice the obvious control
advantages of a taped message for the "looser" live
speaker stimulus. Most of the experimental studies in
Speech, attempting to measure various speaker and message
effects, have sought to control variations in delivery
by putting the message on tape, thus insuring that E
and C groups will have messages "identical" in all
aspects with the exception of the stimulus variable
(See such studies as Haiman (1948), Benedict (1958),
and Moretti (1965) for such use). Good methodology
dictates the control of all relevant variables in order
to justify the inference that any differences in the
dependent variable(s) can be attributed to the presence
of the independent variable(s). However, the gain in
specificity attained by rigid controls has its
compensatory loss in terms of the concept of process.

If we are interested in what happens in the public

speaking situation and must forever infer our statements

35

from studies using tape—recorded messages, perhaps the
price is too high. (How many audiences do we have
sitting in front of tape recorders as compared to those
sitting in front of live speakers?)

Hovland (1959) and Bormann (1965) are among those
who have seriously questioned the disparities between
the findings in the laboratory and field studies. Our
studies in Speech need to implement as much of the
rigor of the laboratory as is feasible while still
retaining the "real life" context of actual communication
situations. Therefore, this study utilized a live
speaker in an effort to bring both visual and auditory
stimuli together as they act and interact in affecting
an audience.

Since the message content was established prior to
treatment, and read from manuscript by the speaker,
adequate message control was assumed. The problem of
delivery control is both difficult and intriguing.

That the message content was held constant, and that the
speaker was the same, resulted in some delivery control.
However, the question, "was there delivery variation?"
must be answered "yes." This raises another question:
Might not differences observed in the dependent variables
be the result of delivery variance? That is, how do

we know that OAR, and not some delivery factor, caused
the response?

Common sense, intuition, observation by Speech

teachers, and considerable experimental evidence (much

36

of it cited in Chapter I of this study) support the
contention that speakers who receive reinforcement,
i.e., who perceive positive or negative feedback during
the presentation of a message, produce responses
significantly different from speakers who do not. One
extrapolation from this data might be the following:
The speaker, in the PR treatment (during which time

he received positive reinforcement) responded to this
reinforcement by increasing the vigor of his delivery,
by visually stimulating the audience, etc. or, while

he was in the NR treatment, he responded to the
negative feedback by increasing his rate (a let-me-out—
of—here syndrome). Undoubtedly some of this type of
behavior did take place. This introduces a different
kind of consideration, a more crucial consideration.

It is this: OAR may consist not just of those responses
perceived by some receivers as being made by other
receivers, but the perception of the source's responses
to them as well. In short, what may be necessary is
the perception of the entire feedback loop, taking

into consideration source responses to audience response.
If this is so, then a taped message would, in addition
to eliminating visual aspects of delivery, fail to
provide an accurate "environment" for OAR. Neither
visual nor oral responses of speakers to audience
response are possible when delivery is controlled by

using the taped message.

 

37

Finally, we can note that this study sought to
manipulate OAR while source responses were allowed to
operate "free;" the study was designed to allow source
responses to vary in order to simulate better the process
concept of communication.

The Ss and Cs were allowed to select their own
seats in the treatment room. No attempt was made to
alternate S, C, S, C. . . etc., either by seats or by
rows. Observation by E indicated 83 and Cs were
seemingly randomly distributed in the room. Thomas
and Ralph (1959) found no evidence that overall seating
arrangement, size of the audience, nor an individual's
position in a room, affect, to a significant degree,
response to persuasive appeals. Therefore, in order
to disguise experimental intent and to minimize
contaminating variables, no set groupings were used.

The speaker was introduced by name as a sophomore
associated with speech activities (See Appendix C for
speaker introduction). The title of the speech was
stated and E left the room. The only persons present
during treatment were the Speaker, 83, Cs, and two
additional students from one of E's advanced classes
who were asked to attend and rate the audience (hereafter
referred to as audience raters IARsI). Since the
experiment was one interested in determining the
effects of OAR from peers, no "authority figures,"

neither E nor any of the 85's professors, were present.

38

Immediately following the conclusion of the
speech the speaker left the room, E entered, distributed
post-test questionnaires, and collected them upon their
completion. Both the speaker and ARs filled out SD
scales rating the concept "This Audience." This
measure was taken in order to allow inferences regarding
the presence and perception of the independent variable
in the experimental conditions (See Appendix F for a
comparison of speaker's and ARs' ratings).

Treatment II

 

Approximately 15 minutes following the conclusion
of Treatment I, Treatment II was conducted. All
procedures were the same with the following exceptions:
(1) CS for this treatment had received no instructions.
Audience response was allowed to operate "free." (2)
The speaker had now presented the speech once before
a live audience, approximately one half of which had
been instructed to provide positive response. (3) ARs
were two different individuals drawn from the same class
as ARs for Treatment I. Again both speaker and ARs were
asked to respond to the scales rating "This Audience."

Treatment III

 

Approximately 15 minutes following the conclusion of
Treatment II, Treatment III was conducted. All procedures
and conditions were the same as those for Treatments I
and II with the following exceptions: (1) CS for this

treatment had been instructed to provide negative

39
response during the message presentation. (2) The
speaker had now presented the speech twice before live
audiences of the types described in Treatments I and
II. (3) ARs were different individuals drawn from
the same class as ARs used in Treatments I and II.
Once more speaker and ARs rated the audience at the
conclusion of the message presentation.

Post-test

 

Attitude. The same measures of attitude toward

 

topic (both LT and SD scales) as $5 received for the
pre-test, were administered as the post-test immediately
following each experimental treatment. The control
group received identical post-test measures, in their
classes, on the same day as experimental groups.

Using the same instrument for pre—test and post-
test can have a "sensitizing" effect on 85. This
presents a control problem in that we do not know
whether 85 are responding to the stimulus of the
independent variable or to an interaction between their
increased sensitivity and the experimental treatment.

A design feature which attempted to control for
any such sensitization was the inclusion of a control
group who received both the pre-test and post—test
"stimulus." If there is a significant difference between
experimental groups and the controlgroup and both have
been sensitized by the pre-test, then we should be able

to attribute such a difference to the independent

40

variable. Various investigators, such as Simpson
(1960), have conducted studies based on this line of
reasoning.

Further, as Kerlinger (1964) points out, "testing
is an accepted and normal part of most school and
college situations, and as such should have no great
sensitizing effect." (p. 311) In this sense, additional
control was utilized in giving the pre-test its "cover"
explanation. Finally, it should be noted, Ss completed
source credibility ratings prior to attitude measures
in their post—tests. The questionnaires for pre-test
and post-test "looked different" in that post—test
forms were bulkier and carried a cover sheet for
personal information not carried by the pre-tests.
These differences could have provided some "cover."

We are thus left with two alternatives: We can assume
that no significant sensitization took place because
of adequate control procedures; or, if sensitization
took place despite control procedures, it would be
detected in the analysis of control group data.

Source Credibility. Both LT and SD scales designed

 

to obtain measures of Ss' attitudes toward the speaker
were administered immediately following each experimental
treatment. These measures preceded the attitude

measures discussed above in the questionnaire form
making up the post-test. Each questionnaire contained

(1) a data sheet requesting personal information,

41

which 85 detached from the questionnaire and submitted
separately, in order to maintain subjective anonymity;

(2) LT scales for Authoritativeness; (3) LT scales for
Character; (4) SD scales for dimensions of Authoritative—
ness, Character, and Dynamism measuring the concept

"This Speaker;" (5) LT scales for the speech topic;

(6) SD scales designed to measure the concept "Federal
Control of Education." The questionnaire packet
completed by Ss consisted of the above measures in the
order indicated. (See Appendices B and D for pre-test

and post-test questionnaires.)
Experimental Design

In summary, three experimental treatments were
devised: (1) positive, (2) "free," and (3) negative.
The design included a control group who received neither
the experimental stimulus (OAR) nor the message
presented by the speaker.

Ss' and CG's attitudes toward topic were obtained
in a pre-test administered in classes on the same date.
Ss' attitudes toward topic and source were obtained
in a post—test immediately following experimental
treatments. CG's attitude toward topic was measured,
for the second time, on the same day as 83 received

their post-tests. The design is diagramed in Table l.

42

TABLE 1

Experimental Design

 

 

 

 

 

 

Pre- Post- Source Credibility
Attitude Attitude Ath. Ch. Dy.
Treatment I
(PR) A. E I L 0
Treatment II
(FR) B F J M P
Treatment III
(NR) C G K N Q
Control Group D H

 

References to Table 1 now enable us to relate the
hypotheses to the design for their test. Hypotheses
were presented at the conclusion of Chapter I which
predicted both between and within-group differences.
Before observing the various cell relationships
representative of the a priori hypotheses it should be
noted that two other cell relationships need to be
present in order to support an assumption of adequate
control: (1) There should have been no significant
differences in pre-test scores among the four groups,
(i.e., A = B = C = D) and (2) there should have been no

significant differences in pre—test and post—test

43
scores for the control group (i.e., D = H).
The within—group predictions take the following
expressions:
A < E; B < F.
The between-group predictions can be expressed
as follows:

E < F < G; I < J < K; L < M < N; O < P < Q.

CHAPTER III

RESULTS

Elimination of Subjects and Data

83 names were obtained on separate sheets for both
pre and post-test measures. 85 separated name sheets
and questionnaires and submitted them separately, having
been told that names were necessary in order to know
who had responded to tests and who had not. Questionnaires,
however, were coded to name sheets with "invisible ink"
(lemon juice) so that Ss' pre-tests and post-tests could
be matched for purposes of analysis.

(Although both LT and SD scales were employed in all
measures taken, initial inspection of the data indicated
that some $5 misunderstood directions for LT questions
and presented responses appropriate for SD scales to LT
forms. (LT measures of source credibility followed SD
scales for attitude in the questionnaire packet.) Add-
itional LT forms were omitted by some 85 who did, however,
respond to SD scales.

McCroskey's (1966) work with the scales employed
in this study indicated very high correlations between
the LT and SD scales for attitude toward topic (.897)
authoritativeness (.981), and character (.980) (p. 130).

On the basis of these findings, and to facilitate

scoring, analysis of data was limited to that provided

44

45

by SD scales for both attitude toward topic and
source credibility.

Before reporting the results of the various
analyses, it should be pointed out that of the 163 55
who received pre-tests and the 151 who received
post-tests, 129 usable questionnaires were obtained.
In order to facilitate analysis, 83 were randomly
eliminated until all cells contained equal Ns.
Analysis was performed on 120 83, each cell with an

N of 30.
Secondary Analysis: The Independent Variable

Before proceding farther, let us discuss the
analysis of the measures mentioned in Chapter II which
were obtained from the speaker (Sp) and the audience
raters (ARs) after each experimental treatment.

If the independent variable (OAR) was present
and Operating there should have been discernable dif-
ferences between those audiences where response was
being manipulated (Treatments I and III) and the
audience where it was not (Treatment II). Recalling
that Sp and AR measures consisted of SD scales by
means of which they responded to the concept "This
Audience," it might now be appropriate to discuss in
some detail the scales selected and the method employed
for their analysis.

The writer used scales designed for measuring

46
source credibility. This decision was made on an
intuitive and arbitrary basis. Twenty-five scales
were selected, representing five dimensions of source
credibility. McCroskey's scales (six for each
dimension) for authoritativeness and character made
up twelve of those used. The Berlo and Lemert scales
(consisting of four scales for each dimension) for
dynamism, competence, and trustworthiness madeup an
additional sixteen. Two extra scales (active-passive,
decisive—indecisive) were added to the dynamism
dimension. (Scales representing the five dimensions
total 26; however, since character and trustworthiness
share one scale set only 25 were used.)

The line of thinking that prompted the scale
selection ran something like this: An audience, half
of which has been instructed to provide positive
responses, will be perceived by its speaker and
"outside observers" as being more authoritative, more
dynamic, more competent, more trustworthy, and having
greater character than an audience where positive,
negative, and "no" response have equal chances of being
present. The same comparison can be made, at least
for authoritativeness and dynamism, between the "free"
and negative conditions. (A provocative question is:
Will the positive audience be rated as more authoritative,
more dynamic, than the negative one?)

These differences should show up if the independent

47

variable has created conditions that are discernibly
different and if the condition in which response was
not manipulated was comparatively "neutral."
Analysis

Osgood, Suci, and Tannenbaum (1957, p. 99-102)
point out that, in analyzing differences between the
meanings of the same concept for two different individuals
or two different groups, we are faced with a multivariate
problem and that the usual univariate tests of significance
(Eftests, for example) are not applicable. Further, that
in examining the differences between the meanings of
two concepts for the same individual we are comparing
only two points in the multidimensional semantic space.
They say (p. 101):

Although each of the two concepts (for

the same individual or for different
individuals) is associated with a series
of scores on k scales or kf factors,

these are not mere replicates and

cannot be treated as a sample over which
the usual univariate tests of significance
can be taken.

What the gentlemen seem to have done with these two
statements is to have closed the door on all of the
statistics available for examining differences between
scores. Neither multivariate nor univariate tests are
applicable to the data we are working with.

However, Osgood and his associates have found that

test-retest deviations by the individual subject judging

the same concept yield the following changes in factor

48
scores (A factor score is obtained by averaging over
the scales representing each factor.): "We find that
a change of more than 1.00 for the evaluative factor,
more than 1.50 for the potency factor, and more than
1.33 for the activity factor is significant at about
the 5 per cent level." (p. 139)
Table 2 displays factor scores for the experimental

treatments.

TABLE 2

Factor Scores for Sp and ARs
Rating the Concept "This Audience" For
3 Experimental Treatments

 

 

Dimension/ Treatment I Treatment II Treatment III

 

 

 

 

 

 

Factor (PR) (FR) (NR)
AR Spy AR Sp AR Sp
Authorita-
tiveness 5.83 5.66 3.83 4.83 4.17 5.00
Character 5.66 5.50 4.50 3.00 2.67 2.83
Dynamism 5.33 5.00 3.67 2.50 6.00 5.66
CompetenCe 5.75 5.00 4.00 5.00 4.50 5.25
Trustwor—
thiness 5.33 7.00 5.00 5.00 2.00 2.00

 

The authoritativeness dimension is made up of six
scales, all of which fall within the evaluative factor.

Character, represented by six scales, likewise falls

49
within the evaluative factor; so do competence and trust-
worthiness. However, these latter two dimensions are
represented by four scales and three scales respectively.
The dynamism dimension is a special case. As has been
pointed out earlier, research evidence raises some
serious questions as to the independence of this
dimension. McCroskey (1968) has equated it with the
activity factor. In examining the data represented by
the Sp's and ARs' audience ratings we will bear in
mind both McCroskey's research and that of Berlo and
Lemert and treat each dimension as independent, even
though four of the five are within the evaluative factor.
Further, we will take changes of 1.00 or greater in
any of the dimensions in the evaluative factor and changes
of 1.33 or greater in the dynamism dimension as
significant at about the five per cent level. Reference
to Table 2 reveals the following:

Authoritativeness. There was no significant

 

difference between Sp's and ARs' ratings of either the
PR or the NR audiences. A significant difference
exists between the Sp's and the ARs' ratings of the

FR audience. In other words, the speaker's perception
of the two audiences in which the independent variable
was being manipulated coincided with that of the
observers, but their perceptions differed when
evaluating the audience where no manipulation was

attempted.

50

Turning now to any differences among Sp's ratings
of the three audiences, we find that there were no
significant differences. However, ARs rated the PR
audience significantly above both the FR and the
NR audiences (their lowest rating being reserved for
the FR audience, as was that of the speaker).

Character. There was no significant difference

 

between Sp's and ARs' ratings for any of the three
audiences. ARs and Sp were closest in their ratings
for the two audiences where the independent variable
was being manipulated (PR and NR audiences) and
farthest apart where it was not (FR audience).

Examining Sp's factor scores between the audiences
we find a significant difference between the PR and the
FR audiences, a significant difference between the PR
and the NR audiences, and no significant difference
between the FR and the NR audiences. In other words,
the speaker rated the character of the positive audience
significantly higher than either the "free" audience or
the negative audience, but found no difference between
the "free" audience's character and that of the negative
one.

The ARs, on the other hand, evaluated the PR
audience significantly higher than the FR audience, and
the FR audience significantly higher than the NR
audience.

Dynamism. There were no significant differences

51

between Sp's and ARs' ratings of dynamism for any of
the three audiences. (Ratings of PR and NR audiences
differed less than 1.00 and differences between FR
ratings differed 1.17; we remember that the dynamism
dimension is being viewed as the activity factor and
thus requires a difference of 1.33 for significance.)

There was no significant difference between Sp's
ratings of the dynamism of the PR and the NR audiences.
There was a significant difference between his ratings
of the PR and the FR audiences, and between his ratings
of the NR and FR audiences. The speaker, then, found
the audiences where the independent variable was being
manipulated more dynamic than the audience where it was
not.

The ARs likewise did not find the PR audience more
or less dynamic than the NR audience, but found both of
them significantly more dynamic than the_FR audience.

Competence. There was no significant difference

 

between Sp's and ARs' ratings of the PR and NR audiences.
However, as they did on authoritativeness, they differed
significantly in their ratings of the FR audience. (This
consistency should be expected in light of the similarity
of the two dimensions.)

Just as he had for authoritativeness, the speaker
found no significant difference between the three audiences
in his evaluation of their competence. The ARs followed

the same pattern as they had for authoritativeness,

52
ranking the PR audience significantly higher in
competence than either the FR or NR audiences, and
finding no significant difference between the FR and
NR audiences. Again, just as they had on authoritative-
ness, they rated the FR audience lowest, NR second, and
PR highest.

Trustworthiness. Although this dimension was

 

represented by only three scales (perhaps because it
was) some interesting differences appeared. Because
Sp's and ARs' ratings of the PR audience had not
differed significantly in any of the other four
dimensions (especially since there was no significant
difference in their ratings for character) we would
not have expected a difference here. Nevertheless, a
significant difference between Sp's and ARs' ratings
of the PR audience's trustworthiness was obtained.

Sp's ratings on this dimension differed
significantly among the three audiences (and differed
to a greater degree than on any other dimension). His
factor score for the PR audience was significantly
greater than for FR or NR audiences, and his evaluation
of the FR audience was significantly higher than for the
NR audience.

There was no significant difference between ARs'
ratings of the PR and FR audiences. There was a

significant difference between AR's ratings of the PR

53
and NR audiences, and the FR and NR audiences. In

other words, ARs found the negative audience significantly
less trustworthy than either the positive or "free"
audiences. We note that this differed from their
evaluations in the character dimension. There they

found a significant difference between the character

of the PR and FR audiences (as did the speaker).

Summary. To summarize the analysis of this
particular set of data we must begin by remembering
that it was not collected in order to support any of
the a priori hypotheses of this study. Rather it was
the result of an effort to seek some support for the
assumption that the independent variable was present
and "functioning."

Caution must be observed in interpreting these data,
for at least three reasons: (1) the problem of analysis,
which has already been considered; (2) the obvious lack
of verifiable scale validity, for this particular use;
and (3) the problems of measurement presented by
differences in variables affecting the speaker and those
affecting the ARs. Further, there are considerations
raised by such questions as: What difference does it
make that the speaker was necessarily evaluating out
of the "base line" established by the first audience
when he evaluated the second and the third? Can any
reliable statements be made about speaker and AR
differences/similarities when ARs were different for

each treatment?

54

With these reservations and various others that
might be raised firmly in mind, it is still interesting
to note what was found.

In those dimensions where we might expect the
presence of the independent variable to be revealed
(i.e., dynamism, with such scale sets as:' aggressive—
meek, energetic-tired, active-passive; and trustworthiness,
with such adjectives as: just-unjust, cruel-kind,
admirable—contemptible) we found clear differences among
the three treatments, differences that could be termed
supportive of the assumption that the independent
variable was operating.

The most exciting aspect of these data is the
possibilities presented for additional research. We will

discuss some of them in Chapter IV.
Secondary Analysis: Control

Prior to reporting analyses performed in order to
examine the hypotheses of the study we should look at the
findings pertaining to control.

An analysis of variance, conducted on pre-test scores
of the 85 making up the three experimental treatments (PR,
FR, and NR) and the control group Ss, produced an F-ratio
of .20 and thus supported the assumption that there were
no significant differences among groups. (See Table 3)
This analysis, in supporting the assumption that external

controls had been sufficient, also indicates that no

55
statistical control procedures (such as analysis of

covariance) were necessary prior to additional analyses.

TABLE 3

Analysis of Variance Summary of Pre-test
Scores for Experimental and Control 85

 

 

 

 

Source d.f. Mean Square F
Between 3 3577.33 .20
Within 116 17376.01

No significant difference (E = —.64, df = 29,

.8 < p < .5) was obtained between pre—test and post-test
scores for the control group, thus supporting the
assumption that no variable outside the experimental
treatments had a significant effect on the experimental
83 between the time of the pre-test and the post-test
measures.

Having reported the results of the secondary analyses,
and having obtained some support for the assumptions that,
(1) the independent variable was probably operating in the
experimental treatments, and that, (2) the controls for
the experiment were probably adequate, we can now turn to
the analysis of the data bearing on the hypotheses of the

study.

56

Major Analysis

The basic statistical test employed was an
analysis of variance in which there were four between-SS
factors and two within-SS factors. One—tailed Eftests
were used to analyze within-group and between—group
differences where significant F-ratios were Obtained.

A confidence level of .05 was established for
statistical significance for all tests; however, obtained
probability levels are reported for all measures represent-
ing tests of a priori hypotheses.

Analysis of variance of post-communication attitude
toward message measures revealed a significant between—
group difference (F = 3.02, p < .05). The obtained

F-ratio is presented in Table 4.

TABLE 4

Analysis of Variance Summary of
Post-test of Attitude Toward Message
for PR, FR, NR, and CG

 

 

 

Source d.f. Mean Square F
Between 3 48801.66 3.02*
Within 116 16114.00

 

*F significant, p < .05

 

57
Hypotheses

 

Now let us look at the results of the analyses
bearing on each of the individual hypotheses of the study.
They fall into three major groupings. The first group is
concerned with within-group differences in attitude toward
message (measures between precommunication and post-
communication attitude toward the message). The second
group is concerned with between-group differences in
postcommunication attitude toward message (Treatment I,

PR vs. Treatment II, FR vs. Treatment III, NR). The third
group is concerned with between-group differences in
attitude toward speaker (PR vs. FR vs. NR).

Hypothesis I-l: Ss stimulated by positive OAR will

 

express a more favorable postcommunication attitude than
precommunication attitude. A significant difference

(p = 2.29, df = 29, p < .025) between Ss' precommunication
and postcommunication attitude was obtained. Thus support
was obtained for the prediction that Ss who had received
the message while observing positive audience response
would shift opinion in the direction of the message.

Hypotheses I-2: 88 where the OAR stimulus is "free"

 

will express a more favorable postcommunication attitude
than precommunication attitude. No significant difference
(3': .56, df = 29, p < .4) was obtained between 85
postcommunication and precommunication attitude. Thus

no support was obtained for the prediction that 85 who

had received the message while OAR was allowed to Operate

58
"free" would shift opinion in the direction of the
message.

Although not a hypothesis of this study, it is
interesting to note that post hoc analysis revealed that
83 who were stimulated by negative feedback did not express
a more favorable postcommunication attitude than pre-
communication attitude. No significant difference
(E_= -.64, df = 29, p < .4) was obtained between Ss'
postcommunication and precommunication attitudes. Thus
it was found that Ss who received the message in the
presence of negative OAR did not shift Opinion in the
direction of the message.

To summarize the analyses bearing on Hypotheses I
l, 2: Only the PR treatment produced significant changes
in Ss attitude toward the message. PR Ss shifted opinion,
as predicted, in the direction of the message. Neither
FR nor NR 85 shifted opinion significantly from pre—test
to post-test. It is interesting to note that although
changes in FR Ss cannot be attributed to other than
chance causes, mean score for post—test was higher than
pre-test, but for NR 85 post—test mean was lower than was
their pre-test.

Hypothesis II—l: Ss stimulated by positive OAR will

 

express more favorable attitudes toward message tOpic than
Ss where the OAR stimulus is "free." A significant difference
(£}= 1.74, df = 59, p < .05) was obtained between 85 who

received the message while being exposed to positive OAR

59

and those who receive the message while OAR operated
"free." Thus Hypothesis II-l was supported.

Hypothesis II—2: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker than
88 where the OAR stimulus is "free." This hypothesis was
broken down into three sub—hypotheses (II-2-a, II-2-b,
and II—2-c) since no “overall" measure of source
credibility could be made (See Chapter II, "Methods of
Measurement"). Analysis of data obtained in measures of
authoritativeness, character, and dynamism are reported
under their respective sub-hypotheses.

Hypothesis II—Z—a: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker's
authoritativeness than 85 where the OAR stimulus is "free."
No significant difference (p_= 1.65, df = 59, p < .06)
was‘obtained between $5 in the positive treatment and Ss

in the "free" treatment as they rated speaker's
authoritativeness. Thus Hypothesis II-Z-a was not
supported at the .05 level.

Hypothesis II—2-b: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker's
character than 85 where the OAR stimulus is "free." A
significant difference (p = 2.19, df = 59, p < .025) was
obtained between these Ss' ratings of speaker's character.
Thus the hypothesis that $5 exposed to positive response
would rate character higher than those in the "free"

treatment was supported.

60
Hypothesis II—2—c: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker's
dynamism than 55 where the OAR stimulus is "free." A
significant difference (p = 3.70, df = 59, p < .001) was
obtained between PR and FR Ss' ratings of speaker's
dynamism, thus confirming the hypothesis.

To summarize the analyses bearing on Hypotheses II
1, 2a, 2b, 2c: Of the four between—group differences
predicted three were supported by the data. 88 exposed to
positive response shifted attitude toward topic significantly
more than $5 in the "free" treatment. 85 in the positive
treatment rated the speaker higher on both character and
dynamism than did FR 85, but there was no significant
difference in authoritativeness ratings between PR and
FR groups.

Hypothesis III: Ss stimulated by positive OAR will
express more favorable attitudes than Ss stimulated by
negative OAR. The hypotheses in this category were
concerned with differences between Ss exposed to positive
OAR during message presentation and those exposed to
negative OAR during the presentation.

Hypothesis III-l: Ss stimulated by positive OAR will

 

express more favorable attitudes toward message than Ss
stimulated by negative OAR. A significant difference

(3 = 2.90, df = 59, p < .005) was obtained between 85 who
received the message in the presence of positive response

and those who received the message while negative response

 

 

 

 

61
was being generated. Thus Hypothesis III—l was supported.

Hypothesis III-2: Ss stimulated by positive OAR will

 

express more favorable attitudes toward speaker than $5
stimulated by negative OAR. This hypothesis, as did
Hypothesis II-2, predicted differences in three source
credibility measures, authoritativeness, character, and
dynamism. These predictions took the form of three sub-
hypotheses, III-2-a, III-2-b, and III-2-c.

Hypothesis III-2-a: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker's
authoritativeness than $5 stimulated by negative OAR. '

A significant difference was obtained (E = 4.36, df = 59,,
p < .0005). Thus the prediction that $5 in the positive
treatment would rate the speaker's authoritativeness
higher than those in the negative treatment was supported.

Hypothesis III-2-b: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker's
character than $5 stimulated by negative OAR. A significant
difference was obtained (E_= 3.26, df = 59, p < .005).

The prediction that Ss exposed to positive OAR would rate
speaker's character higher than $5 exposed to negative

OAR was supported.

Hypothesis III-2-c: Ss stimulated by positive OAR

 

will express more favorable attitudes toward speaker's
dynamism than Ss stimulated by negative OAR. A significant
difference (E_= 2.83, df = 59, p < .005) was obtained.

Thus the prediction that Ss' ratings of speaker's dynamism

 

62

in the positive treatment would be greater than in the
negative treatment was supported.

To summarize the findings bearing on Hypotheses III
1, 2a, 2b, and 2c: All predictions regarding between-
group differences for Ss who received the positive OAR
stimulus and those who received the negative OAR stimulus
were supported.

Hyppthesis IV: 85 where the OAR stimulus is "free"

 

will express more favorable attitudes than Ss stimulated
by negative OAR. Hypotheses IV 1, 2a, b, c make up the
predictions between $5 in the "free" OAR treatment and
those in the negative treatment.

Hypothesis IV-l: 85 where the OAR stimulus is "free"

 

will express more favorable attitudes toward message than
$3 stimulated by negative OAR. No significant difference
(3': .008, df = 59, p below table values) was obtained
between attitude toward message for Ss in the "free" OAR
treatment and those in the negative one; hence the
hypothesis was not supported.

Hypothesis IV—2: 85 where the OAR stimulus is "free"

 

will express more favorable attitudes toward speaker

than 83 stimulated by negative OAR. As in Hypotheses II-2
and III—2, this hypothesis represents three separate
predictions of between—group differences in source
credibility ratings.

Hypothesis IV-2-a: 35 where the OAR stimulus is

 

"free" will express more favorable attitudes toward speaker's

63

authoritativeness than 85 stimulated by negative OAR.

A significant difference (E_= 2.69, df = 59, p < .005)

was obtained. Thus the prediction regarding the first
dimension of source credibility was supported, i.e., Ss

in the "free" treatment rated speaker's authOritativeness
significantly higher than did Ss in the negative treatment.

Hypothesis IV-2—b: 88 where the OAR stimulus is

 

"free" will express more favorable attitudes toward speaker's
character than 85 stimulated by negative OAR. No significant
difference (p = 1.63, df = 59, p < .06) was obtained between
Ss' ratings of the Speaker's character in the "free" and
negative treatments. Thus the prediction that the speaker
would receive higher ratings on character from Ss where
audience response was not manipulated than from those where
negative response was displayed was not supported at the

.05 level.

Hypothesis IV-2—c: 85 where the OAR stimulus is

 

"free" will express more favorable attitudes toward speaker's
dynamism than $5 stimulated by negative OAR. No significant
difference (E = .004, df = 59, p below table values) was
obtained between FR and NR ratings of the speaker's
dynamism. Thus the prediction that $5 in the "free"
treatment would rate the speaker higher on the dynamism
scale than 88 in the negative treatment was not supported.

To summarize the analyses bearing on Hypotheses IV
1, 2a, 2b, and 2c: Only one of the four predictions
regarding differences between the "free" OAR treatment and

the negative feedback treatment was supported by the data.

64
FR 85 did not differ significantly from NR Ss in their

attitudes toward message following the experimental
treatments. It can be noted that mean attitude scores
for NR Ss had decreased from pre—test to post-test while
mean attitude scores for FR 85 had increased. However
interesting it may be to take such a look at direction,
no support from the data was obtained; i.e., no statistically
significant differences in attitude toward message were
present.

FR 85 differed significantly from NR 85 on only one
of the source credibility dimensions, authoritativeness.
A significant difference was "approached" but not achieved
on character; and there was clearly no significant
difference in dynamism ratings. Table 5 displays the mean

scores for Ss attitude toward message topic and speaker.

TABLE 5

Pre and Post-test Mean Scores for Attitude Toward
Message Topic and Speaker for 3 Experimental Groups
and Control Group

 

 

 

Attitude Trtmt I Trtmt II Trtmt III Control Grp
Message Topic:

Pre-test 4.59 4.46 4.47 4.32
Message Topic:

Post-test 5.13 4.57 4.31 4.25
Authoritativeness 5.12 4.67 3.86
Character 5.02 4.60 4.28
Dynamism 4.69 3.70 3.83

 

65

Summary of Results

Analysis of the data consisted of two secondary
analyses and the major analysis bearing on the hypotheses
of the study. Preceding these analyses we discussed
the elimination of Ss and data.

Data, collected by means of the LT scales, was not
used for purposes of analysis. All analysis was
performed on data collected from the SD scales. After
attrition and equating of cells, data from 120 Ss (30
$8 per cell) was analyzed.

Ratings by the speaker and "audience raters" of
the audiences for each treatment were examined and
interpreted as supporting the assumption that the
independent variable was Operating.

Analysis of pre—test scores for the three treatments
and the control group reveal no significant difference.
Analysis of pre and post—test scores for the control
group revealed no significant difference. These two
findings were interpreted as supporting the assumptions
that: (l) The Ss making up the three treatments and
the control group were drawn from the same pOpulation,
and (2) External control had been adequate.

Analysis of variance and appropriate petests
within and between the three experimental treatments
revealed the following:

1. Predictions that there would be within-group

66

differences in attitude toward message were

partially supported.

a. The positive group differed from pre to
post-test.

b. The "free" group did not differ from pre
to post-test.

c. The negative group did not differ from pre
to post-test.*

Predictions that there would be between-group

differences in attitude toward message were

partially supported.

a. The positive group differed from the "free"
group.

b. The positive group differed from the negative
group.

c. The negative group did not differ from the
"free" group.

Predictions that there would be between-group

differences in attitude toward speaker were

partially supported.

a. Between positive group and "free" group
(partial support)
(1) No difference on authoritativeness.
(2) Difference on character.

(3) Difference on dynamism.

 

*A post hoc finding, no directional a priori hypothesis.

67

b. Between positive group and negative group
(complete support)
(1) Difference on authoritativeness.
(2) Difference on character.
(3) Difference on dynamism.
c. Between negative group and "free" group
(partial support)
(1) Difference on authoritativeness.
(2) No difference on character.

(3) No difference on dynamism.

CHAPTER IV

DISCUSSION AND CONCLUSIONS
Design Implications

In discussing this study, we should return once
more to the rationale that produced the hypotheses and
design. The logical extensions of a process View of
communication; the recognition that, for public speaking,
the familiar heuristics were limited; and the resulting
efforts to analyze more precisely the one-to-many
communication event, all led to an examination of the
literature bearing on feedback and intra—audience
effects.

The growing suspicion that investigators of the
past have neglected an important element in this
particular kind of communication gave rise to the
conceptualization of OAR. Once conceptualized, the
notion needed to be tested. If OAR is "out there"
operating in the one-to-many communication event, then
its effect should be discernable by measurement.

The decision to manipulate response in two different
and "opposite" ways was made in order to give its effect
every chance to be observed. Original thinking was
concerned with positive response only, and a means by

which its effects could be ascertained. An early design

68

69

planned for two treatments, positive response and no
response. Here, the idea was to instruct confederates
in the positive treatment just as they were instructed
in this study, and to instruct confederates in the no
response treatment to withhold feedback from the speaker.
The problems of definition, precision, and artificiality
soon became apparent. We can assume, with some appeal
to commonsense and research, that positive response is
rewarding, and that negative response is punishing.
But what assumptions do we make about no response,
even if it were possible to create it (which is doubtful)?
Is it viewed by a speaker as neutral feedback or as
negative feedback? How is it viewed by other audience
members when they perceive it being produced/ not produced
by those around them? Therefore, the decision to create
both a positive and a negative condition, with a free
condition against which to compare them, was reached.

Although we have discussed, in some detail, the
justification for the decision to use a live speaker,
this seems the appropriate place to point out that the
advisability of doing so should be put to the test. If
the speaker's responses to the audience's feedback play
an integral part in the influence attributed to OAR, then
we need a study designed to demonstrate it. A replication
of this study, in which a taped (either audio only or
video) message was used, would give us information on

this issue.

70

With this review of the implications of design and
a recapitulation of the purpose of the study, let us now
turn to a discussion of the results by examining the

hypotheses tested.
The Hypotheses

Within-group Differences

 

We already know that communication is effective.
The literature is filled with studies attesting to the
efficacy of a persuasive message in creating short-term
opinion change. Once more we have an instance of a
message presented to audiences with the expectation
that they will change Opinion on the message topic in
the direction advocated. The speech, in this case,
was one which had changed opinion in the past for
similar audiences (See McCroskey, 1966a). Yet, the
findings in the present study indicated that only one
of the three audiences hearing the speech shifted
opinion significantly. The audience that received the
speech in the presence of positive OAR shifted opinion,
and in the direction predicted.

The audience in the "free" and negative treatments
did not shift opinion significantly. Why? Our first
tendency is to look for some confounding variable
responsible for the lack of change in the "free"
treatment. We are not concerned with the failure of

the negative treatment to produce change in the direction

71

of the message; indeed, we would be concerned if those
83 had done so. This would have meant that either
source or message, or both together (or even some other
variable) had interfered with the predictable effect

of the treatment.

Could the failure of the "free" treatment to shift
be attributed either to some sensitizing or "Hawthorne"
effect? Neither of these alternatives is reasonable.
If Ss in the "free" treatment had been sensitized by
the pre-test, they might attempt to reproduce their
pre-test responses on their post-tests. If they did
do that, then why did not 85 of the positive treatment
do likewise? If a "Hawthorne" effect were operating
(The S says to himself, "I know I'm in an experiment,
and I'm going to give them what they want.") the
obvious thing to do, after hearing a message advocating
X, is to indicate your approval of X. What these
considerations do seem to indicate is that control was
adequate in the experiment. Therefore, let us look
outside the experiment for some variable that might
account for the failure of the "free" treatment to
respond as predicted.

Between the time the speech had been previously
used with success, and the time it was used in the
present study, there were two developments that should
be entertained as possibly causative. Remembering that

the speech topic was "Federal Control of Education,"

72

we saw on the national scene, between 1966 and 1967,
increasing publicity for federal contributions to
education. (For example, it was during this period
that the federal government at last gave its support
to a large funding program for the humanities.) Some
Ss no doubt heard or read of this program, and of other
federal efforts in education. More salient, however,
were the local conditions, i.e., state—wide education
policies in California.

With state-wide cuts in funding of education, cuts
which directly affected the state college system, Ss
who were members of that system could hardly remain
unaffected. As to how they were affected we can only
speculate. They may have reacted to this on-going
dispute over higher education by viewing federal control
as a desirable alternative to the state's program.

The expectation for the speech topic did not take
such a variable into account. The expectation was:
Most persons are opposed to federal control of education.
Such opposition might even be viewed as a cultural
truism. Therefore, pre-test scores should be unfavorable;
following an effective speech, post-test scores should
show significant change. We should note, parenthetically,
that this expectation in no way "loaded the dice" in
this study, for the independent variable was OAR, not
some message variable.

Examination of pre—test mean scores for all groups

73

(See Table 5) indicated that none fell below 4.32
(4.00 is the theoretical neutral point on the seven
point scale). This indicates that prior to hearing
the speech 85 were not significantly opposed to "federal
control of education." Nor were they significantly in
favor of the topic. Given that 55 were affected by
information in the environment regarding education,
they might require more than the presentation of the
message to shift opinion. They might require the
additional force of perceiving it as being approved by
their peers. In short, for this particular topic and
this particular audience, Opinion change may have been
obtainable only with both an effective speech and
positive OAR.

One further consideration needs to be noted. Since
the message was presented live, perhaps some delivery
variable resulted in the positive treatment shifting as
predicted, while the "free" treatment did not. If the
speaker were going to improve as a result of practice,
for example, he should be better the second time than
the first. Further, when the speaker faced the "free"
treatment audience, he had just finished a "successful"
experience, one in which at least half of the audience
had been providing positive feedback. If anything,
his delivery should have been better. Yet there were
significant differences in both character and dynamism

(not in authoritativeness) indicating that his second

 

74
audience did not find him a "better speaker" than
his first.

Perhaps the practice effect did not result in
better delivery. Perhaps instead of being exhilerated
by his "success" the speaker was depleted. He was
tired, and therefore did not give the "free" audience
the same delivery he had presented to the positive
audience. This line of thinking would eXplain a
difference in dynamism ratings, but does nothing to
explain the difference and direction for character
(nor, for that matter, the lack of difference in
authoritativeness).

It seems more reasonable to believe that the
differences found were accounted for by the independent
variable, and that the failure of the "free" audience
to shift opinion should be attributed to the necessary
presence of positive OAR in order to obtain a significant
shift on this topic for this particular population.

Between-group Differences: Message

 

Although two of the three between-group predictions
regarding attitude toward message were supported, we
should examine the prediction which failed to receive
support.

We remember that the positive treatment resulted
in a more favorable postcommunication attitude than the
"free" treatment, and likewise more favorable than the

negative treatment. However, the "free" treatment did

75

not differ significantly from the negative treatment.
We can return to the line of thinking developed in our
discussion of the within-group finding for the "free"
treatment. Here too, the most reasonable explanation
does not seem to lie with control, but rather in the
idea that the presence of the independent variable was
necessary in order to shift opinion.

There is, however, a difference between the
unsupported hypothesis of the within-group prediction
and that of the between—group prediction. Why, if
the independent variable was the cause of shifting
opinion, did it not shift opinion far enough in the
negative treatment to produce a significant difference
between the "free" treatment and the negative one?

We cannot offer our previous rationale because
the independent variable - OAR — was manipulated in
the negative treatment; thus we must turn elsewhere.
It is possible that positive response is "stronger"
than negative response. What the unsupported
hypothesis of this between-group prediction may indicate
is that people are moved more by perceptions of
rewarding feedback than of punishing feedback. In
other words, it may be easier to like when others do
so than it is to dislike; to be more influenced by
praise than by denigration.

To repeat the previous inference drawn in the

discussion of the within-group hypothesis: perhaps a

76

significant shift in attitude toward t0pic could have
been achieved only when the persuasive force of the
message was coupled with the persuasive force of
audience approval (which in this case took the form
of positive OAR); that without them working together,
so to speak, no shift in opinion was obtainable.

Between-group Differences: Speaker

 

Examining the results of the hypotheses on source
credibility, i.e., predictions regarding between-group
differences in attitude toward the speaker, we find
three sets of predictions representing the three
combinations of treatments.

Treatment I vs. Treatment II. Of the three

 

hypotheses pertaining to the positive vs. the "free"
treatment, two were supported; i.e., the positive
treatment resulted in a more favorable attitude toward
the speaker's character and his dynamism than did

the "free" treatment. However, there was no
significant difference between ratings of the speaker's
authoritativeness.

One explanation is to suggest that the author-
itativeness dimension is not an adequate discriminator
for these Ss, hearing this speaker, on this topic.
This is to call into question the validity of the
scales, and to require the refutation of considerable
evidence (most of it cited in Chapter II) supporting

their validity. Further, to question the validity

77

of the authoritativeness scales and then to rely on
that of the character scales would be unexplainable
contradiction. We must look elsewhere for the answer.
We remember that the report of results included
obtained probability levels, regardless of their
failure to meet the criterion set for hypothesis
acceptance. The obtained level for this hypothesis
was < .06 and the obtained E_was 1.65. A p of 1.67
would have resulted in a probability level of .05,
the criterion set. Some investigators argue that
there is "no such animal" as "very significant" or
"highly significant" and certainly not "almost
significant." Either the data are significant or they
are not significant. The results of the analysis of
data pertaining to this hypothesis were reported as
not significant. We might, however, be allowed the
phrase, "approached significance." No other
explanation to account for the failure to support
this hypothesis is offered.

Treatment I vs. Treatment III. The hypotheses

 

regarding between-group differences in attitude toward
speaker, for the positive vs. the negative treatment,
were all supported well beyond the probability level
set. These findings especially, as well as those of
the entire set of predictions pertaining to source
credibility, can be interpreted as failure to find

(or to generate) a "boomerang effect" with the negative

78

treatment. One pertinent design consideration was
just such a possibility. Had there been favorable
ratings by negative treatment Ss viewing the speaker
as an "underdog," as someone being given a "hard-
time," their response might well have taken the form
of high ratings on source credibility. Further, this
"boomerang" could have extended to attitude toward
topic. They might have decided to repay the speaker
for his perceived-as—unjust treatment by indicating
their agreement with his message.

Apparently, neither of these effects occurred. It
does seem more reasonable to expect that they would show
first, or strongest, in measures of source credibility
rather than in attitude toward topic or elsewhere.

Treatment II vs. Treatment III. The final set of

 

hypotheses pertaining to source credibility concerned
differences between the negative and "free" treatments.
Here, only one of the three hypotheses was supported.
Although it was predicted that the "free" treatment
would produce higher ratings on authoritativeness,
character and dynamism, only on authoritativeness did
the speaker receive significantly higher ratings from
the "free" group. The differences between ratings of
character approached significance (p < .06) but did not
meet the .05 criterion. Once more the necessary E_was

1.67, and the 2 obtained was 1.63.

79

It is the "dynamism hypothesis" that was clearly not
significant. Here there are several possible explanations.
we should once more consider control in this live speaker
situation. Could the speaker have been tired, having
presented the speech twice, so that his delivery suffered
in the third treatment? If that were the case, then
dynamism, if any dimension, should have showed it, i.e.,
the speaker should have been rated significantly lower
on dynamism in the negative treatment. The data indicated
no significant difference on this dimension. We must
reject this possibility. On the other hand, could the
speaker have stiffened-under-fire, so to speak, and
become more dynamic in response to the negative feedback?
Again, this should have produced significant differences
(but with the direction reversed) between the two treat-
ments. We must once more conclude that there was no
reason to suspect inadequate control.

There is an explanation that, although not altogether
satisfying, does have the merit of consistency. A
previous interpretation which considered the possibility
that positive OAR could be "stronger" than negative OAR
may find application here also. We might posit that the
three conditions (at least in terms of attitude toward
the speaker) are hierarchical. This hierarchy could take
the following order: positive, "free," negative. We
would then expect our greatest difference between the

positive and negative treatments, our next greatest

80
between the positive and the "free" treatments, and our
least difference between the "free" and negative treat-
ments. The results, at least, do not deny such a
hierarchy. All source credibility dimensions were
significantly different in the positive vs. negative
comparison. Two of the three differed significantly
(with the third approaching significance) in the positive
vs. "free" comparison. One differed significantly, one
approached significance, and one did not differ in the
"free" vs. negative comparison.

Secondary Analysis

 

Now, having discussed the results pertaining to the
hypotheses, with special attention to those that were not
supported, we can turn to some of the questions raised
in the secondary analysis. We recall that measures of
the audiences were obtained from the speaker and observers,
designated "audience raters" (ARs); and that while
reporting the results of analysis we pointed out that
there were areas of research that might be explored.

The possibilities for research designed to identify
and classify audiences go beyond the OAR phenomenon.

Here we enter an area that seems to call for a massive
descriptive research effort.

Perhaps a meaningful way of getting started would be
the creation of scales, similar to McCroskey's and/or
Berlo and Lemert's, by means of which speakers and other

audience members could rate various audiences. By this

 

81
means we might begin to develop sets of audience
profiles.

Even considering the highly limited nature of the
audience rating in this study, we may have been provided
with a provocative research lead: the high agreement
between speaker's and ARs' ratings. Common sense would
tell us that a speaker views an audience much differently
from the way the audience views itself, yet the evidence
from this study, meager as it is, seems to indicate
otherwise. This becomes more interesting when we
remember that the audiences being measured were deliberately
manipulated to differ significantly. Does this mean
that a speaker, scheduled for a particular audience,
might "scout it" by means of ARs whose evaluations would
reflect, or at least come close to, his own? If such
ratings (speaker's vs. ARs) differ, where and how do
they do so? These questions and many others in this
area, are in need of investigation.

The preceding perhaps serves as an apprOpriate
preface for our next considerations, those regarding

future research on the effects of OAR.

Future Research on the Effects of OAR

In discussing the possibilities provided by this
concept for future research, let us begin with the most
obvious areas in which we might look for variations

extrapolated from this study. That is, what variations

 

82

are available in speaker, speech, and audience?
Speaker

Remembering that the speaker for this study was
one selected with deliberate intent to hold source
credibility low, so that it might not "mask" the
presence of other variables, we could design a study
which would manipulate the speaker's credibility.
What would be the results of a high credible source
receiving negative feedback? WOuld the "boomerang"
effect, mentioned previously, take place? Would
there be differences between a genuine authority
figure interacting with OAR and one whose authority
was artificially "created" (the latter might be
attempted by introducing an unknown speaker with
material designed to give him high credibility)?
Other speaker variations, as well as these, could be
devised.
Speech

The possible message variations are many. This
study utilized a message assumed to be strongly attitude
discrepant. As mentioned previously in this chapter,
examination of pre-test scores indicated that 85 were
generally neutral. Studies examining the influence of
OAR interacting with attitude discrepant and attitude
congruent messages seem called for.

Message salience is another variable to investigate.

What would be the effects of OAR when it was varied in

83

conjunction with high-salient and low—salient topics?
There is yet another kind of message variance to
be explored in combination with OAR. What would be the
effects of introducing materials in the speech designed
to off-set the effects of feedback? For example, such
a message might contain such phrases as: "We should think
for ourselves. . . " "As independent thinkers, we . . . "
etc. A recent cigarette commercial seems to make this
sort of "anti-conformity" appeal.
Audience
The variations in audience could take many forms.
This study utilized audiences in which Ss and Cs were
drawn from the same population; 85 were being influenced
by their peers. A study could be designed in which
authority figures provided the manipulated feedback.
What would have been the results of the present study
had the student Ss received treatment in the presence
of their professors? OAR could come from a reference
group as opposed to a membership group, or from what
Newcomb (1950) has called a negative reference group.
(We will explore the possibilities of a negative
reference group below, under considerations of "other
dependent variables.") Another audience variation
might find OAR coming from opinion leaders. (These
opinion leaders could be "created" from a series of
pre-treatment group discussions in which emergent

leaders were identified and then reinforced.)

 

 

84

Closely connected with the preceding types of
variations would he studies designed to determine
whether some personality types were more or less
influenced by the responses of others. If there are some
persons who are more susceptible to the influence of
OAR, what, if any, are their shared characteristics?

Variations in age groups would be possible with
a series of replications using younger, older, or
varying combinations of age groups demonstrably different
from the "young adult" category that made up the
audiences for this study.

Finally, audience attitudes could be manipulated and
"played—off" against OAR in a number of ways. For
example, various sets toward speaker and/or message
could be created prior to treatment. The OAR could
then be manipulated either to support or conflict with
these predispositions.

Three additional types of variation might well be
considered in a continuing examination of OAR: (l)
within-treatment variations, (2) situation or context
variations, (3) the consideration of other dependent
variables.

Within—treatment

 

Some challenging possibilities for studies present
themselves when we consider ways in which OAR could be
varied, other than the positive only and negative only

forms that were used in this study. Positive and

85

negative OAR could be presented in varying proportions
within the same treatment, and this manipulation could
take a variety of forms. For instance, we might have

a treatment in which the majority of the OAR was positive,
but a small minority of the audience was producing
negative OAR. (This is a common occurrence at the typical
political address, at which the majority of the audience
is partisan but a contingent from the opposition shows

up to heckle the speaker.)

Another within—treatment manipulation could take
the form of one type of OAR gradually (or suddenly)
giving way to another type, i.e., rather than having two
conflicting types of OAR operate simultaneously, we could
have a treatment which saw positive OAR at the outset
gradually turning into negative OAR during the latter
portions of the message presentation, or vice versa.

The question raised with this particular example is:
Would 85 be more inclined to respond favorably to
climactic or anti-climactic positive OAR?

Still another consideration would be variations in
intensity of OAR. We recall from Chapter II, that Cs
were encouraged to create OAR, for both treatments, as
they determined. Studies examining strong positive OAR
vs. weak positive OAR, strong negative vs. weak negative,
and various other combinations could be undertaken.

Finally, within-treatment variation could take the

form of using OAR intermittently to reinforce a selected

86

point or points within the message.

These are only some of the studies that could be
undertaken by varying OAR within treatments, rather
than between, as this study did.

Situation or Context

 

In generalizing from this study, we cannot exclude
the possible limitations of context or situation.
Student 85 heard the speech and received the experimental
stimulus in a college environment. We do not know how
important they considered the experience, whether they
took a positive or a negative view toward the situation,
etc. We can assume that their attitude toward situation
or context could have been manipulated. Studies containing
such manipulations would provide additional insight into
OAR. Putting it another way, a major concern of this
study was to hold situation or context constant so that
any changes between treatments could be attributed to
the study's independent variable. However, situation
or context could be manipulated (treated as an independent
variable) in order to observe its interaction with OAR.
For example, the situation for one treatment might be
made more important in order to compare it with another
treatment where situation was not so manipulated. As
Doob (1948, p. 531) has pointed out, when people believe
a speech is being broadcast they consider it important

and, "what is important tends to make them more submissive."

 

87

Other Dependent Variables

 

This study was designed to measure the effects of
OAR on attitude toward a message and on source credibility.
There are numerous other variables that might be measured
in order to determine its effects. One of the most
obvious is comprehension or retention. Do Ss retain
more of a message after the stimulation of OAR? Does
positive or negative OAR cause a speech's theme to be
retained more accurately or for a longer time? These
and other similar questions await answers to be provided
by studies in which OAR is manipulated and different
dependent variables are measured.

Newcomb's (1950) notion of negative reference groups,
mentioned earlier, provides an intriguing application
for an OAR study. A negative reference group, as its
name implies, is one which an individual regards as
"the opposition" or "the enemy." Just as his own
membership group (or one to which he aspires) serves as
a check on how he should behave, the negative reference
group provides a model for how he should not behave.

If 85 were members of an audience composed primarily
of a negative reference group, then the perception of
positive OAR ought to shift their attitudes either toward
the group, or toward the message, or both, depending on
message content. Similarly, the perception of negative
OAR should strengthen their attitude toward the message
or decrease their attitude toward the negative reference

group.

”T'Wi

88
In concluding this portion of the discussion,
relating to possibilities for future research, we must
call attention to at least one more concern. It is
the question of whether the effects of attitude change
induced by OAR are temporary or lasting. The answer to
this question lies in delayed measures to determine if

85 "return" to their original pre—treatment attitudes.

Some Necessary Reservations

At this point, we should muster all available
scepticism, and once more review this study with the
specific goal in mind of focusing on necessary reservations
as to its worth.

It should not be necessary to review again the

control questions raised by the use of a live speaker.
The problems inherent in such use have been recognized
and discussed at some length. Therefore, let us go on
to consider the limitations of OAR intensity that were
touched on previously.

No "base line" existed in the study by which intensity
(or "degree") of perceived OAR could be measured. We
can talk only about kind of OAR generated in the treat—
ments. As to how much positive OAR was being produced
in Treatment I, there is no satisfactory answer, nor can
we say how much negative OAR was being produced in
Treatment III. We can say that approximately one—half

of the audience was probably generating the appropriate

89

kind of OAR for the particular treatment. Whether
this was contagious, spreading from Cs to 88 within
a treatment, is another unknown. Such "unknowns"
present problems for accurate replication.

Even though we may, for this study, argue that
some crude quantification is possible (on the basis of
ratio of Cs to 88), how do we know whether we had weak
(whatever "weak" may mean) positive OAR vs. strong
(whatever "strong" may mean) negative OAR, or vice
versa? We do not. Our ability to specify becomes even
more questionable when we consider the "free" treat-
ment. Did we have equal portions of positive and
negative OAR operating? Did one or the other pre-
dominate? Did we have a treatment that produced
mixed OAR simultaneously, or did OAR take some linear
pattern?

We do know that in both the positive and negative
treatments half the audience probably maintained the
same general kind of OAR throughout the treatment. No
matter what the speaker said or did, the Cs were there
to produce the kind of OAR appropriate for their treatment,
and we assume they did so. This meant that only half
the audience (the 85) were free to produce any kind
(or no kind) of OAR, whereas the entire audience in the
"free" treatment could produce whatever kind of OAR, and
at whatever points in the event, they wished. What

contribution did this make to our results?

90

All of the above questions can rightly be raised
in any consideration of the findings. Are there any
answers to them? There seem to be two responses (if
not answers). First, although it is true that the
"free" treatment could have varied as to kind or degree
of OAR, we have some tentative evidence to consider.
Both the speaker and independent observers (ARs) Viewed
this audience as different from those of the positive
and negative treatments; and, in most cases, supported
the assumption that a "neutral" classification (as
compared with the other two) was a reasonable one.

Further, we can assume that the "free" audience's
responses were made out of the same context, with the
same speaker, presenting the same message, to a group
drawn from the same population, and holding the same
precommunication attitudes, as those of the positive and
negative treatments. We can, in short, assume that
all variables operating in the manipulated treatments
were operating in the "free" treatment, with the
exception of the independent variable.

Second, and perhaps more important then any "defense"
of the experiment, is the observation that out of such
considerations as these grow additional research
hypotheses and experiments and, hence, additional know—
ledge. We must begin somewhere, and we need to begin
by establishing that a variable can have gpgg effect

before we can analyze its effect(s) with any precision.

 

91

We did recognize, in this study, a significant
variable in the communication process; future research
will enable us to analyze that variable further. One
of the first steps in such an analysis would be to
determine a method of specifying "strength" or
intensity of OAR. A promising beginning would be the
development of the audience rating forms we have
discussed previously.

Another reservation refers to the limitations
imposed by the experimental context or situation. We
have already implied this problem in our preceding
discussion. There, we pointed out that context was
at least limited by the time, the place, and the
immediate environment of the experiment. Perhaps the
most significant element operating in this particular
situation was the CUltural norms of audience behavior
in a school environment. We cannot discount the
possible effects of the experimental stimulus interacting
with such norms (whatever they may be).

Further, we could speculate that the amount of OAR

perceived (not the amount generated) is probably the

 

result of an interaction between the amount generated
and the situational variables, or expectations, of a
particular audience. Since we do not know what these
expectations may be, let alone what the effects of
violating or complying with them might be, we can only

call attention to the necessity of recognizing their

92

obvious role in this and other communication events.
One further limitation needs to be mentioned. The
Cs providing the stimulus in the experiment were

"acting,' that is, we assume that despite their held
attitudes they generated the OAR appropriate for their
particular treatment. What effect does this produce?
Would results be different if their OAR was congruent
with their attitudes? Again, we do not know. Once
more, future description and experimentation in this
entire area is called for.

With these reservations firmly in mind, let us now

consider some wider applications that may be made.

Significance of the Research

Now let us discuss, in a broader fashion than that
of the preceding sections, what the results of this
study may mean. First, we can recall that the impetus
for the study came from a consideration, or the "logical

' of a process View of communication. This

extension,‘
view led to speculation in relatively "unknown territory,"
that of intra-audience effects. Perhaps this study,

and others of a similar nature, will reawaken interest

in such concepts as social facilitation, and will indicate
the applicability of the research (past and current) in
imitation, conformity, and vicarious socialization to

public speaking. Such interest, once aroused, could

lead to further extensions of the "boundaries" of the

93

speaking event. Beginning with such a premise, we

may be led to examine pre—event and post-event variables.
"Response" of some sort must take place before a

speaking event. What do people say to each other (both
verbally and non—verbally) as they begin to assemble in

a lecture hall? We have talked previously about

creating precommunication sets, and there is ample concern
shown in the literature for precommunication attitudes,
but how much does pre-event interaction among audiences
account for these things?

Closer to the OAR concept is interaction that
takes place following a speaking event. As an audience
leaves an auditorium, members may react in ways as
"simple" as one man turning to another and asking,

"What did you think of the speech?" and receiving the
reply, "I thought it was very good," or, "Damn boring!"
One the other hand, we may have reactions as "complex"

as a group spending an entire evening in heated discussion
of the speech attended to earlier.

The preceding is an example of something that might
be labeled "delayed OAR." It could well be an important
element in explanations of results on delayed post-
communication measures. It is certainly an area for
study itself.

Whenever we discover, define, or invent a new
concept, or look at an old one in a new way, the

applications and insights provided seem to be virtually

94

limitless. We have touched on them throughout this
study. In Chapter I, we speculated on the relationships
that may exist between OAR and other concepts such as
conformity, imitation, and vicarious learning. Here,

in the concluding portion, we have attempted to indicate

some of the "doors" that the concept and its implications

may begin to Open.

References

Allport, F. H. Social Psychology. Boston: Houghton
Mifflin, 1924.

 

Amato, P. and T. Ostermier. "The Effects of Audience
Feedback on the Beginning Public Speaker," Speech
Teacher. 16: 56—60, 1967.

Asch, S. E. "Studies of Independence and Conformity:
A Minority of One Against a Unanimous Majority,"
Psychological Monographs. 70: 1956.

. "Effects of Group Pressure Upon the
Modification and Distortion of Judgments," in
E. E. Maccoby, T. M. Newcomb, and E. L. Hartley,
Readings in Social Psychology. 3rd ed. 174-182.
New York: Holt, Rinehart, and Winston, 1958.

 

Ashby, W. R. Design For a Brain. New York: John
Wiley, 1952.

 

Bavelas, A. "A Mathematical Model for Group Structures,"
Applied Anthropology. 7: 16—30, 1948.

 

. "Communication Patterns in Task Oriented
Groups," Journal of the Accoustical Society of
America. 22: 725-730, 1950.

 

. '"Communication Patterns in Problem-Solving
Groups," in H. von Foerster, et al. (eds.) '
Cybernetics: Circular Causal and Feedback Mechanisms
in Biological and Social Systems. New York: Macy
Foundation, 1952.

 

Benedict, T. W. "An Experimental Study of Social Status
as a Dimension of Ethos." Unpubl. doctoral
dissertation, University of Southern California, 1958.

Berelson, B. and G. Steiner. Human Behavior. New York:
Harcourt, 1964.

 

 

Berger, S. M. "Incidental Learning Through Vicarious
Reinforcement," Psychological Reports. 9: 477-491,
1961. '

Berlo, D. K. The Process of Communication. New York:
Holt, 1960.

 

95

96

and H. Gulley. "Some Determinants of the
Effects of Oral Communication in Producing Attitude
Change and Learning." Speech Monogrgphs. 24:
10-20, 1957-

 

 

and J. B. Lemert. "A Factor Analytic Study
of the Dimensions of Source Credibility." A paper
presented at the convention of the Speech Association
of America, 1961.

 

and R. J. Mertz. Dimensions for Evaluating
the Acceptability of Message Sources. Research
Monograph, Department of Communication, Michigan
State University, 1966.

 

 

 

Bormann, E. G. Theory and Research in the Communicative
Arts. New York: Holt, Rinehart, and Winston, 1965.

 

Callaghan, J. C. "Social Facilitation in Persuasion,"
Quarterly Journal of Speech. 25: 643-656, 1940.

 

Campbell, D. T.' "Systematic Error on the Part of Human
Links in Communications Systems," Information and
Control. 1: 334—369, 1958.

 

Cherry, C. On Human Communication. Cambridge: M.I.T.
Press, 1957.

 

 

Cohen, A. R. "Upward Communication in Experimentally
Created Hierarchies," Human Relations. 11: 41-53,
1958.

. Attitude Change and Social Influence. New
York: Basic Books, 1964.

 

 

Collins, B. E. and H. Guetzkow. A Social Psychology of
Group Processes for Decision-Making. New York:
John Wiley, 1964.

 

 

Doob, L. W. Public Qpinion and Propaganda. New York:
Henry Holt, 1948.

 

Eisenson, J. et al. The Psychology of Communication.

New York: Appleton-Century-Crofts, 1963.

 

Festinger, L. "Social Psychology and Group Process," in
C. P. Stone and Q. McNemar (eds.) Annual Review of
Psychology. 6: 187—216, 1955.

 

 

97

Frates, R. "An Experimental Study of Attitude Change
as a Function of Dogmatism and Experimental
Environment." Unpubl. M.A. thesis. San Jose State
College, 1968.

Haiman, F. S. "An Experiment in Informative Speaking."
Quarterly Journal of Speech 34: 355-360, 1948.

 

Holtzman, P. D. "Confirmation of Ethos as a Confounding
Element in Communication Research." Speech

Monographs 30: 464—466, 1966.

Houston, J. P. and S. A. Mednick. "Creativity and the
Need for Novelty," Journal of Abnormal and Social
Psychology. 66: 137-141, 1963.

 

Hovland, C. I. "Reconciling Conflicting Results Derived
from Experimental and Survey Studies of Attitude
Change." American P§ychologist 14: 8-17, 1959.

 

Johnson, F. C. and G. R. Klare. "Feedback: Principles
and Analogies," Journal of Communication. 12:
150—159, 1962.

 

Kelley, H. H. "Communication in Experimentally Created
Hierarchies," Human Relations. 4: 39-56, 1951.

Kerlinger, F. N. Foundations of Behavioral Research.
New York: Holt, Rinehart, and Winston, 1964.

 

 

Krasner, L. "Studies of the Conditioning of Verbal
Behavior," Psychological Bulletin. 35: 148—170,
1958.

l
l
Lambert, W. W. and W. E. Lambert. Social Psychology.
Englewood Cliffs: Prentice—Hall, 1964.
l
l
l

 

Layes, E. "An Experimental Investigation of Attitude
Change Under Varying Audience Conditions." Unpubl.
research project. San Jose State College, 1967.

Journal of Abnormal and Social Psychology. 51:
175—183, 1955.

Lazowick, L. M. "On the Nature of Identification."
Leavitt, H. J. and R. A. Mueller. "Some Effects of

Feedback on Communication." Human Relations. 4:

401-410, 1951. .
Maccoby, E. N. et a1. "Social Reinforcement in Attitude

Change." Journal of Abnormal and Social Psychology.

63: 109—115, 1961.
l
l

98

McCroskey, J. C. "Experimental Studies of the Effects
of Ethos and Evidence in Persuasive Communication."
Unpubl. doctoral dissertation, Pennsylvania State
University, 1966a.

"Scales for the Measurement of Ethos."
Speech Monographs. 33: 65-72, 1966b.

 

 

. An Introduction to Rhetorical Communication.
Englewood Cliffs: Prentice—Hall, 1968.

 

and R. E. Dunham. "Ethos: A Confounding
Element in Communication Research." Speech
Monographs. 30: 456-463, 1966.

 

 

Miller, G. R. "Variations in the Verbal Behavior of A
Second Speaker as a Function of Varying Audience
Responses." Speech Monographs. 31: 109-115, 1964.

 

. Speech Communication a Behavioral Approach.
Indianapolis: Bobbs-Merrill, 1966.

 

‘. et al. "The Effects of Differential Reward
on Speech Patterns." Speech Monographs. 28: 9—15,
1961.

 

Minnick, W. The Art of Persuasion. Boston: Houghton-
Mifflin, 1957.

Moretti, Michael. "An Experimental Study to Measure the
Effect that Two Methods of Suggesting Speaker
Intelligence and Competence Have on Audience Ratings
of Speaker Character and on Message Persuasiveness."
Unpubl. M. A. Thesis, San Jose State College, 1965.

Newcomb, T. M. Social Psychology. New York: Dryden
Press, 1950.

 

Nyquist, H. "Certain Factors Effecting Telegra hSpeed."
Bell System Technical Journal. 3: 324, 1824.

 

Oliver, R. Persuasive Speaking. New York: Longmans
Green, 1950.

 

Osgood, C. E., G. J. Suci, and P. H. Tannenbaum. The
Measurement of Meaning. Urbana: University of
Illinois Press, 1957.

 

Reitan, H. T. and M. E. Shaw. "Group Membership, Sex-
Composition of the Group, and Conformity Behavior."
Journal of Social PsycholOgy. 64: 41-51, 1964.

 

99

Sherif, M. "A Study of Some Social Factors in
Perception." Archives of Psychology. 187, 1935.

 

Simpson, R. H. "Attitudinal Effects of Small Group
Discussion." Quarterly Journal of Speech. 46:
413—418, 1960.

 

Sopolsky, A. "Effects of Interpersonal Relationships
Upon Verbal Conditioning." Journal of Abnormal
and Social Psyghology. 60: 241—246, 1960.

 

 

Stoltz, W. and P. Tannenbaum. "Effects of Feedback on
Oral Encoding Behavior." Language and Speech. 6:
218-228, 1963.

 

Stotland, E. "Identification with Persons and Groups."
Final report on Grant M—2423 to N.I.M.H., U.S.
Public Health Service, October, 1961, cited in
Secord, P. and C. Backman. Social Psychology.

New York: McGraw-Hill, 1964.

 

Thibaut, J. W. and J. Coules. "The Role of Communication
in the Reduction of Interpersonal Hostility."
Journal of Abnormal and Social Psychology. 47:
770-777, 1952.

Thomas, G. L. and D. C. Ralph. "A Study of the Effects
of Audience Proximity on Persuasion." Speech
Monographs. 26: 300-307, 1959.

 

Vlandis, J. "Variations in the Verbal Behavior of a
Speaker as a Function of Varied Reinforcing Conditions."
Speech Monographs. 31: 116-119, 1964.

 

Walter, O. "The Measurement of Ethos." Unpubl. doctoral
dissertation. Northwestern University, 1948.

Wiener, N. Cybernetics. Cambridge: M.I.T. Press, 1948.

 

APPENDIX A

INSTRUCTIONS TO SS PRECEDING PRE-TEST

PLEASE READ TO CLASS BEFORE DISTRIBUTING QUESTIONNAIRES:

The Speech Activities Class at SJSC is interested in
gathering opinions on a variety of topics. Your response

to this questionnaire would be much appreciated.
There are two important points:

(1) The questionnaire has Egp_parts. 'In the first part
you are asked to respond to statements using a number
system. In the second part you are asked to respond to
a concept by checking a set of scales. Please read the

directions for both parts carefully.

(2) Please ignore the blank on the questionnaire for
name. We want your honest opinion and do not want you
to feel intimidated by signing the questionnaire.
However, we would like to know who did and who did not
participate. Therefore,-will you please print your name
on the attached piece of scratch paper; remove it from

the questionnaire and hand them in separately.

100

APPENDIX B

PRE-TEST QUESTIONNAIRE

Name:

 

DIRECTIONS: Read each of the following statements carefully.
If you strongly approve of the statement as it stands, enter
the number _H_ in the space provided at the left of the
sentence. If you strongly disapprove of the statement then
enter the number _l_ at the left of the sentence, and so on,
with regard to the other attitudes (approve, undecided,
disapprove).

STRONGLY STRONGLY
APPROVE APPROVE UNDECIDED DISAPPROVE DISAPPROVE
(5) (4) (3) (2) (l)

1. Quality in education can best be guaranteed by
local school boards.

2. Public schools should be financed primarily by
federal taxes.

3. Equality of opportunity in education for all citizens
can be best guaranteed by federal control of education.

4. The curriculum of public schools should be determined
by local school boards.

5. The federal government should have no control of the
curriculum of public schools.

6. Public schools can be better administered by the
federal government than by local school boards.

7. Quality in education can best be guaranteed by the
federal government.

8. The curriculum of public schools should be determined
by the federal government.

9. Equality of opportunity in education for all citizens
can be best guaranteed by local control of education.

10. Public schools should be financed primarily by local
taxes.

11. The federal government should assume complete
financial responsibility for education.

12. Public schools can be better administered by local
school boards than by the federal government.

13. The amount of teachers' salaries should be determined
by the federal government.

14. Teacher employment standards should be controlled by
local school boards.

lOl

15.

l6.

17.

18.

19.

20.

21.

22.

102

APPENDIX B, continued

Public schools should not be primarily financed
by local taxes.

The amount of teachers' salaries should be determined
by local school boards.

The federal government should have no control of
teacher employment standards.

The federal government should substantially
increase its financial support of public education.

The federal government should not determine the
amount of teachers' salaries.

Teacher employment standards should be controlled
by the federal government.

Public schools should not be primarily financed by
federal taxes.

Local school boards should not determine the amount
of teachers' salaries.

103

APPENDIX B, continued

DIRECTIONS: In using the following scales please make your
judgments on the basis of what these measures mean to you.

Here is how to use the scales:
(Concept)

Good : : : : : : Bad

 

If you felt that the concept was in general extremely good,
you would place a check mark in space number 3. If uite
ood (but not extremely good), in 2; if slightly good 1n

1; 1f neither good nor bad, in 0; if slightly bad, in -1;
if quite bad, in -2, and if extremely bad, in —3.

 

 

 

 

 

 

The "0" or neutral space on the scale may also be used for
"I don't know" or "I don't think this scale applies" answers.

Be sure to put a check mark somewhere along each scale. Put
your check within the spaces, not on the lines separating the
spaces. Put one, and only one, check on each scale. Please
do not omit scales.

FEDERAL CONTROL OF EDUCATION

 

 

 

 

 

Harmful : : : : : Beneficial
Good : : : : : : Bad
Wrong : : Right
Fair : . : . : Unfair
Negative : : : ' Positive
Wise : : : : : : Foolish

 

APPENDIX C

SPEAKER'S INTRODUCTION

Good morning,
The speaker that we will hear this morning is Ken Yules,
who will discuss the role of the Federal government in

public education.

Mr. Yules is a sophomore who is associated with speech
activities. The title of his speech is, "The Altar of

Responsibility."

104

 

APPENDIX D

POST-TEST QUESTIONNAIRE

DETACH THIS SHEET FROM YOUR QUESTIONNAIRE AND SUBMIT IT

SEPARATELY

Name :

Age:

Year in School:
Major Field of Study:

Minor Field of Study:

DETACH THIS SHEET FROM YOUR QUESTIONNAIRE AND SUBMIT IT

SEPARATELY

105

106

APPENDIX D, continued

DIRECTIONS: Read each of the following statements carefully.
If you strongly approve of the statement as it stands, enter
the number _H_ in the space provided at the left of the
sentence. If you strongly disapprove of the statement then
enter the number 1 at the left of the sentence, and so on,
with regard to th§_5ther attitudes (approve, undecided,

disapprove).

STRONGLY STRONGLY

APPROVE APPROVE UNDECIDED DISAPPROVE DISAPPROVE
(5) (4) (3) (2) (l)

l. I respect this speaker's opinion on the topic.
2. This speaker is not of very high intelligence.

3. This speaker is a reliable source of information on the
topic.

. I have confidence in this speaker. ‘
. This speaker lacks information on the subject.

This speaker has high status in our society. 1
. I would consider this speaker to be an expert on the topic.
8. This speaker's opinion on the topic is of little value.
9. I believe that this speaker is quite intelligent.

10. The speaker is an unreliable source of information on
the topic.

11. I have little confidence in this speaker.
12. The speaker is well—informed on the subject.
13. The speaker has low status in our society.

14. I would not consider this speaker to be an expert on this
topic.

15. This speaker is an authority on the topic.

16. This speaker has had very little experience with this
subject.

17. This Speaker has considerable knowledge of the factors
involved with this subject.

18. Few people are as qualified to speak on this topic as
this speaker.

19. This speaker is not an authority on the topic.
20. This speaker has very little knowledge of the factors
involved with the subject.

21. This speaker has had substantial experience with this
subject.

22. Many people are much more qualified to speak on this topic
than this speaker.

107

APPENDIX D, continued

DIRECTIONS: Read each of the following statements carefully.
If you strongly approve of the statement as it stands, enter
the number 5 in the space provided at the left of the
sentence. If_you strongly disapprove of the statement then
enter the number 1 at the left of the sentence, and so on,
with regard to thé_5ther attitudes (approve, undecided,

disapprove).
STRONGLY STRONGLY
APPROVE APPROVE UNDECIDED DISAPPROVE DISAPPROVE
(5) (4) (3) (2) (l)
___l. I deplore this speaker's background.
___2. This speaker is basically honest.
___3. I would consider it desirable to be like this speaker.
___4. This speaker is pp; an honorable person.
___5. This speaker is a reputable person.
___6. This speaker is £93 concerned with my well—being.
___7. I trust this speaker to tell the truth about the topic.
___8. This speaker is a scoundrel.
9. I would prefer to have nothing at all to do with this

speaker.

10. Under most circumstances I would be likely to believe
what this speaker says about the topic.

11. I admire the speaker's background.

12. This speaker is basically dishonest.

13. The reputation of this speaker is low.

14. I believe that this speaker is concerned with my well-being.
15. This speaker is an honorable person.

16. I would app prefer to be like this speaker.

17. I do pp; trust the speaker to tell the truth on this topic.

18. Under most circumstances I would not be likely to believe
what this speaker says about the topic.

19. I would like to have this speaker as a personal friend.

20. The character of this speaker is good.

 

108
APPENDIX D, continued

DIRECTIONS: In using the following scales please make your
judgments on the basis of what these measures mean to you.
Here is how to use the scales:

(Speaker)

Good : : : : : : Bad
3 2 1 0 -l —2 -3

 

If you felt that the speaker was in general an extremely good
one, you would place a check mark in space number 3. If quite
good (but not extremely good), in 2; if slightly good in 1; if
neither good nor bad, in 0; if slightly bad, in —1; if quite
RES! in -2, and if extremely bad, in -3.

 

 

 

 

 

 

The "0" or neutral space on the scale may also be used for
"I don't know" or "I don't think this scale applies" answers.

Be sure to put a check mark somewhere along each scale. Put
your check within the spaces, not on the lines separating the
spaces. Put one, and only one, check on each scale. Please
do not omit scales.

THIS SPEAKER

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Reliable : : : : : : Unreliable
Uninformed : . : : : Informed
Qualified : ° : : : : Unqualified
Unintelligent : : : : Intelligent
Valuable : . : : : Worthless
Inexpert : : : : : Expert
Honest : : . . : Dishonest
Unfriendly : : : : Friendly
Pleasant : : : . : : Unpleasant
Selfish : . . : . : Unselfish
Nice : : : . : Awful
Sinful : : : : : : Virtuous
Aggressive : : : : : : Meek
Timid : : : : : : Bold
Energetic : : : : : : Tired
Introverted : : : : : Extroverted

 

 

 

109

APPENDIX D, continued

DIRECTIONS: Read each of the following statements carefully.
If you strongly approve of the statement as it stands, enter
the number _H_ in the space provided at the left of the
sentence. If you strongly disapprove of the statement then
enter the number 1 at the left of the sentence, and so on,
with regard to th5_5ther attitudes (approve, undecided,
disapprove).

STRONGLY STRONGLY
APPROVE APPROVE UNDECIDED DISAPPROVE DISAPPROVE
(5) (4) (3) (2) (l)

1. Quality in education can best be guaranteed by local
school boards.

2. Public schools should be financed primarily by federal
taxes.

3. Equality of opportunity in education for all citizens
can be best guaranteed by federal control of education.

4. The curriculum of public schools should be determined
by local school boards.

5. The federal government should have no control of the
curriculum of public schools.

6. Public schools can be better administered by the federal
government.

7. Quality in education can best be guaranteed by the
federal government.

8. The curriculum of public schools should be determined
by the federal government.

9. Equality of opportunity in education for all citizens
can be best guaranteed by local control of education.

10. Public schools should be financed primarily by local taxes.

11. The federal government should assume complete financial
responsibility for education.

12. Public schools can be better administered by local school
boards than by the federal government.

13. The amount of teachers' salaries should be determined
by the federal government.

14. Teacher employment standards should be controlled by
local school boards.

15.

16.

17.

18.

19.

20.

21.

22.

110

APPENDIX D, continued

Public schools should not be primarily financed by
local taxes.

The amount of teachers' salaries should be determined
by local school boards.

The federal government should have no control of
teacher employment standards.

The federal government should substantially increase
its financial support of public education.

The federal government should not determine the amount
of teachers' salaries.

Teacher employment standards should be controlled by
the federal government.

Public schools should not be primarily financed by
federal taxes.

Local school boards should not determine the amount
of teachers' salaries.

111

APPENDIX D, continued

DIRECTIONS: In using the following scales please make your
judgments on the basis of what these measures mean to you.

Here is how to use the scales:

(Concept)

If you felt that the concept was in general extremely good,
you would place a check mark in space number 3. If quite
good (but not extremely good), in 2; if slightly good in 1;
if neither good nor bad, in 0; if slightly bad, in —1; if
quite bad, in —2, and if extremely bad, in —3.

 

 

The "0" or neutral space on the scale may also be used for
"I don't know" or "I don't think this scale applies" answers.

Be sure to put a check mark somewhere along each scale. Put
your check within the spaces, not on the lines separating
the spaces. Put one, and only one, check on each scale.
Please do not omit scales.

FEDERAL CONTROL OF EDUCATION

Harmful ___:___:___:___:___:___:___ Beneficial
Good ___;___:___:___:___:___:___ Bad
Wrong ___:___:___:___:___:___:___ Right
Fair ___:___:___:___:___:___:___ Unfair
Negative : : : : : : Positive

Wise : : : : : : Foolish

APPENDIX E

EXPERIMENTAL SPEECH

Almost two hundred years ago Thomas Jefferson told
the American people that if we expected to remain both
ignorant and free we were expecting what never was and
never will be. If it was true that man could not remain
ignorant and free in the 18th century, it is even truer
today. Thus it is not surprising that almost every
American will tell you that he is "all for the best
educational system possible." With this historical
support and apparently favorable modern attitude we could
be led to the assumption that the United States has the
best possible public educational system already in
operation. Before we accept this assumption as fact
we should determine just what are the criteria for the
best possible educational system for the 20th century and
how well our present system measures up to this ideal.

It will be my purpose this evening to do just that.

While there may be some disagreement on the order of
importance, most people concerned with our educational
system would suggest four criteria for the first class
program. First, the quality of the instruction must be
high. Second, there must be adequate finances available
to provide for all legitimate educational needs. Third,
the school system must provide equal opportunity for all

children in the nation. Finally, qualified people must

112

113

APPENDIX E, continued

control the operation of the system. Let's look at
these criteria of the ideal school system to see what
they really mean and how our present school system in
the United States meets or fails to meet them.

Probably the most difficult thing to define in
relation to education is quality. But I think we can
assume that whatever quality is, it will be present
if students take the right courses from well trained
teachers. Of course, who is to say what are the "right"
abilities and interests. Thus, different courses are
needed by different students in every school. The right
courses for some students are rigorous college preparatory
subjects, while for others vocational training courses
are what are most appropriate. The important thing in
assessing the quality of an educational system is
whether or not individual students, whatever school they
must attend, are able to study the courses that are right
for them. Unfortunately in many of our nation's schools,
students are not able to take the right courses--simply
because they aren't even offered. According to figures
released by the Department of Health, Education, and
Welfare of the Federal Government, over one—third of the
nation's high schools do not offer such essential college
preparatory subjects as chemistry or physics, and about I
the same number don't offer even one foreign language.
Only a small fraction of our public schools offer a broad
program of vocational education. From this we must

conclude that many of our students are not obtaining the

114

APPENDIX E, continued
quality education we desire.

But, for a moment, let us assume that every student
in the United States has the opportunity to take the
right courses. We still will be forced to conclude
that the quality of American education is not accept-
able because of that second characteristic of a quality
educational system that I mentioned a few moments ago-—
well trained teachers. I'm sure I don't have to tell
anyone about the tremendous shortage of adequately
trained teachers. Almost every state is presently
forced to accept substandard teachers. In their

report entitled The Financial Status of Public Schools

 

the Committee on Educational Finance of the National
Education Association reports that last year the nation
was short 118,000 qualified teachers just to meet
minimum standards. We can only guess what that figure
would be if we tried to eliminate all of the incompetent
teachers in the classrooms today and replace them with
thoroughly trained and qualified individuals. I would
suggest 500,000 as a very conservative starting figure.
But, whatever the figure is, since many needed courses
are not even offered students in many of our schools
and we face a serious shortage of competent teachers,
we must conclude that our present educational system falls
far short of our ideal of a quality educational system.
Now let us turn our attention to the criterion of

finance. We can't set down an exact figure and say this

115
APPENDIX E, continued

or that amount of money is adequate for a first class
educational system. No one is in a position to be that
exact. However, we can say that if our schools have
enough money to provide educationally acceptable physical
plants, to pay professional salaries to our teachers,
and to cover costs of operating expenses and equipment,
that could be called adequate finance. Unfortunately,
many of our school districts do not have that kind of
money.

Let us look first at physical plants. According
to figures released by the United States Office of
Education in January of this year, 25.4% of the nation's
classrooms are, in their words, "obsolete and unacceptable"
for public schools because of such things as extreme
fire hazards. Translating that percentage into numbers
of classrooms, we find that over 375,000 classrooms are
presently unacceptable. In terms of students, this
means that over nine million American children are
presently attending substandard schools, some of which
the U.S. Office of Education calls "fire hazards." But
one may ask, "Isn't this problem being overcome?"
Unfortunately it isn't in many areas. I needn't point
out that most schools are built by finances derived by
selling municipal bonds. These bond issues must be
voted on by the peOple in the communities involved. If
the bonds are voted down, the new school facilities are

not built. The U.S. Office of Education reports that

c.1‘ III-I’M-

 

116
APPENDIX E, continued

28% of the bond issues for such facilities were defeated
between 1957 and 1963. The figure rose to 31% last year.
It is apparent from these figures that not only are there
numerous school buildings in completely unacceptable
condition, but even in those communities where an attempt
is made to remedy the problem, over a fourth of the
attempts are unsuccessful.

And how about adequate financing to provide for
professional salaries for our teachers? Well, let's
look at the facts and then decide for ourselves. The
most recent national study of teachers' salaries was
released January 3, 1965, by the National Education
Association. This material was included in the NEA

Research Report 1964 R17. We find that the average

 

elementary school teacher in the United States earns just
over $6,000. The average in California is over $7,500.
However, in Mississippi and South Dakota it is only
$4,000. Figures for secondary school teachers are
similar. The national average is $6,500 with California
leading the country with an average of $8,400 and
Mississippi bringing up the rear with an average of

under $4,300. Probably many of us have enough information
already to draw a conclusion about the adequacy of
financing in some of our states like Mississippi and
South Dakota. But let's get away from state and

national "averages" and look at teachers salaries from

another perspective. Most of us know that the generally

117
APPENDIX E, continued

accepted income level under which people are considered
to be living in abject poverty is $3,000. Certainly a
professional educator should be expected to earn far
more than that. However, according to figures from

that same NEA Research Report, 1964 R17, 23% of the

 

teachers in South Dakota earn less than $3,500. Thirty-
one percent of the teachers in Arkansas earn less than
that figure. Then, of course, there is Mississippi. Only
14% of Mississippi's teachers earn less than $3,500.

This wouldn't seem too bad if it weren't for the fact

that 80% of Mississippi's teachers earn less than $4,500.
This compares with New York, Pennsylvania, Arizona,
California, Nevada, and Alaska which have pp teachers
making less than $4,500.

Now, don't misunderstand me. Some teachers make a
fairly good income, but their salaries just don't compare
with other occupations requiring a college degree. The
average college graduate, according to a survey conducted
by Elmer Roper and Associates, can expect to earn an
annual salary of about $7,500 after three years on the
job. According to figures reported in the NEA research
report I mentioned before, in three states, New York,
California, and Alaska, over 20% of the teachers earn
more than that figure. But on the other side of the
ledger, in 13 states less than 1% of the teachers receive
such a salary, seven of these states have pp teacher

making that amount.

118
APPENDIX E, continued

Thus, while some schools in some areas have
excellent financing, other schools in other areas are
fire hazards staffed by teachers receiving salaries
which force them to live in what our government calls
"abject poverty." I don't know what conclusion you
will draw from these facts, but I can only conclude
that the present financing of our public schools is
very inadequate.

But, we can not complete our evaluation of the
present school system in the United States without
considering the criterion of equality of opportunity
for all of our children. There are several things that
we must consider in determining whether equality of
opportunity is present, some of which I have already
mentioned. For a national educational system that offers
equal opportunity to all of its children, course
offerings must be somewhat similar across the country.

We have already seen that this isn't the case in American
education. Also for equality of opportunity to exist,

the teachers should be reasonably comparable from one

area to another. But it would be stretching the
imagination pretty far to suggest that Mississippi can

get as high a quality of teachers for $3,000 as California
can for $8,000. Finally, we must mention that many of

our children are still prohibited from achieving equality
of opportunity in education because of race. Governor

Johnson of Mississippi during the last election bragged

119
APPENDIX E, continued
that no school in the state of Mississippi was integrated.
There has been some improvement since then. Now only a
little over 99% are segregated.

Truly, equality of opportunity in American education
is nothing but a dream, a dream that will peyep come true
as long as the Johnsons, the Wallaces, and their kind
control education.

This brings us to the last criterion for an ideal
public educational system that I posited early in this
talk, that qualified people must be in control of the
system. Well, what is a qualified person? I would
suggest that three characteristics are essential. Such
a person should, among other things, be well educated,
he should understand the process of curriculum building,
and he should have a thorough understanding of modern
teaching procedures. Let's look at who is actually in
control of our schools. As we all know, the local school
board is in charge of our schools so we need to determine
whether these people are capable of properly running an
educational system.

First, we can consider what the requirements are
for a person to become a school board member. According
to Bulletin 1957—13 of the U.S. Office of Education,
entitled "Provisions Governing Membership on Local Boards
of Education," the picture is not encouraging. Not one

single state requires that a school board member know

120

APPENDIX E , continued
anything at all about education! In 26 states the
only requirement is that the person be a qualified
voter. Ten others have additional residence requirements.
Eleven require an eighth grade education. Four require
that the board member be a taxpayer or parent. And, one
state, Rhode Island, has no requirements at all.

From this we might suspect that our boards are
made up of people totally unqualified to run an educational
system. Such a suspicion is born out in fact. From
that same U.S. Office of Education report that I
mentioned a moment ago we find that the U.S.O.E. national
survey of school boards determined that 23.8% of the
nation's school boards include people who are not even
high school graduates. In the South the figure is
41%. People who haven't even finished high school are
telling our teachers not only what to teach but how to
teach it. Some people are concerned about the future
of high school dropouts. It seems that we have little
cause for concern. They will just grow up to be tomorrow's
school board members!

Of course, some school board members have finished
high school, so let's look at the occupations of school
board members in general. Again citing official U.S.
Office of Education figures, we find that 35% of the
school board members are business owners, officials, and
managers. Twenty-seven percent are in the professional

and technical services——doctors, lawyers, and engineers.

 

I.___¥

121

APPENDIX E, continued
Twelve percent are farmers, 9% are laborers and
craftsmen, 7% are housewives, and 7% are clerks. Did
you notice one group missing from that list? I did,
Educators! There were so few qualified educators that
were members of local school boards that the U.S. Office
of Education did not even report them as a separate
category.

I think it says something significant about our
nation's attitude toward education that we let just
anyone serve on our school boards. It is even state
law in most states that a man must be a licensed
veteranarian to take care of our sick dogs. But our
children's schools? Anyone is capable of taking care
of them. Well, I, for one, refuse to buy that attitude.
I think it is time that we make some drastic revisions
in our American educational system. The place to begin
is right at the heart of the present system, with the
people who are controlling the schools, the ones who
are responsible for the present deplorable state of
American education. These problems can not be overcome
by merely increasing federal aid to education as some
people suggest. Turning money over to states like
Mississippi and Alabama won't solve anything. Neither
will it be of any help to turn over federal money to
local school districts run by school boards composed of
school drop outs. The only answer is to do for our

children's schools what we have done for our dogs-~turn

122
APPENDIX E, continued

them over to trained, qualified, experts and provide
the money needed to properly educate American youth to
take its place in the space age.

More specifically, I suggest that it is time for
the Federal Government to assume ultimate control of the
educational system of the United States. This is not
to suggest that we turn the schools over to the federal
politicians, rather it is to take them away from the
lgggl politicians and turn them over to the educators.
The specific proposal that I recommend is very similar
to the one first suggested by Carl J. Megel, President
of the American Federation of Teachers, in testimony
before the Committee on Education and Labor of the U.S.
House of Representatives when that committee was
considering the late President Kennedy's program for
public education in 1962. This program has three points:

First, it should be established by law that 10% of
all future Federal budgets be devoted to American public
school education. Along with this law, provision should
be made for establishing an absolute priority for
education before all other expenditures of the government.
California now has such a provision and leads the nation
in almost every area of education.

Second, the Federal Government should assume all
present debts of public schools. This would equalize
the program so that communities that have gone into

debt to build present schools would not be penalized for

 

123
APPENDIX E, continued

that actiOn.

Finally, a national council on education should be
formed composed of members of the National Education
Association and the Education Associations of the
fifty states. This council would serve as advisors
to the U.S. Office of Education which would be exclusively
empowered to dispense all funds for education in the
United States. This would not only guarantee
standardization of the educational system across the
U.S. but would also guarantee that the special needs
of the state and locality would be served by that
state's education association representatives. In short,
the ultimate control of education would be in the hands
of the Federal Government, but the operation of the
schools would be left to professional educators hired
on the state and local level.

What would be the effect of this program? Well,
let's turn to the criteria for an ideal school system
that I mentioned a while ago to see how well this
program would stand up. First, we said that the
quality of education must be high. Since under the
program I have recommended our schools would all be
part of one standardized system, each student would have
the opportunity to take the courses most suited to his
needs, wherever he lives. Since there would be no

shortage of funds, top flight people would be drawn

 

124

APPENDIX E, continued
into the teaching profession by truly adequate
professional salaries. As we noted before, if a
student takes the right courses from qualified teachers
we have what can only be described as quality education.
Second, we said that a school system should have adequate
finances available to provide for all legitimate education-
al needs. If the budget and priority for education that
I have recommended is adopted, no educational decisions
will be dependent on financial considerations. The only
important thing will be "Is it needed?" If salaries are
too low, they will be increased. If a classroom building
is a fire hazard, it will be replaced. If a teacher
needs a slide projector or a tape recorder for her class,
it will be provided.

It is important to note one more thing in this
regard. Today, in most communities when the school
budget is increased the property owners are forced to
pay most of the bill through property taxes. These taxes
are excessively high already in most areas and can't be
expected to be increased much more. People on stable
incomes just can't pay these exorbitant taxes. The
people hurt most are the retired peOple and widows who
own their homes but have little or no income. Under the
program I have suggested this oppressive form of taxation
would not be needed and so could be abolished. All funds
for education would come from Federal taxes which are

based on a person's ability to pay, not on where he

125
APPENDIX E, continued

lives or what he owns.

Thus, under this system, the financial needs of
education would be met and at the same time an oppressive
tax would be removed and replaced by the most democratic
type of tax system. Certainly under such a program, we
can say that the criterion of "adequate" financing will
be met.

Our third criterion was that the school system
must provide equal opportunity for all children in the
nation. Under the program I have suggested the facilities
would be equal, the teachers would be relatively equal,
and the course offerings would be equal. But most
significantly, pply under a program such as I offer
can we ever hope to have racial equality in education
in many parts of the nation. It should be abundantly
clear to anyone who is concerned enough to look at the
situation in the South that under state and local control
of education, Negro children will never be truly equal.

Finally, we said that our school system should be
controlled by qualified individuals. Such people should
be well educated, have a thorough knowledge of curriculum
building, and an understanding of modern teaching
procedures. The only people who have these characteristics
are professional educators, precisely those people who
would be administering our public school system under my

program.

 

126
APPENDIX E, continued

But what are the possible objections to this new
program for American education that I have suggested?
The most obvious objection is that it costs a lot of
money. It certainly does. To make up for the neglect
of our schools over the past fifty years is bound to be
expensive. But any country that can afford $40 billion
to put a man on the moon can certainly afford to educate
its children.

Of course, the other objection is that this program
is socialistic. It certainly is. Public schools by
definition are socialistic institutions. The only
question is whether we want this socialistic institution
controlled at the local level by high school dropouts,
the state level by men like Governors Wallace of Alabama
and Johnson of Mississippi, or at the national level by
professional educators. To me that choice is simple. I
think it is to most thinking Americans, no matter what
their political persuasion.

In the final determination we have to decide whether
we, the richest nation on the face of the Earth, wish to
have an educational system capable of meeting the needs
of our youth, or whether we are going to continue to
sacrifice our children's future on the Altar of

irresponsibility.

 

APPENDIX F

COMPARISON OF SPEAKER'S AND ARs'

AUDIENCE RATINGS

Reliable
Informed
Qualified
Intelligent
Valuable
Expert
Honest
Friendly
Pleasant
Unselfish
Nice
Virtuous
Aggressive
Bold
Energetic
Extroverted
Active
Decisive
Experienced
Expert
Trained
Competent
Just

Kind
Admirable

Note:
ARs'

This Audience

Speaker's ratings:

ratings:

solid line
broken line

127

(POSITIVE TREATMENT)

Unreliable
Uninformed
Unqualified
Unintelligent

Worthless

Inexpert
Dishonest
Unfriendly

Unpleasant

_ Selfish

Awful
Sinful

Meek

Timid

Tired
Introverted
Passive
Indecisive
Inexperienced
Ignorant
Untrained
Incompetent
Unjust
Cruel

Contemptible

 

Reliable
Informed
Qualified
Intelligent
Valuable

Expert
Honest
Friendly
Pleasant
Unselfish
Nice
Virtuous
Aggressive
Bold
Energetic
Extroverted
Active
Decisive
Experienced
Expert
Trained

Competent

Admirable

Note:
ARs'

Speaker's ratings:

128

APPENDIX F

COMPARISON OF SPEAKER'S AND ARs'

This Audience

ratings:

 

solid line
broken line

AUDIENCE RATINGS ("FREE" TREATMENT)

Unreliable
Uninformed

Unqualified

Unintelligent

Worthless
Inexpert

Dishonest

Unfriendly

Unpleasant
Selfish
Awful
Sinful

Meek

Timid

Tired
Introverted
Passive
Indecisive
Inexperienced
Ignorant
Untrained
Incompetent
Unjust
Cruel

Contemptible

 

129

APPENDIX F

COMPARISON OF SPEAKER'S AND ARs'
AUDIENCE RATINGS (NEGATIVE TREATMENT)

This Audience

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

   

 

Reliable : : : : : : Unreliable
Informed : : ‘J,’”1;,_H——_H—__ Uninformed
Qualified Unqualified
Intelligent Unintelligent
Valuable Worthless
Expert Inexpert-

Honest Dishonest
Friendly Unfriendly
Pleasant Unpleasant

Unselfish Selfish
Nice Awful
Virtuous Sinful
Aggressive Meek
Bold Timid
Energetic Tired
Extroverted Introverted
Active Passive
Decisive Indecisive
Experienced Inexperienced
Expert Ignorant
Trained Untrained
Competent Incompetent
Just Unjust
Kind : : : : __r___ Cruel
Admirable : : : : : T‘s: Contemptible

 

 

Note: Speaker's ratings: solid line
ARs' ratings: broken line

 

lulllll

083 06

llllllll

03

II"
E "l'
H
T H

3 1293

lllllllll