THE EFFECTS OF PERCEIVED AUDIENCE
RESPONSE 0N SPEAKER ATTITUDES

Thesis for the'Degree of Ph. D.
MICHIGAN STATE UNIVERSITY
JAMES CARL GARDINER
1969

 

 

LIBRA R 1’

Michigan State
University

 

This is to certify that the

thesis entitled
THE EFFECTS OF PERCEIVED AUDIENCE

RESPONSE ON SPEAKER ATTITUDES

presented by

James Carl Gardiner

has been accepted towards fulfillment
of the requirements for

Ph. D degree in Speech

é” //
Emmifiré/

 

 

 

Date April 28, 1969

 

0-169

  
  

amomo av

"DAG & SUNS'
800K BINDERY INC

.ltn-.-_-

   
   

 

WM E W

 

 

ABSTRACT

THE EFFECTS OF PERCEIVED AUDIENCE

RESPONSE ON SPEAKER ATTITUDES

by

James Carl Gardiner

Purpose: This thesis was concerned with: a)
investigating the effects of Expected Audience Response
(EAR) and Perceived Audience Response (PAR) on speaker
attitudes, b) testing the predictability_of cognitive
balance theory in a communication feedback setting, and
c) providing a comprehensive review of the experimental
literature on feedback.

Hypotheses: a) There will be an interaction

 

between expected and perceived audience response on

audience and performance ratings by the speaker. b)

Speakers who perceive positive audience response will

rate the audience and their own performances higher

than speakers who perceive negative audience response.
Design: The hypotheses were tested in a 2x2 ex-

perimental design. Speakers were first given an expec-

tation of the response they might receive from the

James Carl Gardiner

audience (positive or negative). As they delivered per-
suasive messages, they were administered either positive
or negative nonverbal responses by 13 trained audience
members. After the experimental induction, the speakers'
attitudes toward the audience and toward their perform-
ances were measured with semantic differential scales.
The results were submitted to a 2x2 factorial analysis
of variance.

Results: a) There was a significant interaction
between EAR and PAR on performance ratings by the speakers:
speakers in the negative EAR, negative PAR condition
rated their performances significantly higher than
speakers in the positive EAR, negative PAR condition.

b) There was no significant interaction between EAR and
PAR on audience ratings by the speakers. c) Speakers

who perceived positive audience response rated the audi-
ence and their own performances significantly higher than
speakers who perceived negative audience response.

Discussion: Balance theory was generally success-

 

ful in predicting attitudinal outcomes in a communication
feedback setting. Speakers who expected positive re-
sponse but perceived negative response restored balance
by devaluing their performances. Speakers who expected
negative response but perceived positive response re-

stored balance by rating the audience positively. It

James Carl Gardiner

was concluded that speakers must realize the impact that
audience response may have on their attitudes and be pre-
pared to cope with this eventuality. The communication
receiver must also understand the potential negative or
positive effects his responses can have on the communi-

cation source.

THE EFFECTS OF PERCEIVED AUDIENCE

RESPONSE ON SPEAKER ATTITUDES

BY

James Carl Gardiner

A THESIS

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

DOCTOR OF PHILOSOPHY

Department of Speech and Theatre

1969

/"
'3 ,.
I

7——

 

ACKNOWLEDGMENTS

The writer wishes to thank the following in-
dividuals for their contributions to this project:
first, Dr. James McCroskey for his able and efficient
guidance; second, Drs. Brad Lashbrook, Lawrence Messé
Gerald Miller, and David Ralph for their help in formu-
lating the research ideas; third, Dr. Jack Bain for
allowing the Communication 305 students to participate
as Ss; and finally, "the group": Carole, Chris, Joan,
Josie, Madie, Marc, Mike, Mindy, Nancy, SAM, Sarah, Sue,
and Tom, whose response-administering abilities were

superb.

ii

 

TABLE OF CONTENTS

Chapter

I.

II.

III.

IV.

INTRODUCTION . . . . . . . . . . . . . .
Feedback Conceptualized . . . . . . . .
A Model of Communication Feedback . . .

A SUMMARY OF EXPERIMENTAL LITERATURE ON
FEEDBACK O O O I O O O O O O O O O O O 0

Theoretical Bases . . . . . . . . . . .
Methodology . . . . . . . . . . . . . .
IVs studied with Feedback . . . . . . .
DVs Measured . . . . . . . . . . . . . .
Methodological Problems . . . . . . . .
Results found in the Feedback Studies

Conclusions about Feedback . . . . . . .
THEORY AND HYPOTHESES . . . . . . . . .
Balance Theory Predictions . . . . . . .
Hypotheses . . . . . . . . . . . . . . .
METHOD . . . . . . . . . . . . . . . . .
Experimental Design . . . . . . . . . .
Subjects and Confederates . . . . . . .
Procedures . . . . . . . . . . . . . . .
Measurement . . . . . . . . . . . . . .

Statistical Inference . . . . . . . .

iii

Page

11
12
13
17
18
21
25
42
44
47
54
56
56
60
63
65

69

 

Chapter
V. RESULTS . . . . . . . . . . . . . . . . . .
Manipulation Checks . . . . . . . . . . . .
Factor Analysis . . . . . . . . . . . . . .
Test for Homogeneity of Variance . . . . . .
Results of the Performance DV . . . . . . .
Results of the Audience Credibility DV . . .
Supplementary Analysis: SEX X PAR . . . . .
Summary of the Results . . . . . . . . . . .
VI. DISCUSSION . . . . . . . . . . . . . . . .
Balance Theory . . . . . . . . . . . . . . .
Applications of Feedback . . . . . . . . . .
Limitations and Future Research . . . . . .
REFERENCES . . . . . . . . . . . . . . . . . . . .

APPENDIX A: A Methodological Summary of Research
Studies on Feedback . . . . . . . . . . . .

APPENDIX B: Response-Administering Assignments
for the Trained Audience . . . . . . . . . .

APPENDIX C: Persuasive Assignment:
Communication 305 .. . . . . . . . . . . .

APPENDIX D: Audience Profile . . . . . . . . . .
APPENDIX E: Dependent Measures . . . . . . . . .

APPENDIX F: Statistical Tables . . . . . . . . .

iv

Page

75
75
77
78
79
81
84
84
86
86
91
93
97

108

119

121
123
125
129

 

LIST OF TABLES

Table Page

1. Factor Analysis of Audience Ratings . . . . 129

2. Factor Analysis of Performance Ratings . . 129

 

3. Results of Bartlett's Test for Homogeneity
of Variance . . . . . . . . . . . . . . . . 130

4. Tests for Significance of Difference
(PAR) O O O O O O I O O O O O O I I O I O I 130

5. Perceived Audience ReSponse: Group Means,
Standard Deviations, Size . . . . . . . . . 131

6. Analysis of Variance Summary for
Character Ratings (PAR) . . . . . . . . . . 132

7. Analysis of Variance Summary for
Authoritativeness Ratings (PAR) . . . . . . 132

8. Analysis of Variance Summary for

Dynamism Ratings (PAR) . . . . . . . . . . 132
9. Analysis of Variance Summary for

Performance Ratings (PAR) . . . . . . . . . 133
10. Analysis of Variance Summary for

Character Ratings (SEX) . . . . . . . . . . 133
11. Analysis of Variance Summary for

Authoritativeness Ratings (SEX) . . . . . . 133
12. Analysis of Variance Summary for

Dynamism Ratings (SEX) . . . . . . . . . . 134
l3. Analysis of Variance Summary for 134

Performance Ratings (SEX) . . . . . . . . .

Table

14.

15.

16.

17.

Analysis of Variance Summary for
Character Ratings (OAR) . . . .

Analysis of Variance Summary for
Authoritativeness Ratings (OAR)

Analysis of Variance Summary for
Dynamism Ratings (OAR) . . . .

Analysis of Variance Summary for
Performance Ratings (OAR) . . .

vi

Page

134

135

135

135

9 l 1)!) 3 v W‘A‘A‘k‘t‘x‘ ne'-

 

LIST OF FIGURES

 

Figure Page

1. A Model of Communication Feedback . . . . . . 8

2. Balance Theory Predictions . . . . . . . . . 50
\rii

CHAPTER I
INTRODUCTION

Communication theorists in the past two decades
have begun to place more emphasis on feedback as a vari-
able in the communication process. Beginning with the
Leavitt and Mueller study in 1951, over 50 experimental
studies have investigated the effects of receiver re-
sponse on the attitudes and behavior of communication
sources.

This study was concerned with three major pur—
poses: a) investigating the effects of Expected Audi—
ence Response (EAR) and Perceived Audience Response (PAR)
on speaker attitudes, b) testing the predictability of
cognitive balance theory in a communication feedback
setting, and c) providing a comprehensive review of the
experimental literature on feedback.

Answers to the above questions will be presented
by discussing the following: 1) The theoretical liter-
ature on feedback. 2) A model of communication feed-
back. 3) The experimental literature on feedback. 4)

Balance theory as it leads to the hypotheses of this

study. 5) The methodology used to test the hypotheses.

 

6) The results of the experimental study. 7) The impli-

cations and limitations of the results.

Feedback Conceptualized

 

Integration of physical and human components at the
technical level has led, in recent years, to anal-
ogies which compare the theory of machines and human
behavior. In human communication research . . . use-
ful machine-human analogies have been applied to the
feedback function. Still further applications might
be profitably encouraged for human communication re-
search. (Johnson and Klare, 1962)

The purpose of this section is to develop a com-
munication feedback model that is analogous to feedback
as developed by cybernetics, consistent with the process
View of communication, and heuristic in nature. A
selected portion of the theoretical literature from
speech, communication, and cybernetics will be drawn
upon to accomplish this purpose.

Perhaps the earliest modern theorist to deal with
a phenomenon called "circular response" was Overstreet
(1925). He described the sensitive speaker as one "who is
so mindful of what is happening in his audience-~he gets
it from their facial expression, their nods, their
blankness, their scowls, their interrupting questions--
that his own thought and expression are influenced (p.
'77) o it

More recently, Johnson and Klare (1962) note that

feedback, as applied to communication theory, is an

 

analogy that compares the process of receiver response
and consequent source adaptation to feedback as applied
in mechanical engineering. Before proceeding further,
a brief explanation of feedback as it has found its way
from control engineers to human behaviorists via cyber-
netics is necessary.

Cybernetics, as developed by Norbert Wiener
(1954), deals with "techniques that. . .enable us. . .to
attack the problem of control and communication in gen-
eral." Wiener states his main thesis as follows: "The
physical functioning of the living individual and the
operation of some of the newer communication machines
are precisely parallel in their analogous attempts to
control entropy through feedback." He defines feedback
as “a method of controlling a system by reinserting into
it the results of its past performance (p. 61)."

Feedback, in its pure mechanical sense, is a
regulating function for a mechanical system. When the
system departs from its course, negative feedback re-
stores it to its intended pattern of behavior. An ex-
ample of this is the "cruise control" option available
on most modern automobiles. The driver simply selects
a desired speed, sets a control at that speed, and allows
the automobile to automatically regulate the speed.
When the automobile encounters a gravitational disrup-

tion such as a hill, the governing mechanism informs the

 

acceleration system to increase the power. The auto-
mobile is able to maintain a constant speed because of
its feedback system. Wiener calls this function "nega—
tive" feedback, since the feedback function operates
only when the system departs from its designated pattern
of operation. Consequently, there is no need for a
”positive" function in feedback. Johnson and Klare
(1962) point out that "the engineer works, for the most
part, with negative feedback for controlling servosys—
tems, and tries to keep positive feedback from disrupt-
ing the system.fi

With the above ideas in mind, it is difficult to
apply the feedback principle, in its true mechanical
sense, to communication behavior. In the communication
setting, the feedback mechanism possesses a positive—
negative dimension, whereas the cybernetic concept of
feedback does not. Recognizing this, Wiener (1954) ex-
panded feedback to include systems that are not as
tightly regulated as machines. "Feedback may be as simple
as that of the common reflex, or it may be a higher order
feedback, in which past experience is used not only to
regulate specific movements, but also whole policies of
behavior (p. 32)." He goes a step further by stating
that "if. . .the information which proceeds backward from

the performance is able to change the general method and

 

 

 

pattern of performance, we have a process which may well
be called learning (p. 61)."

While Wiener has expanded the concept of feed-
back to include higher order policy-regulating functions
as well as simple corrective functions, he has not aban-
doned the idea that feedback adjusts future conduct
through the influence of past performance. Many contem-
porary communication theorists, however, have weakened
the feedback analogy by diluting the essential function
of the feedback system. Johnson and Klare (1962) point
out that some of the more recent communication litera-
ture does not follow the basic principles of cybernetic
feedback.

An example of the above is Bettinghaus' (1968)
discussion of feedback. He defines feedback as "any
information that the source gains from his receivers
about the probable reception of his message (p. 207)."
While this definition is correct in describing the re-
ceiver response portion of the feedback process, it
fails to include the corrective process that results
from the receiver's response.

Berlo (1960) deals with feedback differently
than Bettinghaus by stressing that "feedback provides
the source with information concerning his success in
accomplishing his objective. In doing this, it exerts

control over future messages which the source encodes

(pp. 111-112)." While discussing the shortcomings of
feedback as a concept, Berlo warns against taking a
source-oriented viewpoint, rather than a receiver-
process-oriented view. The communicator, he stresses,
must not concentrate too heavily on "getting feedback"
and thus ignore the interdependence that the concept of

feedback correctly implies.

5:
a .

McCroskey (1968) also defines feedback as a
message from a receiver to a source. He emphasizes,
however, the role the source must play in correctly
interpreting the receiver's reaction, and adjusting his
subsequent communicative behavior according to that
interpretation.

The intent at this point is not to play semantic
games or quibble with terminology. Rather it is to
point out that the cybernetic formulation of feedback
is a broader concept that mere audience response. The
only danger in using the term "feedback" synonymously
with "receiver response" is that the reader may miss
the process implications of the concept of feedback.
This writer, therefore, shall employ the term feedback

only to refer to the total response-adjustment cycle.

A Model of Communication Feedback
Definition of Feedback. Feedback is defined as

an adaptive process in which the source adjusts his

input (attitudes, purpose, strategy, message and perform-
ance) as a result of output (observed response) from a
receiver.

The above definition and subsequent feedback
model, shown in Figure l, are in keeping with the re-
quirements stated at the outset of this section. First,
the model is analogous to feedback as discussed by
Wiener. It features the function of feedback in changing
future communication behavior. The model is consistent
with what Wiener calls higher order feedback, or "feed—
back of policies."

Second, the model is consistent with the process
view of communication. It is dynamic, in that it
stresses continuous changes in the source, receiver,
purpose, strategy, message, and performance.

Third, the model is heuristic. It features
variables and relationships that are operable and ex-
perimentally testable. For example, it is possible to
control receiver output and observe the subsequent ef-
fect on such source variables as attitude toward the
receiver or change in communication strategy.

While many of the features of the model are
self-explanatory, a number of aspects need special con-
sideration. The primary elements of the model are as

follows:

 

S O U R C E
Attitudes: ‘\\\
Toward Topic \\\
Toward Receiver ‘*\
Toward Performance \oooooooooooooooooooo0000000..-
Communicative Purpose ADJUSTMENT ADJUSTMENT’

/.......OIOOOOOOOC.OOIO 0......

Communicative Strategy//
Performance Quality ,.

/
Message C3ntent_//

 

Input / Input' I

CHANNELS '
\\ Output \\ Output'

 

 

Attitudes : \ \

Toward Topic \‘\\

Toward Source \\................................

Toward Performance ADJUSTMENT ADJUSTMENT'
Interest in Topic //................................
Knowledge Level’///

*n—u—

 

 

Figure l. A Model of Communication Feedback

l. The source is the purpose-generating ele-
ment of the model. He initiates the communication pro-
cess, and is responsible for the inputs. The source
develops the communicative purpose, the strategy, and
the message content. He possesses attitudes toward his
topic, toward his ability as a communicator, and toward
the receiver of his message.

2. The receiver is the recipient of the communi-
cative purpose. He initiates the feedback output. Im-
portant to the model are the various attitudes the re—
ceiver possesses. These attitudes are directed toward
the source (credibility), as well as toward the topic
of the communication. Two other important variables
the receiver adds to the communication model are inter-
est in the topic and level of knowledge about the topic.

3. Adjustment is the primary feature of the

 

model. It stresses the adaptation made by both the
source and the receiver as a result of the inputs and
outputs of the communication event. The source can ad-
just his attitudes, his communicative purpose and strat—
egy, the quality of his performance, or his message con-
tent. The receiver can also adjust his attitudes, as
well as his interest and knowledge of the topic. Ad—
justment is a continuous, process variable.

4. The input-output feature of the model re-

presents the messages sent in the communication process.

10

The source encodes a message (input) and sends it to the
receiver. The receiver adjusts to the message, then
reacts by returning a message (output). Perceiving the
output, the source adjusts his future inputs as a result
of the receiver's reaction. Input—output is also a con~
tinuous variable. The source does not normally give in-
put, stOp, wait for output, adjust, give more input,
etc. The source may be continually giving input, con-
tinually receiving output, continually adjusting, etc.
Also, the receiver may be continually receiving input,

adjusting, and giving output.

5. The channels shown in the model are the
various media through which the input-output messages
are transmitted. Three dimensions of the channel con-
cept are important to this discussion of the feedback
process. These dimensions are time (immediate vs. de—
layed), proximity (face-to-face vs. interposed) and
manner (verbal vs. nonverbal vs. mixed).

In summary, this chapter has briefly reviewed a
selected portion of the theoretical literature on feed-
back. It has proposed a theoretical model of feedback
behavior that is consistent with cybernetic definitions
of feedback, in keeping with the process view of communi-

cation, and heuristic in nature.

 

CHAPTER II
A SUMMARY OF EXPERIMENTAL LITERATURE ON FEEDBACK

The purpose of this chapter is to provide a
major summary of the experimental studies dealing with
communication feedback. Two types of studies have been
selected for this review. First, all studies which
focus on communication receiver response and its effect
on the communication source have been included. Second,
the reviewer chose for consideration several studies
from the areas of verbal conditioning, small group inter—
action, psychology of success and failure, and task
performance. Each of the studies chosen has a direct
bearing on feedback as it relates to communication be—
havior.

This review of experimental literature on feed-
back will be divided into the following seven sections:
1) The theoretical bases used to study the effect of
feedback on human behavior. 2) The methodologies used
in the studies, including the response agents employed,
the recipients of the responses, the channels through
which the responses were administered, and the experi-

mental settings. 3) The independent variables (IVs)

11

 

12

studied in connection with feedback. 4) The dependent
variables (DVs) measured in the studies. 5) The methodo-
logical problems observed by the writer. 6) The results
of the studies reviewed. 7) The conclusions that can be
drawn from the results.

For convenient reference, the first four sections

mentioned above, the theory, methodology, independent

.n- I
' .

variables, and dependent variables, are summarized in
greater detail in Appendix A. The 61 studies reviewed

are arranged in chronological order.
Theoretical Bases

Thirty—eight of the 61 studies reviewed reported
using a theoretical basis for predicting the experimental
outcome. Of these 38, 28 employed learning-reinforce-
ment theory as a model (e.g., Greenspoon, l955, or Ver-
planck, 1955). Ten studies reported using a particular
type of consistency theory as a basis. Five used coqni-
tive dissonance theory (e.g., Wallace, 1966), four
employed cognitive consistency theory (e.g., Harvey,
Kelley, and Shapiro, 1957), and one study reported using
cognitive balance theory (Engbretson, 1964). The par-

ticular studies using each of these theories are noted

in Appendix A.

13

Methodology

 

The Response-Administering Agents

Response in the feedback studies was administered

by two types of agents: a single experimenter, or an

-_
u
s

audience.

Experimenter as response agent. The response-
administering agent appeared as experimenter in 25 of
the studies reviewed (e.g., Lucas, 1963). There were
several variations of this role, however. In a number
of studies the experimenter posed as an interviewer
(e.g., Gergen, 1965). In others he appeared as a class
instructor (e.g., Cameron, 1963). In the Verplanck
(1955) study, the experimenter was a participant in a
casual conversation.

Audience as response agent. The remaining 36
studies introduced the response-administering agent as
an audience containing one or more individuals. This
type of agent was varied four ways:

First, four studies used fictitious audiences.
The responses were controlled by the experimenter. This
type of audience was used in the Berger (1968) study,
where §S wrote essays that were read by a "group of stu-
dents down the hall," and judged as persuasive or not

persuasive. The response of that "group" was manufac-

tured by the experimenter.

14

Second, 17 studies employed live audiences that
were trained to administer a predetermined response.
The Hylton (1968) experiment involved an audience trained
by §,to give either positive or negative responses,
according to the assigned condition.

Third, eight studies administered response *

‘—

through a live, untrained audience, with the response
manipulated by the experimenter. This technique was
used by Scott (1957), who staged debates in which the
audience voted on the winners. The experimenter then
"rigged" the results according to pre-assigned win or
lose conditions.

Fourth, seven experimenters used an untrained
audience in order to utilize free, natural feedback.
An example of this is the Leavitt and Mueller (1951) study,-.
which observed the effects of free feedback on the
effectiveness of message transmission.

The Response Recipients

This section will categorize the experimental
83 by the roles they played in the various feedback

studies. The largest category of response recipients

was a source encoding a continuous message. This

category involved 29 studies. Twenty-five of these were

oral encoding (e.g., Ring and Kelley, 1963) and four

were written (e.g., Bostrom, Vlandis, and Rosenbaum,,1951).

In 12 studies the Ss functioned as members of a group,

15

either in a social or task situation (e.g., Smith and
Knight, 1959). Eleven experiments involved gs as inter—
viewees (e.g., Insko, 1965), four involved individuals
performing a task (e.g., Frye, 1966), three involved
classroom teachers (e.g., McNeil, 1962), and two used T
participants in a casual conversation (e.g., Cieutat,

i “ "'T'WTEV

 

The Response Channels

The channels used to administer the responses
can be separated into verbal and nonverbal.

Verbal channels. The majority of the studies

 

reviewed employed a verbal-oral channel to administer
the responses. An example is the Miller (1964) study,
which employed a single E’to reinforce each §fs be-
havior either positively ("good") or negatively ("no").
Another variation of the verbal channel is the verbal-
written. The Videbeck and Bates (1966) study utilized
a computer to administer a written response to gs.
Finally, in three cases the telephone was used as a
verbal-interposed channel (e.g., Insko, 1965).

Nonverbal channels. Twelve studies used non—

 

verbal channels in a face-to-face situation. An example
of this is the Sarbin and Allen (1964) study, where a
trained audience administered nonverbal cues to speakers

delivering a belief-discrepant message.

16

In seven studies, nonverbal response was admin-
istered through interposed channels. Lott, Schopler,
and Gibb (1955) presented task progress reports in graph
form to gs. Amato and Ostermeier (1967) and Combs and
Miller (1968) trained audience members to signal respon-
ses with colored cards. The "General Electric Opinion
Meter" was employed by Stolz and Tannenbaum (1963),
and various light arrangements were used to signal audi-
ence reactions in studies by Huenergardt (1967a), Cahn

(1966), and McNeil (1962).

The Experimental Settings

 

Eight different experimental settings were used
for administering feedback responses. In 24 studies,
a public speaking situation was used (e.g., Barwind,
1969). Twelve experiments placed the Se in a small
group setting (e.g., Molof, 1963). Eleven studies used
interviews to administer the responses (e.g., Krasmer,
Knowles, and Ullman, 1965). Four studies involved in-
dividual task performance (e.g., Hill, 1965), four in-
volved essay writing (e.g., Berger, 1968), three were
conducted in a classroom atmosphere (e.g., Leavitt and
Mueller), two were staged as casual conversations (e.g.,
Verplanck, 1955), and one was an oral examination

(Stolz and Tannenbaum, 1963).

i - use Jug-1*." T?

17

Independent Variables Studied with Feedback

Two classes of independent variables were studied
in connection with feedback. First, a number of cate-
gorized variables were measured and Se were divided
into groups according to certain observed characteris-
tics. Second, a number of manipulated variables were
used, in which gs were assigned to different types of
experimental treatments. The discussion below will
list the categorized and manipulated variables used in

the experimental feedback studies.

Categorized Variables

 

The category variables studied with feedback
were test anxiety (Hill, 1965), number Of years of
teaching experience (Tuckman and Oliver, 1968), type of
motivation: self, feeling, task, interaction (French,
1965; Frye, 1966; Gibb, Smith, and Roberts, 1955;
Lott, gt. 31., 1955), personality types (Lucas, 1963),
need for affiliation (Mikawa, 1963), need for influence
(Berger, 1968), need for social influence (Sarbin and
Allen, 1964), familiarity with the individual adminis-
tering the response (Harvey, at 31., 1957). attitude
toward the topic (Mattox, 1967), self—esteem (Marshall,
1958), sex of the experimenter (Cieutat, 1962), skill

level of the speakers (Barwind, 1969).

.‘W 7:7! m-o-‘u-p—au...
a x T

18

Manipulated Variables

The manipulated IVs studied with feedback were
task responsibility treatments: individual, group, ob-
server, group leader (Lewin, 1968), order of argument
presentation: primacy, recency (Rosnow, 1966), exper-
tise of the audience (Stotland, 1956; Huenergardt, 1967a),

type of self-description given by g: accuracy hypocracy

(Jones, Gergen, and Davis, 1962; Gergen, 1965), type of
argument: same as belief, opposite to belief, off~neutral
Scott, 1959b), publicness of failure (Stotland, 1956),
intensionality of audience (welke, 1967), order of re-
inforcement: superior, inferior (Sereno, 1964), type

of feedback: continual vs. terminal (Molof, 1963),

type of feedback: personal vs. impersonal (Gergen,
1965), type of speech: informative, persuasive (Bostrom,
1963), effort in message preparation (Gardiner, 1968),
expectancy: high, low (Cohen, 1960), audience credi-
bility: high, low (Engbretson, 1964), expected audience
response (Goldstein and McGinnies, 1964), differential
reward to a confederate and the subject (Miller, Zavos,
Vlandis and Rosenbaum, 1961; Miller, 1964), and incen-

tive magnitude (Berger, 1968).
Dependent Variables Measured

Four distinct categories of DVs have been mea-

sured in the feedback studies: the performance of the

19

source, the attitudes of the SS receiving the response,
the attitudes of the communication receivers, and various
group measures. The following sections will expand each
of the four categories, citing examples of studies where

the various DVs have been measured.
Performance of the Source

The largest single category of DVs in the studies
reviewed measured various performance characteristics
of communication sources, including both delivery and
content measures. The delivery measures observed the
following: fluency (e.g., Miller, gt 31., 1961), utter-
ance rate (e.g., Miller, 1964), voice loudness (Stolz
and Tannenbaum, 1963), nervousness, eye contact, and
bodily movement (Amato and Ostermeier, 1967), delivery
ratings by a speech judge (e.g., Miller, gt_§l., 1961),
delivery ratings by an audience (e.g., Combs and Miller,
1968), and speaker estimates of speaking time (Miller,
at al., 1961).

The content measures dealt with changes in
speech content (Karns, 1964), ratings of content (Combs
and Miller, 1968), quantity of content (Cieutat, 1962),
ratings of language (Combs and Miller, 1968), frequency
of opinion statements (Verplanck, 1955), frequency of
plural nouns (Greenspoon, 1955), frequency of sentences

beginning with first person pronouns (Videbeck and Bates,

. I‘ ""|l\un 7'

ha...‘
"Jul.

20

1966), accuracy of message transmission (Leavitt and
Mueller, 1951), and ratings of overall effectiveness

(Combs and Miller, 1968).

ceiver (e.g., Gardiner, 1968), attitude toward self:

as a public speaker (e.g., Bostrom, 1963), self-credi-
bility (Engbretson, 1964), self-ratings (Harvey, gt 31.,
1957), self-evaluations (Stotland, 1956), self-hostility
(Mikawa, 1963), self-esteem (Frye, 1966), self-report
(Welke, 1967); concealment of opinion (Ring and Kelley,

1963), and attitude toward task performance (Engbretson,

Attitudes of the Receiver

 

The DVs dealing with receiver attitudes measured
the following: attitude toward the topic about which
the source communicated (e.g., Hylton, 1968) and atti-
tudes toward the source (e.g., Mehrley and Anderson,

1968).

Group Measures

fl

The various group measures observed in the

 

21

feedback studies were: defensive feelings (e.g., Gibb,

SE 31., 1955), task efficiency (e.g., Lott, gt 31.,

1955), group productivity (Smith and Knight, 1959),

group ratings (e.g., Molof, 1963), group hostility

(Mikawa, 1963), group attractiveness (Frye, 1966),

openness of group relationship (Lucas, 1963), and em- 1

pathy measures (Reddy, 1968). E
A complete summary of the dependent variables

used in the 61 studies reviewed can be found in

Appendix A.
Methodological Problems

The studies involving feedback have not, for
the most part, been hallmarks of methodological excel-
lence. Even before examining the methodological issues,
it was noted that over one-third of the studies reviewed
offered no theoretical basis for predicting the results.
Also, over one-fourth of the studies failed to state
hypotheses. Without a theoretical or predictive base,

it is difficult to establish sound methodological pro-

cedures for scientific investigation. Several of the
weaknesses that appeared in the methods will be dis-

cussed below.

Control of Feedback Response

One of the primary concerns in experimental

22

feedback research should be the manipulation and control
of receiver response. In a number of the studies reviewed,
the response was not well controlled. In several eX*
perimental designs where the response was administered
through verbal channels, the accompanying non—verbal
channels were not controlled. Consequently, nonverbal
cues may have confounded the results. Studies by Ver- f
planck (1955), Miller (1964) and Vlandis (1964) are ex-
amples in which response was administered to the gs
through verbal-oral channels without controlling the
non-verbal channels.

Conversely, a number of studies have controlled
the feedback response so tightly that the manipulation
may have lost its intended effect on the S8. An example
of this possibility is the Huenergardt study (1967a),
where the §_was told that 20 audience members were
listening over an intercom in the next room, and that
their responses would be signalled via a panel of 20
lights. While this setup eliminated all verbal cues
the gs may have doubted whether an actual audience was
present.

Probably the best control of feedback was dis-
played in the Videbeck and Bates (1966) study. In the
experimental setting, gs were reinforced by printing on
a computer console, which flashed "very good" each time

the desired response was produced.

23

The solution to the problem of adequate control
is to choose a response channel that gs will believe,
yet is capable of being controlled. A promising pos—
sibility is the isolated verbal-oral channel (e.g.,
Engbretson, 1964). The speaker can be briefly intro-
duced to his audience, then taken to an isolated room
where he is connected to the audience by an intercom.
He can speak to them and hear their responses through
an interposed verbal-oral channel, eliminating the non-
verbal confounding possibility. This procedure, how-
ever, may make generalization of the results more dif-

ficult.
Manipulation checks

A manipulation check on feedback is necessary
to confirm that each S has perceived the response
(positive, negative, neutral, task-oriented, feeling-
oriented, etc.) as the experimenter intended. Few
studies that manipulated feedback have reported such a
check. Examples of studies were careful manipulation
checks have been made and reported are those by Huener-

gardt (1967a), Karns (1964), and Lucas (1963).
Other Methodological Weaknesses

Other weaknesses appeared in the experimental
designs of the studies and the measurement techniques

employed. A gross example of design weakness is the

24

Amato and Ostermeier (1967) study, which employed a trained
audience to both administer the predetermined response
and rate the speakers on delivery. Combs and Miller
(1968) argue that Amato and Ostermeier's findings deserve
a "response role" interpretation, and cannot be considered
accurate audience ratings. Combs and Miller conducted
a study which provides support for their assertion. The
Combs and Miller criticism seems well founded, except
for two points: first, the Amato and Ostermeier data do
not bear out a "response role" pattern, and second, the
Combs and Miller study was enough unlike the Amato and
Ostermeier study to raise serious doubts about any com-
parisons made between the two studies.

Another example of design weakness can be found
in the Verplanck (1955) study, in which the feedback
channels were mixed. A number of interview situations
were face-to—face, verbal and nonverbal, while others
were interposed and verbal only. Yet, these different
conditions were not separated in the statistical analysis.

The Goldstein and McGinnies (1964) study provides

another example of faulty design control. Since gs in
this study both spoke to an audience, and discussed the
issue with them, it was impossible to determine whether
reinforcement was mediated by the Speaking experience or
the discussion experience.

Finally, weaknesses in measurement were observed

25

throughout the studies reviewed. In many studies, new
measuring techniques were developed. Yet, there were
few instances where the issues of reliability and valid-
ity were met.

The methodological problems in the communication
feedback research have centered around operationalizing
feedback, controlling the experimental design, and mea-
suring the effects of feedback manipulations. Before
these problems can be eliminated, researchers will have
to meet such issues as manipulation checks, design con-

trol, and reliability and validity of measurement.
Results Found in the Feedback Studies

Performance of the Resppnse Recipient: 'Delivery

Fluency. Of the 11 studies dealing with fluency,
seven report that feedback produced a significant effect
on fluency. Vlandis (1964) found that nonfluencies
significantly increased when punishment was introduced
in a speaking situation. Stolz and Tannenbaum (1963)
report that negative feedback produced a significant
increase in relative frequency of unfilled pauses.
Blubaugh (1968) also found that total nonfluency was
significantly greater with negative audience response
than with positive. Miller (1964) reports that when a
speaker received the same response as a speaker immedi—

ately preceding him, his nonfluencies differed

26

significantly from those of a speaker who received better
or worse responses than the speaker preceding him.
Miller, gt_a&, (1961) report that when a speaker per-
ceived that the speaker immediately preceding him was
rewarded, he produced significantly more nonfluencies 1
than if the speaker preceding him was not rewarded. In
a study by Sereno (1964), approval from a listener pro- f
duced a decrease in nonfluencies, while disapproval pro-
duced an increase. Amato and Ostermeier (1967) report
that fluency was the speaking characteristic demonstrat-
ing the greatest change due to unfavorable response.
Results contrary to those cited above are re-
ported by Karns (1964) who found that "aversive stimuli"
(negative responses) were not highly related to non—
fluency. Similarly, Mehrley and Anderson (1968) found
that a speech receiving negative reactions did not con-
tain significantly more nonfluencies than a speech re-
ceiving positive reactions. However, the fact that the
speakers in the Mehrley and Anderson study were experi-
enced academic debaters may account for this result.
Miller, §t_al. (1961) found no difference between nega—
tive and positive response when measuring total fluency
count means. In a finding somewhat related to audience
response, Welke (1967) reports that verbal nonfluency did
not increase as the degree of audience intensionality

(closeness) increased.

27

In summary, support is offered for the two op-
posite views that fluency is and is not affected by
audience response. However, a greater amount of evidence
supports the view that audience response does have an
influence on speaker fluency.

Utterance Rate. Studies measuring utterance

 

rate in connection with feedback generally report that

 

rate decreased with negative response and increased with

 

positive response. Welke (1967) found that speakers
spoke significantly slower as audience intensionality
increased. Vlandis (1964) reports that total utterance
was significantly decreased when the speaker received
negative treatment. Sereno (1964) found that giving a
Speaker less favorable reinforcement for a second speech
than for a first, produced a significant decrease in
total utterance during the second speech. He also re-
ports that approval produced a significantly higher
total utterance, and that superior treatment (when rein—
forcement for speech two was more favorable than rein-
forcement for speech one) produced a nonsignificant in-
crease in total utterance. Blubaugh (1968) found that
rate of speaking and verbal output were significantly
lower under negative audience conditions that under
positive conditions.

Two studies report that feedback had no signifi-

Cant effect on utterance rate. Miller, gt 31. (1961)

28

found that utterance rate means were not significantly
different in positive and negative conditions. Also,
Karns (1964) reports that utterance rate was not highly
related to aversive stimuli. A majority of the studies,
however, report that utterance rate was significantly
influenced by positive and negative audience response.

Time Measures. Stolz and Tannenbaum (1963)

 

found that negative feedback significantly impaired both
relative pause times and qualitative and quantitative
production rates. Testing for the Ss' ability to judge
the amount of time used in speaking, Miller, 33 31.
(1961) found that Speakers who received reward perceived
the time of the speech as significantly longer than
speakers who were not rewarded.

Voice Loudness. In the only study measuring

 

voice loudness in connection with feedback, Stolz and
Tannenbaum (1963) discovered that negative responses
exerted a negligible influence on loudness, whereas
positive responses produced a significant incarese in
loudness.

Stage Fright. Stolz and Tannenbaum (1963) found

 

that positive responses influenced stage fright only
slightly, negative feedback produced a pronounced dis-
ruptive effect.

Nervousness, Eye Contact, Bodi1y_Movement. Amato

and Ostermeier discovered that unfavorable audience

 

29

responses produced lower audience ratings of speaker
delivery, nervousness, eye contact, and bodily movement.
Delivery Ratings by a Speech Judge. In a study
by Mehrley and Anderson (1968), judges did not rate
speeches that received positive feedback significantly

better than speeches that received negative feedback. ‘1
Performance of the Response Recipient: Content

Changes in Content. Karns (1964) reports that

 

aversive stimuli produced noticable changes in speech
content. "Essential changes," he reports, occurred
significantly more often under aversive audience condi-
tions.

Accuracy of Message Transmission. Leavitt and

 

Mueller (1951), who tested communication effectiveness
in varying feedback conditions, found that when no feed—
back was allowed, message transmission was significantly
less accurate than when free feedback was allowed.

Frequency of Opinion Statements. After reinforc—

 

ing opinion statements in a conversation setting, Ver-
planck (1955) found an increase in the frequency of

opinion statements during positive reinforcement and a

 

reduced frequency of opinion statements during extinction.

Frequency of Plural Responses. Reinforcing only

 

plural responses in an experimental situation, Greenspoon

(1955) found that "mm-hm" increased the frequency of

30

plural responses, and "huh—uh" decreased the frequency

of plural responses.

Attitudes of the Communication Source

 

Attitude toward the Topic under Consideration.
Studies dealing with attitude toward the topic under con-

sideration can be divided into three categories: a)

‘.' .r- _ .lI-___‘-a 5
»

situations in which sources encoded belief-discrepant
messages, b) situations in which sources encoded belief—
congruent messages, and c) situations in which sources
responded to questionnaires.

Results from studies involving belief-discrepant
messages are in general agreement. Scott (1957, 1959b),
who studied attitude change in winners and losers of
counter-attitudinal debates, found that winners changed
their attitudes toward the side they debated more than
losers or controls. He also noted that the effects of
response reinforcement were preserved up to periods of
at least ten days. In another study by Scott (1959a),
gs who were rewarded for arguing a position contrary to
their initial opinions altered their attitudes in the

direction of the expressed arguments.

 

Bostrom, 33 31. (1961) report that 3s who re-
ceived "A's" for writing counterattitudinal essays changed
their attitudes significantly more than those who re-

ceiVed "DIS. ll

31

In a recent study by Wallace (1966) 3s were re-
warded for their performances after presenting counter-
attitudinal arguments. Results showed that 3s rewarded
for their performance showed greater attitude change
toward the belief-discrepant view than gs who were not
rewarded. !

Buckhout and Rosenberg (1966) found that 3s who

 

were given positive reinforcement for agreeing with
belief-discrepant statements and negative reinforcement
for agreeing with belief-congruent statements, has sig-
nificantly more affective and cognitive changes than
gs who were not reinforced, or SS who were only positively
reinforced for agreeing with belief-discrepant statements.

Goldstein and McGinnies (1964), who studied the
effects of reinforcement on speakers who delivered
belief-discrepant speeches to favorable and unfavorable
groups and then discussed the issues with the groups,
found that all speakers showed attitude change consistent
with the arguments they delivered. However, speakers
who discussed the issues with groups agreeing with their
belief-discrepant speeches changed more in attitude than
those speakers who discussed the issues with groups who
disagreed with the position they presented.

Berger (1968) reports that after taking a belief-
discrepant position, 35 with a high need-influence who

successfully persuaded an audience to adopt that position

 

32

reported greater attitude change in the belief-discrepant
direction than 3s with a low need-influence. However,
there was little difference in amount of attitude change
between high and low need-influence gs who failed to
persuade the audience.

Rosnow (1966) studied the contiguity effects of

reward and punishment. As §s delivered belief-discrep-

 

ant arguments, he administered reward or punishment

 

either immediately following a belief-discrepant agru-
ment, or immediately following a belief-congruent argu-
ment. He found that when the two opposing arguments
preceded and followed reward, the gs‘ opinions changed
in the direction advocated in the arguments closer in
time to the reward. When the two opposing arguments
preceded the followed punishment, gs’ opinions changed
in the direction advocated in the arguments further in
time from the punishment. He reports, however, that the
effects were not long—lasting.

Sarbin and Allen (1964) report a unique finding,
in which gs who were given negative reinforcement for
encoding belief—discrepant arguments changed in the
belief-discrepant direction more than 35 who were given
positive reinforcement. A11 3s who were reinforced
changed more than the 3s who received no reinforcement.

Turning to studies in which sources encoded

belief-congruent arguments, Huenergardt (1967a) reports

 

 

33

that with speakers who argued in favor of their beliefs,
reward strengthened their attitudes toward the topic,
while punishment produced a weakening effect. Huener-
gardt also varied the expertise of the reinforcing agents.
He found that punishment had a weaker effect on-the
speakers' attitudes when the reinforcing agents were

low in expertise than when they were high.

 

MW: cavalry

In a study by Gardiner (1968) 3s were positively
or negatively reinforced for encoding belief-congruent
arguments. It was found that negative reinforcement
strengthened attitudes already held by the speakers.

Mattox (1967) reports that speakers who strongly
supported or strongly opposed a highly controversial
proposition did not reverse their attitudes after re~
ceiving negative feedback. He states that "listener
feedback does significantly shift strongly held convic-
tions toward a strengthening of_the initial opinion."

Cahn (1966), who studied the effects of simulated
audience feedback, found no differences in attitude
change among groups of speakers receiving positive, nega-
tive, or no reinforcement. Barwind (1969) also found
that positive and negative audience responses had no
significant effect on speaker attitudes toward the
topic.

Finally, a number of studies were conducted that

reinforced attitudes in telephone interviews. Hildum

34

and Brown (1956) report that "good" effectively rein-
forced both positive and negative attitudes. Insko (1965)
found that Opinions reinforced by telephone were still
in force when measured a week later.

In summary, the results seem to indicate that
when a source receives positive response for encoding a

belief-discrepant position, he will change his attitude

m-I-J- li.l'T-_‘. IV.

in the direction of his stated position. However, if
the source receives negative response for encoding a
belief-discrepant message, or receives positive or nega-
tive response for encoding a belief—congruent argument,
his initial attitude will either not be affected at all
or will be strengthened.

Attitude toward the Communication Receiver.

 

Studies dealing with feedback and the source's attitude
toward the receiver nearly all report that positive
response induces positive receiver ratings, while nega-
tive response induces negative receiver ratings.

Engbretson (1964) found that positive feedback
produced an increase in perceived credibility of the
response agent, and that negative feedback generally
produced a decrease in §fs perceived credibility of the
response agent.

Harvey, 3£_31. (1957) report that as the evalu—
ation of a source became more negative, the source tended

to devaluate the initiator of the evaluation. The above

35

researchers also studied evaluations of the §s by peers.
They found that extremely negative evaluations from ac-
quaintances were handled in different ways, depending
on how positively the acquaintance was valued at the
outset: if he was highly valued, his evaluations were
distorted; if he was less well regarded, he was devalu-
ated.

Marshall (1958) found that attitudes toward
others were not changed by 3s after receiving favorable
appraisals. However, less favorable attitudes toward
.others were observed when the gs were subjected to un-
favorable appraisals.

Huenergardt (1967), who varied reward (10% to
90%) from an audience to a Speaker, found that increased
reward from the audience resulted in higher audience
ratings by the source.

In a recent study, Gardiner (1968) found that
speakers who received negative audience response rated
the audience Significantly lower than speakers who re—
ceived positive audience response.

Closely related to attitude toward audience are
two studies that measured empathy and social approval.
Reddy (1968) reports that gs' levels of empathy were in-
creased as a consequence of receiving immediate feedback.

Faules (1967) found that interviewers with a high need

0-!

36

for social approval were low in ability to predict inter-
viewee attitudes.

The research results on speaker attitudes toward
the receiver (response agent) are in strong agreement:
sources who receive positive response will develop a
favorable attitude toward the receiver, whereas sources
who elicit negative response will develop a more nega-
tive attitude toward the receiver.

Attitude toward Self. Researchers studying the

 

self-attitudes of communicators have used a variety of
measures; self~evaluation, self-credibility, self-
ratings, attitude toward self, attitude toward perform-
ance, self-esteem, etc. Most research studies dealing
with self attitudes report that a favorable response will
elicit a positive self-attitude, and that an unfavorable
response will elicit a negative self—attitude.

Marshall (1958) reports that 3s showed no change
in self-attitudes after receiving favorable appraisals.
However, unfavorable appraisals of the 3s elicited less
favorable self-attitudes.

Cohen (1960) and Stotland (1956) both studied
self-evaluation under experimentally-induced success and
failure conditions. Cohen reports that success led to
increases in self-evaluation and failure led to decreases

in self-evaluation. Stotland found that public awareness

 

37

of failure generated lower self-evaluations by §s than
private failure.

Engbretson (1964) studied self credibility under
varying conditions of experimentally induced self credi-
bility, source credibility, and type of response from a
task partner. He found that perceived self credibility

increased under conditions of: a) high self credibility

 

and positive response from a high credible source, b)

 

high self credibility and positive response from a low
credible source, and c) low self credibility and posi-
tive response from a high credible source. He also

found that self-credibility decreased under conditions

 

of low self—credibility and negative feedback from a
high credible source. Engbretson reports that self-
credibility was a relatively stable variable, and that
three of the four significant changes in self-credibility
occurred under positive feedback conditions.

Two studies have reported using self-ratings as
a dependent measure. Harvey, 32 31. (1957) found that
§s who received unfavorable evaluations tended to shift
their self-evaluations in an unfavorable direction.
Jones, 33 31. (1962) asked interviewees to describe them-
selves either accurately or falsely. They found that
§s described themselves more favorably under hypocracy
conditions than under accuracy conditions. Also, the
3s became more self-deprecatory after receiving negative

feedback than after receiving positive feedback.

38

Huenergardt (1967) reports that a high percentage
of approval from the audience resulted in high ratings
by speakers of their own speaking ability.

Frye (1966), who studied self-esteem and its
relation to an individual's influence on his group, re-
ports that self—esteem was dependent on the degree of
successful influence of the group member. If the group
member successfully influenced the group decision, his
self-esteem increased regardless of the effect of his
influence.

Gergen (1965) reports that gs' self-descriptions
became more positive during feedback than in conditions
where no feedback was present.

The single study reporting results contrary to
the above group of studies was conducted by Mikawa (1963).
He reports that significantly higher self-hostility
scores were obtained when 3s were given approval feed-
back than when they were given rejection feedback.

Attitudes toward Task Performance. Two studies
have dealt specifically with attitudes toward task per-
formance. Lewin (1968) induced conditions of success
and failure and observed the effects on each 3's level
of aspiration. He found that levels of aspiration under
failure lagged behind levels of achievement, whereas no
such lag was observed under induced success conditions.

Engbretson (1964), who studied source credibility,

 

39

self credibility, and perceived task difficulty, reports
that perceived task difficulty increased under condi-
tions of a) high self credibility and negative feedback
from a high credible source, b) high self credibility
and negative feedback from a low credible source, and

c) low self credibility and negative feedback from a
high credible source. He also found that perceived task

difficulty was decreased under conditions of a) high

 

self credibility and positive feedback from a low cred-
ible source, b) low self credibility and positive feed-
back from a high credible source, and c) low self credi-

bility and positive feedback from a low credible source.
Attitudes of the Communication Receiver

A very limited amount of research has investi-
gated the effects of audience response on the attitudes
of other audience members. Only two studies have been
reported on this issue. Hylton (1968) explored the
effects of a trained group of listeners (one-half of the
audience) on the remaining half of the audience, who
were naive listeners. He found that when naive gs were
mixed with confederates who gave positive responses to
the speaker, the naive 33' attitudes toward the topic
and toward the speaker were significantly more positive
than the attitudes of §s mixed with confederates who

gave negative responses.

 

ff. 4-

40

Mehrley and Anderson (1968), who also studied
the effects of audience response on other audience mem-
bers, attempted to separate the audience effects from
the speaker effects. They accomplished this by video-
taping speeches that received positive and negative re-
sponces from a trained audience. The videotapes were

played to naive gs, in order to observe the effects of

‘...a—n—h— .m—

positive and negative responses administered by a former
audience. Mehrley and Anderson report that speeches
having received positive response from a former audience
did not produce a significantly greater amount of atti-
tude change than speeches having received negative

audience response.

Group Measures

 

Several studies have explored feedback and its
effects on group behavior. Variables that have been
measured in connection with feedback are attitude toward
the group, defensive feelings, and group productivity.

While studying 35' attitudes toward the group,
Mikawa (1963) observed significantly higher group-

hostility scores with rejection feedback than with

approval feedback.

Three studies report the effects of feedback on
defensive feelings of group members and group produc-

tivity. Gibb, 33 31. (1955) report that positive

41

feedback produced Significantly fewer defensive feelings
and significantly greater task efficiency than negative
feedback. They also found that feeling feedback pro-
duced Significantly fewer defensive feelings and sig-
nificantly higher task efficiency than task feedback.
Lott, 33 31. (1955) also report that positive feedback
produced significantly greater task efficiency than
negative feedback. A third study by Smith and Kight
(1959) reports that personalized feedback consistently
improved group problem-solving efficiency.

Molof (1963) investigated the effects of success
and failure on judgments in small decision-making
groups. He found that experimentally induced success-
failure and continual-terminal feedback variables did
not provide differential contexts for predicting judg-
ments.

Cieutat (1962), who studied social conversation
in a small group situation, found it possible to alter
the quantity of social conversation with nonverbal rein-
forcement.

From the experimental results involving feedback

and group measures, it can be concluded that feedback

has pronounced effects on individual attitudes, defen—

sive feelings, and task efficiency. Positive feedback

improves attitudes, feelings, and efficiency, while

negative feedback produces a deteriorating effect.

42

Conclusions about Feedback

From the results cited above, it is possible to
draw a number of general conclusions about feedback as
a communication variable. This section will present
seven such conclusions.

1. There is strong evidence that negative audi-

 

ence response inhibits speaker delivery. This effect
has been observed in fluency, utterance rate, voice
loudness, nervousness, stage fright, eye contact, and
bodily movement.

2. Studies investigating the effects of feed—
back on speech content demonstrate that certain message
characteristics can be strengthened by reinforcement,
that free feedback increases the accuracy of message
transmission, and that "essential changes" in content
are effected by aversive receiver responses.

3. Feedback's impact on a source's attitude
toward his topic depends on whether he presents a belief-
consistent or a belief—discrepant message. When the
source receives positive responses for defending a be-
lief-discrepant position, he will likely change his
attitude in the belief-discrepant direction. When he
receives negative responses for defending a belief-dis—
crepant position, he will likely maintain his initial

position. However, if the source argues from a

43

belief—consistent position, both positive and negative
audience re3ponse may have either a strengthening effect
or no effect on his original attitude.

4. Sources who receive positive audience re-
sponse will develOp more favorable attitudes toward
themselves and toward the audience than sources who re—

ceive negative audience response.

 

5. Negative feedback will cause 35 to raise
their perceptions of task difficulty and to lower their
levels of aspiration.

6. Studies investigating the effects of audi-
ence feedback on other audience members report that
positive feedback presented in the presence of naive
audience members will move their attitudes in a favor-
able direction, while negative feedback presented in
their presence will influence them in the opposite di-
rection. Other results suggest that the effects of one
audience member's response on another audience member's
attitude is direct in nature, and does not depend on
modifications produced in a speaker's behavior for
mediation.

7. From studies investigating feedback and
group behavior, it can be concluded that positive feed-
back reduces group hostility, produces fewer defensive
feelings among group members, and increases group task
efficiency, while negative feedback produces the opposite

effects.

CHAPTER III

THEORY AND HYPOTHESES

The purpose of this chapter is to develop a

 

theoretical rationale, based on cognitive balance theory,
for predicting attitude change in varying feedback situ-
ations. A definition and theoretical rationale will be
presented first, followed by the hypotheses tested in
this study.

For the purposes of this study, feedback is de-
fined as an adaptive process in which the source adjusts
his attitudes as a result of expected and perceived out—
put from a receiver.

The communication source brings a number of
salient cognitions into the communication Situation.
First, he has an attitude toward the topic about which
he plans to communicate. Second, he has a number of
expectations toward his performance, based on the extent
to which he has prepared and the success or failure of
his previous performances. Third, he has an attitude
toward the receivers. The salience of this third atti-

tude depends on the source's knowledge about the

44

45

receivers, particularly knowledge of their attitudes
toward the issue about which he intends to communicate.

While the communicator enters the situation
with the above attitudes, it is highly unlikely that
they will remain unchanged during the communication
process. It is likely that the situation will add new
cognitions. Important among these are the receivers'
dynamic reactions, which express their acceptance or
rejection of the message. The receivers' responses
will likely affect the source's attitude toward the re-
ceivers, his expectations of success or failure, and
his attitude toward the concept about which he is com—
municating. In the Scott studies (1957, 1959a, 1959b)
reviewed above, it was hypothesized that the receivers'
reactions to the message would significantly affect the
source's attitude toward the tOpic. The results sup-
ported this prediction.

Another important cognition added by the com-
munication situation is the source's judgment of the
success or failure of his communicative effort. This
judgment is influenced by the source's expectations of
success or failure and by the receivers' responses.
Selectivity also plays an important role in determining
the source's judgment of his success or failure. The
source may receive a negative response from a receiver

and perceive it as positive, or vice versa. The Amato

46

and Ostermeier (1967) study provides an example where
selectivity was likely in operation. A speaker who had
received negative responses from the audience reported
getting a positive reaction.

Three particular types of selectivity were dealt i
with in the study reported here. Selective exposure was t

controlled by assigning 3s to either positive or nega- ;

 

tive audience conditions. Selective attention was con— L§
trolled by placing the audience around a large table
with each audience member directly facing the speaker.
This minimized the chances of the speaker "tuning out"
the audience response. Finally, selective perception
was measured in the study by asking each speaker what
type of responses he had received while speaking. The
speakers' responses were compared to the type of reac-
tions the audience actually gave. Any discrepancies
found in the comparisons were considered to be evidence
of selective perception.

Once the communication event has taken place,
the source will likely emerge with the following salient
cognitions: 1) An attitude toward the concept about
which he communicated. 2) An attitude toward his per-
formance. 3) A perception of the receivers' attitudes
toward his performance. 4) A perception of the receivers'
attitudes toward the concept about which he communicated.

5) An attitude toward the receivers.

47

Whether or not the source's attitude twoard the
concept (item one above) is changed during the communi-
cation event depends a great deal on the source's initial
attitude toward the topic. Studies by Scott (1957,
1959a, 1959b) and Bostrom, 33 31. (1961) have found that
when the source is initially opposed to the concept he }

advocates, there is potential for a significant amount i

 

of attitude change. However, when the source is initially L
favorable toward the concept about which he is communi-
cating, as in the Gardiner (1968) study, no significant
attitude change may take place. Since the study de-
scribed here involved a source encoding a belief-con-
gruent message, no attitude change toward the topic was
expected.

The major focus of this study was on the source
and his attitudes toward his performance and toward the
recievers. These two attitudes were measured as the
dependent variables of the study.

A secondary concern was placed on the source's
attitudes toward the topic, along with the source's per—
ceptions of the receivers' attitudes toward his per-

formance and toward the topic.
Balance Theory Predictions

Cognitive consistency theory posits that an in—

dividual is in a natural COgnitive state when all salient

48

cognitions are consistent (Festinger, 1957). When two
or more inconsistent ideas become salient, the cognitive
state becomes unbalanced. The individual then behaves
in a manner designed to restore balance.

Balance theory, a type of consistency theory,
predicts cognitive outcomes in relations between indi- i

viduals and events. Newcomb (1953), who adapted balance ]

 

theory to communication events, developed “systems of

 

orientation," consisting of the orientation of A (a
communication source) toward B (a communication receiver)
and X (the object communicated about), as well as A's
perception of B's orientation toward X. Each of the
three orientations in the system can vary from positive
to negative. An orientation can be described as an
attitude, a predisposition, or an evaluative judgment.
Such a system of orientation can be depicted as follows

X
+ + . In the system just shown, A has a positive

A?B
orientation toward both E and X. Also, according to
A's perceptions, B has a positive orientation toward X.
This is a balanced system, as will be explained below.
According to Heider (1946, 1958), systems with
three positive relations, or with one positive and two
negative relations, are balanced. Conversely, systems
with one negative and two positive relations, or with

three negative relations, are not balanced. As a type

of consistency theory, balance theory predicts that when

49

the system is in a state of imbalance (i.e., it has one
negative and two positive relations, or three negative
relations), the individual will change one or more of
the orientations in order to restore balance.

The reader is asked to recall the communication
situation mentioned above, in which a source delivers a

message, perceives a dynamic response from the receivers,

 

and emerges with a number of new cognitions. It is
possible to view the above situation in a balance theory
paradigm, where A represents the communication source,

B represents the receiver, and X represents A's perform-
ance. The A+B relationship corresponds to the source's
attitude toward the receiver. The A+X relationship
represents the source's attitude toward his performance.
Finally, the B+X orientation is the source's perception
of the receiver's attitude toward his performance. These
variables will be operationalized in more detail in
Chapter IV.

Figure 2 presents four balance theory paradigms
that were tested in this study. The chart will include
the following: 1) The manipulation of expected audience
response (EAR). 2) Perceived audience response (PAR).
3) The experimental inductions, expressed as balance
theory systems. 4) The predicted experimental outcomes,
also expressed as balance theory systems. 5) The Spec—

ific predictions which led to the study's hypotheses.

50

 

 

 

 

 

 

Con— Expected Perceived Initial Predicted*
di- Audience Audience Experimental Cognitive
tion Response Response Induction Outcome
X X
1 Positive Positive + + + +
as. a.
+
X X
2 Negative Positive + + + +
Afl RB AQB
X X
3 Positive Negative +f’\ - —/\.\ -
A? B A :7 B
X X
4 Negative Negative + - + -

 

* Specific Predictions:

l) Attitude Toward Audience (A+B): 3 and 4 < 1 and 2
35::
2) Attitude Toward Performance (A+X): 3 and 4 < 1 and 2
iii

 

Figure 2. Balance Theory Predictions

51

Condition One describes a situation where a
source with a positive attitude toward his performance
and a positive attitude toward the receiver, perceives
that the receiver also has a positive attitude toward
his performance. According to balance theory as devel-
oped by Heider (1958), Condition One represents a bal- ?

anced system. Therefore, in this study it was predicted '

 

that individuals performing under positive EAR and posi-
tive PAR conditions would emerge with positive attitudes
toward their performances and toward the receivers.
Condition Two represents a situation where the
source has a positive attitude toward his performance,
a negative attitude toward the receiver, and perceives
that the receiver has a positive attitude toward his
performance. The above system, according to balance
theory, is cognitively inconsistent, and a change in
one or more of the attitude states is expected. Two
possible changes can be made. First, A can change his
attitude toward B from negative to positive, creating a
balanced state. Second, A can change his attitude toward
X from positive to negative, and restore consistency.
As shown in Figure 2, it was predicted that the source,
perceiving a receivers' positive reactions to his per—
formance, would develop a more positive attitude toward
the receivers, rather than a more negative attitude

toward his performance. It was reasoned that the

52

speaker's attitude toward his performance was more ego-
involved than his attitude toward the audience. Since

the audience, which was initially negative, displayed

a positive attitude toward his performance, it would be
easier for the speaker to increase his esteem of the

audience than to devalue his performance. Also, since

1. If

it was evident to the speaker that he had been a success )

‘3‘

 

(i.e, he had convinced a hostile audience), he would not L
likely derogate his performance. Therefore, positive
audience ratings and positive performance ratings

by the speaker were predicted in Condition Two.

Condition Three describes a situation where the
source has positive attitudes toward both his performance
and the receivers. However, the receivers are perceived
as having negative reactions toward his performance.
According to balance theory, this state is cognitively
inconsistent. Again, two possible changes can be made.
The source can develop a more negative attitude either
toward his performance, or toward the receiver. As
shown in Figure 2, it was predicted that the source
would devalue his performance, rather than the audience.
It was reasoned that since the speaker knew the audience
was favorably disposed toward his topic before the per-
formance, he would blame his performance for the nega—

tive reaponse. While the source's attitude toward the

53

audience in Condition Three was expected to be positive,
it was predicted to be significantly less positive than
the attitudes of the sources in Condition One, who ex—

pected positive response and received positive audience
response. The reason for this seeming contradiction is
that two personality types were operating in all condi- L
tions. However, in Condition Three personality differ-

ences become crucial. It was reasoned that speakers

 

 

with high self-esteem would restore balance in Condition
Three by devaluing the audience, while speakers with low
self—esteem would resolve the imbalance by derogating
their performances. In the present study, a lack of 3S
kept these personality differences from being measured
and observed. Nevertheless, as stated above, the pre—
dictions made allowances for both personality types to
function.

In summary, it was predicted in Condition Three
that some of the speakers (those with low self-esteem)
would restore cognitive balance by devaluing their per-
formances, while the high self—esteem speakers would re-
store balance by derogating the receivers. It was pre—
dicted that the results would produce strongly negative
performance ratings and weakly negative audience ratings.

Condition Four represents a situation where the
source has a positive attitude toward his performance,

a negative attitude toward the receiver, and perceives

54

that the receiver has a negative attitude toward his
performance. According to balance theory, this state

is cognitively balanced. As shown in Figure 2, it was
predicted that the source would develop a highly negative
attitude toward the receivers, while his attitude toward

his performance would remain positive. It Should be I
noted that although the source's attitude toward his L

performance was predicted as positive in Condition Four, F

 

it was expected to be significantly less positive than
the source's attitude toward his performance in Condi-
tion Two. In summary, two attitudes were predicted for
the speakers in Condition Four: 1) A highly negative
attitude toward the receiver. 2) A weakly positive at-
titude toward the performance.

From the above rationale and the predicted cog-
nitive outcomes shown in Figure 2, the following hypo-

theses were formulated for this study:

Hypotheses

 

I. There will be an interaction between EAR and
PAR on performance ratings by the speaker, to the effect
that: Speakers in the negative EAR condition who receive
negative PAR will rate their own performances higher than
speakers in the positive EAR condition who receive nega-

tive PAR.

55

II. Speakers who receive positive PAR will rate
their own performances higher than speakers who receive
negative PAR.

III. There will be an interaction between EAR
and PAR on audience character and authoritativeness
ratings by the source, to the effect that: Speakers in
the positive EAR condition who receive negative PAR will
rate the audience higher in character and authoritative-
ness than speakers in the negative EAR conditions who
receive negative PAR.

IV. Speakers who receive positive PAR will rate
the audience higher in character and authoritativeness

than speakers who receive negative PAR.

 

CHAPTER IV

METHOD

This chapter will discuss in detail, the method
used to test the hypotheses of this study, including the
experimental design, the subjects (3s) and confederates
(gs) involved, the procedures by which the above method-
ological elements were implemented, the techniques of
measurement employed, and the methods of statistical in-

ference used.

Experimental Design

 

Kerlinger (1964) describes research design as a
strategy of scientific investigation, formulated for two
purposes: a) "to provide answers to research questions"
and b) "to control variance (p. 275)." According to
Kerlinger, an adequate design will aid the researcher in
deciding which observations to make, which variables to
control and observe, and which types of statistical tests
to use.

The design employed for this study was a 2x2 fac-
torial. Winer (1962) points out that "factorial experiments

permit the evaluation of interaction effects (p. 140)."

S6

57

The evaluation of an interaction enables the researcher to
observe how experimental variables operate in combination
with one another, as well as how they operate in isolation.
By utilizing a 2x2 design in this study, it was possible
to observe the main effects for PAR and the predicted in-

teraction effects between PAR and EAR.

 

According to Kerlinger (1964), the second purpose

of experimental design is to control variance. The adequate i

 

research design, he notes, will maximize the variance of

the variables in the substantive research hypothesis, control
the variance of extraneous or unwanted variables, and
minimize the error or random variance.

In the present study, two variables (Expected Au-
dience Response and Observable Audience Response) were
manipulated as a part of the design, two variables (prior
attitude toward the topic and sex) were controlled, and
three variables (attitude toward the receiver, attitude
toward the performance and PAR) were measured. The major
purpose of the design was to maximize the variance produced
by EAR and PAR, and control the variance caused by attitudes
toward the topic. The following section will offer an
operational definition of the variables involved in the

study.

58

Operational Definitions

 

Expected Audience Response (EAR): The gs' expecta-
tions of the audience's response, based on information
given the 3s by the experimenter (3) regarding the au—
dience's attitudes toward their topics.

Positive EAR: A condition in which the gs were t

informed that the audience held a favorable attitude toward

“IE £v‘r

their position.

 

Negative EAR: A condition in which the 3s were
informed that the audience held an unfavorable attitude
toward their position.

Observable Audience Response (OAR): The nonverbal
response given by the receivers as the 3 delivered his
message.

Positive OAR: A pre-determined set of favorable
receiver reactions which included smiling, agreeing with
the source, listening attentively, applauding at the end
of the message, etc.

Negative OAR: A pre-determined set of unfavorable

 

receiver reactions which included frowning, head-Shaking,
slouching, looking around the room, disagreeing with the
source, etc.

Perceived Audience Response (PAR): The gs' post-

 

communication perceptions of the responses given by the

audience as they delivered their messages.

59

Positive PAR: The gs' postcommunication percep-
tions that the responses given by the audience were
favorable.

Negative PAR: The gs' postcommunication percep-

 

tions that the responses given by the audience were
unfavorable. l

Attitude toward the receiver: The gs' postcommu-

I

“- 'T—T
.

nication responses to semantic differential credibility

 

scales directed at evaluatively describing the receiver of
their messages.

Attitude toward performance: The §s' postcommuni—

 

cation re5ponses to evaluative semantic differential scales
describing the relative quality of their performances.

Attitude toward tOpic: The gs' evaluative predis-

 

positions toward the tOpics of their messages. This var-
iable was presumably controlled by asking each S to choose
a topic toward which he was strongly favorable.

To review the design, sex and prior attitude toward
the §fs topic were controlled. EAR and OAR were manipulated
to form four experimental conditions:

1. Positive EAR and Positive OAR: Each 3 had been
informed that his receivers favored his position, and while
encoding a persuasive message he was provided with favorable
OAR.

2. Negative EAR and Positive OAR: Each 3 had been
informed that his receivers opposed his position, but he

was given positive OAR during the performance.

60

3. Positive EAR and Negative OAR: Each 3 had been
informed that his receivers favored his position, but he
was provided negative OAR while encoding the message.

4. Negative EAR and Negative OAR: Each 3 had been
informed that his receivers opposed his position, and he

was provided negative OAR.

Subjects and Confederates

Subjects

The 3s involved in this study were 53 students
enrolled in Communication 305--Persuasion, at Michigan
State University during the winter quarter of 1969. There
were certain advantages to using this particular group as
gs. First, the students were not, for the most part, be-
ginning public speakers. The majority of them had taken
as a prerequisite, Communication lOl—-Public Speaking,
where they had given six speeches.

Second, it was possible to introduce the exper-
imental study as a course assignment. The students were
not aware of the experimental nature of the project until
debriefing took place.

Third, Communication 305 met as a mass group. The
researcher did not have to settle for "sampling of conven-

ience" by taking intact sections. Dealing with only one

group enabled E to randomly assign the 3s to conditions,

61

keep all instructions constant, and debrief the group in

a single session.

the g. In the debriefing, many 3s expressed that they
considered participation in the project a worthwhile
endeavor.

The 33 were assigned to the four experimental
conditions by stratified random sampling, according to
sex. "Stratified sampling," according to Anderson (1966),
is used to "stratify, or classify, the population on the
variable one wishes to control (p. 68)." As noted earlier,
sex was a variable the experimenter wished to control in
the design.

Thirty-two males and 27 females were originally
included in the design. However, one male and three
females failed to appear for the experimental session.
Also, two female 3s who participated had to be discarded
from the results. One reported that she perceived the
purpose of the project and could not give a true reaction.
A second accidentally overheard one of the experimental
assistants talking about the nature of the experiment.

After attrition, the total number of 3s was 53, including

31 males and 22 females.

“Vin-F. ‘_‘

62

ggnfederates

O
0
g...
l-'
(D
O
(D
m
(D
H
<
(D
Q:
n)
m
d-
:3"
(D
D)
c:
Q:
5.:
(D
:5
O
(D
H)
O
H
9)
l-‘
l—'
(D
K
"O
(D
H
p
B
(D
:3
('1'
DJ
[.4
O
O
D
Q:
'7

r1-
5.:
0
:3
m
n!
:3
0.:
DJ
0.:
B
1.;
:3
p
m
f.-
(D
H
(D
0.:
r1-
.‘J‘
(D
m
'0
'U
H
O
'U
:1
P
(D
d-
(D
H
(D
m
'U
0
5
m
(D
n-
O
(D
n:
O
:3"
(I)

sponse to the speakers.
The '
folloWing steps were taken to insure that the
audie ' '
nce, a critical element in the experiment wa
, s a
constant variable:
1. '
During two one-hour sessions prior to the ex-
periment '
,‘g trained the ES to administer both positive and
negative response to speakers. \
2. '
Each g_was given two specific assignments:l
one se ' ' ' '
t of instructions indicated his role in administeri
O 0 ng
OSlt
p ive response, and the other indicated his negative
res o
p nse role. There were 13 different sets of positive
and '
negative roles,--one for each C. Each C was trained
to ‘ ' -
perform an identical role for each positive condition

an ’ ‘
d an identical role for each negative condition.

¥

1
For a summary of the assignments, see Appendix B.

3. ‘ES were seated in the same order for all exper-

imental sessions.

4. Each 2 was instructed to wear the same clothing

to each session.
5. A triple check was made, to insure that the ES
administered the correct OAR induction:

a. Each 9 had a list indicating the type of

‘» .\m_'_

response assigned to each 3. .
b. One 9 was designated as leader. She sig- i

naled the proper response before each speech began, and

was instructed to signal the group if any improper responses

were given.
0. 3 viewed each persuasive performance from

outside the room through a small window directly behind

the speaker. All responses, according to the judgment of

E, were consistent with the assigned responses.

Procedures

The procedures of the experiment were carried out
as follows:

Week 1. The experimenter was introduced to the
Communication 305 class as the instructor in charge of the
first speaking assignment. At this time, 3 announced that
each student was to prepare a five—minute persuasive speech
on a tOpic he strongly believed in. A handout was given

the students, entitled "Persuasive Assignment #1: Speaking

. . 2 . .
to a Live Audience." At this time two ideas were intro-

duced to the class. First, it was stressed that innovative

would be measured. Second, it was pointed out that the
first speech would be given to a live audience, arranged

by the instructor. §_promised to help the class analyze

(‘1’
33‘
(D
D)
I:
Q:
l-'
(D
:1
O
(D
U
l<
"O
H
O
<
p
Q:
l-'
:3
LG
9)
'U
H
0
t.“
l-‘
p...
(D
0
HI
rt.
5‘
(D
n)
(3
Q:
g...
(D
:3
0
(D
m
n)
(1'
fl
5..-
I
Why?

tudes, social makeup, professional goals, etc.

Week 2. gs handed in a statement of their topic
and purpose to the instructor. He promised to test the
prospective audience's attitudes toward each topic, and
report the results a week later.. The gs were also in-
structed not to change their speech topics.

Week 3. The 3s were handed a profile3 of the
audience. The information concerning the audience atti-
tudes was manipulated according to the experimental condi-
tion to which each 3 had been assigned. The 3s were in-
structed to complete their message preparation, utilizing
the audience information. Also during the third week the
98 were trained to administer audience responses to the 33.

Week 4. The speeches were delivered during four

three-hour evening sessions. Each 3 signed up for a

—¥

2For a copy of the assignment, see Appendix C.

3For a copy of the profile, see Appendix D.

65

specific time to speak. At his assigned time, the S re-

ported to 3'8 office. The g was then ushered to a nearby

istered to other 3s. 3 introduced each §lto the ES and
promptly left the room. When the 3 had finished, he was
led to another classroom. There he was administered the

dependent measures, which were presented as a "Speaker

‘11..- “av-”h

Reaction Inventory."4 Once each 3 had completed the ques-

tionnaire, he was dismissed.

Measurement

In order to test the hypotheses of this study, four
separate measures were necessary: 1) a manipulation check
on EAR, 2) a manipulation check on OAR (i.e., a measure of
PAR), 3) a measure of the §5' attitudes toward the audience,

and 4) a measure of the §$' attitudes toward their

performances.

Mgnipulation checks

The purpose of a manipulation check is to determine
Whether a given value of a variable has been perceived by
the Ss as E intended. For example, the E who approaches £5

with a strong fear-arousing message will need to measure

Whether or not the Ss perceived the message as fear-arousing.

- ——

4For copies of the dependent measures, see Appendix E.

66

In this experiment, two manipulation checks were
needed. First, it was necessary to observe whether the 3s
had built the expectation of audience response that 3 had
intended. The following scale measured the 3s' perceptions

of EAR:

What type of attitude did your audience have toward
your topic before you spoke to them?
___’Favorab1e
'___ Neutral
Unfavorable

Second, and even more crucial, was the check on
each §fs perception of OAR. Amato & Ostermeier (1967)
reported finding that a speaker who had been given obvious
negative responses reported receiving positive response.
The manipulation check on OAR was used to expose any in-
stances of selective perception in the study. The scale
used to check OAR was as follows:

What type of response did you receive from the audience
as you spoke?

Highly positive

Moderately positive

Slightly positive

Neutral

Slightly negative

Moderately negative

Highly negative

Dependent Measures

The first DV measured the gs' attitudes toward the
audience, which can also be termed "audience credibility."
Therefore, a credibility measure was employed as an index

of the 3s' attitudes toward the receivers.

67

for determining the acceptability of message sources. In

ingful and statistically independent dimensions" for source
credibility (p. 18). These dimensions are safety, qualifi-
cation, and dynamism. Berlo and his researchers worked !
primarily with semantic differential scales. They suggest g
15 scales, five for each dimension, as a reliable measure ‘
of source credibility. I
McCroskey (1966) also tested a number of semantic
differential and Likert-type scales for measuring source
credibility. After testing for validity, he concluded that
his scales are a representative sampling of the universe of
items pertaining to ethos (source credibility). Unlike
Berlo, McCroskey found only two dimensions of source credi-
bility: authoritativeness and character. Lately, however,
he has reported finding weak evidence of another dimension
which he, like Berlo, calls dynamism.
For the purpose of measuring the source's attitude
toward the receiver (audience credibility) it was decided
to combine the scales developed by Berlo, 3£_31., and

McCroskey. The scales used were as follows:

Character: Cruel--Kind, Valuable--W0rthless, Just--

(Safety) Unjust, Unpleasant--Pleasant, Selfish--
Unselfish, Awful--Nice, Honest--Dishonest,
Reliable--Unreliable, Virtuous--Sinful,
Safe--Unsafe, Unfriendly--Friendly.

68

Authoritativeness: Skilled--Unskilled, Qualified--

(Qualification) Unqualified, Expert-~Inexpert, Un-
informed—~Informed, Trained--Un-
trained, Intelligent--Unintelligent,
Experienced--Inexperienced

Dynamism: Active--Passive, Energetic—-Tired, Timid--
Bold, Emphatic--Hesitant, Aggressive—-Meek.

Since the above scales had not been tested as a I
measure of "audience credibility," they were submitted to ‘
a factor analysis to determine whether the same dimensions

held as when Berlo, et a1. (1966) and McCroskey (1966) (

 

tested them on sources. The number of 3s being small (53)
made the factor analysis a questionable procedure. However,
it was decided that proceeding with the analysis would be
better than using only subjective judgment to group the
data.

The second DV measured the gs' attitudes toward
their performances. Fifteen semantic differential scales
were constructed for this purpose by the experimenter.
Eight of the scales were evaluative, and the others were
employed as distractors. In order to determine which
scales to use as the dependent measure, the 15 scales were
submitted to factor analysis. The scales used to test the
gs' attitudes toward their performances were:

Evaluative: Good--Bad, Fearful-~Fearless, Well De-
livered-—Poorly Delivered, Beneficial--
Harmful, Perfect--Imperfect, Unfriendly--

Friendly, Poor in Content-~Good in Con-
tent, Adaptive--Nonadaptive.

69

Distractors: Forcefu1-—Forceless, Self-Conscious--
Confident, Direct--Indirect, Strong--
Weak, Timid--Bold, Speaker Centered--
Audience Centered, Message Centered--
Audience Centered.

Statistical Inference

 

It was necessary to complete the statistical anal- I
ysis of the data in two phases. First, a preliminary anal-

ysis was run. Second, the data were submitted to a major

/ «Swan 2.
in

statistical analysis. (.1

 

PreliminaryrAnalysis

 

Manipulation checks.--The purpose of the first

 

manipulation check was to determine whether the 3S had
perceived the EAR variable as the 3 had intended. The
method used to analyze this check was simply looking at
each 3's results and comparing his perception of EAR with
his assignment to an EAR condition.

The purpose of the second manipulation check was
to determine whether the §§ had perceived OAR as the 9s
had administered it. The results of the OAR check were
submitted to a chi square analysis for significance of
difference. Also, the experimenter checked each 3's re-
sponse by comparing his PAR with his assignment to OAR.

Factor Ana1ysis.-—The individual scales for each

 

of the DVs were submitted to a factor analysis. Twenty-
three items were analyzed from the audience credibility

DV, and 15 items were analyzed from the attitude toward

70

performance DV. Results of the factor analysis were in-
terpreted as follows:

Each item in the analysis was considered indepen—
dently. A judgment was made about which factor, if any,
each item “loaded" onto. This judgment was based on the
relative magnitude of the factor loadings for the item on
each of the factors. The standard established as a cri-
terion was a loading of .60 or above for the factor on
which the item received its highest loading, and a loading
of .40 or below for the remaining factors. Consider, for
example, the following hypothetical factors, items, and

factor loadings:

 

Item Factor

1 2. 2 i
l .2 .8 .1 .1
2 .6 .1 .6 .2

The first item above, according to the criterion
used for the factor analysis, would be judged as loading
on the second factor, since its loading under factor two
is above .60 and its loadings on all the other factors are
below .40. The second item, however, is split between the
first and the third factors, and cannot be considered part
of either factor.

Once an independent judgment had been made for each
item on each of the DVs, the scores in each factor were

summed and prepared for the major statistical analysis.

71

Major statistical ana1ysis

Before conducting a statistical analysis of the
data, the researcher had to decide whether to employ para-
metric or nonparametric statistical tests. Siegel (1956)
describes parametric tests as those statistics which have
stronger or more extensive assumptions. Nonparametric
statistics, he points out, restrict the researcher to
making more general conclusions about his data. Before
parametric statistics can be used, the following assump-
tions must be met:

1. The observations must be independent.

2. The observations must be drawn from normally dis-
tributed populations.

3. The populations must have the same variance.

4. The variables involved must have been measured in

at least an interval scale.

5. The means of the pOpulations must be linear com-
binations of effects due to columns and/or rows

Siegel, 1956, p. 19).

Siegel adds that with the exception of the third
assumption (homogeneity of variance), the conditions out-
lined above are not normally tested in an experiment.
Rather, they are "presumptions which are accepted (PP.
19-20)." However, from experimental results cited by
Boneau (1960), it has been shown that even the homogeneity
of variance assumption can be safely violated in most cases.
However, this is true only when the experimental groups

are relatively equal in size and the number in each group

is 15 or more.

72

Since the smallest number of 3s in any condition
of this experiment was less than 15, it was decided to test
for homogeneity of variance, and base the decision of
whether to use parametric or nonparametric statistics on
the outcome of that test. Bartlett's test for homogeneity
of variance (Winer, 1962) was employed.

Another assumption of concern to §_was whether the
scales used in the measuring instruments constituted equal
intervals. Kerlinger (1964) states that “it is probable
that most psychological . . . scales approximate interval
equality fairly well." He recommends that "the best pro-
cedure would seem to be to treat ordinal measurements as
though they were interval measurements, but to be constantly
alert to the possibility of grggg inequality of intervals
(p. 427)." It was decided that the semantic differential
scales used in measuring the DVs of this study probably
contained no gross inequality of intervals, and that the
assumption of equality of intervals had not been seriously
violated.

As stated at the outset of this chapter, one of
the primary functions of a research design is to aid the
researcher in deciding which statistical tests to use. In
this particular study, which utilized a 2x2 factorial
design, the 2-way analysis of variance was appropriate.

In the cases where all the assumptions for parametric sta-

tistics apparently were met, the F test for analysis of

73

variance (Winer, 1962) was appropriate. If parametric
assumptions clearly were not met, the Wilson (1956) non-
parametric analysis of variance was appropriate.

Analysis of variance tests were used to indicate
significant differences between combinations of experimental
conditions, as well as significant interactions. However,
for estimating differences between individual groups, addi-
tional tests were needed. In the cases where parametric
tests were possible, the 3 ratio (Winer, 1962) was approp-
riate to estimate the significance of between—group differ-
ences. If parametric assumptions had not been met, the
Mann-Whitney U (Siegel, 1956) was appropriate. The prob-
ability level set for significance for all statistical
tests was .05.

All statistical tests were calculated by means of
library computer programs on a Control Data Corporation

3600 computer, in Operation at Michigan State University.5

 

5The library programs employed were:

Williams, A. Factor analysis. Technical Report No. 34.
Computer Institute for Social Science Research.
Michigan State University, 1967.

Morris, J. Nonparametric chi—square tests and analyses of
variance. Technical Report No. 42. Computer In-
stitute for Social Science Research. Michigan
State University, 1966.

Morris, J. Mann-Whitney and Wilcoxon tests, randomization
and sign tests, parametric Eftests and F-tests.
Technical Report No. 45. Computer Institute for
Social Science Research. Michigan State University,
1967.

74

Ruble, W. L. Analysis of covariance and analysis of var-
iance with unequal frequencies permitted in the
cells-~(LS routine). STAT Series Description No.
18. Michigan State University Agricultural Exper-

iment Station, 1968.

CHAPTER V

RESULTS I

This chapter will include seven major divisions.

The first section will cite the results of the preliminary

flu “an-:1. ‘ -_
A i E: as;

“a

manipulation checks. Section two will present the results
of the factor analyses. The third section will report the
results of Bartlett's test for homogeneity of variance.

The fourth and fifth sections will present the major results
of the study. Section six will describe a supplementary

analysis, and the final section will summarize the results.

Manipulation Checks

 

AS stated in Chapter IV, two manipulation checks
on the independent variables were needed. First, a check
on EAR was necessary to determine whether the 3s had built
an expectation of the audience's response. After the ex-
periment each 3 was asked on a questionnaire, "What type of
attitude did your audience have toward your topic before
you Spoke to them?" After examining each of the question-
naires, it was discovered that each § who had been assigned
to positive EAR still recalled the audience's prior attitude

75

76

as positive. Similarly, each 3 assigned to negative EAR
recalled the audience's prior attitude as negative. A
better manipulation check on EAR might have been to ques-
tion the gs before the experimental induction. However,

it was feared that a pretest would sensitize the 3s. The
experimenter therefore decided to combine the EAR check
with the post measurement of the DVs. Results of the check
suggest that the EAR manipulation lasted through the ex-

perimental treatment.

931:

The second experimental induction, OAR, was checked
by measuring perceived audience response (PAR). gs were
asked, "What type of response did you receive from your
audience?" A seven-point scale, ranging from "highly
positive" to "highly negative" was used to measure PAR.
The results were analyzed by two methods. First, the data
were submitted to a chi square test to determine whether
there was a significant difference between the Ss' percep-
tions of positive and negative OAR. A chi square value of
27.69 (p < .05) was obtained.

In the second method of analysis each of the ques-
tionnaires was examined to determine if each 3 receiving
positive OAR reported positive PAR and each 3 receiving
negative OAR reported negative PAR. It was found that all

gs in the positive OAR condition perceived the response

_:-E'_—_ -

flirt _

77

as positive. However, in the negative OAR condition, seven
3s perceived the response as “slightly positive" and one 3

reported receiving a "moderately positive" response. These
findings could not be dismissed as mistakes, since these 3s
also reported that the audience was slightly or moderately

positive toward their topic after the performances. Since

PAR did not equal OAR, it was concluded that the OAR induc-
tion was not entirely successful.

Because selective perception was likely in opera-
tion, and since the focus of this study was on source be-
havior, it was decided to abandon the OAR IV and substitute
PAR. Another reason for making this substitution was that
the balance theory paradigm introduced in Chapter II defined
the B+X relationship as A's perception of B's attitude
toward X. Therefore, the change from OAR to PAR was con-
sistent with the rationale and the a priori hypotheses of
the study. While the major analysis of this study was the
EAR X PAR results, the EAR X OAR data is also reported in

Appendix F, Tables 14-17.

Factor Analysis

 

The two major DVS of the study, audience credibility
and attitude toward performance, were submitted to a factor
analysis. Results of the factor analyzed audience credi-
bility data suggest that three dimensions of credibility

(character, authoritativeness, and dynamism) existed.

78

Table 1, Appendix F, lists the scales which had factor
loadings above .60 on one factor and below .40 on all other
factors. Six scales: Qualified-—Unqua1ified, Reliable--
Unreliable, Expert-~Inexpert, Informed--Uninformed, Trained--
Untrained, and Experienced--Inexperienced loaded on the
authoritativeness factor. Two scales: Timid——Bold, and
Aggressive--Meek loaded on the dynamism factor. Five
scales: Kind--Cruel, Unpleasant--P1easant, Selfish-~Un-
selfish, Awfu1--Nice, and Friendly-—Unfriendly loaded on
the character factor.

Turning to the second DV, speaker attitude toward
his performance, five scales produced high commonefactor
variance. These sclaes were: Good—-Bad, Well Delivered-—
Poorly Delivered, Beneficial--Harmful, Forceful--Forceless,
and Perfect--Imperfect. Table 2, Appendix F, lists these
scales and their factor loadings.

In summary, the factor analysis reduced the data
to four variables. Three variables, authoritativeness,
character, and dynamism measured the 3s' attitudes toward
the audience. The fourth variable, performance, measured

the 3s' attitudes toward their performances.

Test for Homogeneipy of Variance

As explained in Chapter IV, Bartlett's test for
homogeneity of variance (Winer, 1962) was used to test the

homogeneity assumption. As Shown in Table 3, Appendix F,

 

 

all
chi
the
the
two

the

79

DVS in the EAR X PAR analysis produced nonsignificant
square values on the homogeneity test, indicating that
assumption of homogeneity of variance was tenable for
EAR X PAR data. The OAR analysis, however, contained
measures (character and performance) which did not meet

homogeneity of variance assumption. This raises the

interesting possibility that the apparent selective percep-

tion found earlier in the analysis may account for the

serious imbalance of variance in the EAR X OAR analysis.

The

homogeneity check on OAR also indicated that the deci-

sion to use PAR for the major analysis was a wise one.

Since homogeneity of variance was found in the PAR

results, a parametric analysis of variance and 3 tests were

used to analyze the EAR X PAR data. For the analysis of

the

EAR X OAR data, parametric statistical tests were used

to analyze authoritativeness and dynamism, and alternative

nonparametric tests were used to analyze character and

performance.

Results of the Performance DV

Analysis of Variance Results

A two—way analysis of variance of the EAR X PAR

data revealed a significant effect for PAR (F = 40.19,

p<
p<

.05), a significant EAR X PAR interaction (F = 3.71,

.05), and no significant EAR effect (F = 1.22, p < .30).

 

n.-

 

80

Table 9, Appendix F, presents a summary of the analysis of
variance results for the Performance DV. The next section
will relate the performance results to the hypotheses of

this study.

Performance Hypotheses
I. There will be an interaction between EAR and PAR on
performance ratings by the speaker, to the effect
that: Speakers in the negative EAR condition who
receive negative PAR will rate their own perfor-
mances higher than speakers in the positive EAR
condition who receive negative PAR.

As shown by the analysis of variance, the data
supplied full support for Hypothesis I. First, there was
a significant interaction between EAR and PAR on speaker
performance ratings. While there was overall support for
the interaction, it was also necessary to examine the
second part of Hypothesis I for an explanation of the
interaction. It must be stressed that unless the explan-
atory portion of the hypothesis is supported, overall sup-
port for the interaction cannot be claimed.

Results for the second portion of Hypothesis I
showed that 3s in the negative EAR, negative PAR condition
produced a mean performance rating of 18.18, compared to
14.00 for SS in the positive EAR, negative PAR condition.
As shown in Table 4, Appendix F, a 3 test of the difference
between the two means produced a 3 of 1.84 (p < .05), indi-

cating full support for the explanatory portion of Hypo-

thesis 1. Thus, the first hypothesis, which predicted and

 

 

81

explained an interaction between EAR and PAR, was strongly

supported by the data.

II. Speakers who receive positive PAR will rate their
own performances higher than Speakers who receive

negative PAR.
Hypothesis II was tested by observing the main

r..- _<
i ‘7
‘ rm

effect for PAR in the analysis of variance of speaker per-

formance ratings. As reported in the analysis of variance

'1‘. _..= ‘_‘ I ‘r

results, there was a significant PAR effect for the perfor-

mance DV. It must be noted that since there was a signif-

icant interaction between EAR and PAR, a careful analysis

of the PAR main effect was necessary. A close examination

of the means in the four experimental conditions indicated

that the interaction did not appear to confound the main

effect for PAR. Therefore, Hypothesis II was strongly

supported.

Results of the Audience Credibilipy DV

Analysis of Variance Results
An analysis of variance was computed for the results

of each of the three credibility DVS (character, authorita-

tiveness, and dynamism). The character analysis of variance

(summarized in Table 6, Appendix F) revealed a significant
PAR effect (P = 67.53, p < .05), no significant EAR effect

(P = 1.22, p < .30), and no significant EAR X PAR interac-

tion (F = 0.01, p > .90).

82

The authoritativeness analysis of variance (sum—

marized in Table 7, Appendix F) also showed a significant

PAR effect (F = 12.47, p < .05), no significant effect for

EAR (F = 1.28, p < .30), and a lack of significance for the

EAR X PAR interaction (F = 0.37, p < .60).

The dynamism analysis of variance (found in Table 8, E

Appendix F), revealed no significant effect for PAR (F

0.21, p < .70), no significant effect for EAR (F = 1.28,

p < .30) and no significant interaction between EAR and PAR I."
(F: 1.28, P < .30).

CredibilipyVHypotheses

III. There will be an interaction between EAR and PAR on

audience character and authoritativeness ratings by
the source, to the effect that:

Speakers in the positive EAR condition who receive
negative PAR will rate the audience higher in char-
acter and authoritativeness than speakers in the
negative EAR condition who receive negative PAR.

Support for Hypothesis III was tested by a two-step

analysis. First, an analysis of variance tested for an

overall interaction. Second, a p_test for significance of

differences between the positive EAR, negative PAR condi-
tion and the negative EAR, negative PAR condition was com-

puted to test the explanatory portion of the interaction

hypothesis.

The audience character ratings did not support

Hypothesis III. The analysis of variance for interaction

revealed no significant interaction between EAR and PAR on

character ratings. In the explanatory analysis of the

83

interaction, 3s in the positive EAR, negative PAR condition
and in the negative EAR, negative PAR condition produced
mean character ratings of 16.42 and 14.90, respectively.
While the difference between the two means was in the direc-
tion predicted, a‘E test revealed no significant difference
between the group means (E.= 0.61, p < .30). F
The authoritativeness ratings also did not support t
Hypothesis III. The analysis of variance showed no signif— é
icant interaction between EAR and PAR. 3s in the positive ’
EAR, negative PAR condition produced a mean authoritative-
ness rating of 23.57, while 3s in the negative EAR, nega-
tive PAR condition produced a mean authoritativeness rating
of 20.72. As in character ratings, the differences in
authoritativeness were in the direction predicted, but were
not significant (3 = 1.06, p < .20)-
Therefore, no support was found for Hypothesis III,
either in character ratings, or in authoritativeness ratings.
IV. Speakers who receive positive PAR will rate the
audience higher in character and authoritativeness
than speakers who receive negative PAR.
Hypothesis IV was tested by observing the main
- effect for PAR in the analysis of variance of character and
authoritativeness ratings. Results on both character and
authoritativeness ratings supported this hypothesis. No
significant PAR effect was predicted for the dynamism dimen-
sion of credibility, and no PAR effect was observed for

dynamism. Therefore, Hypothesis IV was strongly supported

by the results of this study.

84

Supplementary Analysis: SEX X PAR

Although the 3s had been assigned to groups with
sex equalized in the experimental conditions, it was de-
cided to re-analyze the data in terms of SEX X PAR. This
was done because some 3s were reassigned as a result of the
PAR measurement and because some 35 were lost due to

attrition.

A two—way analysis of variance for SEX X PAR re-

 

vealed a significant PAR effect (p < .05) for character,
authoritativeness, and performance ratings. There was no
significant PAR effect for dynamism (p < .80). No signif-
icant SEX effects were found for any of the DVS, and no
Significant interaction effects were discovered in any of
the analyses. The fact that no significant sex or interac-
tion effects were found suggests that sex was probably not

a confounding variable in the results.

Summary of the Results

1. Strong support was found in the results for
Hypothesis I. First, there was a significant interaction
between EAR and PAR on performance ratings by the Speaker.

Second, speakers in the negative EAR condition who
received negative PAR rated their own performances signif-
icantly higher than speakers in the positive EAR condition

who received negative PAR.

85

2. Hypothesis II was also strongly supported in
this study. Speakers who received positive PAR rated their
own performances significantly higher than speakers who
received negative PAR.

3. The data did not support Hypothesis III. No
significant interaction was found between EAR and PAR on
audience character or authoritativeness ratings by the
source.

Also, speakers in the positive EAR condition who

 

received negative PAR did not rate the audience signif-
icantly higher in character and authoritativeness than
speakers in the negative EAR condition who received nega-
tive PAR.

4. Hypothesis IV was solidly supported by the
data. Speakers who received positive PAR rated the audience
significantly higher in character and authoritativeness
than speakers who received negative PAR.

5. A supplementary analysis for SEX X PAR showed
a highly significant PAR effect, no significant interaction
between SEX and PAR, and no significant difference between
males and females who received negative and positive re-

sponse from the audience.

CHAPTER VI

DISCUSSION

This final chapter will discuss the results of this
study in light of balance theory, practical applications of

feedback, and future research on feedback.

Balance Theory

 

The experiment described in this thesis tested the
predictive ability of balance theory in a communication
feedback setting. Public speakers were placed in four
experimental conditions. Two of these conditions were
designed to create cognitive imbalance, while two others
were designed to maintain cognitive balance. The speakers'
attitudes were measured in order to test the predictions
made from balance theory. The results of each of the four
experimental communication settings will be discussed below,
along with their support or lack of support for balance
theory.

The first experimental condition involved speakers
who expected positive audience response and received posi-
tive response as they spoke. It was assumed that this
situation would produce no cognitive imbalance for the

86

 

87

speakers, and consequently their performance ratings and
audience ratings would be strongly positive. This predic-
tion was supported. First, the speakers' performance
ratings and their attitudes toward the audience were higher
than those from speakers in the other three groups. Second,
all measures were consistently on the positive side of the
midpoint. Third, the speakers who expected and received
positive audience response showed significantly higher

performance ratings, audience character ratings, and au-

 

dience authoritativeness ratings than speakers who expected
positive response and received negative response (Hypo-
theses I and IV). Therefore, the outcome of the first
experimental condition was successfully predicted by balance
theory.

The second experimental condition involved Speakers
who expected negative audience response but perceived the
response as positive. According to balance theory, this
was a cognitively inconsistent situation. Speakers could
restore balance either by improving their attitudes toward
the audience or by derogating their own performances. It
was predicted that the speakers would resolve the imbalance
by rating the audience positively. It was reasoned that
since the source perceived his effort as successful, he
would not likely derogate his performance, but would in-
crease his esteem of the audience. The results supported

this prediction. The audience character and

88

authoritativeness ratings were well above the midpoint on
the credibility scales. The audience ratings were signif-
icantly higher when speakers expected negative response and
received positive response than when the speakers expected
negative response and received negative response (Hypo-

thesis IV). Also, the audience credibility ratings from E
speakers who expected negative audience response and re- 1

-1“: '.'I h.

ceived positive response were not significantly different

'1- '1-

from audience ratings from speakers who expected positive '
response and received positive response. Therefore, in the
second experimental situation, solid support was found for
balance theory's ability to predict the cognitive outcomes.
Speakers in the third experimental condition ex-
pected positive audience response but received a negative
reaction from the audience. It was predicted that this
situation would be inconsistent. The speakers in this
condition could resolve the situation either by derogating
the audience or by devaluing their performances. As dis-
cussed in Chapter III, both audience derogation and perfor-
mance derogation were expected in this condition. This
seeming contradiction was attributed to differences in
self-esteem. It was reasoned that speakers with high self-
esteem would restore balance by devaluing the audience,
while speakers with low self-esteem would resolve the sit-

uation by derogating their performances. It was predicted,

89

however, that the overall outcome would be a slightly
stronger derogation of performance.

The results indicated that the speakers did down-
grade both their performances and the audience. Speakers

who perceived negative responses rated their performances

 

significantly lower than speakers who perceived positive a;
responses (Hypothesis II). Also, speakers who expected I
positive responses and received negative responses rated

their performances significantly lower than speakers who i

expected negative responses and received negative
responses (Hypothesis I).

While the speaker performance ratings were low in
this experimental setting, the speakers also derogated the
audience. Speakers rated their performances six points
below the midpoint on performance scales, while they rated
the audience only 3.58 points below the midpoint on char-
acter ratings and 0.43 points below the midpoint on author-
itativeness ratings. Thus, speakers derogated their per-
formances more than they devalued the audience, supporting
the balance theory prediction.

The fourth experimental setting involved speakers
who expected a negative audience response and perceived a
negative response as they spoke. It was predicted that
this setting would be balanced, and that speakers would
give comparitively high performance ratings and low au-

dience ratings. The results supported this prediction.

90

The speakers who expected negative response and received
negative response rated their performances significantly
higher than speakers who expected positive response and
received negative response (Hypothesis I). The speaker
performance ratings in this fourth condition were 1.82
points below the midpoint. They were weakly negative,
rather than weakly positive, as predicted.

The audience ratings in the fourth experimental
condition were low as predicted. Audience character and
authoritativeness ratings from speakers who received nega-
tive response were significantly lower than character and
authoritativeness ratings from speakers who received posi-
tive response (Hypothesis IV). Audience ratings from
speakers who expected negative response and received nega-
tive response were also lower than those from speakers who
expected positive response and received negative response.
However, there was not a significant difference between the
two groups (Hypothesis III).

Therefore, the outcome of the fourth experimental
situation was predicted reasonably well by balance theory.
While the differences predicted were not significant in
all cases, they were consistent in direction and pattern.

In summary, the results allow two generalizations
about the experimental situations that were designed to
create cognitive imbalance. First, the speakers who ex-

pected positive response but received negative response

van:
0

'..' 'afin:

"" '- I

w

91

restored balance primarily by devaluing their performances.
Second, the speakers who expected negative response but
received positive response restored balance by raising
their evaluations of the audience.

This discussion has attempted to relate the findings

of this study back to the theoretical rationale used to

. r hie" a")?!

make the predictions. Balance theory was generally success-

ful in predicting attitudinal outcomes in the four communi- t

 

cation feedback situations.

Applications of Feedback

 

The Communication Source

It is generally accepted among communication
theorists that in order to successfully persuade, the
communication source must maintain a relatively high degree
of rapport or credibility with his audience. If the Speaker
perceives, through negative audience response, that his
message is not successful, he will develop (according to
the results of this study) a negative attitude toward his
performance, toward the audience, or toward both. Assuming
the speaker develops a negative attitude toward his perfor-
mance and the audience, it will be difficult for him to
build up his credibility and accomplish his purpose.

The results of this study suggest that a speaker
must either determine how to avoid negative audience re-

aponse or overcome the tendency to derogate when negative

92

response is given. The rhetorically skilled speaker who
receives negative responses will modify his persuasive
strategy, or if necessary, his persuasive purpose. He will
adapt his message to the receiver. He will plan alterna—
tive means for accomplishing his persuasive goals (Williams,

1964), and will use these means when faced with negative

 

rSSponse. Thus, a negative message from the receiver will i ‘
initiate an adjustment in the source's communication strat— ;
egy, and he will produce a new input designed to better i

accomplish his purpose.

The Communication Receiver

The communication receiver also must realize the
important role he plays in the feedback cycle. Negative
and positive response, initiated by the receiver, can have
strikingly differential effects on the attitudes of the
communication source (e.g., when the receiver acts as
evaluator). If the results of this study can be applied
to speech evaluation in the classroom, it is possible that
negative criticism for the novice speaker may produce nega-
tive attitudes toward his speaking and toward the evaluator.
Conversely, positive criticism may generate positive atti-
tudes toward speaking and toward the individual evaluating
the speech. Other studies (e.g., Bostrom, 1964) dealing

with this problem have reached similar conclusions.

93

In summary, the feedback phenomenon can be applied
from two perspectives. First, the speaker must realize the
possible impact of audience response on his communication
attitudes and be prepared to cope with this eventuality.
Second, the receiver who acts as evaluator should understand
the potential negative or positive effects he can have on

the communication source.

Limitations and Future Research

 

This final section will simultaneously deal with
the study's limitations and suggest ideas and procedures
for future research on feedback.

The first improvement suggested for future research
is a more rigorous induction of EAR. NA few 3a in this
study reported having doubts about the validity of the
"Audience Profile." When questioned as a group, however,
over 75% of the 3s believed that the profile was the actual
result of an audience survey. There are a number of pos-
sible methods for strengthening the induction of EAR. One
possibility is measuring the audience's attitude in the
presence of 3, This can be done either orally, with brief
written scales, or by a nonverbal vote. The audience can
be instructed on how to respond before 3 arrives. This
induction would likely be more believable than the one

employed in the present study.

 

94

Future research on feedback should also eliminate
potentially confounding variables that may contaminate OAR.
When audience response is administered by a live audience,
as in the present study, a number of potentially confounding
variables are present. Some of these are: 1) Different
movements and expressions by the 3s between treatments.

2) Different movements and expressions by the 3s within
treatments. 3) Possible failure of the gs to execute OAR
as instructed. 4) Fatigue on the part of the 3S as the
experimental session progresses. 5) Failure of both 3s and
33 to distinguish between person-oriented and content-
oriented responses. 6) Different clothing worn by 3s
during different experimental sessions. 7) The tendency
for the ES to give subconscious approval cues during dis-
approval conditions. While the research method employed
was designed to control the above variables, absolute con-
trol of such elements probably is impossible.

Three possible alternatives for controlling OAR
are suggested. First, the 3s can be trained to act with
robot-like precision in each positive and negative condi-
tion, yet appear as natural as possible. Second, the re-
sponse channel can be limited to verbal-oral-interposed,
by placing a barrier between the speaker and the audience
so that the audience can respond only verbally. Third, 3

can limit the response channel to nonverbal-interposed,

 

95

by manipulating the response through nonverbal means such
as lights (e.g., Huenergardt, 1967).

While the above alternatives are not ideal, they
are a step toward controlling a number of the confounding
variables that possibly influence the results of feedback
studies. In each case, however, the increased control is
accompanied by increased movement from a "real life" com-
munication situation, and greater difficulty in generalizing
the results.

A third area where improvement is needed is pre-
dictability. While balance theory was quite successful
for predicting the outcome of this study, the predictions
could have been more precise, had a more thorough knowledge
of the personality characteristics of the 3s been available.
Future research should study possible interactions between
feedback and such personality variables as self-esteem,
need-affiliation, need-influence, need—achievement, intro-
version, dogmatism, and authoritarianism.

Fourth, a variety of independent variables should
be tested with feedback to study the possible interaction
effects. Some of these variables are: ego involvement
with the topic, familiarity with the audience, prior
credibility of the audience, delayed vs. concurrent re-
sponse, personal- vs. message- vs. topic-oriented responses,
sex, and amount of effort involved in preparing for the

performance.

 

 

96

Finally, and most important, future research should
examine the total feedback cycle as it relates to communi-
cation behavior. The sophisticated feedback study should
observe, control, or manipulate: a) the original input by
the source, b) the consequent adjustment by the receiver,

c) the output by the receiver, d) the consequent adjustment ED

..<..

in the source's message, purpose, strategy, performance,

attitudes, and finally e) the source's new input, which

“mums.“ ‘ -.

marks the completion of a single feedback cycle. This type
of research necessarily is microScopic, and requires care-
ful control. However, it must be done before feedback can
take its place as a part of what will some day become

formalized communication theory.

REFERENCES

REFERENCES

Amato, P. P. Potential utility of automated instruction,
feedback, and auxiliary feedback devices in speech
education. Unpublished thesis, Emerson College,
1961.

Amato, P. P., & Ostermeier, T. H. The effect of audience
feedback on the beginning public speaker. The
Speech Teacher, 1967, 13, 56-60.

 

Andersch, E. G., & Staats, L. C. 3peech for everyday use.
New York: Holt, Rinehart, & Winston, 1960.

 

Anderson, B. F. The_psycholggy experiment; an introduction
to the scientific method. Belmont, CalifBrnia:
WadSWorth, 1966.

 

Ashby, W. R. An introduction to cybernetics. New York:
Wiley, 1956.

Barwind, J. A. The effects of varied positive and negative
audience feedback on the attitudes and behavior of
skilled and unskilled speakers. Doctoral disserta-
tion in progress, Bowling Green State University,
1969.

Berger, C. The effects of influence feedback and need
influence on the relationship between incentive
magnitude and attitude change. Unpublished doc-
toral dissertation, Michigan State University,
1968.

Berlo, D. K. The process of communication; an introduction
_. m
to theory and practice. New York: Holt, Rinehart,

& Winston, 1960.

 

Berlo, D. K., Lemert, J. B., & Mertz, R. J. Dimensions for
evaluating the acceptability of message sources.
Unpublished manuscript, Michigan State University,
1966.

Bettinghaus, E. P. Persuasive communication. New York:
Holt, Rinehart, & Winston, 1968.

97

98

Blubaugh, J. A. The effects of positive and negative au-
dience feedback on selected variables of speech
behavior of normal-speaking college students.
Unpublished doctoral dissertation, University of
Kansas, 1966.

Blubaugh, J. A. Effects of positive and negative audience
feedback on selected variables of speech behavior.
Paper presented at the meeting of the Speech Asso-
ciation of America, Chicago, December 1968.

Boneau, C. A. The effects of violations of assumptions
underlying the E_test. Psychological Bulletin,
1960' 27., 49-640

Bostrom, R. N. Classroom criticism and speech attitudes.
Central States Speech Journal, 1963, 13, 27—32.

Bostrom, R. N., Vlandis, J. W., & Rosenbaum, M. E. Grades
as reinforcing contingencies and attitude change.
Journal of Educational Psychology, 1961, 33, 112-
115.

Buckhout, R., & Rosenberg, M. J. Verbal reinforcement and
attitude change. Psycholggical Reports, 1966, 18,
691-694. '—

Cahn, D. D. The effect of simulated audience response on
speaker attitudes. Unpublished thesis, Northwestern

University, 1966.

Cameron, H. K. A study of the effectiveness of feedback in
teaching principles of educational psychology.
Unpublished doctoral dissertation, Michigan State
University, 1963.

Cherry, C. On human communication; a review, a surve , &
a criticism (2nd ed.). Cambridge: M. I. T. Press,
1966.

 

Cieutat, V. J. Sex differences and reinforcement in the
conditioning and extinction of conversational be-
havior. Psychological Reports, 1962, 13, 467-474.

Clevenger, T. Audience analysis. New York: Bobbs-
Merrill, 1966.

99

Cohen, A. I. Changes in self concept evaluations as a
function of expectancy and experimentaIl induced
success, failupe and neutral conditions (DoctoraI
dissertation, SouthernEIIIinois University). Ann
Arbor, Mich.: University Microfilms, 1960. No.
61-2080.

Combs, W., S Miller, G. R. The effect of audience feedback
on the beginning public Speaker, continued: a
counter view. The 3peech Teacher, 1968, 11, 229-
231.

Eisenson, J-r Auer, J. J., & Irwin, J. V. The psychology
9f gommunication. New York: Appleton—Century-‘"
Crofts, 1963.

 

Engbretson, R. 0. Cognitive adjustment of perceived self
credibility, perceived source credibility, and per-
ceived task difficulty as a result of feedback in
task oriented dyads. Unpublished doctoral disser-
tation, Michigan State University, 1964.

Faules, D. F. Communicator ability and the ability to
elicit and interpret feedback under four conditions
of stimuli. Unpublished doctoral dissertation,
Ohio State University, 1963.

Faules, D. F. The relation of communicator skill to the
ability to elicit and interpret feedback under four
conditions. Journal of Communication, 1967, 17,
362-371. 7‘

Feezel, J. D. The effects of preparation time and feedback
upon speech hesitations. Unpublished thesis, Uni-
versity of Wisconsin, 1965.

Festinger, L. A theory of cpgnitive dissonance. Stanford:
Stanford Uhiversity Press, 1957.

French, E. G. Effects of the interaction of feedback and
motivation on task performance. American Psychol-
ogist, 1956, 11, 395.

 

Frye, R. L. The effect of feedback of success and effec-
tiveness on self, task, andinteractiop oriented
gppup members (Doctoral dissertation, Lofiisiana
State University). Ann Arbor, Mich.: University
Microfilms, 1961. No. 62-46.

 

T-‘ffi-fi—R’TV— 31h. rm..- ga-

100

Frye, R. L. The effect of orientation and feedback of
success and effectiveness on the attractiveness
and esteem of the group. Journal of Social Psye
chology, 1966, 13, 205—211.

Gardiner, J. C. Effort, audience feedback and attitude
change. Unpublished manuscript, Michigan State
University, 1968.

Gergen, K. J. The effects of interaction goals and per-
sonalistic feedback on the presentation of self.
Journal of Personalipy & Social Psychology, 1965,
1, 413—424.

Gibb, J. R., Smith, E. E., & Roberts, A. H. Effects of
positive and negative feedback upon defensive be-
havior in small problem-solving groups.) American
Psyphologist, 1955, 13, 335.

Goldstein, I., & McGinnies, E. Compliance and attitude
change under conditions of differential social
reinforcement. Journal of Abnormal and Social
Psychology, 1964, 33, 567-570.

Gray, G. W., & Wise, C. M. The bases of speech (3rd ed.).
New York: Harper & Row, 1959.

Greenberg, B. S., & Tannenbaum, P. H. Communicator per-
formance under cognitive stress. Journalism
Quarterly, 1962, 33, 169-178.

(J!

=~'.

-d‘

Greenspoon, J. The reinforcing effect of two spoken sounds
on the frequency of two responses. American Journal

Guilford, J. P. Psychometric methods (2nd ed.). New York:
McGraw-Hill,gl954.

Harvey, O. J., Kelley, H. H., & Shapiro, M. M. Reactions
to unfavorable evaluations of the self made by
other persons. Journal of Personality, 1957, 13,
393-411.

 

Heider, F. Attitudes and cognitive organization. Journal
of Psychology, 1946, 31, 107-112.

Heider, F. The psychology of interpersonal relations.
New York: Wiley, 1958.

101

Hildum, D. C., & Brown, R. W. Verbal reinforcement and
interviewer bias. Journal of Abnormal and Social
Psychology, 1956, 33,108-111.

Hill, K. T. The effects of social reinforcement and non-
reinforcement following success and failure as a
function of test anxiety of sub'ect, sex of sfib-
jéct, and sex of experimenter Doctoral disserta-
tion, University of Minnesota). Ann Arbor, Mich.:
University Microfilms, 1965. No. 65-15,265.

Huenergardt, D. W. An experimental stugy of the effects of
increasing percentages of Simultaneous ppncontln ent
audience approval on speaker attltudes (Doctora
dissertation, NorthWestern Universityl. Ann Arbor,
Mich.: University Microfilms, 1967. No. 68-3192.

Huenergardt, D. W. Effects of audience response on speaker
attitudes. Paper presented at the meeting of the
Speech Association of America, Los Angeles,
December 1967.

Hylton, C. The effects of observable audience response on
attitude change and source credibility. Unpublished
doctoral dissertation, Michigan State University,
1968.

Insko, C. A. Verbal reinforcement of attitude. Journal
of Personality and Social Psychology: 1965, Z, 621-
623.

Insko, C. A. Theories of attitude change. New York:
Appleton-Century—Crofts, 1967.

Johnson, P. C., & Klare, G. R. Feedback: principles and
analogies. Journal of Communication, 1962, 13,

Jones, E. E., & Gerard, H. B. Foundations of social_psy-
chology. New York: Wiley, 1967.

Jones, E. E., Gergen, K. J., & Davis, K. E. Some determin-
ants of reactions to being approved or disapproved
as a person. Psychological Monographs, 1962, 76
(Whole No. 521). '_'

Kaplan, A. The conduct pf inquipy: methodology for behav-
ioral science. San Francisco: Chandler, 1964.

102

Karns, C. F. The verbal behavior of a speaker as a function
of certain non-verbal aversive stimuli presented by
an audience in a public speaking situation. Unpub-
lished doctoral dissertation, University of Pitts-
burgh, 1964.

Kerlinger, F. N. Foundations of behavioral research; edu-
cational and psychological Inquiry. New York:
Ho t, Rine art, an Winston, .

Krasner, L. Studies of the conditioning of verbal behavior.
Psychological Bulletin, 1958, 33, 148-170.

Krasner, L., Knowles, J. B., & Ullmann, L. P. Effect of
verbal conditioning of attitudes on subsequent
motor performance. Journal of Personalipy & Social
Psychology, 1965, 1,407-4l2.

Krasner, L., Ullmann, L. P., & Fisher, D. Changes in per-
formance as related to verbal conditioning of atti-
tudes toward the examiner. Perceppual & Motor
Skills, 1964, 13, 811-816.

Krauss, R. M., & Weinheimer, S. Concurrent feedback, con-
firmation, and the encoding of referents in verbal
communication. Journal of Personality & Social
Psychology, 1966, 3"3433346.

 

Leavitt, H. J., & Mueller, R. A. Some effects of feedback
on communication. Human Relations, 1951, 3, 401-
410.

 

Lewin, A. Y. The effects of success and failure and task
responsibility on the level of aspiration and per-
formance. Unpublished doctoral dissertation,
Carnegie-Mellon University, 1968.

Lewis, J. D. Feedback in mass communication: its nature
and use in decision making. Unpublished doctoral
dissertation, Michigan State University, 1966.

Lott, A. J., Schopler, J. H., & Gibb, J. R. Effects of
feeling-oriented and task-oriented feedback upon

defensive behavior in small problem-solving groups.

American Psychologist, 1955, 13, 335-336.

Lucas, C. A. Task performance and group structure as a
function of personalipyg3nd relevance of feedback

(Doctoral dissertation, Syracuse Unlversity). Ann

Arbor, Mich.: University Microfilms, 1963. No.
64-7466.

 

103

Lyon, C. E. Audience consciousness. QuarterlyJournal of

 

Marshall, R. J. Variation in self-attitudes and attitudes
toward others as a fhnctibn of peer group appraisals
(Doctoral dissertation, UniverSity of‘Buffan).

Ann Arbor, Mich.: University Microfilms, 1958.
No. 58—1947.

Mattox, P. R. An experimental studyyof the effect of lis-
tener feedback on speaker attitude (Doctoral dis-
sertatibn, University of Southern California).

Ann Arbor, Mich.: University Microfilms, 1967.
No. 68-1689.

:3?

1"

McCroskey, J. C. Scales for the measurement of ethos.
Speech Monographs, 1966, 23, 65-72.

 

what-.- 1:'-_;i’u‘--.'. E.

McCroskey, J. C. An introduction to rhetorical communica-
tion: the theory and_practice of_public speaking.
EngIeWOOd Cliffs: Prentice-Hall, I968.

McNeil, J. An experimental effort to improve instruction
through visual feedback. Journal of Educational
Research, §§, 1962, 283-285.

McNemar, Q. Psychological statistics (3rd ed.). New York:
Wiley, 1962.

Mehrley, R. S., & Anderson, L. An experimental study of
feedback. Paper presented at the meeting of the

Speech Association of America, Chicago, December
1968.

Mikawa, J. K. The effect of affiliation need and approval-
rgiection feedback on feelings of’hostility toward
self and others (Doctoralidissertation, University
of_Texasi. Ann Arbor, Mich.: University Micro-
films, 1963. No. 63-5252.

 

Miller, G. A., Galanter, E., & Pribram, K. H. Plans and

the structure of behavior. New York: Henry Holt,
IQGO.

 

 

Miller, G. R. The effect of varied incentive conditions on
the oral presentation of a second speaker. Unpub-
lished doctoral dissertation, University of Iowa,
1961.

104

Miller, G. R. Variations in the verbal behavior of a
second speaker as a function of varying audience
reSponses. Speech Monographs, 1964, 2;, 109-115.

Miller, G. R., Zavos, H., Vlandis, J. W., & Rosenbaum,
M. E. The effect of differential reward on speech
patterns. Speech Monographs, 1961, SS, 9-15.

Molof, M. J.- The effects of success versus failure and
continual versus terminal feedback on_judgment in
small, decision—making groups (Doctorai_disserta-
tion, Stanford University). Ann Arbor, Mich.:
University Microfilms, 1963. No. 64—5585.

Morris, J. Nonparametric chi-square tests and analyses of
variance. Technical Report No. 42. Computer In—
stitute for Social Science Research. Michigan
State University, 1966.

Morris, J. Mann-Whitney and Wilcoxon tests, randomization
and sign tests, parametric t-test and F-test.
Technical Report No. 45. Computer Institute for
Social Science Research. Michigan State University,
1967.

Mulac, A. An experimental study of the relative effective—
ness of three feedback mechanisms for teaching
fundamentals of speech. Unpublished doctoral dis-
sertation, University of Michigan, 1968.

Newcomb, T. M. An approach to the study of communicative
acts. Psychological Review, 1953, SS, 393—404.

Newcomb, T. M. Individual systems of orientation. In S.
Koch (Ed.), Psychologyi_ a study of a science.
Vol. 3. New York: McGraw-Hill, 1959. pp. 384-422.

Osgood, C. E., Suci, G. J., & Tannenbaum, P. H. The meas-
urement of meani_g. Urbana: The University of
Illinois Press, 1957.

 

Ostermeier, T. H., & Amato, P. P. The effect of audience
feedback on the beginning public speaker, continued:
a rejoinder to Combs and Miller. The Speech Teacher,
1968, ll, 231-234.

Overstreet, H. A. Influencingphuman behavior. New York:
Norton, 1925.

105

Powers, W. T., Clark, R. K., & McFarland, R. L. A general
feedback theory of human behavior. Perceptual and
Motor Skills, 1960, ll, 71—88, 309-323.

 

Pryer, M. W., & Bass, B. M. Some effects of feedback on
behavior in groups. Sociometry, 1959, SS, 56-63.

 

Reddy, W. B. The effects of immediate and delayed feedback
on the learning of empathy YDoctoral dissertation,
University of Cincinnati). Ann Arbor, Mich.:
University Microfilms, 1968. No. 68—15,l90.

P‘

Ring, K., & Kelley, H. H. A comparison of augmentation and
reduction as modes of influence. Journal of Abnor-
mal and Social Psychology, 1963, SS, 95-102:

 

. -..‘.-_- tr uh.- ._ a (in; u.
p —' 1

Rosnow, R. L. "Conditioning" the direction of opinion
change in persuasive communication. Journal of
Social Psyghology, 1966, SS, 291-303.

mm .-_

 

Ruble, W. L. Analysis of covariance and analysis of var-
iance with unequal frequencies permitted in the
cells--(LS routine). STAT Series Description No.
18. Michigan State University Agricultural Exper-
iment Station, 1968.

Salzinger, K. Experimental manipulation of verbal behavior:
a review. Journal of General Psychology, 1959, 61,
65—94. '—

Sarbin, T. R., & Allen, V. L. Role enactment, audience
feedback, and attitude change. Sociometry, 1964,
S1, 183-193.

 

Scott, W. A. Attitude change through reward of verbal
behavior. Journal of Abnormal & Social Psycholpgy,
1957' -5_5’ 72-750

Scott, W. A. Cognitive consistency, response reinforcement,
and attitude change. Sociometry, 1959, S3, 219-229.

(a)

Scott, W. A. Attitude change by response reinforcement:
replication and extension. Sociometry, 1959, 22,
328-335. (b) —'

Sereno, K. K. Changes in verbal behavior of a speaker
during two successive speech performances as a
function of the sequence of listener responses.
Unpublished doctoral dissertation, University of
Washington, 1964.

106

Siegel, S. Non—parametric statistics for the behavioral
sciences. New York: McGrawIHiII, 19562

Smith, E. E., & Kight, S. S. Effects of feedback on insight
and problem solving efficiency in training groups.
Journal of Applied Psyghology, 1959, 2;, 209—211.

Squier, R. W., Jr. An experimental investigation of a
feedback procedure fbr‘increasing understanding of
other pegple (Doctoral dissertation, Stanford Uni-
versity). Ann Arbor, Mich.: University Microfilms,
1964. No. 65-2901.

Stolz, W., & Tannenbaum, P. H. Effects of feedback on oral
encoding behaviour. Language & Speech, 1963!.ér
218-228.

 

Stotland, E. The effects of public and private failure on
self-evaluation. American Psychologist, 1956, 11,
357. ’-

Tuckman, B. W., & Oliver, W. F. Effectiveness of feedback
to teachers as a function of source. Journal of
Educational Psychology, 1968, SS, 297-301.

 

Verplanck, W. S. The control of the content of conversa-
tion: reinforcement of statements of opinion.
Journal of Abnormal & Social Psychology, 1955, 51,
668-676. ‘—

Videbeck, R., & Bates, H. D. Verbal conditioning by a

simulated experimenter. Psychological Record,
1966, 1S, 145-152.

Vlandis, J. Variation in the verbal behavior of a speaker
as a function of varied reinforcing conditions.
Speech Monographs, 1964, Si, 116-119.

Wallace, J. Role reward and dissonance reduction. Journal

of Personality & Social Psychology, 1966, S, 305-
3l2.

Welke, J. W. The effects of intensional and extensional
audiences on communicator anxigtyi an experimental
study (Doctoral dissertation, Indiana University).
Ann Arbor, Mich.: University Microfilms, 1967.

No. 67-15,175.

107

Westley, B. H., & MacLean, M. S. A conceptual model for
communications research. Journalism_guarterly,
1957, SA, 31-38.

Wiener, N. Cybernetics; or, control & communication in the
animal & the machihe (2nd ed.). New York: M. I. T.
Press, I961.

Wiener, N. The human use of human beings; cybernetics &
society. Boston: Houghton MiffIin,_1954.

Winer, B. J. Statistical_princip1es in experimental design.
New York: McGraw-Hill, I962.

Williams, A. Factor analysis. Technical Report No. 34.
Computer Institute for Social Science Research.
Michigan State University, 1967.

Williams, K. R. Audience analysis and adaptation: a con-
ceptual clarification. Unpublished doctoral dis-
sertation, Pennsylvania State University, 1964.

Wilson, K. V. A distribution-free test of analysis of
variance hypotheses. Psychological Bulletin, 1956,
13'." 96-101.

APPENDICES

108

rl-4‘N!4qnm ..... ..

 

 

 

 

 

 

aocmwowm
1mm xmmu xomncmmm .mmoumoum xmmu nmmmHP
«mmcflamwm UmusmflHOIxmmu Hams» mcwumowpsfl .8H0m ammum fiH .Hm um
o>wmcommn .m> Imswaoom cowumeuomsfl so>wm wumz mmsoum ca mm Inn: .uuoq
.msdo: amudam
mason mcflms How ucmfimouomcflmu A=ns|£sn:v Dame Ammmav
Hondam no .| o>wymmms no A=ESSIEEV w>fluwmom noouom coomm
mosmsvmum III: m o>mm m .mcfluuom 3ma>uousfl cm cH icflmm Icmouw
monoflowm vmusmano l
1mm xmmu Ixmmu .m> .GOADMDDHm xmmu msoum M GA mm AMMQHP
“mmcwamom Imsfiaomm ou mcowumnam>m 0>Hummms cam w>wuflmom .Hm no
0>wmcmmma “mm mo mama cmumumflswficm wum>nomno macho cmcwmua III: .nnflw
sown
Imwfimsouu .omcommou moan
mommmoe .maso oslmwm .maco Hmnum>sos .wmsommmu Aammav
mo mafia a ouou "msowuwpcoo mm v Home: oosmwwfim Hwaamsz
homusood III: so on msumuumm oauuofiomm pmnflnommc mm III: .uuw>mmq
m>o m>H mmmBO aomamz wmomma Mosam
mudmammm zo mmHQDBm mom<mmmm mo NMdzsz A¢UHUOqoaomBmz 4

4 NHQZHQQd

109

 

mumom a maom

am an poumasmflcmfi .mummm Eoum

 

 

 

 

 

 

vumsou Emoumm mcoflumsamtw mpflHMGOmHmm .>mmss Ho .>mm Ammmav
opsuwuud Imaom pm>floomu mm .soauomnousa msoum a mound uuuuu Hamnmnmz
.m an ombudsmfisme
Damon mums muo> mo muasmmn cam .mumssfls usmfi
cumsop on» so ©ou0> mocowcdm ad .moumnom ImOHOM Aummav
modufluufi IIIII Hmcwwsuaubmnnmussoo :H commmco mm IcHom upoom
mswumn
msH>Hm Hoom mcaumn msfl>am .Homm HowawEmMss no Hmaawfimu m an AbmoHP
mo mcflumu noon spas ayH mGOHpMSHm>oIMHom “soap can» Ho3oa Am hocmu .Hm um
nonflumuumaom Inmaafifimu .mm momma: Honuwm pouch onm3 msoum CH mm Imflmcou .%0>Hmm
m mo .336
mmwuuwmxo mcmE.mo me>HHm ummx unbuwm mos
mGOADMSHm>m .oHsHHmw mo mHSHHmm m.m .m unmmxmcfi no u unomxw Ammmav
Imamm mmosoaandm an xmmu manammomfia so so>Hm 0Hm3 mm IIIII psmeoum
Uflmou. : ogflléz s =©OOOIF magma Awm¢Hv
pnm3ou cpw3 Asoo_h\a .onm N\HV noncommou mm nooHOM s3oum
ocsufluu¢ lllll poouomcflou m .m30H>HmusH msonmoaop sH Inflom .Edpaam
mcowumm .Amocoaflm
.cHE ca no omHmMmfipv .mswmu .mms .cflE oa
m mcwusc .Aommucmmumm Ho omnmmv .msaon .mom .GHE
mquEmumum IQH .Ho>ma usmuomo .cflE oa.lumommum usofi
cowcflmo mo m CH mmcmsmumum sowcamo .mm woouom Im0H0m Ammmav
mososwonm IIIII Inflow mm .mswuuom coHDMmHm>coo m cH Inflow xosmHmuo>
m>o m>H mmmao aomamz wmomma NQDBm

 

110

 

mmcHumu

 

 

 

 

 

 

mmvsn «wEHu
mcmemmm .m Scum AmocmHHmv
mo mouse m can 0 cumzousom no A=poom=v cum3mu om>HoooH a usms AHMmHN
nHumm «much mo pumson mxomm mm con» m Eoum omsommmu msH>Hoo ImoHOH .Hm um
«hosmsHm HmHHsmHommHQ Ion can mcmemmm U m co>nomno umHHm mm Ichm .HmHHHz
.ucofiooHOH
UHQOH IchH on no .Ammcmum =a=v o>Hummms usmE Ahmohk
cum3ou .Ammvmum =¢:v 0>HHHmom sm>Hm mum3 whom Imouom .Hm no
mosuHuud IIIII umm HmchsuHuumnnmusnoo muons 0:3 mm IsHom .Eouumom
hosmu .oocmEHOMHom xmmu HHonu so COHu
msoHumsHm>m nommxo xmmu ImEHOHsH mHSHHmm Ho mmmoosm mSOHun AommHv
IHHom 30H no anm IOHH co>Hm muo3 soHHMSUHm xmmu m sH mm lllll sonoo
uanmsH .HoH>mnmn name AmmmHv
IHHmm «>HH>HD maonw HHmnu co mm cmNHHMGOmHmm Honuo Imogen psmHM
nonpoum mnouw IIIII Sumo m>mm mm .maHuumm UHmHH m CH IsHmm .nuHEm
.ucm: .ouHm
0Hmou Iommo .onH .mompdn m Eonm name name
©Hm3ou Ion mm meow ImoHOMGHmH AumoHv m>Hummmc Ho .Aso3w IoOHOH anmmmav
opsuHuud “GOHuHmom o>HuHmom cm>HmooH oumnov m sH mm IGHmm Huoom
nMIEoum omHmum
UHQOH nuH3 pooH0msHon mums mm .musmfismum aocou AcmmmHv
.63» ocsuHuué IIIII HmchsuHuumuHmucsoo msHuo>HHmc Hound ImHmsoo uuoom
m>a m>H mmmao nomamz mmomma unDBm

 

lll

 

mmhfimmwg

.xomnvmmm AcmucmHHOImsHHmmmv usm>mHmHHH

 

 

 

 

 

 

msoum «ovum mommu muH Ho AcousmHMOIxmmuv usm>onu Hmnun AmmmHv
lauomumm xmma IHMGOmHom co>Hm mums s0Hum5uHm xmmu m sH mm nun: muonq
.Axomncmmm
Héaz vsoanzv ousuomH 6cm Axomnpomm usoosum
“Hdmz «muoom mcHHMSHm>m can mchmunoosmv 0>HumnmmucH AmmmHv
meo Hmch lllll "toummsoo mums mcHzomou mo mpocuoe 039 III: coumfimo
m>Hm
msH Imnmuom no .HoumsHm>m
Ixmmmm pum3ou m>HumEHom mco an hHHmuo EmHoHuHHo m>Hummos no name
.mmosuHubm can IcH “nooomm o>HuHmom HonuHm pm>HmomH 0cm .muoum ImOHOM AmomHv
“mmsHumnlmHmm mo mums IsHm>o mo Hmsmm m on noncommm m>mm mm IcHom Eouumom
.mxmmc HHonu sH coHHmum
muonomou mo nsH msouusn uanHIHosmcm nqu moocm AmmmHV
oosmEHOHHom lllll IEHOMHmm .mumnomou cmumsam>m musmcaum III: HHozoz
mGOHumHHommp .mGOHumHHomoU HHocp nuH3
momusoom commmumfiw mHnmuo>mwcc no thmuo>mw
Ho3oH>HousH .m> momuo HkoHm mm3 m umnu cmEHOHGH umHMH mums
mo msflmumsHm>m nomhm .OH mm .msoHumHHomeIMHom HMUHuHuoomwn usoa AmmmHV
.mm «msoHu .Hs “sMHH no mbmusoom uo£UHo m>Hm on omuosnum IoUHOM .mm um
nmsHm>olemm IH0>MH£omz nsH mums GOHuMDuHm 3mH>HmusH cm sH mm nchm .mos0b
.mm on8
now no mHmE an ucmfimoH0msHmH Am an
QOHpmm Umnosqu w>Hummos Ho «m kg 0» nocmu ucoE
|Ho>soo HMHoom I. ImHHv o>HuHmom Hmnuo>sos co>Hm oum3 nooHOM AmmmHv
mo auHucmso N no xmm mGOHummum>coo mchsommoHIoonm cH mm uchm umucoHU
m>o m>H mmmao Dozens wmomms MQDBm

 

112

.dl ‘ . it I; .
F‘lu if Efkliir’

 

.moschsm Honncoo

 

 

 

 

 

 

 

hosous «onmn m An no .mmmmmmE on» msHpsmumnmcss mo
unconsoo comomm xomH .nsofimmanch cm3onm n0H23 mosoHcfim AemmHv
sH momcmco IIIII posHmnu m Am on mmMmmoE m omno>HHm© mm nnn msnmx
Hmnusms
.coo .onm AvmmHv
0Hmou ”UHQOH .3on maonm on» an3 mammH on» commno name moHacHuoz
cnmson opnanum Ich cosy .msonm HHmEm m on mommmom loonOM .sHonm
mcsann¢ mosoHUD< Hmchsannmnnonssoo m pmnm>HHmc mm Ichm IcHou
mnHHHQHUmno .mmsommon
xmmn .monSOm monSOm Hmnnm> o>fiwwmms no o>HuHmom Umnmu AemmHv
.MHom pnms .muHHHnH IchHEUm on3l.o m on oHNNsm UHnnoEomm com
non mmcannn¢ IcanIMHom m poanommU mm .mGMM© consoHnOIxmmu CH mocmHmm lumnnmsm
.nmuoE QOHsHmo .oon .smw m MH> momcsn
uannm mmmum Eonm Amfims ..mos ..momv mmmcommmn AmmmHv
“mmosUDOH «mGOHu so>Hm mnm3 mm .EoonoucH cm no>o mGOHw Edmnsmcsme
ImuHmon «mEHB IIIII Imwsw ponm3mcm mew Hmno cm mstmu mm lllll .NHoum
.m £HH3 .m an sw>Hm
nsofimmnmm «GOH QOHnmano .AncofismHsamv soHuompon name Ammmav
usHmo mo usmm mo unH no Apnmzonv GOHHmucofimsm nmsonnn H mm noon0m moHHoM
IHmoosoo .mm IHHnHmdem pmonowsHon onus noncommmn soHsHmo .mm IsHom .mch
xomn .m an
Icomm Hm soHumEn0msH onsHHmm no mmmOUSm so>Hm
mmcHumnIMHmm IsHEnmu .m> mnoz cam .mchsnm ommo on mcoHusHOm AmmmHv
«mmanmn msonw Hmsansou muons can commnoch masonm xmmn HHmEm lllll moaoz
muonm cnms GOHH .msonm on» Eonm xomnpomm QOHnoonmn no name
no» mnHHHnmos IMHHHmmm Hm>onmmm so>Hm swan .msonm HHmfim m an noonom AmmmHV
“huHHHumosluHom Icooz nmmn muHHmsomnom m sm>Hm mm3 m comm usHom msmtz
m>n m>H mmmao oomamz wmomme MQDBm

 

113

 

.wm cmnCmHno

.mmoCmEnomnom nHoCu naonw.m an COHum

 

 

 

 

 

 

moCmEnom umCHHoom uEnOHCH qu>mHmn|mCHHoom no qu>onm AmmmHv
Inom xmma .m> Ixmma Ixmmn Co>Hm mno3 moonm xmmu m CH mm IIIII COConm
>0C05HHC0C mo .Comomm m mCHn0>HHoU.m nCmE
mquEoE «moCm Co on nCmEmonOHCHmn A=OC=V m>Hummmm loonOH AvmmHV
unmnus nmuoe ..... no A=noom=v m>HuHmom smnmancHsem m .chm mnocmn>
noHnoHCH .momCommon
nCmnCo .nOHnmmCm AcmmC oxmnmv o>HnmmmC no AUOC a oHHEmV uCoE
Hmnno> "nCmEmonomCH m>HnHmom Cw>Hm mnm3 UCm .oOCoHCCm loonOM AvmmHv
“moCoCHHCoz non no nmpno coCHmnn m on mmCooomm o3n cmnm>HHmc mm ICHom 0Cmnow
.mmmmMoE qumwnomHo
0H .HC ImoHHon cono>HHo© mm mm AHm>onmmm
”Hm>onmmm uch .Eovmnonv mHm>HummoC nol AHm>onm nCmE AvmaHV
UHCOU cnm3on HMHOOm Imm .COHnCounmv >H0>HnHmom mm coonOH loon0m CoHH4
oUCann< now @002 ICHon mnmmm coCHmnn no ooCoHcsm CC ICHom .CHnnmm
nCmECmH l
1Com .pnm3 .m Conn A=0C=v Hm>onmmm
IMn AWEMm Ich Hmnno> no .nCmEEoo 0C .A:Uoom=v
MUCmCHHCOC um a 0 Ho Hm>onmmm Hmnnm> ©m>Hwomn UCm oxomm nCoE
mo moCoCHoCH Unm3mn HMHH mm Conn m Eonn omCommmn mCH>Hoomn noonOH HemmHv
«oumn moCmnmnnD ICmanMHn UCm mCmeomm U m Um>nmmno umnHm mm ICHmm anHHz
mmnCmmoE
nonmEoEMCao .onon m Unmsou momcomm nCmE nwmmww
«OHQOn Unm3 Ion mHnmno>mm mCH>Hm now coonOHCHmn Imonom .Hm no
:0» mUCanu¢ IIIII mHo>HuHmom ono3 mm .3wH>nouCH Cm CH ICHmm .nmCmmnM
m>a m>H Cameo Domamz umomma wnDBm

 

114

 

mmoHo .nCovnum

0Hmou H .nommomonm

.munmnn nmamnm

 

 

 

 

 

 

no HoCmm m MH> uCoECmHCCm no pnozon uCoE
pnm3on Cooomm uoOCo ©o>Hooon a .EoononCH Co no>o oOCoHCCm noonOH Aommav
ooCnHunC IHoCo no omma mCOHanon m on moCooomm cono>HHoo mm ICHom CCoU
.omCommon 0C no
.moHHmon uCoCnmCoolonHon on omCommon
o>HummoC mCHm moHHmon HoCHCCannmnonCCoo
0Hmon on momCommon o>HuHmom An .moHHmon HoCH AmmmHv
ono3ou IUCuHuuolnouCCoo on omCommon o>HnHmom Am %0Con mnonComom
ooCnHun4 lllll uo>mm nocuHo.m .nm CnHz 3oH>nonCH Co CH ImHmCou .nCOCxocm
onCmmoE
nonoEOEmCmc .onon oCu onmzou nCoE AMMQHV
“0Hmou Gnosou momCommon oHnmno>om nom boonOHCHon loonOH .Hm no
oUCuHun¢ IIIII MHo>HnHmom ono3 om .3oH>nouCH Cm CH ICHom .noCmonM
.onon
0Hmou loCn cnmzon momsommon oHnono>om n0m uCoE
nnmzou m an n=ooomgv omononcnmn nnm>nnnmom -monom Ammanv
ooCnHuud IIIII ono3 oCOCmoHou an coon>nonCH mm uCHom omeH
muonCo .AmUCoEEoo OCV nCoEoonomCHonCOC no nCoE
oOCmEnom noon A=©oom=v uCoEoonomCHon HMHOOm noCnHo loonom AmomHv
unom xmoa «mmooosm Co>Hm ono3 COHuoCnHm Row» M CH om ICHom HHHm
m an mCOHu
annomoolmHom .m an A.ouo
momnoommn no =.oonmmmHo mo unOm H=V uCoEoonoHCH
l hoonsoom “an Ion o>HnmmoC no A.ono =.©oom .oonmo H uCoE
mm an mCOHn HMCOmnomEH .mowzv uCoEoonmmCHon o>HnHmom noCnHo noonOH AmmmHv
ImHnomoplmHom .m> HoCOmnom Co>Hm ono3 mm .3oH>nonCH Co mCHnCo ICHom Comnou
m>o m>H mmmao Dozens Mmomma unasm

 

115

 

mOCoCHH .nCoE
Io>oE .mmoCmCo>

.moCooomm mCHno>HHoc

 

 

 

 

 

 

InoC .noouCoo mm on AHmnnCoCuCoonm .oHnono>omCCncon AhmmHv
oho “no mmCH .oHnmno>omnouHC3v monoo conOHoo CnHz noHoEnoumo
upon ooCoHUCC IIIII momCommon oonCmHm oOCoHoCm ooCHmnn C IIIII .ouoad
am an ooanCmHCoE oCm .oOCoHosm
Cm an conoumHCHEoo mm3 omCommom .Cuon
0Hmon no .Unm3on oOCoEnOHnom .pnmzon uConCoo
onoson omCoCo noCnHo Co>Hm ono3 ouonov m CH mmoHHom ooCMC AmomHV
ooCnHund IIIII nHoCu ouHmommo 3oH> o cooComop 0C3 mm IOmmHa oooHHmz
mCCOC
Ionm .mnom an .mCCOConm COmnom umH Can mooCouCom
CnH3 mCHCCHm mCHCCHmon now nonsmfioo m an A:©oom uCoE AmmmHv
Ion mooConCoo mno>=v coonomCHon >Ho>HnHmom onoB loonOH monmm
no %0Cocvonm IIIII xmmu CoHnoCnumCoo ooConCom m CH mm uCHom .xoonoUH>
oCmmH COHu «m uCoE
onm3ou nmnComonm an coCmHCCm no coonmson noCnHo onoB noon0m Ammmav
COHCHmo no noono oCmmH Co C0 C00 oCm onm Cnon mCHCmno mm ICHom 30Cmom
.ooCmEnomnom xmmu nHoCn AmmmHv
mononnm Co xomnooom COHnmfimHHCoo no nConnCoCoo noEHoCCHoz
ooConomom lllll nonoumHCHEoo ono3 mm .CoHuoCuHm goon CH IIIII .mmsonm
.nonEoE maonm
HMCpH>H©CH on» no mmoCo>Huoommo an no
Eoonmo xomnvoom .msonm osn mo mmoCo>Huoommo Am "OCH
IHHom «COHu onm>Hnm ICnooCoo xOMQEWom onCHHom no mmooosm Ammev
loonuno maono .m> UHHQCA Co>Hm ono3 mm .COHnmanm xmou m CH IIIII ohnm
m>a m>H mmmao oomemz Mmomma MQDBm

 

116

 

.moCooomm mCHno>HHoo
mm on momCommon A.ono .nomnCoo omo 0C

 

 

 

 

 

 

usanso mUOHnom .COHnCounoCHv o>HnomoC no A.ono .mcos
Hmnno> «onon onCCHE .moHHEm .COHuConnoV o>HnHmom nonou AmmmHv
“>0CoCHMCoz IH «xom ImHCHEco mnonEoE ooCoHUCo UoCHonu o>Hm lllll CosmnCHm
.mnCoE
usmno nHoCn no onCHHMH no mmooosm on»
UHCOH COHnmoHHH Co ouCoEEoo Conans mCmon oonoanCHE
onmzou Inmsfl nooCo loo ono3 .ooCoHch mCOHanon m nom oOCMC AmmmHv
ocanuu4 ICHHCHupooz mUCoECmnm nCmmonomHCIHoHHon ouon3 mm loomHC nomnom
moCoCHHCOC oOCoHGCm
.oumn .nnomon mo mnHHm .AmmoComOHov MHHHoCOHmCouCH mCHmno» AhmmHv
:mHom ”muonCC ICOHmCouCH no moOCoHoCo on moCooomm cono>HHoo mm nnnnn oxHoB
Coo oCo .momCommon Agndlnsszv
onm “0Hmon o>HnmmoC no A=CCCICC=V o>HnHmom
UHQOH onMBOH onmzou oonoanCHEpo 0:3 .noCoumHH ooCHmnu oOCoC Abomav
ooCannd ooCann4 onCHm m on moCooomm cono>HHoU om IOmmHa xouumz
.mnCmHH mo HoCmm m an CoHMC
-mHm Awom .woe .wom .wom .wonv nm>onm
0Hmou .mHom umm ooCoHoCo w mm omCommon co>Hooon nCoE AhmmHv
.ooCoHUCo pnm3 omHnnomxo pCm .ooCoHcCo mCOHOHHOHm m on 500 loonOH nonmm
sou opCuHun< ooCoHcsd InouCH Cm no>o moCooomm cono>HHo© om ICHom InoCoCm
momma“ .mHCo xomncoom
an m3oH>non Conan3 .MHCO xomnooom mnonHoCo
ICH mo mmCHuon .mHCo xomnooom HoCmH> .xoonooom HHCM
uoCCanno ooonw umCoHnHvCoo nCow nooCC moosoH>nonCH
-nmucn mo mm gone connmsnoncn mosunuum pom on “Homnv
an mCoHuoHUonm lllll wouosnnmCH ono3 mno3oH>nonCH mm mm IIIII moHCom
m>o m>H mmmao Conan: umomma MCDBm

 

117

 

.momCOQmon xoonvoom,0C no.hvohoHo©
.onMHonEH noCnHo Co>Hm ono3 mm .umHQ

 

 

 

 

 

 

monCmmoE IonoCn o no cCommon on coxmo .mEHHH. AmmmHv
hCnomfim IIIII amononuonohmm toUMHCEHm CBOCm onoB om lllll hpoom
.moHoCoCHH .om on» ono3 0:3
ICOC noxmomm .ooCoHoCm o>HH ononomom o now po>MHm
«momvfin Cooomm ono3 moCooomm vomouoovH> one .mnoxoomm
an moCooomm on momCommon anouHUCo a HMCmH> A.ouo
no mmCH .Eoon oCu mCH>ooH =.0C= .mHsoom .moooC
lumn «noxmomm mConCmv o>HnomoC no H.0uo .omCon uCoE AmmmHv
oCu onmson umm =.EEm|EEE= .moHHEm .mCHCCOCV o>Hn noonom ComnoCCd
mooCannm .om lllll nHmom oonoanCHEcm mooCoHUCm coCHonB ICHom .moHnCoz
anHHHn .mnn
mHmom xmmu IHmComm IHHom no mmooosm cooCCCHImHHouCoEHnom AmomHv
no uCoEnanod non xmoe Ixo Co>Hm ono3 CoHuoCnHm xmon m CH mm IIIII CHBoH
.Anwv mnonEoE ooCoHpsm o>HMC mo nHm
noxoomm oCm IoCon oCu now momCommon =oonm= no .o>Hu
UHQOH onozbw ImmoC .o>HUHmom oonoanCHEUo oOCoHoso moCou AmmmHv
ooCuHuum .mm lllll coCHmnu o .noxoomm m on mCHCoanH oHHCB ImHmCou Cothm
.m an CoumHCmHCmE ono3 AHHom
ooCoHpCo .m> ooooosmv mnHCmon one .mucofismno
UCo .OHmou cnoz oCu mo mmoCo>Hmmsmnom oCn Co cono> 0C3 ooCMC AmmmHV
Ion oCCnHuud nnomwm ooCoHoCm Co on mnCoECmnm Uono>HHoU om IOmmHo noCHUnmw
HHono>o
.ommCmCmH .uCon .noxmomm m on
ICoo .mno>HH AHmnnCoCuCoonm .oHnono>omCCnoon .oHnm AmmmHV
now “no mmCH Ino>om nouHC3v monoo oonOHoo Can mom noHHHz
unmn ooCoHCCC lllll ICommon poHoCmHm oOCoHoCm coCHonn C lllll .mnEoo
m>n m>H mmmao nomamz wmomma macaw

 

118

.t:. ok‘lf‘a \!

 

MOCoCHmmHv
“oOCoEnOMnom
.oOCoHCCm .mHom
.onon Gnos

Ion movaUHund

. mm on»
no HHme
mConomm

.oxomm moCu no momCommon o>HnmmoC
no o>HuHmom co>Hooon pCm .ooconso
o>HH coCHmnn m on mommmmoa cono>HH
lop mnoxmomm ooHHmeCC UCm voHHme

 

mnosomon
mo mmCH
upon uCoUCnm

ooCoHnomxo
mCHComon
no mnmo»

.xooncoom 0C Ac .mnOmH>
Inomsm oCo mnCocnum Cnon Eonm mmCH
numn A0 .mnOmH>nomCm Conn mmCHnmn An
.muCoanm Eonm mmCHumn Am "mm xomn
upoom Co>Hm ono3 mnoComon EoonmmoHo

oOCMC Ammmav
nommHC CCH3nmm
Ammmav
no>HHo
lllll .CoExoCB

 

m>a

m>H “WEED

QOZBNZ

wmomma MQDBm

119

APPENDIX B

RESPONSE-ADMINISTERING ASSIGNMENTS FOR THE
TRAINED AUDIENCE

POSITIVE REACTIONS

General instructions for the entire audience:

1.

Allow a warmup period, in which you give little or
no reaction. WOrk your feedback in gradually, as
if you are beginning to get the gist of the speech.
Be as natural as possible, as in a meaningful
communication situation.

Try to get the feel of what the speaker is trying
to do, then work to help him out.

Be like a machine...do the same thing for each
speaker in the positive feedback condition.

Keep your eyes on the speaker at least 90% of the
time.

Control all movements, figeting, etc.

Sit erect, but comfortable.
By the end, appear to be enthused with the speech.

Clap at the end. When the Speaker is through,
comment to each other, and to the speaker, "good
jobW, "great speech", "that was cool", etc.

Specific assignments for audience members:

1.

Appear to be concentrating on the content of the
message. Act as though you understand the points

the speaker makes.

Agree with nearly every assertion the speaker makes
by nodding your head.

Smile at the speaker. If he happens to falter, smile
as if you understand.

Sit with your hands on your lap.

Fold your hands in front of you on the table.

Take notes on the speech. Appear to be concentrating
on the message.

Keep time by using time cards. When the time is up,
wave the "0" card briefly until the speaker sees it.

 

120

NEGATIVE REACTIONS

General instructions for the entire audience:

1.
2.
3.

4.
5.

6.

7.

As with positive feedback, allow a warmup period

before you begin administering feedback.
Try to get the feel of the speaker's purpose, then

disagree with it.
Be like a machine--—do the same thing for each speaker

in the negative feedback condition.

Do not give the speaker more than 50% eye contact.
If the speaker pauses, falters, or makes a mistake,
appear disgusted, like he's really out of it.

When the speaker is through, appear relieved.
Start talking with one another with muffled negative

comments.
Do not clap at the end.

Specific Assignments for Audience Members:

1.

Listen for the main assertions, react to each with
a frown. Be skeptical, obviously disagree with the

speech.

Put your elbows on the table, chin in hands, and
appear bored.

Slouch and squirm throughout the speech, yawn,

appear sleepy.
Play with your pen, pencil, or object from your purse.

Tap your fingers on the table.
Look at speaker, but appear bewildered, as though
you don't know what he's talking about.

Shake your head, as though you are disagreeing.
Begin to take notes at the start, but give up about
two minutes into the speech, closing your notebook,

shaking your head.
Keep time by using time cards. When the time is up,
wave the time card until he sees it and then say

"time".

 

121

APPENDIX C

COMMUNICATION 305
PERSUASIVE ASSIGNMENT #1

"Speaking to a Live Audience"

OBJECTIVES:

1. To give the student experience in preparing and
delivering a persuasive message.
2. To provide an opportunity to speak before a "live"

audience.
3. To help the student realize the prime importance

of audience analysis.

DETAILS:

1. Topic: You are to choose a topic which you feel
very strongly about. Narrow and limit this topic
until you have a specific purpose in mind. Pre-
pare a statement of tOpic and purpose to hand in.

2. Preparation: After you have finalized on your topic
and purpose, begin preparing a five-minute persua-
sive message. When your preparatiEn i§ in progress,
you will be given a profile of the audience you will
speak to, including specific attitudes toward the
topic you are speaking on. This profile should be
of great benefit in helping you adapt your message
to the receiver.

3. Delivery: You will deliver your message to a live
audience, arranged by the instructor. You will
only have to come, deliver your speech, check with

the instructor, and leave.
4. Place: 56 Baker Hall.

SEMETABLE:

*January 14 (Tuesday): Hand in a statement of your topic
and purpose at class time. We have to have this in-
formation now in order to test the audience's atti-
tude on the tOpic you are speaking about. From this
point on you may not change your tOpic.

*January 21, 23: You will receive a profile of the
audience you will be speaking to (if all goes well).
Also, you will sign up for a time to speak.

1
E
3

 

122

*January 27, 28, 29, 30: From 7-10 p.m. on each of
these days, the persuasive speeches will be de-
livered. You will only have to arrive at the
time you signed up for, speak to the audience,
check with the instructor, and be on your merry

way.

If you have any questions or problems, check with the
instructor in charge: James Gardiner, 56 Baker Hall,
353—6773.

-A‘b‘h' - .7 mfi um

123

APPENDIX D

AUDIENCE PROFILE

COMMUNICATION 305

ASSIGNMENT 1

 

 

 

SIZE: 13 SEX: 10 female, 3 male. F5
I
AGE DISTRIBUTION: Youngest - 17 Oldest — 24 }
Mean age - 18.5 i
MARITAL STATUS: 11 single, 2 married. i

RESIDENCE: 4 apartment dwellers, 9 dormitory residents.

 

CLASS STANDING: 8 freshmen, 2 sophomores, l junior,
2 seniors.

 

COLLEGE MAJORS: French, Pre-med, Business, English
Literature, German, Elementary Educa-

tion, Secondary Education, Undecided.

 

OCCUPATIONAL ASPIRATIONS: Teaching, Medicine, Foreign
serVice, Business Management, Undecided.

SQME TOWNS: Detroit, Lansing, Pontiac, Flint, Kalamazoo,
Jackson, Milwaukee, Buffalo, Decatur, Ga.,

Boulder, Colorado.

 

POLITICAL LEANINGS: 8 Democratic, 3 Republican,
2 Independent.

RELIGIOUS AFFILIATIONS: 7 Catholic, 4 Protestant,
1 Jewish, 1 No affiliation.

EQEEEES: Skiing, Sewing, Dancing, Fencing, Music, Reading,
Swimming, Horses, Decorating, Painting, Travel,

Cycling.

EXPECTATIONS OF YOUR PERFORMANCE: The audience does not
know who you are, except that you are from a public
speaking class. They will have an agenda with your
name and your tOpic. You will be introduced to the

group by Jim Gardiner.

124

AUDIENCE ATTITUDE:

Explanation: Your general tOpic was presented to
the prospective audience, and each

member responded to the topic on a

scale similar to the one below;

 

 

 

 

Strongly FavoraBle Neutral Unfavorable Strongly
Favorable Unfavorable

s ‘pho. rm; mu... - n—‘T

owe-5.5:. .1

The following is a graph-type summary of the attitude-
responses toward your topic, which was:

 

  
  
 

13
12
11
Number 10

of 9
Audience 8
Members: 7

6
5
4
3
2
l

Strongly Favorable Neutral Unfavorable Strongly

Favorable Unfavor-
able

125

APPENDIX E
DEPENDENT MEASURES
COMMUNICATION 305

SPEAKER REACTION INVENTORY

Name Student #

 

The following is a questionnaire seeking your reactions to
the type of speaking situation you just experienced.
Please record your first impressions, placing a check

mark on each of the scales below.
1. How did you feel about the slgg_of the audience you
spoke to?
[:7'Too large
D About right
[:7 Too small
2. What type of attitude did your audience have toward
your tOpic before you spoke to them?
[:7 Favorable
[:7 Neutral
£7 Unfavorable
3. How did they appear to feel toward your topic after
you spoke to them?
0 Highly favorable
[:7'Moderately favorable
[:7’Slightly favorable
[:7 Neutral
[:7 Slightly unfavorable
[:7'Moderately unfavorable
[:7'Highly unfavorable
4. How would you rate the audience on responsiveness?
[:7 Highly unresponsive [:7 Responsive
[:7 Unresponsive [:7 Highly responsive

[:7’Neutral

w.“ .“l-ol an /._-o _—.-‘-

‘

5. Please rate TH

a check mark on each of the

Cruel __:
Active __:

Valuable __;
Unskilled __3
Just __:
Unpleasant __;
QualifiedI__:
Selfish __:
Awful __:
Honest __:
Energetic __:
Reliable __3
Virtuous __:
Expert __:
Uninformed'__:
Emphatic __:
Safe __:
Unfriendly __:
Aggressive __:
Trained __:
Unintelligent __:

Experienced __;

126

O I O
O O I
_— _'— — _

O
I
—— — —_ —— —-

——*_—

O O I O
O O O O
— _ — — fl

0 I O
O O O
— -— - - —

-’———

‘_-—_

‘————

O O O
O O O
— — fl — ’

. C O
O O O

E AUDIENCE YOU JUST SPOKE TO by placing
following scales:

Kind
Passive

Worthless
Skilled
Unjust
Pleasant
Unqualified
Unselfish
Nice
Dishonest
Tired
Unreliable
Sinful

Bold
Inexpert
Informed
Hesitant
Unsafe
Friendly
Meek
Untrained
Intelligent

Inexperienced

mum-1.211 unnu- ‘1241‘IMfm

127

6. I would rate my persuasive performance before this

 

audience as:

Good

Fearful

Well-Delivered

Beneficial

Speaker-Centered

Forceless

Perfect

Self-Conscious __3__3__3 :
Direct __°

Unfriendlyl__.

Strong __

Poor-in-Content _:__:__:__:_:__:_

Timid __:__: : : : :

Adaptive __.

Message-Centered __

a?

7.

Neutral

mummmmmg

Slightly negative
Moderately negative
Highly negative

Bad

Fearless
Poorly-Delivered
Harmful
Audience-Centered
Forceful
Imperfect
Confident
Indirect
Friendly

Weak
Good-in—Content
Bold
Non-Adaptive

Audience-Centered

type of Response did you receive from the audi-
as you spoke?

Highly positive
Moderately positive
Slightly positive

“may“ QA axruly .‘m 2": m1“,

128

8. I found this persuasive experience to be:

Highly pleasant
Moderately pleasant
Slightly pleasant
Neutral

Slightly unpleasant
Moderately unpleasant
Highly unpleasant

0000000

f I were grading this Speech, I would receive:

\0
O
H

.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

O

10. If my audience were grading this speech, I would
guess that I would receive:

TOO—FITTEE—O'TFT'O'E—EW r5

 

 

STATISTICAL TABLES

129

APPENDIX F

TABLE 1

Factor Analysis of Audience Ratings

 

Scales

Dimensions and Factor Loadings:

Authorita-

tiveness

Dynamism Character

 

Kind--Cruel
Unpleasant--Pleasant
Qualified--Unqua1ified
Selfish--Unselfish
Awfu1--Nice
Reliable--Unreliable
Timid--Bold
Expert-Inexpert
Uninformed--Informed
Unfriendly--Friend1y
Aggressive--Meek
Trained--Untrained
Experienced-—Inexperie

.76

.72

.88
.66

.76
nced .79

TABLE 2

-080

-.79

-.61

-085
.64

-o74
.84

Factor Analysis of Performance Ratings

 

Factor Leadings

 

Scales

Good—Bad .81

Well-De1ivered--Poor1y-Delivered .71

Beneficial--Harmfu1 .76

Forceless--Forcefu1 .78
.64

Perfect--Imperfect

 

 

130

TABLE 3

Results of Bartlett's Test for Homogeneity of Variance

 

 

W Pawn .. .31. REESE:
Character EAR x PAR 3 4.38 <.30*
EAR x OAR 3 10.97 <.05
Authoritativeness EAR x PAR 3 2.05 <.70* F;
EAR x OAR 3 5.84 <.20* '- ~'
Dynamism EAR x PAR 3 1.50 <.70* i
EAR x OAR 3 1.09 <.80* g
Performance EAR x PAR 3 3.03 <.50* I.
EAR x OAR 3 9.57 <.05 y

 

*Indicates that the Homogeneity assumption was held
as tenable.

TABLE 4

Tests for Significance of Difference (PAR)

 

 

. Approximate
Comparison DV ,5, Probability
Positive EAR, Negative PAR Character 0.61 <.30

Vs. Authorit. 1.06 <.20
Negative EAR, Negative PAR Performance 1.84 <.05

Positive EAR, Positive PAR
Vs. Performance 5.43 <.05

Positive EAR, Negative PAR

131

TABLE 5

Perceived Audience Response
Group Means, Standard Deviations, Size

 

 

 

 

 

PAR: Positive EAR Negative EAR DV:
3? = 25.41 1? = 24.27
Positive: s.d. = 3.88 s.d. = 4.29
n = 17 n = 18
_ _ Performance
X = 14.00 X = 18.18
Negative s.d. = 5.35 s.d. = 5.87
n = 7 n = 11
J? = 29.05 i = 27.22
Positive: s.d. = 5.80 s.d. = 4.05
n = 17 n = 18
Character
§'= 16.42 §'= 14.90
Negative: s.d. = 6.52 s.d. = 4.72
n = 7 n = 11
x' = 28.35 1? = 27.50
Positive: s.d. = 5.30 s.d. = 4.25
n = 17 n = 18
Authorita-
_ tiveness
Tc = 23.57 x = 20.72
Negative: s.d. = 7.04 s.d. = 6.70
= 7 n = 11
§'= 10.00 §'= 10.00
Positive: s.d. = 1.93 s.d. = 1.60
n = 17 n = 18
Dynamism
§'= 10.85 §'= 9.63
Negative: s.d. = 2.26 s.d. = 1.68
n = 7 n = 11

 

132

TABLE 6

Analysis of Variance Summary for
Character Ratings (PAR)

 

 

 

 

 

Source of A °
. pproximate
Variance. df “'3' F Probability
EAR 1 32.35 1.22 .27
PAR 1 1737.11 67.53 < .05 E
EAR x PAR 1 0.29 0.01, .92 ;
Error 49 26.46 3
F
TABLE 7 i

 

Analysis of Variance Summary for
Authoritativeness Ratings (PAR)

 

 

 

Source of Approximate
Variance df M.S. F Probability
EAR l 39.26 1.28 .26
PAR 1 383.46 12.47 <.05
EAR X PAR l 11.39 0.37 .55
Error 49 30.74

TABLE 8

Analysis of Variance Summary for
Dynamism Ratings (PAR)

 

 

Source of Approximate
Variance df M.S. F Probability
EAR 1 4.28 1.28 .26
PAR 1 0.70 0.21 .65
EAR X PAR l 4.28 1.28 .26

Error 49 3.33

133

TABLE 9

Analysis of Variance Summary for
Performance Ratings (PAR)

 

 

 

 

 

 

 

 

 

Source of A -
- df . . pproximate
Variance M S F Probability
EAR l 26.68 1.22 .28
PAR 1 880.44 40.19 <.05 F
EAR X PAR 1 81.17 3.71 <.05
Error 49 21.91
i.
TABLE 10 F
Analysis of Variance Summary for
Character Ratings (SEX)
Source of Approximate
Variance df M‘S' F Probability
SEX l 16.72 0.62 .43
PAR 1 1883.83 70.22 <.05
SEX X PAR l 11.65 0.43 .50
Error 49 26.83
TABLE 11
Analysis of Variance Summary for
Authoritativeness Ratings (SEX)
Source of Approximate
Variance df M'S' F Probability
SEX 1 73.31 2.45 .12
PAR 1 479.28 16.02 <.05
SEX X PAR l 30.49 . 1.02 .32
Error 49 29.92

 

134

TABLE 12

Analysis of Variance Summary for
Dynamism Ratings (SEX)

 

 

 

 

 

 

 

 

 

Source of Approximate
Variance df M'S' F Probability
SEX 1 0.77 0.22 .64
PAR 1 0.33 0.09 .76
SEX X PAR 1 0.77 0.22 .64
Error 49 3.44
TABLE 13
Analysis of Variance Summary for

Performance Ratings (SEX)
Source of Approximate
Variance df M'S' F Probability
SEX 1 1.81 0.08 .78
PAR 1 719.81 31.36 <.05
SEX X PAR 1 34.23 1.49 .23
Error 49 22.95

TABLE 14
Analysis of Variance Summary for

Character Ratings (OAR)

. . Approximate
Source of Variance chi square Probability
Total 34.12 <.05
EAR 2.42 .12
OAR 31.79 <.05

0.00 1.00

EAR X OAR

 

 

135

TABLE 15

Analysis of Variance Summary for
Authoritativeness Ratings (OAR)

 

 

 

 

 

 

 

 

 

Source of Approximate
Variance df M'S' F Probability
EAR 1 23.82 0.71 .40
OAR 1 295.40 8.84 <.05
EAR X OAR 1 4.99 0.15 .70
Error 49 33.41
TABLE 16
Analysis of Variance Summary for
Dynamism Ratings (OAR)
Source of Approximate
Variance df M.S. F Probability
EAR 1 2.82 0.85 .36
OAR l 3.04 ' 0.91 .34
EAR X OAR 1 ,2.69 0.81 .37
Error 49 3.32
TABLE 17
Analysis of Variance Summary for
Performance Ratings (OAR)

. A roximate
Source of Variance Chi square ngbability
Total 18.70 <.05
EAR 0.79 .37
OAR 18.51 <.05
EAR X OAR 0.00 1.00

 

 

    
 

’ My

31293 01739 7997