ABSTRACT

AN EXPLORATORY STUDY OF THE EFFECT
OF VAHYING MODES OF POSITIVE REINFORCEMENT
ON STUDENT ANIMATION IN THE BEGINNING SPEECH CLASS
By
Dan Pyle Millar

The purpose of the research was to determine if posi-
tive reinforcement presented by varying modes can result in
the improvement of the beginning speaker's speaking skill.
Particular attention was directed toward light on-set rein-
forcement as applied to a physical aspect of the nonverbal
code --- gesture. Nine intact groups. drill sections of the
beginning speech course. were compared via nonparametric
analysis of variance routine. Improvement in gestural skill
was measured by the difference between evaluations of Judging
panels during the pre-test subtracted from those during the
post-test.

Two dependent measures of gestural skill were
created. The Gestural Evaluation Score was the total of a
three-Judge panel received by an 88 on nine elements of
gesturing. The Physical Performance Score was the total

score received by an 33 on five Dynamism scales. Each

Dan Pyle Millar
dependent measure was assumed to be unidimensional yet rela-
ted (supported by correlational data). The Judging panels
evaluated the speaking of the 88 before and after exposure
to three modes of positive reinforcement. In the Minimum
Written Reinforcement condition. 88 received an instructor's
written critique of a speaking performance and an experimen-
tal evaluator's written scale evaluation of a single element
of the gestural performance. In the Maximum Written Rein-
forcement condition, the 33 received the instructor's
written critique and an experimental evaluator's written
scale evaluation of nine elements of gestural performance.
In the Light On-Set Reinforcement condition. the 83 received
an instructor's written critique and a flash of light for
each and every gesture determined acceptable to the experi-
mental evaluator.

Reinforcement theory would lead to the prediction
that the 38 receiving the immediate positive reinforcement.
found only in the Light On-Set condition. would show greater
improvement than those receiving either of the written rein-
forcements. The greater positive reinforcement. assumed ex-
tant in the Maximum Written condition. also would be predic-
ted to generate greater change than the Minimum Written con-
dition. The chief hypothesis of the research. then. was
that the three positive reinforcement conditions ought to
generate different improvement effects. If improvement was

significant. then the maximum Written should show greater

Dan Pyle Millar
change than the Minimum written. and the Light On-Set should
show greater change than either written condition.

No statistical significance was found in the data.
Trends occur in the predicted direction for the Maximum and
the Light On-Set $8. The Ss in the Minimum Written condition
show a strong trend in the negative direction. While no
significance was found. the trends suggest the theoretic
hypothesis to be sound and that increased control within an
experimental framework would encourage statistical signifi-

cance in a replication.

Accepted by the faculty of the Department of Speech
and Theatre, College of Communication Arts, Michigan State
University, in partial fulfillment of the requirements for

the Doctor of Philosophy degree.

W415“: 5. {W

Director of Thesis

Guidance Committee: W.” 3 [W

Chairman

flZe/Mfldfi $71 Zéxy’
(::)22:;66%} ii ,%Zt{:w¢4f7~_.lu
7( -/ 7 / "

if “1’ I - I ,‘1; k 1. Ir / r
. M/ (/1 :\:’:?/\Nv/M//’

[
(”mtg-I‘d QM

 

AN EXPLORATORY STUDY OF THE EFFECT
OF VARYING MODES OF POSITIVE REINFORCEMENT

ON STUDENT ANIMATION IN THE BEGINNING SPEECH CLASS

By
Dan Pyle Millar

A THESIS

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

DOCTOR CF PHILOSOPHY

Department of Speech and Theatre

1969

ACKNOWLEDGMENTS

The writer expresses thanks to William B. Lashbrook,
the thesis advisor. for his insight and direction of this
project. and the members of the committee. Kenneth G. Rance.
Walter F. Johnson, Gerald R. Miller and Gordon L. Thomas.
for their individual suggestions and encouragement. A
special note of gratitude is extended to Dennis Dunne.
Thomas Droessler and Herschel Mack, who gave up much of
their own study time, as well as Donna Emlich, Greg Gardner
and Carol Shaw. who saw that the research was carried out.
Also, the writer acknowledges his wife who performed the
manual and emotional tasks which brought this project to

completion.

11

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . v
LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . vii
CHAPTERS

I RATIONALE AND HYPOTHESES . . . . . . . . . . . 1

Introduction
Rationale
Relevant Research
Hypotheses

I I PROCEDURES 0 O O O O O O O O O O O O O O O O O 19

Pilot Study

Experimental Conditions

Experimental Evaluators

Subjects

Measuring Instruments

Logistics
Minimum Written Reinforcement Condition
Maximum Written Reinforcement Condition
Light On-Set Reinforcement Condition
Procedures for Data Collection
Classroom Assignments
Administration of the Pre-Test
Administration of the Reinforcement

Treatments

Administration of the Post-Test
Statistical Design
Summary

I II RESULTS 0 O 0 O O O O O O O O O O O O O O O O “1
Preliminary Analysis
Testing the Research Hypothesis

Trends in the Data
Summary

111

IV

CONCLUSIONS AND IMPLICATIONS FOR FURTHER

RESEARCH . . . . .

Potential Sources of Research Error

Further Research
Conclusions
Summary

BIBLIOGRAPHY O O O O O C O O

Questionnaires to Faculty and Teaching

Gestural Evaluation Instrument and Rater

Instruments of the Minimum Written

Instrument of the Maximum Written

0

APPENDICES
A
Assistants . . . .
B
Instructions . . .
C
Reinforcement Condition .
D
Reinforcement Condition .
E Speaking Assignments .
F Instructions to Instructors .
G

Table 20. Rater Reliabilities for
Gestural Evaluation Score, Time 1 and 2 .

Table 21. Rater Reliabilities for

Physical Performance Score, Time 1 and 2 .

iv

61+

73

78

8O

82

85
86
89

92

9h

10.

11.

12.

13.

14.

15.

LIST OF TABLES

Distribution of Subjects by Treatment . . . .
Subject Disqualification by Speech Assignment

Correlations Between Items for the Gestural
Evaluation Score. Time 1 . . . . . . . . . .

Correlations Between Items for the Gestural
Evaluation Score, Time 2 . . . . . . . . . .

Correlations Between Items for the Physical
Performance Score. Time 1 and 2 . . . . . . .

Correlations Between Items for the Gestural
Evaluation Score and Physical Performance
Score. Time 1 and 2 . . . . . . . . . . . . .

Median Rater Reliabilities for the Gestural
Evaluation Score. Time 1 and 2 . . . . . . .

Median Rater Reliabilities for the Physical
Performance Score, Time 1 and 2 . . . . . . .

Bartlett's Test for Homogeneity of Variance
on Gestural Evaluation Score. Time 1 . . . .

Bartlett's Test for Homogeneity of Variance
on Physical Performance Score, Time 1 . . . .

Kruskal-Wallis Test on Gestural Evaluation
Score. Time 1 . . . . . . . . . . . . . . . .

Kruskal-Wallis Test on Physical Performance
Score. Time 1 . . . . . . . . . . . . . . . .

Kruskal-Wallis Test on Gestural EValuation
Score Change Scores . . . . . . . . . . . . .

Kruskal-Wallis Test on Physical Performance
Score Change Scores . . . . . . . . . . . . .

Means and Medians by Conditions, Time 1 and 2

28
29

43

Mb

1&6

“7

49

49

51

51

53

5h

55

55
57

16.

17.

18.

19.

20.

21.

Number and Direction of Evaluative Changes by

Conditions, Gestural Evaluation Score .

Wilcoxon Z-Scores for Gestural Evaluation

Score and Physical Performance Score,

Minimum Written Reinforcement

Wilcoxon Z-Scores for Gestural Evaluation

Score and Physical Performance Score.

Maximum Written Reinforcement

Wilcoxon Z-Scores for Gestural Evaluation

Score and Physical Performance Score,
Light On-Set Reinforcement .

Rater Reliabilities for Gestural Evaluation

Score,

Rater Reliabilities for Physical Performance

Score.

Time 1 and 2

Time 1 and 2

vi

58

59

60

61

92

9b

1.

LIST OF FIGURES

Description of Experimental Conditions .

vii

2?

CHAPTER I

RATIONALE AND HYPOTHESES

Introduction

 

When the college student enters the beginning speech
classroom, he has been speaking to other peOple for at least
18 years. The problem facing the college speech instructor
is to encourage the student to organize, develOp and cor-
rect habits he already possesses. Thus, the teaching of
public speaking seeks two objectives: (1) the understanding
and remembering of a set of theoretic principles, and
(2) the refinement of individual skills. An instructor of
a beginning speech course needs to discover techniques from
within and without the field of Speech which will maximize
the achievement of these two objectives. The focus of this
research is upon the second of the objectives --- the re-
finement of personal skills. The particular interest is
the determination of the feasibility of varying modes of
positive reinforcement as means of increasing the impact of
instruction in one of the fundamental areas of student per-

formance --- animation.

Rationale
A major factor in the refinement of the performance
Skills of a student in the beginning speech course is the

insitructor's critique of that performance. The instructor
1

2

must identify weaknesses, offer constructive means for
eliminating those weaknesses, and reward the removal of
error. This alternating process of approval and disappro-
val is designed to improve the skills of the beginning
speech student. In other words, the consequence of the
critique of performance is the deveIOpment of desirable
speaking behavior. Robinson and Kerikas (1963) suggest

the importance of the teacher's role in skill deveIOpment
when they write that criticism underlies all learning in
the speech classroom. Improvement in the individual skills,
particularly delivery skills, according to Densmore. rests
"first. last. and always upon the teacher in charge (19h6.
p. 67)." Donald Smith (1961) calls the criticism function
of the instructor the "central reason for being" and in-
dicates this function has not changed with speech instruc-
tion over the centuries. With the critique having such
importance, whatever may improve the ability of the in-
structor to communicate his evaluations and recommendations
of the performance to the student should be of value in the
teaching of public speaking.

Given the importance of the instructor's communi-
cations to his students, there is still need for research
in this area of speech education. Donald Dedmon notes that
"much research is still needed on the subject of criticism
of student speeches. The possible relationship between
techniques of criticism and modification of speech beha-
vior needs to be explored (1967, p. 28h)." One way to

carry on such research would be to take the reinforcement

3
findings of the laboratory and place them within the class-
room. a recommendation advanced by Resnick (1963). Con-
cerning laboratory findings. Becker (1963) suggests that the
field of Speech needs to research what events may be rein-
forcing to the student and how these events may be pre-
sented to the student during class periods when the student
is practicing speech-making. To research the instructor's
critique. his chief technique for feeding back to the stu-
dent approval or disapproval of performance. one might in-
vestigate what is reinforcing to the student. how best to
relay the reinforcement to the student. and observe any be-
havioral changes which take place. It is the purpose of
this particular study to apply learning research to the
speech classroom and to determine if positive reinforcement
presented by varying modes can result in the improvement of

the beginning speaker's speaking skill.

Relevant Research

 

 

Since the instructor critique underlies learning
in the speech classroom. this review will begin by examining
some of the means of critiquing student speeches. Many
recommendations have been made concerning the nature of a
good critique. but two appear to be directly related to the
present consideration: (1) the kind of comments that are
made. and (2) the time of the critique in relation to the
student's performance.

Favorable. constructive remarks seem to be re-

garded as the kind of comments composing a good critique.

z.

Dedmon (1967) makes such a recommendation. as do Weaver.
Borchers and Smith (1952). Wellace (1954). and Bostrom
(1963). However. a descriptive survey of the central
united States completed by Lothers (1966) found this
recommendation not followed. From the 186 respondents.
it was found that 50% give equal weight to favorable and
unfavorable remarks: another 25% give greater weight to
unfavorable remarks.

In addition to the content of the critique. the
time relationship between the critique and the performance
is also important. Robinson and Kerikas (1963) suggest
that the instructor may offer criticism after every speech.
at the end of a series of speeches. and/or during the per-
formance of the speech. They support the latter as the
best time for handling errors of delivery. Since this
study is concerned with delivery. their recommendation is
of particular significance. The Lothers (1966) survey in-
dicates that few instructors actually criticize during the
student's speech. The majority of respondents offer their
critiques at the end of the speaking day. while only 22
(11.8%).answered that they critique during the speech it-
self.

It would appear from the preceding discussion that
the typical instructor's criticism. the chief classroom
tool for skill development. contains as much disapproval
as approval and is presented to the student at varying
times after he has spoken. The findings of learning re-

search suggest that both practices may be contrary to

5
conditions of maximum student improvement.
Research stemming from the psychological and educa-
tional laboratories shows the consistent finding that res-
ponses that are accompanied or followed by immediate rein-
forcement will be more likely to reoccur within similar

situations (Becker. 1963). This finding resembles Thorn-

dike's "law of effect." which states that when "a modifiable

connection is made and is accompanied by or followed by a
satisfying state of affairs. the strength of the connection
is increased: if the connection is made and followed by an
annoying state of affairs. its strength is decreased (Hil-
gard and Bower. 1966). Thorndike later modified the "law"
because experimental findings did not show punishment re-
ducing the probability of behavioral repetition (Postman.
19h7). However. the experimental findings supported the
conclusion that positively reinforced behavior tends to be
repeated.

Inherent in the preceding discussion is the rela-
tionship between reinforcement and its timing with respect
to the behavior being reinforced. Contiguity in time is
presupposed as a basis for learning (Shelton. Arndt and
Miller. 1961). Students learn best. writes Byers (1963).
when they have immediate and valid knowledge of their
success or failure. Harms connects the notion of conti-
guity to the classroom when he writes: "Experimental data
irudicate that immediately reinforced responses are neces-

Iséily’for efficient learning. Hence. the more responses a

Student makes which receive immediate reinforcement. the

6

more rapidly and surely he will learn (1961. p. 216)."
Learning. then. may not only be encouraged by the content
of the reinforcement but also by the time relationship to
the event.

While the acquisition of new behavior requires
immediate positive reinforcement. the development of existing
behavior also utilizes this kind of reward. Skinner (i953)
distinguishes acquisition from deve10pment by calling the
latter skill. He offers the example of ball throwing where
certain responses release the ball at precisely the right
instant to achieve maximum distance. Early or late release
will decrease the distance that the ball travels and will
not be reinforced. Reinforcement which develops skill must
be immediate or the precision of the differential effect
will be lost.

Delay of reinforcement may be particularly con-
founding for speech learning. Since speaking is a complex
activity. a delay in reinforcement from the deserving act
may encourage the repetition of a series of intermediary
actions and not the specific action desired (Shelton.

Arndt and Miller. 1961). Yet the time for classroom criti-
quing appears generally to be several minutes after a par-
ticular speaking event is completed. Differential positive
reinforcement enhances skill deveIOpment. provided that the
.reinforcement of the actions occurs immediately after their
appearance .
It is assumed that an ana10gous situation exists

t>€3tween ball throwing and the hand-arm animation of a

7
Speaker. If so. then refinement of the speaker animation
requires immediate. positive reinforcement. The delay. as
has been noted. caused by waiting until a speaker or series
of speakers has finished his or their speeches. may reduce
the effectiveness of the reinforcement. Such a reduction
can then delay the development of gestural skill refinement.
As the Lothers survey indicated. the situation described
approximates the speech classroom.

One problem peculiar to the teaching of speech is
that the student does not enter the classroom with no ex-
perience in the behavior the teacher wishes to encourage.
The student has been speaking all his life. Instruction.
then. may not require that the instructor encourage the be-
havior as much as refine the behavior he finds already being
emitted by the student. "Good" speech behavior may exist
already in the student's repertory of behaviors. What is
required of instruction is improvement of existing patterns
so they more closely resemble acceptable speaking behavior.
This situation resembles Type R conditioning described by
Hilgard (1966).

In Type R conditioning. the emitted responses are
designated as Operants. These Operants. once observed. are
reinforced. Hilgard writes that "reinforcement cannot fol-
low unless the conditioned response appears: reinforcement
is contingent upon response (p. 109)." Skinner relates
this operant conditioning to the "law of effect" by saying
that "we strengthen an operant in the sense of making a

response more probable or. in actual fact. more frequent

8
(p. 65)." The consequence of the reinforcement of an
operant is to increase the rate with which the operant res-
ponse is emitted. The frequency of response. or emission.
becomes a measure of the learning process of the subject
(Bigge. 1964). Applied to this particular study. the number
of gestures emitted and the increase in animation over time
will affect the judges' evaluation of the effectiveness of
the speaker's physical behavior.

How often the learner receives reinforcement is also
an important consideration. Two general schedules of rein-
forcement have been studied within the learning laboratory:
continuous reinforcement (£1221 or n9 response reinforced)
and intermittent reinforcement (some ratio of responses
reinforced. A continuous schedule of reinforcement is the
quickest way to increase the frequency of response (Berelson
and Steiner. 196b). Sidman indicates this schedule is a
powerful technique because the learner will respond even
when deprivation is minimal and reinforcement is small (1962).

In addition to educational and psychological research
findings. at least one communication concept relates to
learning: the concept of feedback. Weiner defines feedback
as a "method of controlling a system by reinserting into it
the results of its past performance (i95h. p. 61)." Feedback.
according to Berlo. can be rewarding or nonrewarding to the
source of the initial message. In other words. reward (po-
sitive feedback) promotes the continuation of current mes-

sages. while nonrewarding (negative) feedback promotes

9

change. The human feedback system is external (1962): that
is. the output passes through the environment and returns
to the system as input. Feedback. then. resembles rein-
forcement in the laboratory. Weiner supplies the linkage
between feedback and learning. He says that information
which proceeds backward from the performance and is able to
change the general method and pattern of performance is a
process called learning.

Most studies of feedback have used the concept as
the independent variable with subsequent changes in verbal
learning. delivery and fluency. or some other source beha-
vior attributed to the manipulation of the amount and kind
of feedback. Miller. Zavos. Vlandis and Rosenbaum (1961)
studied the effect of reward ("good") and nonreward (experi-
menter silence) feedback on two sequential speakers. While
they found that lack of approval of the second speaker did
not significantly effect his fluency. ratio of nonfluencies.
and his estimate of time spoken. they did find that approval
of the first speaker had a disruptive effect upon the per-
formance of the second.

In a later study. Miller (196h) used approval. dis-
approval and "no comment" as the feedback for student
speakers. He found that approval-disapproval variation be-
tween two speakers in sequence resulted in a decrease in
speaking efficiency in the second speaker. Vlandis (196W)
also manipulated reinforcing conditions in the form of feed-
back. Unlike the Miller eXperiment. Vlandis used only one

speaker and manipulated the feedback during the different

1O
partions of the same speech. The punishment group (received
'uua" from the experimenter) showed disruptive effects when
measured on total utterance and number of nonfluencies.
The: reward group showed a slight. but insignificant. increase
in Iitterance and similar results for the moments of non-
fluency.

Sereno (1964) varied feedback between two sequential
spueeches for the same person. Using 90 women as subjects and
mesisuring their total verbal utterance and nonfluencies. he
foxxnd the inferior treatment (less favorable at Time 2 than
at. Time 1) to produce a significant decrease in verbal
uirterance at Time 2. Nonfluencies increased at Time 2 under
the inferior condition. but did not reach statistical signi-
fixzance. The superior treatment (more favorable at Time 2
truan at Time 1) showed an increase in verbal utterance and
a (decrease in nonfluencies at Time 2. but neither result
treached statistical significance.

Blubaugh (1966) studied the effects of positive and
Inegative audience reaction upon fluency and speaker output.
.Like the above authors. he found greater nonfluencies under
Ilegative conditions of feedback than in the positive con-
<iition. Rate of speaking decreased and verbal output was
lessened in the negative treatment group.

Davis (1967) used two females in a discussion situa-
tion with an authority figure (high-low) interjecting "yes"
(reward) or "no" (punishment) into their discussion or re-
Bmaining silent. The reward group had significantly higher

‘total utterance than either of the other conditions. while

11
the "same" group had a significantly higher total utterance
than the punishment group. Nonfluencies were higher for
both the disapproval and approval groups.

Like Blubaugh. Karns (1964) used live audiences to
manipulate the feedback conditions. One audience was to
display an "attentive." "interested." and "accepting pos-
tnire" of the speaker's message while the experimental
anidience. on cue. was to show "disagreement" or "lack of
‘understanding." When comparing the recorded speeches with
tune manuscripts of the prepared speeches. Earns found that
changes in the manuscript were preceded by aversive stimuli
from the audience.

Amato and Ostermeier (1967) studied the effects of
‘neutral and unfavorable feedback upon the delivery of a
student speaker. a study more closely related to this work.
thing colored cards to provide feedback. they hypothesized
that delivery. measured by observers. would deteriorate in
lflne unfavorable condition. There were six characteristics
<3f delivery which they consolidated into bodily action. eye
(Hintact. fluency and nervousness. The hypotheses were sup-
Iported for all delivery characteristics. The unfavorable
feedback condition also prompted a shortening of the
Speeches by the speakers.

The results of the study were challenged by Combs
amid Miller (1968) on the grounds that the audience. who
Euiministered the feedback. also made the Speaker evalua-
1Iions. Combs and Miller concluded that the speaker dif-

ferences were due to the response role of the raters and

12
not the effect of the feedback. To demonstrate their chal-
lenge. they conducted a similar experiment with the addition
of a favorable condition. This study also measured the
audience rating of delivery. content. language and overall
effectiveness. The data. which showed a relationship between
the feedback condition and student rating. was interpreted
as supporting their initial challenge.

In a rejoinder. Ostermeier and Amato (i968) maintain
their study supported the original conclusions. The one
reason they mention that may deny the response-role chal-
lenge is that their subjects were instructed to provide a
"majority" of reSponses in the favorable-unfavorable feed-
back conditions. rather than to respond only positively or
negatively to the speaker. The Combs-Miller study asked the
audience-raters to provide only one kind of feedback.

One characteristic common to the mentioned studies
is that they were concerned with the deterioration of verbal
or physical delivery. The prOposed study will focus upon
the positive feedback condition to determine if such feed-
back will increase the desired delivery behavior of the
student speakers. The research briefs examined earlier
suggest that positive feedback (reinforcement) ought to
generate improvement. yet their accompanying experiments
irregularly support that conclusion.

Two additional studies need to be mentioned.. One.
completed under a Works Progress Administration grant. used
a signaling device to "feedback" to the student instructor

evaluations during the speech (Hayworth. 1939). The device

13
was an elongated box of six double compartments with an
opaque glass front. Speaker-directed comments were printed
upon the glass. The speaker was signaled when the light be-
hind the compartment glass flashed. highlighting the printed
evaluation remark. This device was used during an entire
teaching term. Statistical tests of differences between the
experimental and control groups found no significance in as-
pects of delivery. The results may be due to the directives
of the evaluative comments revealed to the speaker when the
light flashed. Comments like "more vigor." "walk." "louder."
"improve posture." "use gesture." or "extend gesture" all
direct the speaker to new courses of action. The implica-
tion to the receiving student may be that his current beha-
vior is inadequate. The signals. like other feedback de-
vices mentioned. seemed to be negative in intent.

Another study which used feedback over a term period
as a teaching device was done by Dorothy Hinde (1949). She
constructed a collection of cards containing comments upon
the speaker's delivery and flashed them at the Speakers du-
ring performance. This lasted for an entire semester with
ratings done on the first and last speeches in the course.
Ratings were for general effectiveness. bodily communication.
use of voice and articulation and pronunciation. Aside from
the flash-card group. the other treatment group received
comments immediately before the speech. and another received
only written criticism after the Speech. An overall view
Showed no significant differences in improvement in the

three conditions. All three classes showed improvement in

14

bodily communication. but it was not significant. All three
classes did show significant improvement in articulation and
pronunciation. As a sidelight. the students responded ad-
versely to the flash cards. This effect may be due partly
to the adverse nature of the comments on the cards. The
students also indicated that the appearance of the cards di-
verted attention and caused embarassment.

Several conclusions may be drawn from the preceding
discussion of studies relevant to the intended research:
(1) positive reinforcement (approval. positive criticism) of
a speaker is recommended but is not always practiced. (2) pc-
sitive reinforcement (approval. positive feedback. positive
criticism) ought to encourage change in behavior but experi-
mental results with people as Ss do not unanimously support
the assertion. and (3) research is still needed on the pre-
sentation and effects of positive reinforcement within the
classroom setting as a means of developing speaking skills

in the student of public speaking.

Hypotheses

The major purpose of this study is to determine the
feasibility of varying modes of positive reinforcement within
the speech classroom with particular attention given to the
use of mechanical techniques. The personal-performance
skill selected for study is gesture. the movement of the
hands and arms. Deutschmann (1964) refers to gesture as one
of the four codes found in the communication situation.

.Ensenson. Auer and Irwin (i963) identify the importance of

15
gesture to the speaking situation by stating that gestures
enhance the speaker's oral symbols. emphasizing meanings and
underscoring oral words. For this reason. gesture was cho-
sen as the delivery skill to be used in this study.

One mechanical. electrical device that has been
found to be reinforcing is light on-set. Lockard (1963) has
reviewed light as applied to rodent behavior and found it a
reinforcer. It has been found to be a reinforcer for humans
as well. Krasner (i958) identifies light as one experimental
cue used in verbal-learning research. Greenspoon (1954)
also found light to be reinforcing to human subjects. Cere-
kwicki and Grant (1967) found light reinforcing to human
subjects when conditioning eyelid blinking.

In a study of speaker's attitudes toward tapic.
audience and self. Huenergardt (1967) used light as a rein-
forcing instrument. The SS were instructed that a panel of
20 lights was being controlled by a hidden. listening
audience. When a light flashed. a member of the audience
was to be regarded as favorably disposed to what was being
said. In fact. the E manipulated the number of lights. using
2/20. 6/20. 10/20. 14/20 and 18/20 as conditions of.rein-
forcement. A significant linear trend was found between the
percentage of audience approving and the speaker's attitudes.
Cahn (1966) used a similar light panel researching attitude
change of speakers toward their tOpics under conditions of
positive. negative and no reinforcement. The point here is
that light on-set has been used as a positive reinforcer for

college students.

16

In order to accept light on-set as an effective
classroom reinforcer. two criteria must be met: (1) light
must be found to be reinforcing (able to change behavior).
and (2) light ought to encourage the develOpment of gestural
skill at least as well as currently used classroom reinforcers.
In applying these criteria. this study was placed within a
beginning speech classroom. and the effects of light com-
pared with the effects of a mode of reinforcement in current
use.

The current mode used by instructors is the written
critique. Thus. for this study. two forms of written cri-
tique were created. both of which make use of ratings. One
form. Minimum Written Reinforcement. relayed to the student
a positive comment upon a single aSpect of his gesturing.
The.other written form. Maximum Written Reinforcement. re-
layed positive comments on a multitude of aspects of student
gesturing. The Light On-Set Reinforcement condition flashed
a light for each and every gesture found acceptable to the
experimental evaluator. This study will compare the impact
of the two written modes with that of the light reinforcement
condition as measured by the amount of change each condition
generates in the student.

Thus. the basic hypothesis under consideration is:

H1: Minimum Written. Maximum Written and Light

On-Set Reinforcement will have significantly
different effects in terms of the observed
improvement in student animation.

In the above hypothesis. "observed improvement" is

the difference between judges' evaluations of student spea-

kers before and after exposure to the reinforcement conditions:

17
"student animation" is the hand and arm movements displayed
during a classroom Speech.

Should the results of the study demonstrate that
differences have occurred. these differences should appear
in a particular order. Since the Maximum Written Reinforce-
ment will show positive evaluations to the student on
several aSpects of gesturing. it appears that this condition
would be providing the student with (1) more information on
what elements of gesturing are considered important by the
instructor. and (2) more approval of the gestural behavior
than would the single comment of the Minimum Written condi-
tion.

Therefore. it is hypothesized that:

H13: Maximum Written Reinforcement student im-
provement will be greater than student im-
provement in the Minimum Written Reinforce-
ment condition.

Even though the Maximum Written condition will pro-
vide more information to the student than the Minimum Writ-
ten condition. a similar weakness exists. Both conditions
delay the positive reinforcement. and both conditions offer
evaluations of the general gestural behavior of the student.
The Light On-Set Reinforcement. on the other hand. offers
immediate positive reinforcement for a specific act of the
hands and arms. This Light On-Set condition most utilizes
the theoretic conclusions drawn earlier and so should gene-
rate greater change than either written condition:

Hib‘ Light On-Set Reinforcement student im-

provement will be greater than student

improvement in the Minimum Written Rein-
forcement condition.

18
H10: Light On-Set Reinforcement student im-
provement will be greater than student
improvement in the Maximum Written Rein-
forcement condition.

If the major and secondary hypotheses of this study
can be affirmed by the change (improvement) data. then it
may be concluded that immediate. positive reinforcement pre-
sented by a light flash. may be a useful tool for speech in-
structors to use in refining speaking skills of student
speakers.

This chapter has provided justification for a re-
search concerned with communications between instructor and
student. A brief review of relevant research was presented.
‘aS well as a rationale for the study of gesturing as a
speaking skill and for the use of light as a reinforcer.
One major hypothesis and three secondary hypotheses have
been posited to be tested. The procedures used to gather

and.analyze the data under these hypotheses will be ex-

plained in the following chapter.

CHAPTER II
PROCEDURES

The purpose of this chapter is to explain the logis-
tics of this study and the materials and methods used to ga-
ther and analyze the data. Methods for the selection of the
subjects and for the formation of the measuring instruments
are included. as are the descriptions of the several treat-

ments.

Pilot Study

In order to establish procedures for this research.
a pilot study was conducted. The subjects were 14 freshmen
enrolled in Speech 102. the beginning speech course at Bow-
ling Green State University. during the summer quarter of
1968. This particular class was part of the second of two
five-week sessions that comprised the summer quarter.

The course syllabus requested that each student de-
liver five original speeches. one to be given each day of
the week. These speeches could be on any t0pic of the stu-
dent's choosing but could not be a "demonstration" speech.
Each Speech was to be no fewer than two and no more than
four minutes in length. Since the summer classes last two
hours. the time restrictions permitted every student to
speak every day.

The purpose of the pilot study was to determine
i9

20

(1) if light could be conveniently used within the classroom
setting. and (2) if light on-set would be found reinforcing
to the students in a classroom. The answer to the first
question was found by examining judgments of the instructor-
experimenter and a graduate student associate. The answer to
the second question was found by measuring changes in the
frequency of gesturing from speech 1 to speech 5.

In order to measure changes in frequencies over time.
a base-rate for each student's gesturing was established.
This base-rate period included the entire speech 1 and the
first 1% minutes of speech 2 during which time no light rein-
forcement was presented to the student. The associate.
acting as an objective counter. watched and counted every
gesture (hand and arm movements) emitted by the students.
The base-rate ratio of gestures to time was computed by
using 15 seconds as one unit of time. Each student had a
score which was the average number of gestures emitted for
a 15-second time period. This rate of emission was computed
for every speech. but only the ratio at speech 1 and speech
5 was used in analysis.

After the base-rate period. the light was flashed
for each and every gesture emitted by the student in the re-
maining speeches. (The associate had been counting similar
gestures.) The flash came from a green light bulb of 25-watt
power and lasted approximately two-tenths of a second. The
bulb sat in a porcelain base placed immediately in front of
the speaker's lectern at the front of the room. The operator

sat with the counter in the center of the room within easy

21
view of the lectern.

Each day a different Speaking order was followed.
and each day the Ss were reminded that the light meant "good
gesture." Panels of three students provided free-response
comments to the speakers at the end of each speech round.

If time permitted. oral comments were made by the instructor.
usually dealing with some error common to the majority of
the class for a particular speech round.

In order for light reinforcement to be considered
effective in the classroom setting. the students exposed
should increase the number of gestures emitted during their
oral presentations. A perusal of the data collected from the
pilot study suggested that no change had taken place in the
emission rate of the students. In fact. 10 of the.14 stu-
dents showed a decrease in the number of gestures exhibited
during speech 5 compared with speech 1. Analysis of the
data by the Sign test (Siegel. 1956) showed this result to
reach p.<:.090 level of Significance.

Based upon the data. experimenter observations. and
discussions with the counter. four recommendations.for the
main research grew from the pilot exploration:

(1) The light should be shielded from the other
class members. When the light was first flashed during
speech 2. the nonspeaking class members laughed and whis-
pered.. Reactions similar to these occurred for every speaker
during that day. This result can be explained within the
classical conditioning paradigm in which the negative peer

responses may have become associated with the light.

22

(2) Greater control over the types of speeches
should be exercised. While some types of speeches encou-
raged physical movement (a persuasive speech for civil
rights): others did not (a eulogy to Robert and John Ken-
nedy).

(3) More restrictive standards for gestures to be
reinforced should be determined. Counting and reinforcing
every hand and arm movement was not realistic in a beginning
speech class where much of the student's movement springs
from fear and nervousness. When such nervous movements were
reinforced. the watching students seemed confused about the
criteria for evaluation. A "Shaping" procedure seemed to be
more appropriate for beginning speakers.

(4) A content-related gestural evaluation procedure
Should be established. Instead of a counter. some instru-
ment apparently needed to be created to permit an instructor
evaluation of the movements within the frame of the content
and the minimum standards of "appropriate" physical behavior.
The statistical results of the study indicated that the num-
ber of gestures decreased. However. it was the opinion of
the experimenter and the counter that the quality of gestu-
ring (size of movement. relation to content. etc.) did im-
prove during the five Speeches.

How these recommendations were incorporated into the
main research will become apparent in the remainder of this

chapter.

23
Experimental Conditions

As suggested previously. the intent of this research
was to observe and measure differences between classes ex-
periencing dissimilar modes of positive reinforcement. The
written mode was chosen because (1) the classroom setting re-
quired this study to be as realistic to the standard Speech
course as possible. and the written critique is used most by
the teaching assistants of the course: (2) most speech in-
structors use written comments to students (Lothers. 1966):
(3) speech instructors seem to prefer written.remarks as a
means of focusing the student's attention upon strengths and
weaknesses peculiar to him. and as a means of supplying the
student with a permanent record of his own progress (Lothers.
1966).

Besides having current acceptance. the written mode.
particularly written scales. has several other qualities
recommending its use in this study. The content of the
scales. that is. what elements of gesturing are included.
can be standardized. This standardization assists in redu-
cing.variability which arises from instructor idiosyncrasies
concerning gestural elements or from differential nomencla-
ture. Furthermore. phrases expressing the evaluative judg-
ments of the observers can be standardized. which provides a
control of the range of written reinforcement.

A written scale has at least two advantages over
oral comments (the other mode of instructor criticism most
frequently used). The content of the oral comments may in-

teract with verbal and nonverbal codes. strengthening or

24

weakening the reinforcement. depending upon the perceptions
of the student. Such interaction may occur even if the oral
comments have been pre-arranged and standardized eXperimen-
tally (Bostrom. 1963). In addition. public oral comments.
if given immediately after the completion of a single
Speech. may have a disruptive effect upon the speaker(s)
following (Miller. 1964).

However. the written mode in the light of learning
theory suffers from two distinct disadvantages: (1) reward
is delayed. and (2) specific behavior can not be reinforced.

As mentioned previously. light has been found rein-
forcing to college students. Furthermore. a light flash.
because it need not be disruptive to the speaker (as are
oral comments). can be used while a student is speaking.
Light. therefore. permits the instructor to contiguously
reinforce specific speaker behavior. Also. light has an
advantage over other mechanical reinforcers in that its ap-
paratus may be relatively simple and easy to transport and
assemble. The modes of classroom reinforcement to be com-
pared in this research. then. are written scales and light
on-set.

Inherent within each mode are differing amounts of
reinforcement. A written scale. for example. may permit
several different negative and positive values for each item
on the scale. The number of values per item may be varied
according to the wishes of the experimenter. The number of
values and their placement upon a visual continuum may then

appear to be of different amounts of reinforcement. This

25
perception depends upon the terminal points of the continuum
and the placement of values along that continuum. It is
assumed. for example. that the value judgment "excellent"
would be perceived as a greater amount of positive reinforce-

ment than the evaluation "okay" on the following continuum:

Superior Excellent Very Good Good Okay

By standardizing the number of values available for each and
every item on a scale. some control of the perceived amount
of reinforcement may be exercised.

In the structuring of the experimental conditions
for this study. it was assumed that a scale containing a
greater number of items than another will be perceived as
offerring more reinforcement. That is. if a scale contains
six items of five values each. this appears to be greater
than a scale with only one item of five values.

Based upon these apparent differences. a minimum and
a maximum written reinforcement condition was selected.
Both conditions used the same number of values (a range of
four evaluative judgments) and were to be delivered to the
student at the same time (immediately upon the conclusion
of his speech). They differed in the number of items on the
scale presented to the student. The minimum written scale
consisted of only one item. while the maximum written scale
included nine items.

The choice to use only one scale in the minimum
written condition was based on the observation that most in-

structors' critiques often contain only one or two remarks

26
on several variables found in the speaker's performance.
Time. of course. forces the critiquing instructor to limit
the characteristics he wishes to emphasize to the student
and the number of comments he can make concerning those few
characteristics.

How the items of the scales were chosen and the en-
tire instrument was distributed will be discussed later.

In the light on-set condition. the value system was
dichotomous --- present or absent. However. since light
could be flashed for each and every desired gesture. the
amount of reinforcement was related to the number of ges-
tures emitted by the speaker. The more the Speaker movement
was deemed acceptable by the evaluator. the greater was his
reinforcement. It is this characteristic that could make
the light on-set different from either the written conditions:
reinforcement could be presented as often as the speaker's
behavior met certain standards. The written conditions.
however. could be presented to the speaker only once. and
then. some time after the desired behavior had been emitted.

The characteristics discussed above Operated in this
research as three reinforcement conditions: Minimum Written
Reinforcement (MNW). Maximum Written Reinforcement (MXW).
and Light On-Set Reinforcement (LOR). Figure 1. page 27.

provides a brief comparison between the conditions:

27
Written Light

Minimum Maximum

 

1-1tem scale 9-item scale Flashed for all ges-
tures meeting minimum
standards of "accep-
tability"

4 positive values 4 positive values 2 values: present
and absent

Given Ss immediate- Given SS immediate- Given SS immediately
1y after his speech 1y after his speech after gesture emitted

Specific evaluation Specific evaluation General evaluation
referring to the referring to the of specific gestural
general gestural general gestural behavior of the Ss
behavior of the SS behavior of the Ss

 

Figure 1. Description of Experimental Conditions
As can be seen in Figure 1. each condition was an aggregate

of several variables: time. amount and kind of reinforcement.

Experimental Evaluators

The experimental evaluators (EE) who administered
the reinforcing conditions were doctoral students in the De-
partment of Speech at Bowling Green State university. All
had taught at least four years before beginning their
graduate studies. Several discussions took place during
which time the purpose of the research was explained and the
EE'S place within the research clarified. Each EE worked
with the reinforcement instrument prior to the research.
During one such training session. two student Speakers were
invited to prepare and give speeches to be witnessed by the
HE. Each EE completed his written ratings (or flashed the

light) for each Speaker. These evaluations were then discussed

28

and any deviations particularly analyzed. The training
session continued until each EE felt confident in the use of
his instrument and in the similarity of his evaluations with

the other two EE.

Subjects

Subjects (SS) were undergraduate students enrolled
in the beginning speech course at Bowling Green State Univer-
sity during the spring quarter of 1969. The course struc-
ture was divided between a lecture and two drill sections
per week. Nine drill sections with 104 students registered
were chosen for the eXperiment.

Of those students who originally registered. 16
missed one or more of the first seven drill section class
periods and were disqualified as 85. These students were
not informed that they had been disqualified. thus they con-
tinued to complete the course requirements. The 88 students
who completed the experiment and who were eventually included
in the analysis were divided between the experimental condi-

tions as follows:

Table 1. Distribution of Subjects by Treatment

 

Treatment Enrollment Disqualified _‘ggtal
Minimum Written 35 6 29
Maximum Written 34 3 31
Light On-Set 35 7 28

 

Total 104 16 88

 

29
There appeared to be no systematic pattern in the SS
disqualification. Table 2 identifies the times at which stu-

dents were disqualified as SS.

Table 2. Subject Disqualification by Speech Assignment

 

 

pre 1 2 3 4 post Total
MNW 2 1 1 2 6
MXW 1 1 1 3
LOB 4 2 1 7
Total 2 0 6 3 1 4 16

 

Several of the Ss missed more than one speaking assignment.
However. the first time they were absent constituted the
disqualification time. While speech 2. the fourth class day
and second treatment period. shows the greatest loss. there
appears to be no systematic loss within treatments or across

speeches.

Measuring Instruments

 

To determine the effects of differing modes of posi-
tive reinforcement upon the gesturing of student speakers.
an instrument for recording judges' evaluations was created.
In creating the measuring instrument. a free-response ques-
tionnaire was sent to members of the teaching faculty and
graduate students in the Department of Speech at Bowling
Green State University (see Appendix A. page 78). Each was
asked to list characteristics of good gesturing which he

sought to teach in the beginning Speech course. From the

30

responses given. 14 characteristics were identified. These
14 were then randomly listed in a second questionnaire sent
to the same personnel (see Appendix A. page 79). The tea-
chers were asked if there appeared to be overlapping charac-
teristics in the listing. If so. they were to identify the
characteristics that seemed similar. Also. they were asked
if any important characteristics did not appear in the lis-
ting. From these reactions. a total of nine characteristics
of good gesturing were found. The characteristics identified
were: "appropriate to content." "clarity." "completeness."
"coordinated with other bodily action." "definiteness."
"directed toward the audience." "fitting the speaker."
"Spontaneity" and "visibility." The nine characteristics
were used to form part of the total measuring instrument.
They were placed on a scale resembling the semantic diffe-
rential (seven-step). The extreme positive position was
assigned the value "7." while the extreme negative position
received the value "1." The instrument was assumed to be
unidimensional (see Tables 3 and 4. pages 43 and 44).
allowing the combined judge ratings for all nine items to
constitute a student's Gestural Evaluation Score (GES).

Given that gesturing bears relationship to the total
physical behavior of a speaker (posture. walking. etc.).
then improvement in gesturing ought to improve the im-
pressions an evaluator might have of the total physical be-
havior. To test this improvement. an additional measure was
created. Five sets of paired adjectives were selected from

the Berlo. Lemert. Mertz (1961) research on credibility.

31

The adjectives chosen from the Dynamism factor were: "bold-
timid." "aggressive-meek." "energetic-tired." "confident-
unsure" and "colorful-dull." The semantic differential for-
mat was used to present the Dynamism items to the judges.
By adding together the judges' evaluations for each item and
adding the items together. a total Physical Performance
Score (PPS) was computed and analyzed. A sample of the in-
strument is contained in Appendix B.

Since the Speaker evaluations were made by panels of
three judges each. a GES for a single SS could range from a
low of "27" (all judges awarding "1" for all nine scale
items) to a high of "189" (all judges awarding "7" for all
nine items). The PPS could range from a low of "15" to a

high of "105."

Legistics
The characteristics of good gesturing found by

polling the teaching faculty were used not only in the mea-
suring instrument. The nine elements formed the content of
the item scales used in both written reinforcement condi-
tions. However. the number of items (as mentioned earlier)
differed between the two conditions.

Minimum Written Reinforcement (MNW) Condition

In the MNW condition. the characteristics of good
gesturing were divided into nine separate scales. The sepa-
rate scales were attached to the inside of a standard 9-by-
12-inch file folder. When the folder lay open. all scales
were clearly visible to the EB. As he watched the speaker.

he decided which characteristic of good gesturing was best

32

performed. After deciding which characteristic he wished to
reinforce. the BE then circled the evaluative comment (of the
four available) which most clearly expressed his judgment of
that particular characteristic. When the student had finished
speaking. but before he had received the instructor's written
comments. he was given the gestural scale completed by the
HE. A sample minimum reinforcement instrument can be found
in Appendix C. page 82.

For each class he evaluated. the ER had a folder con-
taining the nine scales and a class roster.

Maximum Written Reinforcement (MXW) Condition

In the MXW condition. the characteristics of good
gesturing were combined into a single evaluation sheet of
nine scales. The evaluation sheets were attached.to the in-
side of a.standard 9-by-12-inch file folder.. As the EE
watched the speaker. he judged him on all nine characteris-
tics. The EE circled the evaluative comment.which expressed
his judgment of the speaker's performance of each and every
characteristic. A sample maximum reinforcement instrument
appears in Appendix D. page 85.

For the three classes he evaluated. the EB had a
folder containing the scales and a class roster.

Light On-Set Reinfgggement (L9H) Condigigg

The task of the HR for the LOR condition differed
from the tasks of either EE for the written conditions.
Whereas the HE for both written conditions made evaluations
of specific gestural characteristics for a completed speech.

the BE for the LOR made an evaluation of each and every hand

33

and arm movement of the speaker within a single speech. The
nine items of good gesturing were applied to each gesture.
an evaluation made. and the light was flashed or not. If
the gesture was considered "acceptable." the light flashed;
if the gesture was considered "unacceptable." the light was
not flashed.

The light apparatus was 18 inches wide. 11 inches
high and 12% inches deep. The base consisted of a piece of
Z-by-o-inch lumber chosen to ensure a solid mounting for the
light fixture and to give stability to the whole apparatus.
The Sides. tOp and back were made of cardboard. When resting
upon a desk, the apparatus appeared to be a cardboard box
with one long side missing. The light was mounted on the
board in the center of the box. The fixture itself was a
lightbulb encased in a green glass cone resting in a black
metal base. A ié-by-Zfi-inch Opening covered with green
gelatin was located two inches from the rear-center of the
tOp of the box. This arrangement enabled the BE to see the
light as it flashed. yet did not permit light to escape.

The power for the light was a single six-volt bat-
tery encased in a box 6-by-6%-by-3 inches. To flash the
light. the BE pushed a button on tOp of the battery casing.
While the bulb was small and powered by a mere six volts.
its flash produced 25 candlepower. sufficient light to be
seen from several feet. However. the bulb's power was not
enough to produce light reflections off objects in the room.
which might distract the audience or cause them to react.

As a further precaution. the ER sat closest to the Speaker's

34
lectern. alone in the front row.

The EE for the LOR also had a 9-by-12-inch file fol-
der for each class he worked. It contained a class roster
and a listing of the nine characteristics of good gesturing
used in the study.

Prpcedures for Data Collection

In order to minimize classroom differences. only
three drill instructors were chosen. They were selected
with the assistance and recommendation of the coordinators
of the beginning course. All three had been undergraduate
speech majors: all were graduate assistants majoring in the
area of Rhetoric and Public Address. Each instructor taught
three drill sections. the normal graduate teaching load du-
ring the spring quarter. To further control for instructor
variability. the three experimental conditions were randomly
assigned to the sections of each instructor. Thus. every
instructor had one section in each of the three conditions.
To prepare them for the research and their place in it. the
three instructors met with the researcher in.an extended
session prior to the beginning of the spring quarter. Indi-
vidual discussions also were held prior to and during the
running of the experiment.

Classroom_Assignment§

Before the start of the spring quarter. the drill
instructors and the experimenter structured the course. It
was decided that the first seven drill sections would be de-
voted to the administration and execution of the experiment.

The intent of the common syllabus was to discourage variability

35

due to differences in speaking assignments. A complete
assignment sheet appears in Appendix E, pages 86-88,

During the first class period of the quarter. the
procedural matters of the course and the eXperiment were
clarified. Each student received a cOpy of the syllabus.
and each assignment. including those of the SXperiment. was
explained. and questions were answered. Grading and atten-
dance were also discussed. and the class roster corrected.

It was explained that grades would be given by the instruc-
tor. but that the SS would not receive them until after the
first Six speeches were completed. The SS were told that two
grades would be recorded --- the highest grade given and an
average of the other five speech grades. The SS were told
that they would be treated differently from other sections
of the course. but that such "eXperimentation" was common to
a multi-sectioned course like Speech 102.

Administration of the Pre-Test

Sixteen faculty and teaching graduate assistants
were randomly assigned into nine panels of three judges each.
While several judges evaluated together on more than one panel.
no two panels were identical. They were instructed to evalu-
ate the hand and arm movements of the students. using the
first nine scales of the evaluation Sheet. The final five
Dynamism scales were to be used to indicate their impressions
of the general physical behavior of the speakers. Posture.
facial expression and walking were to be evaluated by these
scales. These instructions and a c0py of the evaluation

sheet were given to the judges prior to the time that they

36
were to hear speeches.

The students Spoke in alphabetical order. The in-
structor called each student by name and spelled the last
name of the student for the benefit of the judges. Such a
procedure kept the instructor in control of the class and
allowed him to set the pace of the speaking order. Since
the instructors were also evaluating. critiquing and grading.
it was necessary to permit them to determine the amount of
time needed between Speakers.

After the 88' name was called. he delivered his
speech from the front of the classroom. Upon completion of
the speech. the SS received his written critique from the
drill instructor. The panel of judges independently marked
the evaluation sheet but did not inform the SS. the instruc-
tor. or the other judges of their evaluation. When forms
and critiques were given to the SS. the instructor called
for the next SS. and the process was repeated until all SS
had Spoken. If time permitted. oral remarks were made by
the instructor on any aspect of Speaking except that under
investigation.

Administration of the Reinforcement Treatments

During the next four drill section periods. other
speaking assignments were performed. and the reinforcement
treatments administered. The instructor called the SS and
spelled the last name for the benefit of the eXperimenters.
While an alphabetical order was maintained. each day the in-
structor began at a new. predetermined point in the alphabet.

At the conclusion of a speech. the SS received a

37

written critique from the instructor and a completed evalua-
tion sheet from the BE. The material from the EE preceded
the material from the instructor. Once the SS had received
his material. the next Speaker was called. If time permitted.
an oral critique by the instructor was given. The HR said
nothing to the students or the instructor.

In the LOR group. the procedures were the same ex-
cept that the SS were told the light meant "good gesture"
(see Appendix F. page 89). Only the instructor wrote com-
ments to the SS. The ER in the LOR group sat in the first
row in front of the lectern. The light mechanism was placed
before him so he could clearly see the SS and the light in
the box. A light flash lasted approximately two-tenths of a
second --- the time needed to press and release the button.

Administration of the Post-Test

Seventeen judges were randomly assigned into nine
panels of three judges each. No panels were the same. nor
were any post-test panels identical to any pre-test panels.
Twelve judges evaluated during both time periods. but only
two heard in the post-test students from classes they had
heard in the pre-test. The same evaluation sheet was used.
leading to a second GES and PPS for each student.

Classroom procedures used in the preceding five
speaking periods remained constant for the post-test. As in
the earlier sections. the instructors were not informed of
the judges' evaluations. AS each panel returned its evalua-

tions. the experimenter checked them for errors. Fortunately.

38
no judge's errors caused a 33 to be disqualified from
analysis.

Upon completion of the experiment. the drill sec-
tions returned to more normal procedures. That is. fewer
Speeches were given. and persons from outside the class were
not present.

After the judges had completed the post-reinforce-
ment ratings. the differences between the evaluations of the
pre- and post-test speeches were computed. These difference

scores constituted the basic data for this study.

Statistical Design

Since the purpose of positive reinforcement is to
improve the gestural skills of the students receiving the
reinforcement. it was decided to use improvement scores as
the basic data for analysis. That is. the improvement score
was the difference for each student who participated in the
research as shown by changes in both the GES and PPS from
Time 1 (T1) to Time 2 (T2). The relative effect of the
modes of reinforcement presented to the students would be
reflected in variations of these differences.

Using intact educational groups (speech classes) for
SS. there was a possibility that significant differences in
gesturing ability existed before the reinforcement condi-
tions were applied. To detect such differences. a one-way
analysis of variance was computed on the pre-test GES and
PPS evaluations. If a significant F-value was found. a
single-factor covariance design was to be used in which the

pre-test GES and PPS for each SS would be the covariant for

39
analysis of the post-test GES and PPS. On the other hand.

if no significant F-value was found for the pre-test data.
the basic statistical design was to be a one-way analysis of
variance. using difference (improvement) scores as the de-
pendent variable for both the GES and PPS. Should the analy-
sis of the improvement scores (by either covariance or one-
way analysis of variance) have resulted in a significant
F-value. then t-tests were to be applied to make individual
comparisons between groups.

The above statistical design was to hold if the data
did not violate the mathematical assumptions of the parame-
tric analysis of variance routines. If. however. those data
did not satisfy the assumptions. then nonparametric tech-
niques would be used in analysis. Specifically. the Kruskal-
Wallis one-way analysis of variance would be applied to the
difference scores. If a significant difference was found
between the three modes of reinforcement. the Mann-Whitney
g test would be computed to make individual comparisons be-
tween the conditions. These two statistics approximate the
parametric tests described above.

Regardless of the statistical procedures employed in
analysis. the major null hypothesis for the study of the GES
may be stated:

Hock”! MNW fi/\.wa =/R.LOR
The major null hypothesis for the PPS. using improvement
score for analysis. is:

HOPE/\.MNW “IL mxw ”)1.Lon
If the major null hypotheses were rejected. then further

4O
statistical analysis would test the following hypotheses and
their alternatives regarding the direction of the differences

for both the GES and PPS evaluations:

HOa‘ Ix mm =/‘\ nxw Hla‘fi xxw <4 MNW
HOV/K MNW 3/1 LOB Hlb‘A LOR<A MNW
Hog/“k LOR =/\ xxw ch‘fl Loa<lwxw

Summary
This chapter has explained the pilot study. the se-

lection of subjects and the way in which the instruments for
reinforcement and measurement were created. The speech class
course structure was explained. as well as the procedures for
gathering and analyzing the collected data. The results of
the statistical comparison and an explanation of the results

of the experiment are reported in the next chapter.

CHAPTER III
RESULTS

The analyses reported in this chapter were conducted
to test the hypotheses (expressed previously) concerning dif-
ferences between classes in a beginning public speaking
course after each had been exposed to varying SXperimental
conditions. The object of the statistical tests was (1) to
determine if there were differences between the treatment
groups. and (2) to determine (by cellular comparisons) where
those differences occurred. As preliminaries to the actual
testing of the hypotheses. several questions had to be an-
swered: Was there justification for adding across items of
the measuring instruments to produce a total GES and PPS
score? Did the judging panels reach some minimum level of
rater reliability? Have the assumptions underlying the F-
and t-tests been satisfied to warrant the application of
these two statistics? Since the eventual analyses of re-
sults depend upon satisfactory answers to these questions.
they will be investigated first before proceeding to the

testing of the major hypotheses of the research.

Preliminary Analysis
The individual items of the GES scale must bear some
positive relationship in order to add across the items and

arrive at a Single total score. Using the judges' responses
41

42

on the nine gesturing items from the pre-test. product-moment
correlations were computed between each item and every other
item and between each item and the summated score. The cor-
relation between the separate items and between the items and
the total score indicate the assumption of unidimensionality
to be warranted. Table 3 (see page 43) contains the inter-
correlation matrix for the GES items and the GES total score
at T1. The correlations are judged sufficiently high to con-
clude that the items share a positive relationship.1

Since several judges evaluated only at T2. correla-
tions were computed for the GES items and summated at that
time as well. It seemed possible that the additional judges
might not use the instruments as those who had already eva-
luated Speeches. If so. their inexperience might change the
relationships between the items. The intercorrelation matrix
for the GES at T2 appears in Table 4. pagelH4. Once again.
the high correlations suggest the assumption of unidimen-
sionality to be merited. Apparently. judges who rated only
at T2 did not negatively influence the scale.

In using the PPS as a measure of general physical
behavior. the assumption of unidimensionality was made again.
The correlations reported in Table 5. page446. support the

assumption that the PPS also is a unidimensional measure of

 

1While additional statistical procedures are available
for testing unidimensionality (factor analysis). because of
the small number of evaluations given plus the confounding in-
fluence of repeated measures. it was decided to support the
assumption with the correlation matrix alone (Guildford. 1954).

43

 

 

--- mmm wee ems mma cam mmm new 3mm mam Apoev Hence
--- 0mm mam oem was mam Sea mam Hoe Apnea
cocoaess ens cacao» nonconaa
-n- cam one saw New emu mam mam xooov
eoapoe aaaeom can: eouuedesooo
Ill me mmm Nmm ORB m+~® mow afloufloo 063 Cu. oud%mmw.wdg¢
nu: mom new mom new mom Loewe msooceceoem
-u- nee new Sea mom gauze Hewzsez
-u- mmm Ham use kaoov oucaegoo
nu- New new gauge cascades
nu: Hmm lance apaaeao
::- Ama>v oapama>
mmm mmm mum _mm¢ .mmm mmm mam mum mam mm»
H cede .onoom zoapmsHm>m HSHSpmow on» How maan nookpom msoapsHoahoo .m cance

 

44

8

0mm New mam mom ewe com mam mam Hmm “page Hence

--- oma see was new new mam mam mmw Andes
oozcduss on» Samson pepoehdm

-u- Ham sow omm How sea Nma new Aooov
sodpo< aadeom and: eopaeaesooo
--- mmm mmm mam Sam mom oom peopeoo on» op ouewmmwwee<
an: wmm mam mmm 3mm 2mm Aoamv msoesepSoam
-n- new man mam Ham gauze assessz
--- use one was laces opnseaoo
--- use see icons oaseseoa
nu- omm Aeaov sadneflo

nu- Ama>v cabama>

Ham 000 mm< 0mm 062 800 Mom 6H0 ma>

 

N cede .eaoom soapesHs>m Hmhnumoo one How maouH soozpom msoapmaohhoo .3 cance

45
gestural behavior. Table 5 (see page 46) reports the inter-
correlations for both T1 and T2 and. at both times. the cor-
relations seem sufficiently high. The decision to add across
items of both measures in order to use total GES and PPS
scores was sustained by the correlation data.

As reported earlier. it was assumed that hand and
arm movements do bear some relationship to the general physi-
cal quality of a speaker. Given that assumption. the PPS
items were intended as a second measure of gestural perfor-
mance. However. for the PPS to be a reliable measure (for
the assumption to be supported) required that the items of
the PPS Show positive correlations to the items and total
score of the GES. The results of product-moment correlation
computations for both time periods are reported in Table 6.
page 47. The figures on the left report the relationship of
the GES items at T1 to the PPS items at T1: the figures on
the right report the correlations between the GES and PPS
items at T2. While the correlations are not as high as
those computed for the GES and PPS as separate units. the
two measures appear sufficiently related to warrant using
them as two measures of related. but not necessarily the
same. characteristics. It Should be noted that the total
GES correlated with the total PPS at T1 = .840 and T2 = .803.
accounting for over 60% of the variability. The decision to
use both the GES and PPS as measures of gestural performances
appeared reasonable in light of the correlational data re-
ported.

The next question raised concerned the reliability

46

 

nu: mmm mmm
nu: mom

mmm
mam
0mm

osm

mmm
mam
New

0mm

awn

mmm
0mm
enm

new
Ham

Hom

«:8

use one
--- was
new .mmu

mom
mmm
«mm

mom
0mm
mew

03m

was

Laney H : Hence
Aaooe Hacsoaoo

kneel
secedeeoo

Aosmv awacsm

Awwev
cbdmmohwws

Aflomv eaom
Hem Aanev a : Hence
new “Hove Hacnoaoo

sow Asoov
peoefleeoo

mam loamy kahuna

mam Amman

obdmmehmw<
-u: “Home efiom
Adm

 

N use H osae .oaoom cosmahouaom Hmodmhsm

use you maouH zeosuom msodusHohhoo .m cance

a?

 

mom
mmm

mmm

owe

mam
ans
New
cam
0mm
0N5

Nte
mmm

wee
own

mmo

Now

non
mmm
mmm
new
mmu

new
Hoo

mmm
mmm

«No

mmm

mmm
mom
035
was
mam
Son

200

mum
mom

can

mmo

mmn
own
amp
own
son
mmm

was
mmm

mum

How

cow
was
cam
mam
mma
see
was

man
nan

man

Now

an“
“we
was
was
«mm
mom
Hem

03m
emu

amw

0am

mes
new
new
new
mi.
was

«:8
mmm

mam
emu

3mm

new
now
won
moo
«mm
mum

new
:3

moo

mom
$05

was

emu

mom
was
Nam
moo
Nam
Nam

Saw
«mm

mum

ans

«am
now

mom

How

Hdpoa mmm

00:0«e54
on» Samson

eopoouaa
coasoe aadeom
gusto spa:
concedesooo

unannoo on» on
oumaaaoada<

uncondusoam
HeHSusz
opoaaaoo
ou«:«moa
huahmao
sandma>

 

N Una a made

.oaoom eosmaaomacm
Hcoamhsm use ewoom soapausbm amazemcu on» pom maepH soezpcm msoupsaoaaoo .w cance

48
of the judging panels. In other words. did the raters reach
a sufficient level of reliability to warrant the addition of
the judges' scores for each item? Guilford (1954) recommends
the use of intraclass correlations as a method for determining
reliability among raters. Essentially. the formula below
provides for an average intracorrelation:

rkk = VP ' ve
P

where rkk = variance for the people rated

Vp

variance for the peOple rated

V

e variance for error
The rkk was computed for each panel of three judges for each
item on both instruments for T1 and T2. The median relia-
bility at T1 and T2 for each instrument item of the GES is
reported in Table 7. page 49.

The rater reliabilities for the PPS are reported in
Table 8. page 49.

The rkk computations for each panel used in the re-
search and for each item in the GES and PPS scales are col-
lected in Appendix G. pages 92-94. While there is expected
variability between the panels. the reliabilities appear
high enough to warrant adding across judges of an evaluating
panel in order to arrive at a single score. Furthermore.
examination of panel reliabilities shows no treatment group
systematically affected by rater variabilities.

Having examined the a priori assumptions of this re-
search and found reasonable justification for using two mea-

sures (adding across items and judges). the next preliminary

analysis tests the underlying assumptions of the F- and

49

Table 7. Median Rater Reliabilities for the Gestural
Evaluation Score, Time 1 and 2

 

Igggg Time 1 Time 2

Visible 801 700
Distinct 742 714
Definite 668 648
Complete 586 753
Natural 760 557
Free 660 826
Appropriate

to the Content 606 576
Coordinated

with Other Bodily Action 639 671
Directed

toward the Audience 663 714
Total 836 678

 

Table 8. Median Rater Reliabilities for the Physical
Performance Score. Time 1 and 2

 

1292 2122.2. 2.1.22.2.
Bold 770 747
Aggressive 639 763
Energy 613 673
Confident 616 742
Colorful 766 814

Total 517 779

 

5O
t-tests. Both tests make the assumption of a normal distri-
bution. Examination of the raw score data at T1 shows a po-
sitively skewed distribution for each reinforcement condition.
Winer (1962) writes that the F- and t-tests are robust regar-
ding the normality assumption. even when the distribution is
"markedly skewed (p. 34)." Lindquist concurs. suggesting
that. for the F-value. the distribution is "insensitive to
the form of the distribution (1956. p. 86)." Given these
comments. the skewed form of the distribution seems insuf-
ficient reason to reject the parametric tests.

Both tests also assume homogeneity of variance. As
noted above. the F- and t-tests seem relatively uninfluenced
by failure to satisfy this assumption. Moderate amounts of
deviation may be accounted for by setting more rigid limits
for rejection. according to Lindquist (1956). However.
examination of the data suggests that heterogeneity in this
study might be extreme. This deviation was detected in the
standard deviations computed for the total GES and total PPS
scores at T1. Since these were the scores to be eventually
used in the analysis. it seemed necessary to determine if
the assumption of homogeneity was violated. Bartlett's Test
for Homogeneity of Variance. as described in McNemar (1962).
was performed. This test is appropriate for samples of un-
equal sizes. The results for the GES are reported in Table 9.
page 51. With such variability. the mathematical assumption

underlying the F and t routines appears to be violated.

51

Table 9. Bartlett's Test for Homogeneity of Variance on
Gestural Evaluation Score. Time 1

 

 

 

3.0. 3.13.2 n - 1
MNW 24.24 587.5776 28
MXW 43.42 1885.2964 30
LOR 22.82 520.7524 27
x2 15.3133 with df = 2,E<.001

 

While the initial observation of the PPS did not re-
veal such extreme variability as appeared in the GES. diffe-
rences did appear. and so these differences were also tested.
The PPS variability (reported in Table 10) approaches statis-
tical significance and So recommends that analysis of variance
might be suspect if used for analysis.

Table 10. Bartlett's Test for Homogeneity of Variance on
Physical Performance Score. Time 1

 

 

 

sgpg 5.0.2 n - 1
MNW 14.73 216.9729 28
MXW 20.60 424.3600 30
LOR 14.29 204.2041 27
x2 4.9505 with df = 2. p_<.10 but > .05

Kerlinger (1965) and McNemar (1962) both argue to
use parametric statistics because they are more sensitive.

and so. more powerful than are nonparametric statistics.

52
However. both also admit that serious violations of the
assumptions underlying the F- and t-tests may justify turning
to the distribution-free statistics.

In view of the departures from the assumption of homo-
geneous variances between the groups comprising the experi-
mental conditions. coupled with the deviation from normality
of the distributions. the decision was made to turn from the
F- and t-tests to applicable distribution-free statistics.

The test chosen to replace the parametric F was the
Kruskal-Wallis one-way analysis of variance (H statistic)
described by Siegel (1956). This statistic tests the hypo-
thesis that k independent samples had been drawn from the
same population. like the parametric F. but it uses ranks as
the variable. The H test assumes that the variable under
study has a continuous distribution. an assumption apparently
met by Speaker animation. A second assumption requires or-
dinal measurement which seems satisfied by the semantic dif-
ferential format used to collect the judges' evaluations.
The Kruskal-Wallis test. writes Siegel. "seems to be the
most efficient of the nonparametric tests for k samples
(p. 194)." When. in fact. the assumptions of the F-test are
met by the data and the H test is used. the power efficiency
of the H statistic is 95.5% when compared to the F-test.
This statistical test. applicable to groups of unequal size.
appears to be the most appropriate substitute for the para-
metric F.

Should difference be found in the improvement score

analysis. the Mann-Whitney U test. which also uses ranks as

53

the variable. would be computed to make individual compari-
sons between reinforcement conditions. This statistical
procedure for data analysis roughly approximates the para-
metric t-test when further analysis of a Significant F is
required.

The final preliminary analysis was to test for group
difference at T1. In order to minimize the possibility that
differences at T2 result from initial differences between
the groups and were not the result of the experimental ex-
periences. the H test was computed for T1. With the level of
rejection .05 and df2 = 2. the chi-square distribution re-

quires 5.99 to be significant.

Table 11. Kruskal-Wallis Test on Gestural Evaluation Score.

 

 

Time 1
Sum of Ranks g;
MNW 1444.5 29
MXW 1248.5 31
LOR 1222.5 28
R 1.6811 with df = 2. andcx;= .05 n.s.

 

No such level is reached by the data reported in
Table 11. The H statistic. however. is influenced by ties
in the ranks. Of the 88 rankings. 51 (57.95%) were ties.
The recommended correction for tied ranks was computed.
using the formula

Hcorr =

54

where T = t3 - t: t is the number of tied observations in a
group of tied scores

The result of the computation was that Hcorr = 1.684. which.
with df = 2 and the significance level = .05. was not sig-
nificant.

Similar computation was made for the PPS at T1. with
the conclusion that no significant differences existed be-
tween the groups at that time.

Table 12. Kruskal-Wallis Test on Physical Performance Score.
Time 1

 

 

Sum of Ranks 34 H d; 52$ 53 p
MNW 1526.5 29 5.033 2 .05 5.99 n.s.
MXW 1355 31
LOR 1034.5 28

 

However. with the H approaching a significant chi-
square and the H usually increased when corrected. the cor-
rection for tieS was computed. The resulting H increased to
Hoorr = 5.038. which was not Significant. From the results
of testing at T1 for the GES and PPS. it seemed reasonable
to conclude that the groups were drawn from the same popula-
tions. and that any differences which might appear at T2
would be a result of the experimental conditions. Therefore.

change score data may be used to test the research hypothesis.

Testing the Research Hypothesis
Since improvement was hypothesized for all groups at

T2. the nonparametric H was applied to difference scores

55

computed for each Ss by subtracting his evaluation at T2
from his judges' evaluations at T1. The differences for
each SS were then rank-ordered for the combined groups.
Those ranks were used in the H statistical computation with
the results reported in Table 13. While the H does not
reach Significance. the value falls between a chi-square of
3.22 (p = .20) and 4.60 (p - .10).

Table 13. Kruskal-Wallis Test on Gestural Evaluation Score
Change Scores

 

 

Sum of Ranks 25 H d; £25 33 p
MNW 1054 29 4.163 2 .05 5.99 n.s.
MXW 1467 31
LOR 1395 28

 

Forty of the scores (45.45%) were involved in tied
ranks. so the correction was computed. The value of H in-
creased to Hcorr = 4.164. which did not reach the necessary
X2 = 5.99 to be considered statistically significant. With
no Significance in the GES. it was suspected that the PPS
would also Show nonsignificance. The results of the Kruskal-
Wallis test for the PPS are reported in Table 14.

Table 14. Kruskal-Wallis Test on Physical Performance Score
Change Scores

 

215$ 2

ICE

Sum of Ranks 24
MNW 1109 29 3.007 2 .05 5.99 n.s.
MXW 1387.5 31
LOB 1419.5 28

56

With 50 of the 88 ranks tied. the correction for
ties was computed. The Hcorr for the PPS became 3.0089.
which fell between 2.41 (p = .30) and 3.22 (p = .20). but
was not significant.

Despite the fact that neither the GES or PPS reached
statistical Significance. apparently changes were taking
place within the treatment groups. For both dependent mea-
sures. but particularly the GES. the data tends toward the
predicted direction. Given that tendency. it was decided to

look at the data for evidence of change.

Trends in the Data

 

In Table 15. page 57. the means and medians. by ex-
perimental condition. are reported. The differences that
appear between these two measures of central tendency may be
due to a few extremely high scores which appear in each group
and "pull" the means upward. For example. with a potential
GES of 189 at T1. three SS in the MXW received scores higher
than 155. and four other SS received scores of 101 or higher.
In the MNW. two Ss scored over 120. while in the LOR. two 33
scored over 100. At T2. the total number of scores over 100
was 13: two in the MNW. Six in the MXW and five in the LOR.
The point to be made is that the median may be more represen-
tative of the groups' scores than the mean. It is to be
noted. however. that both measures change between the two
experimental times.

It is apparent from Table 15. page 57 . that changes
were taking place in the various groups. At T1. for both

evaluation scores. the means and medians of the MNW were

5?

larger than those of either of the other groups. However.

at T2. the MNW means and medians are lower than either group
in the GES and are lower than the LOR in the PPS. The dis-
placement in central tendency measures. particularly in the
GES. may be due to the limited amount of reinforcement given
the SS in the MNW condition. That is. one evaluative comment
may not relay sufficient. useful information to the beginning
Speech student for him to be able to improve his gestural be-
havior. Furthermore. the data seem to indicate that the MNW
condition differed from the MXW condition. supporting (post

hoc) a pre-research assumption.

Table 15. Means and Medians by Conditions. Time 1 and 2

 

 

 

§§§ §£§
1 7; 1 2.
64.03 53.79 mean 48.55 44.76
MNW
60 47 median 46 43
63.39 71.81 mean 44.84 44.65
MXW
54 58 median 40 41
L0 58.39 71.57 mean 38.71 45.39
R
51 67 (computed) median 35 (computed 41

 

But. a countering increase in the evaluations of the
MXW and LOR was also occurring. In fact. the evaluations of
87 Ss changed between T1 and T2 : 43 Ss showed positive

change. and 44 SS Showed negative change. The description

58
of the SS movements between the two time periods appear in
Table 16. The nonparametric sign test (Siegel. 1956) was
applied to the data. It is to be noted that the direction
of change in the MNW is in the wrong direction. Sign test
analysis shows statistical Significance for the number of SS
moving in the negative direction. The direction of movements
of the groups (seen in Table 16) seems to relate to the mag-
nitude of the changes reported in Table 15. page 57.

Table 16. Number and Direction of Evaluative Changes by
Conditions. Gestural Evaluation Score

 

my; mxw LOR

Sign Test 2 .0445 .2358 .2877

 

However. despite these apparent changes. no statis-
tical support for the hypothesis that differences will ap-
pear between the experimental groups after exposure to the
treatments has been found. On the other hand. it is not
possible to conclude that the treatments had a random effect.
Tendencies appear. particularly in MNW. where the direction
of change was opposed to that predicted. which seems to
indicate the experimental modes of reinforcement had a selec-
tive impact upon the Ss.

, Since no significant differences appeared between
conditions. it was decided to look within groups to see if

changes had taken place. The information presented in Tables

59
15 and 16. pages 57 and 58. would encourage the suspicion that
the various modes of reinforcement did not improve the gestural
performance of the 33. To test that suspicion. the nonpara-
metric Wilcoxon Match-Pairs Signed-Ranks described by Siegel
(1956) was computed for each group. With large samples
(N = 25), the Wilcoxon approximates the Z-distribution with

the formula

2 a T - N(N + 1)

NIN + ilIZN = 1)
24
where T a smaller sum of like-signed ranks
N = number of pairs less those which Show d = 0
When compared with the t-test. the Wilcoxon has the efficiency
of 95.5%. In Table 17 the results are reported for both the
GES and PPS of the MNW condition. AS already noted. the changes

Table 17. Wilcoxon Z-Scores for Gestural Evaluation Score and
Physical Performance Score. Minimum Written

 

Reinforcement
a 122.
T (9+)* 146 (11+)): 143.5
N 28 29
Z 1.29 1.60
p .0985 n.s. .0548 n.s.

*Smaller sum of like-sized ranks

in the MNW were in the nonpredicted direction for both mea-
surements. While there are no significant differences within
the MNW between T1 and T2. the probabilities reported in
Table 17 approach the alpha probability of .05.

The negative results found in the MNW suggest that

60

the MNW condition seems unable to encourage positive change
in the gestural behavioral of the SS. In this condition. the
SS received a single gestural comment supplemented by the in-
structor's critique of other physical and contextual matters.
This experimental condition seemed most like the speech cri-
tique of the classroom where only a few comments may be made
to the student in any one critique. Since the majority of
students in each class changed in a negative direction. it
may be concluded that limited reinforcement may be a poor
method for generating Speaking improvement. at least. gestu-
ral improvement. With the PPS (a measure of general physical
behavior) changing negatively. the MNW may not only have
failed to encourage gestural improvement. but may have ac-
tually generated a general deterioration of behavior.

Earlier. the MXW condition Showed an increase in the
mean and median of the GES over the two evaluation periods.

Table 18. Wilcoxon Z-Scores for Gestural Evaluation Score and
Physical Performance Score. Maximum Written

 

Reinforcement
982 22.3.
T (13-)* 198 (14+)* 237.5
N 31 31
z .9798 .2058
p .1635 n.s. .3591 n.s.

*Smaller sum of like-signed ranks

 

Table 18 confirms that trend for the GES and the rather ran-

dom movement of the PPS evaluation. Apparently. the Ss in

61

the MXW were provided with more information than the SS in
the MNW about their gestural performance by the BE and were
able to convert that information into improved gesturing.

It is interesting to note that the variability of the MXW
group decreased between T1 and T2. while the mean and median
were increasing. The standard deviation for the GES at T1
was 43.42 and changed to 38.84 at T2. while the variability
of the PPS decreased from 20.60 to 17.66. Apparently. the
MXW condition was able to improve the performances of some
of the Ss. particularly those with the lowest T1 evaluations.
Perhaps this condition. with evaluations on several elements
of gesturing. may have been able to supply useful information
to the Ss which aided him in improving his performance beha-
vior.

The LOR. which earlier demonstrated the greatest
amount of positive change. also failed to reach statistical
significance. However. the approach to the .05 level. repor-
ted in Table 19. on the Wilcoxon suggests that the magnitude
of the changes in the LOR were greater than in the other two
conditions.

Table 19. Wilcoxon Z-Scores for Gestural Evaluation Score and
Physical Performance Score. Light On-Set

 

Reinforcement
282 22.3.
T (12-)* 142 (13-)* 138.5
N 28 27
Z 1.389 1.21
p .0823 n.s. .1131 n.s.

*Smaller sum of like-signed ranks

 

62

While the Sign Test (testing only direction) Showed
a probability of .2877 (see Table 16. page 58). the Wilcoxon
(testing direction and_magnitude of change) produced a proba-
bility of .0823. It may be reasonable to conclude that the
LOR had selective impact upon the SS. That is. those scores
which changed in a positive direction made greater increase
than those scores which changed in the unpredicted direction.
This conclusion may also be inferred from the increased
variability that appears when comparing the GES and PPS at
T1 and T2. The standard deviation of the GES before the ex-
perimental treatment was 22.82. while afterward the deviation
rose to 32.14. On the PPS the standard deviation increased
from 14.29 to 17.65. Apparently. then. the light was able
to have positive impact upon some of the SS without having

serious negative effect upon other SS.

Summary

This chapter has reported the results of the various
statistical tests applied to the collected data. The purpose.
of course. was to determine whether the statistical hypo-
thesis could be rejected and the theoretic hypothesis con-
firmed. Prior to the actual testing of the hypotheses.
several preliminary assumptions were examined. All prelimi—
nary investigations affirmed the pre-research assumptions
except for the heterogeneity of variance found at T1 in both
the GES and PPS. Therefore. nonparametric statistics were
computed for change scores between conditions and for raw
scores within conditions. In either case. no statistical

significance was found. and the theoretic hypothesis could

63
not be affirmed. However. it seemed apparent that changes
were taking place within the various conditions. Examination
of the measures of central tendency and variability seemed to
confirm movement within the groups. as did directional and
cnange analysis. That is. the positive reinforcements sup-
plied in the MXW and the LOR had an impact upon some of the
Ss. Further research seems necessary to draw definitive con-
clusions concerning the use of positive reinforcement in the
classroom. It would seem that light could be used as supple-
mental positive reinforcer for gesturing and other delivery

skills by the classroom instructor.

CHAPTER IV

CONCLUSIONS AND IMPLICATIONS FOR FURTHER RESEARCH

This study has compared three modes of positive rein-
forcement within a beginning speech class. Research in the
field has concluded that reinforcement. positive and conti-
guous to the behavior desired. enhances learning of that be-
havior. It was hypothesized in this study that the MNW. MXW
and LOR would Show differential effects upon the SS when im-
provement scores were compared. However. since only the LOR
met the above criteria. it was suspected that change in the
gesturing of SS would be greater in the LOR than in either
written condition.

When the nonparametric analysis of variance was ap-
plied to the improvement score data. no statistical signifi-
cance was found between the Ss of the three positive modes.
Finding no significance. the experimenter faces alternative
reasons for the results: either the theoretic basis for the
predictions was faulty. or the research itself contained
errors. The latter seems the more reasonable conclusion in
this research for two reasons: (1) the impact of positive
reinforcement has been too well documented in education psy-
chology and irregularly supported in communications research.
and (2) the trends in the data seem to support the original

prediction. Therefore. possible research errors need to be
64

65
identified and recommendations for eliminating those errors
suggested before definitive conclusions may be drawn about

the original predictions.

Potential Sources of Research Error

 

One obvious source of error may be found in the
rating panels. While the median reliabilities for each scale
appear sufficient to warrant adding the panels tOgether.
variabilities are readily apparent (see Appendix 5. page‘92).
An examination of the variability between the panels suggests
that the majority of raters performed their rating tasks with
some agreement. However. there were three raters who. when
they appear in a panel. depress the total reliability of that
panel. Since the 88 scores were based upon adding across
these ratings. it is possible that an individual Ss score may
be inflated or deflated by the appearance of these raters
upon a panel.

While unreliability of a single rater may have affected
the ratings of the SS. the panels themselves may have had an
effect. That is. the appearance of a rating panel in the
classroom may have had a depressing effect upon the perfor-
mance of the 83. If so. this depressing effect might have
greater influence upon SS performance at rating T2 than at
T1. At T1. the SS were in only their second class meeting
and were giving their first speech. Fear and uncertainty
may have been working to reduce the quality of the physical
performance. The appearance of a panel of strange raters

may only heighten that 83 fear and uncertainty. However.

66

the second time the panels appear in the classroom. the SS
were in the seventh drill section and were giving their sixth
speech. If the fear and uncertainty had been diminishing
(because of increased familiarity with peers. instructor and
BE» the appearance of another panel of strange raters may
have encouraged a temporary increase in fear with a conse-
quent reduction in amount and quality of physical behavior.
The result. of course. would be a lower rating at T2 than
the SS deveIOpment at that point merited. Perhaps even a
lower rating at T2 than at T1 would result. particularly if
gestural development had been slow during the treatment
speeches.

Another source of error may have been the BE. Each
was a qualified. experienced teacher working toward a docto-
ral degree in the area of Rhetoric and Public Address. How-
ever. it is possible that the HE may have inadvertently en-
couraged differences to arise between $3 and between treat-
ment groups. The HR could encourage Ss variability by pro-
viding positive or negative nonverbal cues to the 38 receiving
the reinforcement. Such cues may add approval to an already
approving evaluation. thereby increasing the perceived reward.
Of course. a negative display could reduce the impact of a
positive evaluation. Whichever direction. the impact of the
ER would be selective. Some of the 38 within each condition
may be changing because of their interaction with the BE.

The EE may have encouraged differences between groups
if the interaction was systematic. That is. if the HE ap-

peared uninterested in his assignment or the SS. an entire

6?

class might react negatively to him. Or. if the evaluator
seemed particularly interested in his tasks and the SS. the
class might positively respond to him with a resultant in-
crease in their gestural development. It is possible. al-
though only slightly. that the negative reaction of the MNW
may have been due to systematic disapproval Shown the classes
by the BE. Post-research conversations do not suggest such
a conclusion. But these conversations may not have been ac-
curate checks. What may have seemed unimportant (and thus.
unreported to the experimenter) to the instructor and the EB
after the research may have been given great significance by
the class members. The result of such a circumstance might
be a dislike for the EB and an ignoring of his efforts to im-
prove their gesturing.

A fourth area for potential error in this research.
with a resulting failure to Show significance. may have
arisen from an interaction between gesture and the reinforce-
ment mode. One continuing complaint of the instructors. EE
and raters was that several of the SS simply did not move
during their performance. The reinforcement modes used in
this research required that the SS present physical movement
in order to receive the reinforcement. For example. in the
MNW and MXW conditions. if the SS did not move at all. he
received "No basis for comment" as the EE evaluation. In
the LOR. no gesturing would result in no light flashes. In
essence. if the SS failed to move. he selectively placed him-
self in a "no reinforcement" condition. Such a situation

would not provide an appropriate test of the effectiveness of

68
positive reinforcement.

Finally. another potential for error is a consequence
of using the classroom --- peer feedback. The approval or
disapproval shown a speaker by his peers may have greater im-
pact upon the physical behavior than the positive reinforce-
ment administered by the BE. Should the peers act bored or
uninterested in the speech and speaker. the result may be a
deterioration in the behavior of the speaker. Such disappro-
ving feedback might (1) reduce the amount Of positive rein—
forcement received during the treatments (because disappro-
ving peer feedback acted as a depressant upon delivery skills).
and (2) reduce the raters' evaluations (because the SS cur-
tailed their movements in response to the disapproval of the
peers). Such a depressing effect upon physical delivery in
general. and upon gesturing in particular. may have occurred
more during T2 evaluations than in previous Speeches. The
Sixth speech was considered the most difficult by the instruc-
tors because it required the SS to read and formulate Opinions
upon current affairs. While the intent of the assignment was
to ensure audience interest in the topic (tOpics had to be
chosen from the university newspaper). it is possible that
the Opposite effect resulted. If so. audience disinterest
may have depressed gestural behavior. lowering Judges' evalua-
tions at T2 and resulting in the inability of the improvement
score data to reach statistical significance.

or course. the preceding remarks are only specula-
tions. There are no data to affirm or deny the assertions.

However. all or any one of the factors (if actually Operating)

69
could discourage the gesturing of the SS. or encourage the
inaccuracy of the raters' observations. In either case.
greater control of these extraneous. potentially Operating
variables needs to be exerted in order to diminish the effect

upon the Judging panel evaluations.

Further Research

Given that the theoretic base of this study seems
sound. but the research control of error-producing variables
weak. then a next research step would be a replication with
greater control. Such a study might well remove the research
from the classroom and place it within an experimental frame-
work. Ss would be selected from the general population of
students and not taken as part of intact groups. Each SS
would be randomly assigned to one of four reinforcement con-
ditions: the three described in this study and a group in
which no reinforcement was given for gesturing. Comments
would be made to the students concerning other aspects of
delivery and content for every speech. The pre- and post-
test design used in this study would divide the treatment
speeches from the rated speeches.

Each speech by each Speaker would be videotaped.
then coded and stored. The people witnessing the speeches
would be the content evaluator. the EE administering the rein-
forcement and the production crew necessary to record the
Speech. After the completion of a series of six speeches
similar in assignment to those of this study. panels of

raters would view the tapes and evaluate the performances.

70
The raters would be selected carefully and subjected to
several training sessions. Furthermore. the panels would
see the tapes of Speakers from the four conditions randomly
mixed. In other words. a panel of Judges might rate the
taped performances of 10 Speakers in a single session: two
from the MNW. two from the MXW and three each from the LOR
and the no reinforcement condition. Also. some panels would
see Speakers in the pre-test. post-test order. while other
panels would see speakers in a reversed order. NO two panels
would evaluate the same collection of speakers. nor would the
panels know which speaking time (pre- or post-) they were
evaluating.

As in the present research. difference score data
using an analysis of variance statistical design seems appro-
priate. If overall differences should appear between the
four groups. then individual cell differences could be mea-
sured by t-tests. Intracellular improvement might be mea-
sured by correlated t-tests. Such a research project Should
help to encourage differences to appear between the groups
if. in fact. the varying modes of positive reinforcement do
produce differential effects upon the gesturing that is done
by student Speakers.

Aside from the prOposed study described above. addi-
tional research might be undertaken to examine the theoretic
hypothesis: (1) Other mechanical reinforcers might be used
on college students. For instance. a money-drOp might be
installed in a classroom lectern to drop coins for each

gesture found acceptable to the experimental evaluator. It

71
is possible that some collegiate speakers in the current re-
search may have found the flash of light unrewarding.
(2) Some other speaking variables also might be studied under
varying reinforcement conditions. For example. verbal varia-
bles. like total output and number of nonfluencies. which
have been studied already in other research. could be examined.
Pitch. rate and volume could be variables possible for study
under conditions Similar to those described in this research.
(3) To positively reinforce the student by removing some
noxious stimulus might be a possible method. For example. a
disruptive noise might be presented to the SS until he ges-
tured correctly. Such a disruptive noise could be a continuous
bell or buzzer. The use of noxious stimuli might be particu-
larly effective for subjects who are reluctant to move the
hands and arms while speaking. The presentation of a disrup-
tive stimulus might provide sufficient motivation for the
student to move to reduce the irritation. Of course. this

kind of reinforcement might not be acceptable in the classroom.

Conclusions

While no statistically signigicant results were pro-
duced by the reinforcement conditions of this study. several
tentative conclusions may be drawn which supply a basis for
further research: (1) positive reinforcement is recommended
for speech classroom evaluation. but does not appear to be
widely used; (2) immediate reinforcement is recommended to
maximize skill development. but does not appear to be widely

used in the Speech classroom; (3) the gestures of those

72

students eXposed to a mode of reinforcement (light on-set)
meeting these recommendations tended to show greater im-
provement than the gestures of students exposed to other
positive reinforcement modes; (4) a single evaluation of
gesturing appears to be an insufficient mode for admini-
stering positive reinforcement. and (5) additional research
within an experimental setting is needed to examine the im-
pact of positive reinforcement as a teaching tool or a sup-

plemental critique technique for the Speech instructor.

Summar

The four potential studies previously mentioned and
the tentative conclusions drawn indicate additional study is
needed of the theoretic hypothesis of this exploratory re-
search. With the continuous growth of higher education and
the increasing enrollment size of many basic courses. there
appears to be real need to implement new ways for encouraging
the acquisition and refinement of students' knowledge and
skills. Perhaps the variables studied herein may eventually
assist speech instructors in developing the abilities of the

individual students they teach.

BIBLIOGRAPHY

BIBLIOGRAPHY

Amato. Philip P.. and Terry R. Ostermeier. ”The Effect Of
Audience Feedback on the Beginning Public Speaker."
The Speech Teacher. XVI (January. 1967). 56-60.

Becker. Sam. "Research on Speech PedagOgy" from Dimensions
of Rhetorical Scholarship. ROger E. Nebergall (ed. ).
publication of the Department of Speech. University
of Oklahoma. 1963. 31-h8.

Berelson. Bernard. and Gary Steiner. Human Behavior: An In-
ventory of Scientific Finds. New York: Harcourt.
Brace. and World. Inc.. 1964.

Berlo. David C. The Process of Communication. New York: Holt.
Rinehart and Winston. 1960.

Bigge. Morris L. Learning Theories for Teachers. New York:
Harper and Row. Publishers. 19

Blubaugh. Jon A. "The Effects of Positive and Negative
Audience Feedback on Selected variables of Speech
Behavior of Normal-Speaking College Students." Un-
published Ph. D. dissertation. University of Kansas.
1966. abstracted by C. E. Larson for Speech Mono-

ra hs. XXXIV (August. 1967). p. 2&3.

Bostrom. Robert N. "Classroom Criticism and Speech Attitudes."
Central States Speech Journal. XIV (February. 1963).
27-32 a

Byers. Burton H. "Speech and the Principles Of Learning."
The Speech Teacher. XII (March. 1963). 136-ih0.

Cahn. Dudley. "The Effect of Simulated Audience Response on
Speaker Attitudes." Unpublished Master's thesis.
Northwestern University. 1966.

Cerekwicki. Louise E.. and David Grant. "Delay of Positive
Reinforcement in Instrument Eyelid Conditioning."

Journal of Experimental Psychology. LXXV (1967).
3 0-3 a

73

7h

Combs. Walter. and Gerald Miller. "The Effect of Audience
Feedback on the Beginning Public Speaker. I. A Coun-
ter-view." The Sppech Teacher. XVII (September. 1968).
229-231 0

Davis. James W. "Variations in Verbal Behavior in Dyads as a
Function of Varied Reinforcing Conditions." Speech
Monographs. XXXIV (November. 1967). ##3—447.

Dedmon. Donald N. "Criticizing Student Speeches: Philosophy
and Principles." Central States Speech Journal. XVIII
(November. 1967). 276-284.

Densmore. G. E. "The Teaching of Speech Delivery.“ gparterly
Journal of Speech. XXXII (February. 19h6). 7-71.

Deutschmann. P. J. "The Sign-Situation Classification of Ru-
man Communication." The Jouppal of Communication.
(Mimeographed).

Eisenson. Jon. J. Jeffery Auer. and John V. Irwin. The Psy-

cholo of s eech. New York: Appleton-Century-Crofts.
1963.

Ferguson. George A. Statistical Analysis in Psychology and
Education. New York: McGraw-Hill Book Company. Inc..
1959.

Greenspoon. Joel. "The Effect of Two Nonverbal Stimuli on
the Frequency Of Members of Two Verbal Classes."
American PsychOIOgist. IX (August. 1954). p. 38“.

Guilford. J. P. Psychometric Methods. New York: McGraw-Hill
Book Company. Inc.. 1955.

Harms. L. S. "Programmed Learning for the Field of Speech."
The Speech Teacher. X (September. 1961). 215-219.

Hayworth. Donald (ed.). A Research into the Teachin of Pub-
lic Speaking. Official Project #7255. Works Progress
Administration. published by the Committee on Coopera-
tive Research of the National Association of Teachers

of Speech. Michigan Department of Public Instruction
and Michigan Society Of Teachers of Speech. 1939.

Hilgard. Ernest R.. and Gordon H. Bower. Theories of Learnin .
New York: Appleton-Century-Crofts. 3rd. ed.. 1966.

Hinde. Dorothy Mae. "An Investigation into the Effectiveness
Of Criticism of Speech Delivery When Given Before.

During. and After Speech Performance." Unpublished
Master's thesis. State University of Iowa. 1949.

75

Huenergardt. Douglas W. "An Experimental Study Of the Effects
of Increasing Percentages of Simultaneous Noncontin-
gent Audience Approval on Speaker Attitudes." Unpub-
lizhed Ph.D. dissertation. Northwestern University.
19 7.

Johnson. F. Craig. and George Klare. "Feedback: Principles
and Analogies." Journal of Communication. XII (1962).
150-159 a

Karns. Charles F. "The Verbal Behavior of a Speaker as a
Function of Certain Non-Verbal Aversive Stimuli Pre-
sented by an Audience in a Public Speaking Situation."
Unpublished Ph.D. dissertation. University of Pitts-
burgh. 1964.

Kerlinger. Fred N. Foundapgons of Behavioral Research. New
York: Holt. Rinehart and Winston. Inc.. 1965.

Kimble. Gregory A. H. (rev. ed.). Conditioning and Learning.
New York: Appleton-Century-Crofts. 2nd. ed.. 1961.

Krasner. Leonard. "Studies of the Conditioning of Verbal Be-
havior." Psychological Bulletin. LV (1958). 1&8-170.

Lindquist. E. F. Design and Analysisgof Experiments in Pay:
chpiogy and Education. Boston: Houghton Mifflin Com-
pany. 1956.

Lockard. Robert B. "Effects Of Light Upon the Behavior of

Rodents." PsychOlOgical Bulletin. LX (November. 1963).
509-529.

Lothers. William Theron. "A Survey Study of Methods and Ra-
tionales in the Criticism of Classroom Speeches in
the Beginning College Course." unpublished Ph.D.
dissertation. Michigan State University. 1966.

McNemar. Quinn. Psychological Statistics. New York: John
Wiley and Sons. Inc.. 3rd. ed.. 1962.

Miller. Gerald R. "Variations in the Verbal Behavior of a
Second Speaker as a Function of Varying Audience
Responses." Speech Monographs. XXXI (June. 1964).
109-115.

. John W. Vlandis. Harry Zavos. and Milton E. Rosen-
baum. "The Effect of Differential Reward on Speech
Patgerns." Speech Monographs. XXVIII (March. 1961).
9"]. O

Ostermeier. Terry H.. and Philip P. Amato. "The Effect of
Audience Feedback on the Beginning Public Speaker.
II. A Rejoinder." The Speech Teacher. XVII (Septem-
ber. 1968). 231-235. ” '

76

Postman. Leo. "The History and Present Status of the Law of
Effecg." PsychoTpgicaT BulTetin. XLIV (November. 19h7).
089-5 3.

Resnick. Lauren B. "Programmed Instruction and the Teaching
of Complex Intellectual Skills: Problems and Pros-
pects." Harvard Educational Review. XXXIII (Fall.
1963). 439-571.

Robinson. Karl F.. and E. J. Kerikas. Teaching Speech: Methods
and Materials. New York: David McKay CO.. Inc.. 1963.

Sereno. Kenneth K. "Changes in Verbal Behavior of Females Du-
ring Two Successive Speech Performance as a Function
of the Sequence of Listener Responses." Unpublished
Ph.D. dissertation. University of Washington. 196#.

Shelton. Ralph L.. William B. Arndt. and June B. Miller.
"Learning Principles and Teaching of Speech Language."

JournaT_Of Speech and Hearing Disorders. XXVI (Novem-
ber. 1961;: 3 '37 0

Sidman. Murray. "Operant Techniques." Chapter 6 of Sxperimen-
tal FOundatTons of Clinical Psychology. Arthur J.
Bachrach.6(ed.). New York: Basic Books. Inc.. 1962.
171-210.

 

Siegel. Sidney. Nonparametric Statistics for the Behavioral
Sciences. New York: McGraw-Hill Book Company. Inc..

1956.

Skinner. B. F. Science and Human Behavior. New York: The
Free Press. 1953.

Smith. Donald K. "What Are the Contemporary Trends in Tea-
ching Speech?" The Speech Teacher. X (March. 1961).
87-9 0

Smith. Raymond G. "The Criticism of Speeches: A Dialectical

Appgoach." The Speech Teacher. X (January. 1961).
59- 2.

Vlandis. John W. "Variation in the Verbal Behavior of a
Second Speaker as a Function of Varying Audience Res-
ponses." Speech MonOgrSphS. XXXI (June. i96h). 116-119.

Wallace. Murray R. "Techniques and Devices in Teaching
Speech." The Bulletin of the National Association of

Secondar -School Princi als. XXXVIII (January. 1955).
210-215.

Weaver. Andrew T.. Gladys L. Borchers. and Donald K. Smith.

The Teaching of Speech. New York: Prentice-Hall.
Inc.. 1952.

77

Wiener. Norbert. The Human Use of Human Beings. New York:
Doubleday Anchor Books. i95h.

Winer. B. J. Statistical Principles of_§xperimental Design.
New York: McGraw-Hill Book Company. 1962.

APPENDICES

APPENDIX A

Questionnaires to Faculty and Teaching Assistants

78
Dear Colleague:

Although you are often asked to assist in research. I
hope you will give me a few moments Of your time.

We are interested in the gestural behavior of beginning
public address students. In your evaluations of beginning
speakers. what quaTities of gestural behavior do_you_Took fog:
To ask the same question in a different manner. what qualities
of gesturing do you wish to develop in your beginning speech
students?

Please indicate as many of these qualities as you think
important in the consideration of the beginning Speech student.
And. why not do it soon? Remove this little task to be free
from tensions during the holidays.

Thank you for your cOOperation. Have a delightful
Christmas Holiday. You may place these completed forms in
my mailbox in the Speech Office.

Dwn P. Millar

79
Dear Colleague:

Once again I ask you to assist me in my research. I
know these requests take time. but I hope you will spend a
few moments and answer the following questions.

The following is a set of words and phrases used to
describe good qualities of Speaker gesturing (the hand
and arm movements of the speaker). Please look care-
fully at the list and determine if each represents a
separate and distinct category of gestural evaluation.

1. DO you perceive the qualities listed below as
separate and distinct categories? (please circle)

Yes NO

2. If you answered "NO." please look at the words and
phrases again. If you see similarities or over-
lappings. group the overlapping (similar) words and
phrases together in the Space below.

LIST OF QUALITIES
visibility apprOpriate to content

natural coordinated with other
bodily action

fits the speaker free

clarity definite

forceful complete

emphatic directed toward the
audience

spontaneous relaxed

Are there any qualities of gesturing you find important
and missing from this list? If so. please indicate:

Thank you very much for your continued cooperation. Please
place the completed form in my mailbox.

Dan P. Miller

APPENDIX B

Gestural Evaluation Instrument and Rater Instructions

80

INSTRUCTIONS TO RATERS

Dear :

Attached is a sample of the instrument to record your
evaluations of the physical behavior Of student speakers.
The upper portion of the sheet refers only to the gesturing
of the student (hand and arm movements). The lower portion.
beginning with "General Impressions. . . ." asks you to
evaluate the total physical performance (walking. posture.
head and trunk movements. etc.) of the student speaker.

Remember. these are students in the beginning speech
course and should be evaluated accordingly. FUrther. you
should not relay your evaluations to other judges. the in-
structor or the students. This sheet should be an indepen-
dent measure Of your evaluation.

The class instructor will call and spell the name of
the speaking student. Please write the last name in the
upper right-hand corner. Wait for the speech to be concluded
before you mark your evaluation. Keep your evaluations to-
gether and return them to me in 110 South Hall when the
class period has ended. If the speeches are completed before
the period ends. please wait until the instructor asks you to
leave before you go.

Please meet in 110 South Hall approximately 10 minutes
before the start of the class period. Any questions you may
have can be answered at that time. or earlier. if necessary.
YOu may carry these instructions with you into the class.

Thank you very much for assisting me in this research.
I am sure you realize how important this work is to me.

(TTme and PTgce)

(Instructor's Name);

81

GESTURE EVALUATION
Speech 102 Spring Quarter

The following are elements of gesturing. You are being asked
to carefully evaluate the gestural behavior of the speakers
you will see and hear. Be sure you evaluate the speaker on
every element and then mark your evaluation in the appropriate
blank for each and every element.

 

 

VISIBILITY
visible invisible
: z : : x : : :
CLARITY
distinct indistinct
: : : : : : : :
DEFINITE
indefinite definite
: : : : : : : :
COMPLETE
complete incomplete

natural unnatural

restricted free

APPROPRIATE TO THE CONTENT
appropriate inapprOpriate

COORDINATED WITH OTHER BODILY ACTION
coordinated uncoordinated

undirected directed

bold timid
: : : : : : : :

aggressive meek
: : : : : : : :

tired energetic
: : : : : : : :

confident unsure

colorful. dull

APPENDIX C

Instruments of the Minimum Written Reinforcement Condition

82

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

 

VISIBILITY
Excellent-- Good--could Okay--could NO basis for
your gestures see most of see a few of comment
were easily your gestures your gestures
seen
Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

 

CLARITY
Excellent-- Good--most Okay--a few No basis for
your gestures gestures were gestures were comment
were distinct distinct distinct

O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 0

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

DEFINITE
Excellent-- Good--most Okay--a few NO basis for
your gestures gestures were gestures were comment
were forceful forceful and forceful and
and emphatic emphatic emphatic

O O O O O C O O O O O 0 O O O O O O O O O O O O O O O O O 0

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

COMPLETE
Excellent-- Good--most Okay-~a few NO basis for
each gesture gestures were gestures were comment

was finished finished finished

83

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

FITS THE SPEAKER

Excellent-- Good-~most Okay-~a few NO basis for
each gesture gestures were gestures were comment

was natural natural natural

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your Speech have been
evaluated and this element is considered the best element of
your gestural performance.

 

SPONTANEOUS
chellent-- Good--most Okay-~a few NO basis for
each gesture gestures were gestures were comment
was relaxed relaxed and relaxed and
and free free free

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your Speech have been
evaluated and this element is considered the best element of
your gestural performance.

APPROPRIATE TO THE CONTENT

Excellent-- Good-~most Okay--a few No basis for
each gesture gestures gestures comment
related to related to related to

the thought the thought the thought

8h

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

COORDINATED WITH OTHER BODILY ACTION

Excellent-- Good--most Okay-~a few NO basis for
each gesture gestures were gestures were comment

was coordinated coordinated coordinated

with other with other with other

movements movements movements

Speech 102

The following is an element of good gesturing. The hand
and arm movements that you made during your speech have been
evaluated and this element is considered the best element of
your gestural performance.

DIRECTED TOWARD THE AUDIENCE

ExceTlent--
each gesture
was directed
to us

No basis for
comment

Good--most
gestures were
directed to us

Okay--a few
gestures were
directed to us

APPENDIX D

Instrument of the Maximum Written Reinforcement Condition

85

GESTURE EVALUATION
Speech 102 Spring Quarter
The following are elements of good gesturing. The hand and
arm movements that you made during your speech have been

evaluated on the basis of

these elements and the evaluation

indicated.

1. VISIBILITY

Excellent-- Good--cou1d Okay-~could No basis for
your gestures see most of see a few of comment

were easily your gestures your gestures

seen

2. CLARITY

Excellent-- Good--most Okay-oa few No basis for
your gestures gestures were gestures were comment

were distinct distinct distinct

3. DEFINITE

Excellent-- Good--most Okay-~a few No basis for
your gestures gestures were gestures were comment

were forceful forceful and forceful and

and emphatic emphatic emphatic

A. COMPLETE

Excellent-- GOOd--most Okay--a few No basis for
each gesture gestures were gestures were comment

was finished finished finished

5. FITS THE SPEAKER

Excellent-- GOOd--most Okay-~a few No basis for
each gesture gestures were gestures were comment

was natural natural natural

6. SPONTANEOUS

Excellent-- Good-~most Okay-~a few No basis for
each gesture gestures were gestures were comment

was relaxed relaxed and relaxed and

and free free free

7. APPROPRIATE TO THE CONTENT

Excellent-- Good--most Okay-~a few No basis for
each gesture gestures gestures comment
related to related to related to

the thought the thought the thought

8. COORDINATED WITH OTHER BODILY ACTION

Egcellent-- Good--most Okay-~a few NO basis for
each gesture gestures were gestures were comment

was coordinated coordinated coordinated

with other with other with other

movements movements movements

9. DIRECTED TOWARD THE AUDIENCE

E3cellent-- Good--most Okay--a few No basis for
each gesture gestures were gestures were comment

was directed
to us

directed to us

directed to us

APPENDIX E

Speaking Assignments

86

ASSIGNMENTS
Speech 102 Spring Quarter

During the first few class periods of this quarter. you will
give six Speeches. This will be different from the other
sections of Speech 102. However. the differences will not
affect you (or your grade) and are simply part of the conti-
nuing changes made necessary by growing enrollments and tele-
vision teaching.

The following are the assignments for your drill section
periods only:

 

 

grill Section Period Assignment
first Introduction to the course.
second A speech of self-introduction.

Focus upon things you like to
do. future plans. things you
have done. etc. Do NOT talk
just about your high school.
Example: your major and your
plans for jobs; a hobby and why
you like that specific activity.

third A speech of information based
upon your reading of current
news magazines. Read a major
article from one of the following
news magazines: NationaT_Obser-
ver. Newsweeky_Saturd§y Review.
U.S. News & WOTTd Report and
Time. Look only at issues from
within the past 3 weeks. The
Object is to take this article
and report upon the topic in
class.

fourth A Speech of problem-solution on
some local (campus or city) pro-
blem. Take a local problem.
think about it and attempt to
formulate a constructive prOgram
for solving it. This is not a
research assignment but a thin-
king one. Example: campus par-
king. lack of married housing.
student government.

fifth A Speech Of information based
upon your vocational field.
Look at your vocational field
and discover the influences and
trends and directions. Read
from scholarly journals. talk
to your professors to discover

87

what trends exist in your field
and how these trends will influ-
ence your vocation in the future.
Example: the current learning
research and what effect this
knowledge will have upon the
teaching of retarded (or accele-
rated) students in the next 10-
20 years; or. the impact of com-
puter technology upon business.

sixth A speech of entertainment based
upon your summer eXperiences.
Using your job. travels. dates.
play-time as a base for this
speech. build an oral presenta-
tion that the class will find
enjoyable and amusing.

seventh A speech Of information or per-
suasion based upon current events
as reported in the BG News.
Read an article from the BG News.
eXpand upon the tOpic. develOp
it for the class and present it.
You may use national. state.
city events or problems. campus
events and problems including
play reviews. sporting news.
movie reports. etc. But this
is a research assignment so you
will want to go beyond the SS
News to find additional views
and facts. Notice that you may
simply report to the class on
your topic or you may try to
convince them to take a new stand
(adopt a new Opinion) on the
tOpic. Example: the success of
the baseball team. Talk with
players. coaches. etc.. and then
deliver a persuasive speech advo-
cating the campus support the
team by attending the games.

VERY IMPORTANT--READ. The following regulations are placed
upon the speeches you will give during the first seven drill
section periods. And those sections only. Other assignments
will be made for later drill section periods. Please Observe
these regulations.

1. Minimum time Of 2% minutes and a maximum of 4 minutes.
If you speak less. evaluation of your performance is
difficult and inaccurate. If you speak longer. some
one else will be limited in his time because of the

88

number of students in the class.

2. Only those speeches meeting the above assignments
will be acceptable.

3. All speeches must be original. A great part of pub-
lic speaking is learning to create topics and orga-
nize ideas.

A. You may use notes while you speak. Practice using
them so you can relax and speak directly to the
audience.

5. NO written assignment is required for the speeches
given during the first seven drill section periods.

Each speech you give will be evaluated and critical comments
made concerning your performance.

APPENDIX F

Instructions to Instructors

89
Carol. Donna. Greg:

(MNW)
Please begin your class on Tues. this way: "The gentleman
visiting our class is a member of the teaching faculty of
the Department of Speech. He is here to use a new form for
gestural evaluation. I will evaluate your Speech in its en-
tirety: he will evaluate only gestures. Please go to him at
the end of your Speech to receive his written evaluation."

(MXW)
Please begin your Wed. class this way: "The gentleman visi-
ting our class is a member of the teaching faculty of the
Department of Speech. He is here to use a new form for ges-
tural evaluation. I will evaluate your speech in its en-
tirety: he will evaluate only gestures. Please go to him at
the end of your speech to receive his written evaluation."

(LOR)
Please begin your Wed. noon class this way: "The gentleman
visiting our class is a member of the teaching faculty Of
the Department of Speech. He is here to use a new method
for gestural evaluation. The box in front contains a light.
Whenever you see that light flash. it means that the gesture
you have just made is considered good --- the gesture that
you just made was considered good. I will evaluate your
speech in its entirety."

Thanks. Questions?

9o
MEMORANDUM May 6. 1969

T0: Carol Shaw. Donna Emlich. Greg Gardner
FROM: Millar

Thank you for permitting me to use your classes for my dis-
sertation experiment. I know this was an inconvenience and
may yet cause you considerable discomfort with your classes.
I hope any damage done will not be irreparable. however. and
that your classes may assume some semblance of normalcy be-
fore the end Of the quarter.

One last favor. Please see me at your convenience this week.
I would like to talk with each of you for a Short time about
(1) what you observed in the classroom. (2) any student feed-
back. and (3) your own feelings concerning what you saw and
the intent of the experiment.

The following you may read to your classes if you wish.
They may be somewhat relieved to hear what was going on:

"To the 102 classes involved in the experiment:

Thank you for giving your time and energies to help 102 con-
tinue to grow and develop. You were subjected to written
and mechanical devices designed to help you better understand
your arm and hand movements while speaking. The objectives
of the experimentation were the following:

(1) to see if a many-speaking experience. without prior
training. was a superior approach to 102 than other
approaches.

(2) to see if teachers of 102 could reasonably agree on
the characteristics of good gesturing.

(3) to see if 102 could use two evaluators in order to

_ improve the performance of speakers: one to look at
some specific area. in this case gesture. but freed
from the subjective stigma of giving a grade. and
the other to look and grade.

(A) to see if written forms dealing with some Specific
aspect of speaking are superior to the rather general
comments made by an instructor.

(5) to see if some mechanical means of criticism. in
this case the light. would encourage the generation
of some specific desirable speaking trait. in this
case gesturing.

(6) to see if 'Strangers' to the classroom influence
negatively or positively the behavior of beginning
students. From a speaking concern. it might be de-
sirable to have a single class evaluated by a va-
riety of instructors during the course of a quarter.

91

If you. as students. have comments of either a negative or a
positive nature concerning your experience during the first
few class meetings. please pass them along to your instructor.
Or. if you wish to remain anonymous. write them to Professor
Delmer Hilyard. Co-Director of 102. Department of Speech."

Thanks.

APPENDIX G

Table 20. Rater Reliability
Evaluation Score. Time

and

Table 21. Rater Reliability
Performance Score. Time

for Gestural
1 and 2

for Physical
1 and 2

92

Table 20. Rater Reliability for Gestural Evaluation
Score. Time 1 and 2

 

VISIBLE CLARITY
Time 1 Time 2 Time 1 Time 2
1.0000 .9137 .9259 .9005
.9014 .8660 .8430 .7719
.8778 .7976 .7760 .7466
.8180 .7167 .7714 .7148
.8012 .7005 .7424 .7136
.7241 .6935 .7222 6365
.4329 .4405 .6313 5570
.4006 .3613 .2846 .5123
.2522 .0000 .0004 .0456
DEFINITE COMPLETE
Time 1 Time 2 Time 1 Time 2
.9326 .9482 .9773 .8459
.8542 .9248 .9739 .8200
.7545 .8851 .8551 .8088
.7187 .6790 .8324 .7659
.6676 .6481 .5856 .7533
.6134 .5222 .4837 .6470
.1685 .4459 .4832 .6382
.1117 .4455 .3363 .6382
.0034 .0053 .1630 .6307
NATURAL
Time 1 Time 2 Time 1 Time 2
1.0000 .9050 .9961 1.0000
.8670 .8572 .8930 .9629
.7971 .6722 .8525 .8923
.7878 .6434 .6745 .8847
.7595 5566 .6598 .8265
.7592 4923 .6235 .7709
.6204 3784 .5837 .7359
.5898 .2755 .0005 .4347
.3823 .2561 .0000 .3874

 

93
Table 20. (cont'd.)

 

APPROPRIATE TO THE CONTENT COORDINATED VITH BODILY ACTION

Time 1 Time 2 Time 1 Time 2
.9966 .8152 1.0000 .9329
.9096 .8032 .9813 .8751
.8100 .8006 .7539 .8562
.6370 .6364 .7473 .7544
.6058 .5760 .6390 .6715
.5329 .5499 .6333 .6689
.5152 .4331 .6148 .6378
.4355 .3070 .4273 .5881
.4271 .2311 .3867 .2555

DIRECTED TO THE AUDIENCE TOTAL GES

Time 1 Time 2 Time 1 Time 2
.9362 .8338 .9987 .9381
.9033 .7682 .9880 .9127
.8701 .7401 .9650 .8032
.6761 .7224 .8790 .7321
.6633 .7142 .8355 .6777
.5788 .6059 .7075 .6377
.5525 .5411 .6405 .5426
.5465 .5329 .6041 .3316
.4467 .4300 .5544 .2041

 

94

Table 21. Rater Reliability for Physical Performance

Score.

Time 1 and 2

 

Time 1

.9022
.8674
.8541
.8168
.7702
.6083
.2463
.1117
.0750

Time 1

.8993
.8762
.7403
.6795
.6128
.6078
.4435

.3773
.1122

Time 1

.9533
.8379
.8288
.7987
.7663
.7660
.6900
.5058
.3029

BOLD

Time 2

ENERGETIC

Time 2

8500
8047
7985
7631
7469
7334
7280
6931
6344

8670

.8503
.8437
.8304
.6728
.7269
.5548
.5433
.4587

COLORFUL
Time 2

.8686
.8428
.8269
.8184
.8140
.7288
.7020
.6962

.8993
.8142
.7853
.6923
.6391
.6315
.4002
.2558
.0840

Time 1

.9032
.8400
.7864
.6164
.5504
.4764
.1386
.0910

Time 1

.9388
.8771
.8592
.7495
~5175
.4069
.3889
.3313
.2596

AGGRESSIVE

Time 1

CONFIDENT

TOTAL PPS

Time 2

.8775
.8596
.8132
.7820
.7632
.6616
.6544
.5602
.5459

Time 2

.8801
.8223
.7560
.7499
.7421
.7383
.7266
.7211
.4851

Time 2

.9125
.8526
.8433
.8249
.7790
.6984
.5879
.4202
.3802