University of Iowa

Iowa Research Online

Theses and Dissertations

2011

Application of principles from motor-learning

theory to the studio voice lesson: effects of
feedback frequency on retention of classical singing
technique
Lynn Milo Maxfield
University of Iowa

Copyright 2011 Lynn Milo Maxfield

This dissertation is available at Iowa Research Online: http://ir.uiowa.edu/etd/1021

Recommended Citation
Maxfield, Lynn Milo. "Application of principles from motor-learning theory to the studio voice lesson: effects of feedback frequency
on retention of classical singing technique." PhD (Doctor of Philosophy) thesis, University of Iowa, 2011.
http://ir.uiowa.edu/etd/1021.

Follow this and additional works at: http://ir.uiowa.edu/etd

Part of the Music Commons

 APPLICATION OF PRINCIPLES FROM MOTOR-LEARNING
THEORY TO THE STUDIO VOICE LESSON: EFFECTS OF FEEDBACK
FREQUENCY ON RETENTION OF CLASSICAL SINGING TECHNIQUE

by

Lynn Milo Maxfield

An Abstract

Of a thesis submitted in partial fulfillment of the

requirements for the Doctor of Philosophy degree
in Music in the Graduate College of
The University of Iowa

May 2011

Thesis Supervisor: Associate Professor John Muriello

 1

ABSTRACT

Over the past several decades, cognitive and behavioral scientists have been researching

the most effective practices for training muscles to produce specific movements consistently and

accurately. That research has led to relatively wide acceptance of several best practices for the

training of motor skills. One such practice is the reduction in the frequency with which

augmented (external) feedback is provided by the instructor/trainer during skill-acquisition. This

theory of low-frequency feedback has been examined by research in a wide variety of fields

ranging from exercise and sport to voice therapy and rehabilitation. Prior to the study reported

here, however, this theory had not been applied to the acquisition of vocal skills associated with

classical singing techniques. The current research consisted of an alternating treatment single-

subject study, which was conducted on a college campus over the course of a 15-week semester.

8 college voice students (3 male and 5 female) ranging in age from 18 to 25 participated in voice

lessons provided by the researcher and aimed at improving the overall quality of the voices of the

participants. Over the course of the15 weeks, the instructor alternated between providing a high-

frequency feedback (HFF) instruction condition and a low-frequency feedback (LFF) instruction

condition. At the beginning of each session, a vocal sample was recorded to test the retention of

the skills trained in the previous lesson. Those recordings were evaluated by a panel of five

college voice instructors who provided a numerical score (out of a possible 100 pts.) for each

sample on the basis of tone quality, breath management, and intonation. The results of this study

indicated that three of the eight subjects retained more vocal skill ability during the LFF phases

of the study, while the remaining five subjects retained less vocal skill ability during the LFF

phases of the study. It was also seen that the three subjects who responded favorably to the LFF

instruction condition were also those whose scores were higher throughout the duration of the
 2

study. These findings would appear to indicate that an LFF instruction condition may be more

beneficial to more experienced or more skilled singers, while an HFF instruction condition may

be more beneficial to more novice singers. In the final chapter of this report, several

modifications to this study are suggested along with suggestions for future research regarding the

application of other principles from motor-learning theory to the acquisition of new vocal skills.

Abstract Approved: __________________________________________________________

Thesis Supervisor

__________________________________________________________
Title and Department

__________________________________________________________
Date
 APPLICATION OF PRINCIPLES FROM MOTOR-LEARNING
THEORY TO THE STUDIO VOICE LESSON: EFFECTS OF FEEDBACK
FREQUENCY ON RETENTION OF CLASSICAL SINGING TECHNIQUE

by

Lynn Milo Maxfield

A thesis submitted in partial fulfillment of the

requirements for the Doctor of Philosophy degree
in Music in the Graduate College of
The University of Iowa

May 2011

Thesis Supervisor: Associate Professor John Muriello

 Graduate College
The University of Iowa
Iowa City, Iowa

CERTIFICATE OF APPROVAL
___________________________

PH.D. THESIS
____________

This is to certify that the Ph. D. thesis of

Lynn Milo Maxfield

has been approved by the Examining Committee

for the thesis requirement for the Doctor of
Philosophy degree in Music
at the May 2011 graduation

Thesis Committee: _______________________________

John Muriello, Thesis Supervisor

_______________________________
Eileen Finnegan

_______________________________
Katherine Eberle-Fink

_______________________________
David Puderbaugh

_______________________________
Stephen Swanson
To Ellen

ii
 ACKNOWLEDGMENTS

I would like to thank Dr. John Muriello, Dr. Eileen Finnegan, and Dr. Katherine Eberle-

Fink for their time, patience, and expert advice in the preparation of this thesis. Special thanks

are extended to Dr. Muriello for his keen eye in reading and proofing chapter drafts. I would

also like to thank Dr. Katherine Verdolini and Dr. Ingo Titze for their willingness to allow me to

read and cite material from their as yet unpublished text on vocology, and the music faculty and

administration at Monmouth College for allowing me to use their facilities to conduct this

research. Additionally, I would like to thank my entire family for their seemingly endless

encouragement. Finally, I must thank my wife, Ellen, for her love and supporting

encouragement throughout my academic career and especially through the preparation of this

thesis.

iii
 TABLE OF CONTENTS

LIST OF TABLES.. vi

LIST OF FIGURES...viii

CHAPTER 1 INTRODUCTION TO PRINCIPLES OF MOTOR LEARNING 1

Jack A. Adams Closed-Loop Theory.. 2

Richard A. Schmidts Schema Theory. 9
Motor Learning Defined and the Role of Feedback... 12
Application to Voice Science and Need for the Current Study. 23

CHAPTER 2 LITERATURE REVIEW 27

Feedback Frequency and Simple Arm-Positioning/Timing Motor Tasks..28

Feedback Frequency and the Acquisition of Voice Tasks. 42
Conclusion.. 48

CHAPTER 3 METHODOLOGY.. 50

Hypothesis.. 50
Theoretical Framework...51
Subjects...51
Instrumentation... 53
Skill Acquisition Phase...54
Retention/Transfer Tests 56
Expected Results 58

CHAPTER 4 REPORTING OF RESULTS...59

Subject One 61
Subject Two 68
Subject Three.. 75
Subject Four72
Subject Five 89
Subject Six.. 96
Subject Seven... 103
Subject Eight.110

CHAPTER 5 DISCUSSION OF RESULTS116

Results That Support the Hypothesis... 116

Results That Do Not Support the Hypothesis...118
Results That Are Mixed120

iv
 Interpreting the Mixed Results. 125
Initial Summary of Results... 126
Variation of Skills Being Trained.128
Conclusions...137

CHAPTER 6 IMPLICATIONS FOR TEACHERS AND FUTURE RESEARCH.140

Implications for Teachers of Singing... 140

Implications for Future Research. 143
Conclusion 163

APPENDIX A INFORMED CONSENT DOCUMENT. 165

APPENDIX B SUBJECT INTAKE FORM 170

APPENDIX C SAMPLE VOCAL EXERCISES 171

APPENDIX D PERCEPTUAL AUDITORY ASSESSMENT FORM... 173

BIBLIOGRAPHY 174

v
 LIST OF TABLES

Table

3.1 Demographics of eight study participants..53

4.1 Subject one tone quality results. 62

4.2 Subject one breath management results. 64

4.3 Subject one intonation results 66

4.4 Subject two tone quality results. 69

4.5 Subject two breath management results. 71

4.6 Subject two intonation results 73

4.7 Subject three tone quality results... 76

4.8 Subject three breath management results... 78

4.9 Subject three intonation results 80

4.10 Subject four tone quality results.. 83

4.11 Subject four breath management results.. 85

4.12 Subject four intonation results. 87

4.13 Subject five tone quality results... 90

4.14 Subject five breath management results.. 92

4.15 Subject five intonation results.. 94

4.16 Subject six tone quality results 97

4.17 Subject six breath management results 99

4.18 Subject six intonation results..101

4.19 Subject seven tone quality results.. 104

4.20 Subject seven breath management results. 106

vi
4.21 Subject seven intonation results. 108

4.22 Subject eight tone quality results... 111

4.23 Subject eight breath management results... 112

4.24 Subject eight intonation results.. 114

5.1 Scenarios supporting the hypothesis 126

5.2 Scenarios contradicting the hypothesis 126

vii
 LIST OF FIGURES

Figure

1.1 Categories of feedback receive by the learner... 15

2.1 Subject performance on acquisition-phase and retention trials. 41

4.1 Subject one tone quality results. 62

4.2 Subject one average tone quality results by phase. 63

4.3 Subject one breath management results. 64

4.4 Subject one average breath management results by phase 65

4.5 Subject one intonation results 66

4.6 Subject one average intonation results by phase67

4.7 Subject two tone quality results. 69

4.8 Subject two average tone quality results by phase.70

4.9 Subject two breath management results. 71

4.10 Subject two average breath management results by phase.. 71

4.11 Subject two intonation results.. 73

4.12 Subject two average intonation results by phase. 74

4.13 Subject three tone quality results. 76

4.14 Subject three average tone quality results by phase.77

4.15 Subject three breath management results. 78

4.16 Subject three average breath management results by phase 79

4.17 Subject three intonation results 80

4.18 Subject three average intonation results.. 81

4.19 Subject four tone quality results.. 83

viii
4.20 Subject four average tone quality results by phase.. 84

4.21 Subject four breath management results.. 85

4.22 Subject four average breath management results by phase..86

4.23 Subject four intonation results. 87

4.24 Subject four average intonation results by phase. 88

4.25 Subject five tone quality results... 90

4.26 Subject five average tone quality results by phase...91

4.27 Subject five breath management results...92

4.28 Subject five average breath management results by phase.. 93

4.29 Subject five intonation results.. 94

4.30 Subject five average intonation results 95

4.31 Subject six tone quality results 97

4.32 Subject six average tone quality results by phase 98

4.33 Subject six breath management results 99

4.34 Subject six average breath management results by phase 100

4.35 Subject six intonation results..101

4.36 Subject six average intonation results by phase. 102

4.37 Subject seven tone quality results.. 104

4.38 Subject seven average tone quality results by phase..105

4.39 Subject seven breath management results..106

4.40 Subject seven average breath management results by phase. 107

4.41 Subject seven intonation results. 108

4.42 Subject seven average intonation results by phase 109

ix
4.43 Subject eight tone quality results... 111

4.44 Subject eight average tone quality results by phase...112

4.45 Subject eight breath management results... 113

4.46 Subject eight average breath management results by phase.. 113

4.47 Subject eight intonation results.. 114

4.48 Subject eight average intonation results by phase. 115

5.1 Subject two overall average results by phase...129

5.2 Subject six overall average results by phase 130

5.3 Subject eight overall average results by phase.131

5.4 Subject one overall average results by phase... 132

5.5 Subject three overall average results by phase.133

5.6 Subject four overall average results by phase.. 134

5.7 Subject five overall overage results by phase...135

5.8 Subject seven overall average results by phase... 136

x
 1

CHAPTER 1

INTRODUCTION TO PRINCIPLES OF MOTOR LEARNING

Frank is a 30 year old tenor who has been studying classical singing at
a post-secondary level for the past 12 years. He has a capable voice,
but has experienced difficulty in maintaining consistency in his range
and tone quality. He has studied with four different voice teachers and
performed in several master classes. Frequently, he has experienced
elation in a lesson or master class as a tidbit of instruction from the
teacher has allowed him to perform exercises or passages that he had
previously thought were above his abilities. Just as frequently,
however, that elation was quickly washed away as he returned to the
practice room only to find that his new-found ability was just as
unattainable as it had previously been.

In Franks case, it is reasonable to determine that, because his recently-acquired vocal

skills were fleeting in nature, he had not actually learned the behavior. Rather he had seen a

momentary improvement in his performance ability. Franks experience is certainly not

uncommon, and teachers of singing and students alike can undoubtedly readily recall multiple

examples of similar situations. The task for instructors, then, is to determine how they can best

structure the learning environment in their studio to facilitate the development of relatively

stable, long-term improvements in the abilities of their students, which are the very definition of

learning.

Theories pertaining to how humans process information in order to produce long-term

changes in skilled behavior have been developing for centuries and continue to develop as new

information becomes available. In the 20th century, educational theorists and psychologists such

as John Dewey, Robert Gagne, Jean Piaget, Burrhus Skinner, Lev Vygotsky, and Edward
 2

Thorndike1 all contributed to a mountainous volume of research and theory regarding human

learning and understanding, much of which is still referenced today in teacher education

programs around the world. Since learning to sing involves the coordination of hundreds of

muscles throughout the body, those theories that pertain to the learning of motor behavior are of

particular interest to the studio voice instructor. The research study reported in the following

chapters has examined two particularly paramount theories, those put forth by Jack A. Adams in

19712 and by Richard Schmidt in 1975.3 Additionally, subsequent research and publications

furthering the understanding of motor-learning theory are included in attempt to bring further

clarity to the question of how frequently instruction (feedback) should be provided to students

during the learning process. Adams and Schmidts theories are summarized below, followed by

a more thorough discussion of the qualities of feedback.

Jack A. Adams Closed-Loop Theory

In the 1960s and 70s, the majority of theories pertaining to motor-learning were what

Adams considered to be open-loop theories4 based primarily on understandings developed

1
Edward L. Thorndike, Education Psychology: Briefer Course, (New York: Teacher College,
Columbia University 1923).
2
Jack A. Adams, A Closed-Loop Theory of Motor Learning. Journal of Motor Behavior 3
(1971): 111-150.
3
Richard A. Schmidt, A Schema Theory of Discrete Motor Skill Learning, Psychological
Review 82 (1975): 225-260.
4
Adams, 117.
 3

early in the 19th century.5 Chief among these were Thorndikes Empirical Law of Effect,

which states, To the situation, a modifiable connection being made by him between [a

stimulus] and [a response] and being accompanied or followed by a satisfying state of affairs,

man responds, other things being equal, by an increase in the strength of that connection.6 In

other words, following a response with a reward will increase the likelihood of repeating that

response and, inversely, following a response with punishment will decrease the likelihood of

repeating that response.7

Adams took issue with applying this simplified stimulus-response approach to an

overarching understanding of all skilled-behavior acquisition. While most of the research

completed by Thorndike and his associates in support of his theory was completed using animal

subjects, Adams points out that in many key ways animal behavior is dissimilar to human

behavior. The difference between animal and human behavior form the backbone of Adams

concerns, which he articulated with five shortcomings of the Empirical Law of Effect.8

1. Delaying the reward (positive feedback) does not elicit the same response from animals and

humans. Delaying the reward for animals does in fact depress performance, while doing the

same for humans produces little or no effect on performance at all.

2. When the reward or reinforcement is withdrawn from animal subjects, performance of the

desired behavior declines. However, the same is not true for humans.

5
Certainly, additional research was being conducted throughout the century. However, Adams
cites primarily works from the early 19th c., indicating that they significantly influenced the
contemporary theories, and in response, his own theory.
6
Thorndike, 71.
7
Adams, 113.
8
Ibid., 114-115.
 4

3. Human subjects do not simply repeat behaviors just because those behaviors are rewarded.

Instead, Adams claims, humans attempt to correct their error in an effort to produce a more

accurate response in relation to the desired behavior. Adams points to research by Elwell and

Grindley as support for this claim.9

4. Humans covertly guide their motor behavior with verbal responses, at least in the early

stages of learning.10 This verbal guidance does not fit into the stimulus-response model that

follows from Thorndikes law.

5. Human performers can be aware of whether or not their motor movement was correct, often

without an outside reward provided by an external source.

Adams responded to the perceived shortfalls of Thorndikes theories by developing his

own theory of motor learning. As noted earlier, nearly all of the theories of learning at the time

of Adams research were open-loop theories. This meant that, in order for a motor response to

occur, some sort of outside stimulus response was needed. Furthermore, as long as the necessary

motivational and habitual states were achieved along with said stimulus, that same response

would always be produced.11 Adams noted that this explanation did not take into account the

effect that feedback, and specifically feedback regarding error, had on guiding the completion of

the motor task, both during and after an attempt was made to complete that task. He therefore

determined that his theory would need to be closed-loop. More precisely, his closed-loop theory

9
J. L. Ellwell and G. C. Grindley, The Effect of Knowledge of Results on Learning and
Performance, British Journal of Psychology 29 (1938): 39-54.
10
Adams, 115.
11
Ibid., 117.
 5

must be error-centered, with a reference mechanism against which feedback from the response

is compared for the detection and correction of error.12

Adams began the exposition of his closed-loop theory with a discussion of the nature of

knowledge of results (KR),13 which he initially defines simply as information provided to the

subject regarding his or her performance of the desired task that eventually leads the subject to a

correct performance attempt.14 However, he does further distinguish between qualitative KR and

quantitative KR. According to his definition, qualitative KR is dichotomous. This type of KR

would only provide the learner with information regarding which direction away from the

desired task his or her attempt varied (e.g. too long or too short).15 In the case of Frank the tenor

from the beginning of the chapter, if his instructor simply told him, You are flat, the instructor

would be providing qualitative KR. Alternatively, Adams defines quantitative KR as providing

the same information as qualitative KR, but with some scaled numerical information regarding

the degree to which the error was made.16 According to this definition, Franks instructor could

provide quantitative KR by saying, You were a quarter-tone flat. Of special importance to

Adams theory is that KR differs from feedback. Adams seems to refer to KR only in regard to

12
Adams, 120.
13
It should be noted that Adams definition of KR varies rather significantly from Schmidts
definition, which will be discussed later in the chapter.
14
Adams, 122.
15
Ibid.
16
Ibid.
 6

information received from outside the subjects body. Information received from and perceived

by the performers body is what he refers to as feedback.17

After his discussion of KR, Adams made the case for two states of memory, referred to as

the perceptual trace and the memory trace, both of which he claimed to have a determining role

in the performance of any motor task. The memory trace would be responsible for initiating the

movement, choosing its initial direction and determining early portions of the movement.18 The

perceptual trace, on the other hand, would be responsible for guiding the movement to the

correct location along the prescribed pathway.19 For clarity in the distinction between the two

traces, Adams discusses the latter first. In attempt to retain the same clarity, the traces will be

discussed in the same order here.

Adams concept of the perceptual trace is rooted in the notion of a trace or image which

perception has used for a long time to account for the recognition of exteroceptive stimuli.20 In

this sense, when a stimulus is experienced for the first time, whether that stimulus is seeing an

image or performing a motor task, a perceptual trace is created and imprinted. When that

stimulus is encountered again, it is recognized when it matches the imprint of the original

perceptual trace.21 For example, the first time that Frank sang an A4, an imprint of the motor

skills necessary to produce that pitch, referred to by Adams as a perceptual trace, was created.

Subsequently, every time that he attempted to reproduce that pitch again, he would access that

17
Adams, 123.
18
Ibid., 125.
19
Ibid., 124.
20
Ibid.
21
Ibid.
 7

perceptual trace and weigh it against the current feedback stimuli he was receiving in order to

determine the correctness or error of that attempt.

What if, however, Frank had never successfully negotiated the A4? How would he ever

develop a correct perceptual trace for producing that pitch? Adams accounts for this in his

theory by distinguishing between two stages in perceptual trace development: the verbal-motor

stage, and the motor stage. Early on in a subjects training, when mistakes are frequently being

made, the subject cannot simply repeat movements that the perceptual trace would recognize as

having made before or the subject would repeat the errors from past attempts. In this stage (the

verbal-motor stage) the subject must instead rely on KR to make each attempt different from the

previous ones.22 Once the KR has been reporting relatively small errors for some time, it is

safe to assume that, since the correct movement has been made fairly frequently, the perceptual

trace is sufficiently strong that the subject has moved into the motor stage. During the motor

stage, KR can be dropped out completely and conscious behavior eventually becomes

automatic.23

With his theory relying on the perceptual trace for guidance of motor tasks, Adams

admits that it would be tempting to stay with a single-trace theory with the perceptual trace also

being responsible for initiating the task. However, Adams notes three strong reasons that a

second trace, which he terms the memory trace, is necessary for a complete understanding of

motor skill acquisition. First, if one trace were responsible for both initiating a response and

testing its correctness, the response would always be judged against itself, meaning that it would

always necessarily be judged as having occurred with zero error. Second, the perceptual trace

22
Adams, 124
23
Ibid., 125.
 8

requires feedback, which can only occur after a response begins. Consequently, something else

must be responsible for initiating the response in the first place.24

The third justification for the memory trace requires a distinction between recall and

recognition. Adams defines the terms: Recall is a response production and recognition is

identification of a stimulus or a response.25 Accordingly, the initiation of a motor response is an

example of motor recall, whereas judging the correctness of that motor response is an example of

motor recognition. If no one trace can be responsible for both recall and recognition, then there

must be a need for a second trace. The memory trace, therefore, is responsible for initiating a

response, while the perceptual trace, as discussed earlier, is responsible for judging the

correctness of that response.

It is important to note that, according to Adams theory, the perceptual trace can not be

formed without experiencing the correct location at least once.26 Additionally, Adams

postulated that, because of its closed-loop nature, learning could take place, once the perceptual

trace had been established, without the subject being aware of the accuracy of his or her

performance (i.e without KR).27 These were key points with which Richard R. Schmidt took

issue, ultimately bringing him to develop his own theory.

24
Adams, 126.
25
Ibid., 126.
26
Ibid., 125.
27
Ibid., 124.
 9

Richard A. Schmidts Schema Theory

While Schmidt appreciated much of the work that Adams had done (most importantly,

that Adams theory attempted to test relatively long-term learning, whereas previous studies had

focused primarily on improvements in performance28), he took issue with several points of

Adams closed-loop theory. In particular, Schmidt disagreed that the subject needed to

experience the correct location in order to learn to be able to move accurately to that position.29

Additionally, Schmidt took issue with the idea that a subject would be able to continue learning

without subsequent KR.30 It was these issues that Schmidt hoped to address in his own theory.31

The primary focus of Schmidts theory concerns how the development of a schema32 (a

theory long accepted in verbal memory studies) applies to motor-learning. Schmidt uses the

process of learning to recognize a dog as an example of forming a schema.

The idea is that in order to perceive a set of visual stimuli (e.g., a

dog) and to classify these stimuli correctly in the category dog, we
need not have previously received the particular set of stimuli in
question. Through our past experiences with seeing dogs, we store
these stimuli in recognition memory and also abstract these stimuli
into a concept related to dogs for additional storage. This concept
forms the basis of a schema or rule of determining whether a new set
of visual stimuli should be classified into the category dog or not.
Thus to recognize an animal as a dog, we need not have ever seen that

28
Schmidt, 227.
29
Ibid., 228.
30
Ibid.
31
Ibid., 230.
32
S. H. Evans defines a schema as a characteristic of some population of objects [which]
consists of a set of rules serving as instructions for producing a population prototype. Evans, S.
H. A Brief Statement of Schema Theory, Psychonomic Science 8 (1967): 87-88.
 10

particular animal before, and with the use of the schema for dogs, we
correctly identify the animals category.33

Before a schema for a set of motor stimuli can be developed, a goal-oriented movement must be

produced several times, and sensory information from that movement must be stored. Schmidt

outlines four sequential steps of storing information included in the production of a goal-oriented

movement:34

1. Storing the initial conditions - consists of information received from receptors prior to a

response such as proprioceptive information about the positions of the limbs and body in

space.

2. Storing the response specifications for the motor program - consists of information

specifying alterations to the general motor program. These alterations include changes in

speed, force, etc. made to the motor program before it is run off.

3. Storing the sensory consequences of the response produced consists of actual feedback

stimuli received from the sensory organs while the motor program is being run. As a result,

these consequences would be a copy of the afferent information received by the brain during

completion of the task (movement).

4. Storing the outcome information consists of the success of the response in relation to the

outcome originally intended. This information consists of information received after the

movement. Therefore, absence of feedback would result in no outcome information (a major

difference from Adams theory).

33
Schmidt, 233.
34
Ibid., 235.
 11

Information from these four steps are stored together after the movement and, after a number

of these movements (Schmidt does not specify how many), the subject begins to abstract the

information about the relationship among these four sources of information, forming a schema of

that particular movement. Once the schema is formed, Schmidt outlines how that schema

operates to produce a movement in the order in which those steps occur:35

1. A desired outcome is specified/chosen.

2. Initial conditions that may affect the production of that outcome are determined.

3. New specifications for motor program are selected, which accommodate variances in the

initial conditions and, using relationships between outcomes and sensory responses stored

from previous attempts, expected feedback is selected.

4. Movement is initiated by running off the motor program.

5. Within 200 milliseconds, sensory receptors begin providing information about the

movements occurring.

6. Feedback information is conveyed to feedback states selected earlier and compared to the

expected outcomes. From this comparison, a response error is created.

7. Information regarding the response error is fed back to the schema for additional

adjustments. At the same time, error information is fed to the labeling system, resulting in

subjective reinforcement and fed back to the schema as subjective information.

8. Final error information is provided by the experimenter or teacher in the form of Knowledge

of Results (KR). This information stems from the measured outcome of the subjects

response (movement).

35
Schmidt, 239-240.
 12

Motor Learning Defined and the Role of Feedback

After publishing several studies and articles seeming to affirm his theories,36,37,38,39 Schmidt

went on to write a textbook along with Timothy D. Lee, the fourth edition of which is referenced

here.40 In this text, Schmidt finally amalgamates all of the main points of his theory into a clear

definition of motor learning: Motor learning is a set of processes associated with practice or

experience leading to relatively permanent changes in the capability for movement.41 An

important element of this definition is the inclusion of the words relatively permanent changes,

which set motor learning apart from fleeting and inconsistent improvements in performance. To

highlight this point, in a chapter of a book currently in press, Katherine Verdolini dicusses three

core issues providing further distinction between learning and performance. First, the nature of

learning is dynamic. Acquiring new abilities does not result in accumulation of bits of

information in some discrete location in the head.42 Instead, learning is a change in the

36
Richard A. Schmidt, Control Processes in Motor Skills, Exercise and Sport Sciences
Reviews 4 (1976): 229-261.
37
Richard A. Schmidt, Movement Education and the Schema Theory, in Report of the 1976
Conference: National Association for Physical Education of College Women held in Cedar
Falls, IA 3-8 June 1976, ed. E. Crawford.
38
Richard A. Schmidt, Schema Theory: Implications for movement education, Motor Skills:
Theory into Practice 2 (1977): 36-38.
39
Richard A. Schmidt, Past and Future Issues in Motor Programming, Research Quarterly for
Exercise and Sport 51 (1980): 122-140.
40
Richard A. Schmidt and Timothy D. Lee, Motor Learning and Performance: From Principles
to Practice, (Champaign, IL: Human Kinetics Books 1991).
41
Ibid., 264-265.
42
Katherine Verdolini, Motor Learning Principles: How to Train, in Vocolgy, by Ingo R. Titze
and Katherine Verdolini, (currently in press; used by permission), 3.
 13

probability that a certain response will be given in the future as a result of practice.43 Second,

learning cannot be observed directly. It can only be inferred by observing changes in individual

performances. The third issue is what Verdolini may argue to be the most important: a

temporary improvement in performance ability does not indicate that learning has taken place.

Only after a change in performance ability has proven to be relatively stable over time can it be

said that the new skill has been learned.44

Once a clear definition of motor learning had been developed, Schmidt turned his efforts

to determining what factors most greatly influenced the achievement of motor learning. Among

other factors including attention and motivation on the part of the learner, Schmidt realized that

feedback from the instructor greatly influenced motor learning. More specifically, extrinsic

feedback regarding errors is one of the more important sources of information.45

In terms of motor learning theory, as defined by Schmidt, feedback refers to information

received by the learner before, during, and after a task has been attempted.46 This feedback can

be divided into feedback that is received from either intrinsic or extrinsic sources.47 Feedback

from intrinsic sources would include sensory information that arises as a natural consequence of

attempting to perform a task.48 This intrinsic feedback can be further divided into categories of

either proprioceptive or exteroceptive feedback. Proprioceptive feedback refers to sensory

43
Verdolini, 4.
44
Ibid.
45
Richard A. Schmidt, Motor Learning and Performance: A Problem-based Approach, 3rd ed.,
(Champaign, IL:Human Kinetics Books, 2004), 305.
46
Ibid., 276.
47
Ibid., 277.
48
Ibid.
 14

information received from sources within the learners own body with its primary sources being

sensory receptors imbedded within the body tissues. Exteroceptive feedback refers to sensory

information received from sources outside the body, the primary sources of which are vision and

audition.49

Extrinsic feedback, on the other hand, consists of information, other than sensory,

provided to the learner from any source outside of the learners own body, such as a mechanical

device or an instructor.50 Ideally, this feedback should provide information that the learner

cannot receive on his/her own without the aid of the outside information source, and should

supplement the intrinsic feedback that the learner has already received.51 Extrinsic feedback can

provide information regarding the learners performance in one of two ways.

First, it can provide information regarding the result of the performance attempt, such as

how close the attempt was to the target behavior. To use Schmidts terminology, this type of

feedback can be referred to as Knowledge of Results (KR).52 Often, KR is redundant, simply

restating the information that the learner has already gleaned through intrinsic feedback. In this

case, the feedback is of little value to the learner. However, KR can be beneficial, even

necessary, when the learners ability to receive intrinsic feedback is diminished or distorted.53

Second, extrinsic feedback can provide the learner with information regarding the quality

of his/her performance attempt, such as whether the attempt was made in the most efficient or

49
Schmidt, 2004, 92.
50
Ibid., 279.
51
Ibid.
52
Ibid.
53
Ibid., 280.
 15

effective means possible. Again using Schmidts terminology, this type of feedback can be

referred to as Knowledge of Performance. In most cases, Knowledge of Performance (KP) is

more beneficial to learners than KR, as it provides information that the learner would be unable

to receive through any other means.54 Figure 1.1 is an adaptation of a figure in Schmidts text

depicting the breakdown of all information received by the learner into the different categories of

feedback discussed here.55

All Information Received by Learner

Related to the Action Not Related

Available Before Available After

Movement Movement

Intrinsic Feedback Extrinsic Feedback

Proprioceptive Exteroceptive KR KP
Muscles Vision
Audition

Figure 1.1 - Categories of feedback received by the learner

54
Ibid., 281.
55
Ibid., 277.
 16

To clarify each of these types of feedback, let us return to the example of Frank the tenor.

As Frank performs a phrase from an aria, intrinsic feedback would include proprioceptive

information Frank receives regarding whether or not the high notes felt tight in his throat as well

as exteroceptive information received by his ears regarding the intonation and tonal quality of the

phrase attempt (however, it may be noted that because of auditory factors such as bone

conduction, listening to ones self may border on proprioceptive intrinsic feedback). Once Frank

finishes his attempt, extrinsic information would include information provided by his singing

instructor or a vocal coach regarding what he or she noticed about Franks attempt.

Additionally, information provided by a computer display using voice analysis software,

electronic tuning equipment, and sound pressure metering devices all would fall under the

category of extrinsic feedback. Furthermore, if the instructor told Frank, You were flat on the

high A, that feedback would be classified as KR. Assuming that Frank was able to discern

intonation during his/her performance, this KR would be redundant and of little value. On the

other hand, if the instructor said something to the effect of The high A had a pressed tone

quality, that information would be classified as KP. Since Franks own aural perception of his

own sound can be distorted, this KP could be useful information.

While it may be interesting and useful to investigate how to improve a learners ability to

perceive and interpret his/her own intrinsic feedback, this study is aimed at informing training

practices from the instructors point of view. Consequently, the use and quality of extrinsic

feedback is more pertinent. When considering extrinsic feedback, Schmidt suggests several

questions to ask regarding what its properties are, when to provide it, how much information to

provide, and how precise to be when giving it.56

56
Schmidt, 2004, 283-306.
 17

Properties of Extrinsic Feedback

In terms of the properties of extrinsic feedback, Schmidt states that it can serve to

motivate a learner, reinforce a behavior, inform the learner, and/or produce a dependency on the

feedback.57 The first of these properties, motivating a learner, may not be the primary objective

for the instructor when providing feedback. However, the motivational properties of that

feedback often provide an added benefit to the learner by encouraging them to continue to give

their best effort, even when faced with multiple repetitions and monotonous training sessions.58

Reinforcing feedback will, as its name implies, reinforce a certain behavior in one of two

ways. First, positively reinforcing feedback will provide the learner with an experience, which

due to its pleasant nature, would increase the likelihood that the desired behavior will be repeated

(similar to Thorndikes stimulus-response relationship).59 The second possibility, negatively

reinforcing feedback, will provide the learner with an experience consisting of the removal of an

unpleasant stimulus, thereby increasing the likelihood that the desired behavior will be

repeated.60 To clarify, return again to Frank. If Frank performs a certain difficult passage

correctly for the first time and the instructor says, That was it. Good job! this would be an

example of positively reinforcing feedback. Alternately, after the previous five wrong attempts

the instructor could have said with increasing impatience, NO! THAT IS NOT CORRECT. DO

IT AGAIN! If, then, after the first correct attempt the instructor said nothing but let Frank

finally continue to the next phrase, this would be an example of negatively reinforcing feedback,

57
Schmidt, 2004, 282.
58
Ibid.
59
Ibid., 284.
60
Ibid.
 18

because the unpleasant experience of being berated by the instructor was removed. In his 1978

study, Adams found that positively reinforcing feedback is more effective than negatively

reinforcing feedback. Incidentally, the same study also found that both positively and negatively

reinforcing feedback were more effective motivators than was the use of punishment.61

When instructors think about feedback to their students, providing information is often

their primary objective. Indeed, the very definitions of feedback listed by both Adams and

Schmidt state that feedback is inherently information. Schmidt, however, indicates a slightly

more pointed definition of informative feedback as feedback that provides the learner with the

direction they need to correct their errors and to modify their future performance.62 This

definition should be kept in mind in the following discussion of how much and what information

to provide during feedback.

Throughout the learning process, the instructor must be wary of providing feedback in

such a manner that the learner becomes dependent on that feedback in order to produce the

desired behavior. It has recently become evident that producing frequent augmented, verbal

feedback to a learner can guide the learner to the correct movement in much the same way that

physical guidance aids do. However, Schmidt points63 to a 1959 study by J. Annett where it was

found that learners who practiced with a physical guidance aid could not perform the task once

the guidance aid was taken away.64 This was a result of the fact that the learners had become

61
Jack A. Adams, Theoretical Issues for Knowledge of Results, in Information Processing in
Motor Control and Learning, ed. G. E. Stelmach (New York: Academic Press, 1978), 229-240.
62
Schmidt, 2004, 286.
63
Ibid., 225.
64
J. Annett, Feedback and Human Behavior (Middlesex, England: Penguin Books, 1959).
 19

dependent on the guidance aid. Schmidt maintains that learners can develop the same

dependency on verbal feedback when it is provided too frequently.65 Thus, the instructor must

ask the question of when to provide feedback.

When to Provide Extrinsic Feedback

The question of when to provide feedback actually consists of at least two questions: a)

whether or not to provide feedback at all, and b) if it is determined that feedback is necessary,

how frequently should it be provided? When deciding whether or not to provide feedback, the

first consideration should be whether or not feedback is necessary.66 While this consideration

may seem intuitive, instructors frequently provide feedback almost as an instinctual response to

the completion of a learners attempt. Rather than providing this type of reactionary feedback,

the instructor would be better off to consider the complexity of the task weighed against the

ability and experience of the learner.67

The question of how frequently to provide feedback is the very subject of the study

discussed in the following chapters of this paper. Schmidt states that it is important to make a

distinction between absolute feedback frequency and relative feedback frequency. Absolute

feedback frequency is simply a statement of how many times feedback was provided during a

training or instruction session. Alternatively, relative feedback frequency refers to the number of

times feedback was provided relative to the total number of attempts made during the session.68

65
Schmidt, 2004, 287.
66
Ibid., 289.
67
Ibid.
 20

For example, if in a lesson Frank attempted to sing a phrase 15 times and the instructor provided

feedback 5 times, then the absolute feedback frequency for the session would be 5 and the

relative feedback frequency would be 33%. Schmidt claims that increasing absolute feedback

frequency will result in increased learning and cites several studies to support that claim.

However, he also states that decreasing the relative feedback frequency will also result in

increased learning. If both of these assertions are true, then the ideal situation for learning would

be a large number of attempts made by the learner, with feedback being provided by the

instructor only after every few attempts. Several studies have examined these assertions and

they will be discussed in further detail in chapter 2.

How Much and What Information to Provide

Since the memory capabilities of humans are limited, instructors must give careful

consideration to how much information should be provided to the learner during each instance of

feedback. Schmidt recommends focusing on one feature of the task that is most fundamental to

its successful completion.69 He also suggests a method of feedback called summary feedback,

which is again directly related to the question of frequency of feedback. In summary feedback,

the instructor would withhold feedback for a number of attempts, and then provide the feedback

in a summary form. For support of summary feedback, Schmidt points to studies conducted by

J.J. Lavery70 and by Schmidt, Lange, and Young.71 In both studies, it was found that subjects

68
Schmidt, 2004, 303.
69
Ibid., 295.
70
J. J. Lavery and F. H. Suddon, Retention of Simple Motor Skills as a Function of the Number
of Trials by Which KR is Delayed, Perceptual and Motor Skills, 15, 1962, 231-237.
 21

who received summary feedback performed better on a retention test of a practiced task than

those subjects who received immediate feedback following each attempt. However, Schmidt

also notes that, as indicated by the study that he and his colleagues conducted, as the complexity

of the task being learned increases, the number of attempts being summarized should decrease in

order to maximize learning.72 A variation of summary feedback is to average all of the attempts

being summarized to reveal a single trend that can then be addressed by feedback. This practice

is known as average feedback.73

When deciding what information to provide, of primary importance to the instructor is to

ensure that the information he or she is providing is addressing elements of the task that are

under the learners control.74 In order to ensure that an element being addressed by an

instructors feedback is indeed under the control of the learner, it is helpful for the instructor to

have at least some understanding of how people control their movements. In the realm of

singing, this may well be a convincing argument for the need for teachers of singing to have at

least a rudimentary understanding of the anatomy and physiology of the singing apparatus.

It may also be beneficial for an instructor to consider whether the feedback they are

providing is descriptive or prescriptive. Descriptive feedback simply restates the result of the

attempt, whereas prescriptive feedback provides information that will be more helpful in guiding

the learners subsequent attempts.75 To clarify, imagine that Frank the tenor attempted to sing a

71
Richard A. Schmidt and others. Optimizing Summary Knowledge of Results for Skill
Learning, Human Movement Science, 9, 1990, 325-348.
72
Schmidt, 2004, 299.
73
Ibid.
74
Ibid., 289.
 22

word that begins with a vowel and he started the sound with a hard, glottal onset. If his

instructor simply said, The onset of that word was too glottal, that would be an example of

descriptive feedback. If, however, the instructor told Frank how to produce a more balanced

onset and avoid the glottal attack, he or she would be providing prescriptive feedback. This

division may bring to mind Adams distinction between qualitative and quantitative KR or

Schmidts distinction between KR and KP.

How Precise to Be When Providing Feedback

It may seem that determining how precise to be with feedback may be the same as

deciding what information to provide. However, there is a distinction, which can be clarified by

returning once again to the example of Frank. When Frank sings a flat note, his instructor has

two questions he or she must decide to pose. First, will he or she tell Frank that he was flat?

This would be an example of deciding what information to provide. Second, will he or she tell

Frank just how far flat he was? This is an example of how precise to be. Schmidt notes that, in

general, very precise feedback is not necessarily more effective than less precise feedback.76

Furthermore, Magill and Wood found that learners at an early stage will not benefit from highly

precise feedback because the magnitude of their errors is so great.77 Instead, it may be more

beneficial to provide information pertaining only to the direction (sharp or flat) and magnitude

(quarter-tone) of the error. Even then, information regarding the direction of the error is more

75
Schmidt, 2004, 294.
76
Ibid., 300.
77
R.A. Magill and C.A. Wood, Knowledge of Results Precision as a Learning Variable in
Motor Skill Acquisition, Research Quarterly for exercise and Sport 57 (1986): 170-173.
 23

important than information regarding the magnitude.78 One method of feedback that deals with

the matter of precision is the bandwidth feedback method. In this practice, the instructor would

withhold feedback unless the attempt falls outside the realm of some acceptable result (i.e.

outside an acceptable bandwidth). This practice is especially well-suited to tasks in which there

is a very clearly defined desired outcome.79

Application to Voice Science and Need for the Current Study

While Schmidt has clearly outlined his theory in terms that prove its basis in cognitive

science, the application of that theory to practical training in singing and voice therapy has not

yet been widely practiced. To this end, Katherine Verdolini and Timothy Lee published a

chapter in Vocal Rehabilitation for Medical Speech-Language Pathology aimed at applying

schema theory to the speech clinic or voice studio.80 Through their discussion, along with

review of relevant research, Verdolini and Lee conclude that augmented feedback is necessary

for motor learning, but that less frequent feedback may enhance learning more.81 Verdolini and

Lee define augmented feedback as feedback that is received from sources that have been

augmented by an instructor or therapist"82 as opposed to feedback that a person normally

receives as a natural, sensory attribute of an action.83 In other words, augmented feedback

78
Schmidt, 2004, 301.
79
Ibid.
80
Katherine Verdolini and Timothy D. Lee., Optimizing Motor Learning in Speech
Interventions, In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine
M. Sapienza and Janina K. Casper (Austin, TX: Pro-ed, 2004), 403-446.
81
Ibid., 435.
82
Verdolini and Lee, 417.
 24

simply refers to feedback that is provided by an outside source. Augmented feedback may

include verbal and/or graphic reports of the results of a previous performance attempt

(knowledge of results) and can be provided either by an instructor or by various types of

instrumentation.84 For the purpose of the following study, knowledge of results was

supplemented by additional instruction during augmented feedback.

Verdolini and Lee also spend some time clarifying the distinction between learning and

performance. They use an old, clear definition from early research in experimental psychology

that emphasizes differences between performance, skill, and skill acquisition (i.e. learning).85 In

this definition, performance refers to an action that results in a measurable outcome. Skill refers

to an individuals capability to perform. Finally, skill acquisition (learning) is defined as a

process, which leads to relatively long-term changes in an individuals skill, or ability to

perform. Simply put, learning results in permanent or semi-permanent changes in ones ability

to perform. Verdolini and Lee, then, mandate that any claims about learning must be based on

transfer or retention tests conducted some period after a training session has been concluded.86

Without a transfer or retention test, it would be impossible to test the permanence of the changes

in performance. It is important to understand clearly the difference between performance and

learning as studies87,88,89 suggest that, especially in regard to augmented feedback from the

83
Ibid.
84
Ibid.
85
Edward Tolman, Purposive Behavior of Animals and Men, (New York: Century, 1932).
86
Verdolini and Lee, 411.
87
E. A. Bilodeau, Ima McD Bilodeau, and Donald A. Schumsky, Some Effects of Introducing
and Withdrawing Knowledge of Results Early and Late in Practice, Journal of Experimental
Psychology 58 (1959): 142-144.
 25

experimenter or teacher, those practices that most enhance learning may actually depress

immediate performance in lessons or therapy sessions.

While Verdolini and Lee, and subsequently Lynn Helding,90 have all indicated that these

principles of motor learning theory should be applicable to studio voice instruction, very little

research has been conducted to provide quantitative evidence that the principles gleaned from

years of research in other fields can be effectively applied in the voice studio. The current study,

as outlined in Chapter 3, has been designed to test just such an application, especially in regard

to the frequency at which feedback is provided by the voice instructor. Accordingly, this chapter

has provided an overview of the principles of motor learning theory and has focused most

heavily on the theory regarding feedback. For more thorough review of other elements effecting

motor skill acquisition, such as individual differences, motivation, and practice variability, the

reader is encouraged to investigate several textbooks written by Schmidt, each of which examine

the issues of motor skill acquisition from different perspectives.91

88
Timothy Lee, Margaret A. White and Heather Carnahan, On the Role of Knowledge of
Results in Motor Learning: Exploring the Guidance Hypothesis, Journal of Motor Behavior 22
(1990): 191-208.
89
Darl W. Vander-linden, James H. Cauraugh and Tracy A. Greene, The Effect of Frequency of
Kinetic Feedback on Learning an Isometric Force Production Task in Nondisabled Subjects,
Physical Therapy 73, no. 2 (1993): 79-87.
90
Lynn Helding, Voice Science and Vocal Art, Part Two: Motor Learning Theory, Journal of
Singing 64, no. 4 (March/April 2008): 417-428.
91
Richard A. Schmidt, Motor Learning and Performance: From Principles to Practice
(Champaign, IL: Human Kinetics Books, 1991); Richard A. Schmidt, Motor Control and
Learning: A Behavioral Emphasis 3rd ed. (Champaign, IL: Human Kinetics Books, 1999;
Richard A. Schmidt, Motor Learning and Performance: A Problem-Based Learning Approach,
4th ed. (Champaign, IL: Human Kinetics Books, 2004).
 26

The amalgamation of the research discussed in this chapter has led the author to believe

that there are several truths regarding feedback and learning in the classical voice studio.

1. Extrinsic feedback (referred to as augmented feedback by Verdolini and Lee), including

instruction from the voice instructor, visual guidance from devices such as a hand mirror, and

even aural guidance from a piano or other pitch-producing device, is necessary for the

acquisition of the motor skills required by classical singing technique.

2. Extrinsic feedback can help reinforce desirable motor skills while inhibiting undesirable

skills.

3. Extrinsic feedback will often guide a student singer to an immediate improvement in the

ability to perform a vocal task.

4. Excessive extrinsic feedback, however, will interrupt a student singers cognitive processes

as he or she attempts to compare the outcome of an attempt at a vocal task with the expected

or desired outcome of that task.

5. True acquisition (learning) of the skill required to perform a vocal task will only be

evidenced by the ability of the student singer to perform that task consistently after a period

during which there was an absence of instruction and/or the ability to transfer the requisite

skills to a related, but unpracticed vocal task.

If these five statements are indeed true, as a great deal of research suggests that they are,

then it stands to reason that reducing the frequency with which extrinsic feedback is provided to

a student singer, while not removing the feedback altogether, will lead to more effective

acquisition of the motor skills required by classical singing technique. This statement provides

the hypothesis that the current study attempts to test.

 27

CHAPTER 2

LITERATURE REVIEW

Over the past several decades, many studies have been conducted that were aimed at

applying the principles of motor learning theory, as outlined in chapter one of this paper, to

various disciplines. Of particular relevance to the current research are those studies regarding

feedback frequency. Six such studies have been summarized below. These studies have been

categorized as those dealing with feedback frequency and simple arm-positioning/timing motor

tasks, and those dealing with feedback frequency and vocal motor tasks.

In deciding which and how many studies to include in this chapter, the researcher

continued to gather articles and reports until the citations started to become circular (i.e. they all

started citing the same research studies). The most frequently cited articles are those reported

below. The study outlined in chapter three, then, was based on two criteria: how close it could

be to the types of studies that were most frequently reported, and how well it could be adapted to

the voice studio with minimal disruption to the learning environment.

Of course, more studies have been conducted and published that focus on application of

motor learning theory to various motor tasks. Each of the studies reported below, however, has

provided guidance in the development of the current research design, which is presented in

chapter three. While the studies in the second category are directly concerned with the role of

various elements of feedback on the training of voice skills, none of them are centered in the

realm of classical singing technique. The current study is designed to begin to fill this gap in the

research.
 28

Feedback Frequency and Simple Arm-Positioning/

Timing Motor Tasks

Reduced Frequency of Knowledge of Results

Enhances Motor Skill Learning92

In their paper, Carolee J. Winstein and Richard A. Schmidt began by noting that in

previous research regarding the efficacy of reducing the relative frequency of feedback, absolute

frequency of feedback was allowed to co-vary along with the relative frequency of feedback.93

Additionally, they note that previous research had been conducted using relatively simple tasks.

The researchers claim that perhaps enhanced learning [resulting from decreased relative

frequency of feedback] cannot be demonstrated with such simple tasks.94 This claim, according

to the researchers, poses some doubt as to whether or not the results from the previous studies

they cite95 can be generalized. In an attempt to more thoroughly and accurately test their

hypothesis that less-frequent feedback would result in greater long-term improvement in the

ability to perform a motor task, Winstein and Schmidt conducted three experiments, all of which

used a more complex motor task. In each of these experiments, subjects were required to learn

and replicate a four-stage elbow-extension/flexion movement, performed to a set degree of

92
Carolee J. Winstein, and Richard A. Schmidt, Reduced Frequency of Knowledge of Results
Enhances Motor Skill Learning, Journal of Experimental Psychology: Learning, Memory, and
Cognition 16, no. 4 (1990): 677-791.
93
You may recall from chapter one (page 17) that absolute frequency refers to the total number
of times that feedback was provided, whereas relative frequency refers to the number of times
feedback was provided in relation to the number of attempts made.
94
Winstein and Schimdt, 1990, 679.
95
e.g. Linda Ho and John B. Shea, Effects of Relative Frequency of Knowledge of Results on
Retention of a Motor Skill, Perceptual Motor Skills 46 (1978): 859-866.
 29

extension/flexion and within a set time parameter. Subjects were required to extend their elbow

to a prescribed position, reverse the movement and flex the elbow to another prescribed position

(different from the starting position), reverse again and extend the elbow to a different prescribed

position, and finally flex the elbow once again to a final prescribed position, all within 800

milliseconds. A plywood sheet was placed such that subjects were not allowed to directly view

their arm during their movement attempts. Instead, a video monitor was provided on which

subjects could view their movements remotely. In this way, the researchers could manipulate

how frequently the subjects received feedback (visualization of their movement attempts)

regarding the accuracy of their movements.

Experiment One

The authors note that many previous experiments regarding feedback frequency used

retention tests where no KR was provided during the testing trials. Because of this, any evidence

in these experiments that relatively low feedback frequency resulted in better performance on

retention tests could be interpreted to mean that subjects who had been training with relatively

low feedback frequency performed better simply because the testing environment was more

similar to their training environment. This interpretation of the results was referred to as the

specificity hypothesis. In order to test the specificity hypothesis, experiment one utilized four

different retention-test conditions where subjects received feedback on 0%, 33%, 66%, or 100%

of their retention test attempts to perform the extension-flexion task described above. During the

skill acquisition phase of the experiment, however, subjects received feedback only on either

100% or 33% of their attempts. To this end, 136 undergraduate student volunteers (74 female
 30

and 62 male) were divided into eight groups (two acquisition-phase feedback conditions for each

of the four retention-test feedback conditions) of seventeen students each.

The acquisition phase of the experiment consisted of two thirty-minute sessions, during

which each subject made 99 attempts to perform the task. Each subject was allowed 15 to 20

seconds between each attempt, regardless of feedback condition. Those subjects in the 100%

feedback groups received feedback after every trial. Subjects in the 33% feedback groups

received feedback following only two of every six attempts, though not predictably after every

third attempt. An immediate retention test of 27 attempts was conducted ten minutes following

the completion of the second acquisition session. Accuracy data was collected for each attempt

during both the acquisition and retention test phases.

Analysis of the skill-acquisition phase data revealed that subjects in both groups of

feedback frequency reduced their error (i.e. performed the task more accurately) throughout the

phase. Predictably, the 33% feedback group exhibited higher levels of error throughout the

acquisition phase. However, the difference in error between the two groups was not statistically

reliable, indicating that less frequent feedback may not inhibit performance during skill

acquisition quite as much as had been previously thought.

As noted earlier, the 27-trial retention test was administered providing 0%, 33%, 66%, or

100% feedback. Consistently, the subjects who had received 33% feedback during the skill-

acquisition phase tended to perform slightly better than those subjects who received 100%

feedback during skill-acquisition (though the difference was not statistically significant)

regardless of the retention test condition. Therefore, the experiment provided no evidence in

support of the specificity hypothesis.

 31

Experiment Two

As experiment one was designed to test the specificity hypothesis, experiment two was

designed to test what the researchers termed the guidance hypothesis, which predicts that error

information should be most useful early in practice to drive the performer toward the goal

response.96 In other words, this hypothesis indicates that in the early stages of skill acquisition,

more frequent feedback will be beneficial, while less frequent feedback will be more beneficial

during the later stages of skill acquisition. In order to test the guidance hypothesis, Winstein and

Schmidt designed an experiment in which feedback was provided more frequently early on in the

skill acquisition phase and was reduced gradually in later stages.

In experiment two, 58 undergraduate student volunteers (42 female and 16 male) were

divided into two groups: a 100% feedback condition group, and a 50% feedback condition group.

Subjects from both groups were required to perform the same arm flexion/extension task as was

used in experiment one. Similar to experiment one, experiment two consisted of a skill-

acquisition phase, consisting of two practice sessions of 96 practice attempts each, followed by

two no-feedback retention tests. One retention test was performed approximately five minutes

following completion of the skill-acquisition phase, while the second retention test was

performed approximately 24 hours later. During the skill acquisition phase, the 100% feedback

group received feedback following every practice attempt, while the 50% feedback group

received feedback following only half of the practice attempts. Differing from experiment one,

however, feedback for this group was not evenly distributed throughout the phase. Instead,

feedback was provided following each of the first 22 attempts in both of the two practice

sessions. Following these attempts, which were essentially 100% feedback conditions, feedback

96
Winstein and Schmidt, 1990, 682.
 32

was systematically reduced throughout the remaining attempts to such a point that subjects in

this group received feedback following only 4 of the final 12 attempts. As with experiment one,

the absolute feedback frequency was allowed to co-vary with the relative feedback frequency,

meaning that as the relative frequency of feedback decreased, so did the absolute feedback

frequency.

Error data from the skill-acquisition phase indicated that both groups reduced their mean

error throughout the duration of the phase. While not statistically significant in amplitude, the

50% feedback group tended to have slightly lower error scores during the second practice

session. Similarly, the 50% feedback group maintained a lower level of error during the delay

between the two practice sessions than did the 100% feedback group. This indicated that the

50% feedback group did not slide back in ability to perform the task nearly as much as the 100%

feedback group. Error data from the immediate retention test, conducted five minutes following

the completion of the acquisition phase, showed that the 50% feedback group performed with

less error than the 100% feedback group, though again to a statistically insignificant degree.

During the delayed retention test, however, the 50% feedback group performed with a

statistically significant 35% lower rate of error than did the 100% feedback group.

These findings indicate that providing feedback on 50% of attempts during skill

acquisition led to a higher level of retention of the ability to perform the motor task when

compared with providing feedback on 100% of the attempts. The researchers do note that

supporters of the specificity hypothesis would argue that, because the subjects in the 100%

feedback group had never practiced without feedback, it would come as no surprise that they

would perform less accurately on a no-feedback retention test. While this hypothesis cannot be

completely ruled out, the more accurate performance by the 50% feedback group on the delayed
 33

retention test combined with the higher level of retention during the delay between the two

practice sessions would imply that lower frequency feedback resulted in enhanced learning.

Experiment Three

While the results of experiment two did imply that lower frequency feedback resulted in

enhanced learning, it was unclear whether those positive effects were due to the similarity

between the testing and training conditions of the 50% feedback group or if the lowered

feedback frequency was truly beneficial to learning. In an attempt to address this question, the

researchers designed experiment three wherein the feedback frequency schedules were kept the

same during the skill acquisition phase as in experiment two. However, in experiment three KR

was provided during the delayed retention test, whereas KR was not provided during the delayed

retention test in experiment two. 46 undergraduate student subjects (26 female and 20 male)

participated in experiment three, practicing the same elbow flexion/extension task as in

experiments one and two.

Similar to the results of experiment two, subjects in the 50% feedback group performed

the task with less error in the initial trials of the second practice session, though the results were

not statistically reliable (meaning simply that the variance in performance between the two

groups was not of sufficient consistence to be considered reliable). The authors note that the

these initial trials of the second practice session can be viewed as mid-acquisition phase retention

test with KR provided since they were performed after a period of time during which no

feedback was provided and no practice trials were performed (please see Chapter 1 pp. 18 and 22

for more discussion on retention tests). Additionally, the 50% feedback group performed the

task with significantly less error on the delayed retention test performed 24 hours following the
 34

completion of the skill acquisition phase, and in which 100% KR was provided. Since the 50%

feedback subjects were able to perform the task with less error even though they had not

practiced as in the 100% KR retention test conditions, these results provide some evidence that

reduced relative frequency of feedback during skill acquisition is indeed beneficial to learning.

Overall, the findings of all three studies combined seem to lend merit to the guidance

hypothesis of KR as formalized by Alan Salmoni and Richard Schmidt in their 1984 review of

KR in motor learning.97 This guidance hypothesis suggests that KR plays two roles in motor

learning.98 First, it benefits the learner by guiding the learner toward the desired outcome (i.e.

movement or motor task). Second, however, it is detrimental to the learner as it encourages, or

at least allows, the learner to rely on and even become dependent on KR in order to achieve the

desired outcome.

In terms of reducing the relative frequency of KR, the guidance hypothesis would seem

to suggest that any negative effects from reducing the amount of guidance received by the

learner from KR would be offset by the beneficial effects of reducing the dependency-creating

effects of KR. The findings of experiment one, which found little positive effect from reducing

the relative frequency of KR, would support this view of the guidance hypothesis. The guidance

hypothesis is further examined by the researchers of the following study.

97
Alan W. Salmoni, Richard A. Schmidt and Charles B. Walter, Knowledge of Results and
Motor Learning: A Review and Critical Reappraisal, Psychological Bulletin 95, no. 3 (1984):
335-386.
98
Ibid., 380.
 35

On the Role of Knowledge of Results in Motor Learning:

Exploring the Guidance Hypothesis99

Later in 1990, after the previous research was published, Timothy D. Lee, Margaret A.

White, and Heather Carnahan published this research in which the guidance hypothesis claims

of the detrimental effects of KR were examined. The authors first note that early research

conducted by Jack A. Adams100 and Edward A. Bilodeau101 claim that learning takes place most

effectively when KR is provided frequently, which is in direct contradiction to the guidance

hypothesis. In order to examine the discrepancy between these two claims, the researchers

designed three experiments in which a reciprocal movement task was either guided using a

metronome (experiment one) or using augmented auditory KR (experiments two and three).

All three experiments employed the same methodology throughout, with the exception of

the manner in which the augmented feedback was provided. In each of the studies, 30

undergraduate students (15 male and 15 female) were asked to use a stylus to tap one of two

metal plates that were connected to a computer to calculate the timing of the subjects taps. The

subjects were instructed to perform seven consecutive taps, alternating between the two plates

with 500 milliseconds (ms) between each tap. All subjects performed this seven-tap trial at least

50 times with an intertrial interval of 5 seconds.

99
Timothy D. Lee, Margaret A. White and Heather Carnahan, On the Role of Knowledge of
Results in Motor Learning: Exploring the Guidance Hypothesis, Journal of Motor Behavior 22,
no. 2 (1990): 191-208.
100
Jack A. Adams, A Closed-loop Theory of Motor Learning, Journal of Motor Behavior 3
(1971): 111-150.
101
Ina M. Bilodeau, Information Feedback, in The Acquisition of Skill, ed. Edward A.
Bilodeau (New York: Academic Press, 1966) 255-296.
 36

For all three experiments, each of the 30 subjects was randomly assigned to one of three

feedback conditions. The first group, referred to as the 100% group, received feedback

following each of their 50 trials. The second group, referred to as the 50% -50T group, received

feedback following only half of the trials. This group still performed only 50 trials, meaning that

they received feedback in 25 instances. Consequently, for this group absolute feedback

frequency was allowed to co-vary with relative feedback frequency. The third group, referred to

as the 50% - 100T group, again received feedback following only half of the trials. However,

this group performed 100 trials, allowing the absolute feedback frequency to remain consistent

while the relative feedback frequency was reduced to 50%.

Following the completion of the acquisition phase consisting of either 50 or 100 trials,

subjects were allowed a 5-minute rest period before completing a retention test consisting of 20

trials, during which no feedback was provided. Statistical analyses were conducted on how close

the actual inter-tap time gap was to the desired 500 ms goal, providing an absolute constant error

for each groups performance of the task.

Experiment One

For experiment one, subjects were provided with feedback in the form of constant

guidance. In this experiment, subjects were guided to the correct inter-tap timing by a

metronome set at 120 beats per minute (bpm). During the acquisition phase, all three groups

appeared to perform at an equal level of timing accuracy, though the 100% group did present a

statistically insignificant trend to perform at a slightly higher level of accuracy during the first

ten guided trials. At the onset of the retention phase, all three groups again performed with

similar accuracy. However, as the retention test progressed, accuracy for the 100% group
 37

declined, while the 50% - 50T group maintained its level of accuracy and the 50% - 100T group

increased in accuracy. Again, the difference in performance accuracy was not significant.

Experiment Two

In experiment two, subjects were provided aural KR feedback following each tap. If the

inter-tap time gap was too large (525 ms or more), the computer immediately provided 200 Hz

tone, 80 ms in duration. If the inter-tap time gap was too small (475 ms or less), the computer

immediately provided a 680 Hz tone, again 80 ms in duration. If the inter-tap time gap was

between 476 and 524 ms, then no tone was provided. This absence of aural feedback was in fact

KR in itself, as the subjects were able to interpret the absence as an accurate performance.

During the acquisition phase, subjects were between three and six times more likely to produce

error than in experiment one. Similar to experiment one, however, there appeared to be no

significant effect from the relative frequency with which KR was provided. In the retention test,

error was reduced considerably across all feedback groups. However, a statistically significant

difference between any of the three feedback groups failed to emerge.

One possible explanation provided by the authors for the lack of significant difference

between the groups was in terms of the bandwidth of error allowed before a tone was produced.

In experiment two, subjects were allowed to produce a timing variance of 25 ms before a tone

was produced, resulting in 5% around the goal, or a 5% bandwidth. Previous research by D. E.

Sherwood has shown that the bandwidth size, in itself, has an effect on learning.102

102
D.E. Sherwood, Effect of Bandwidth Knowledge of Results on Movement Consistency,
Perceptual and Motor Skills 66 (1988): 535-542.
 38

Experiment Three

Citing the research by Sherwood, the authors designed experiment three to replicate

experiment two in every aspect except that the acceptable error bandwidth was decreased from

5% to 1%. Thereby, a low frequency tone was produced if inter-tap time gap exceeded 505 ms

and a high frequency tone was produced if the inter-tap time gap was less than 495 ms. During

the acquisition phase of experiment three, there again appeared to be no difference between the

three feedback groups in performance error of trials on which KR was provided. In trials on

which KR was not provided (for the two 50% feedback groups only), performance appeared to

be more consistent (not more accurate, but with a lesser degree of fluctuation) than in trials with

KR. This increase in consistency was significant for both 50% feedback groups. Additionally,

during the retention test, both 50% feedback groups performed with lower percentages of error

than did the 100% feedback group, indicating as experiment one did, that lower frequency of

feedback was better for retention, and therefore learning.

The findings of experiment one, where subjects were guided to the correct timing using a

metronome, and experiment three, where subjects were provided with KR within a low

bandwidth of acceptable error, both lead to the conclusion that 50% feedback was more effective

for learning. These results indicate that performance guidance and KR are both used in a similar

manner by learners during skill acquisition, namely that KR is used to guide the learner to the

desired outcome. The similarity between performance guidance and KR lends credibility to the

guidance hypothesis.
 39

The Effect of Frequency of Kinetic Feedback on Learning an

Isometric Force Production Task in Nondisabled Subject103

In this study, Vander Linden, Cauraugh, and Greene sought to apply the concept of less-

frequent feedback to the field of physical therapy. In so doing, it was their hope to address two

issues regarding feedback and motor learning that had not been addressed in previous studies.

First, the researchers sought to make the distinction between two types of feedback: knowledge

of results (KR) and knowledge of performance (KP).104 The researchers contended that, while

physical therapists did provide KR when their patients were able to complete a task, they often

worked with patients who were unable to complete a given task, meaning that much of the

feedback they provided was necessarily in the form of KP. Since the bulk of previous research

had been conducted using KR, there was justification for the researchers study in that they

would be examining the efficacy of varied frequency of feedback using KP. Second, the

researchers noted that previous research had been conducted using feedback that was provided

only after the prescribed task had been completed. Alternatively, the feedback provided in the

course of a typical physical therapy session is often provided both during (concurrent feedback)

and after the task. To fill this hole in the existing research, the authors sought to test the efficacy

of both forms of feedback while examining both the timing of feedback and the relative

frequency with which that feedback was provided (i.e. 100% or 50%).

103
Darl W. Vander Linden, James H. Cauraugh and Tracy A. Greene, The Effect of Frequency
of Kinetic Feedback on Learning an Isometric Force Production Task in Nondisabled Subject,
Physical Therapy 73, no. 3 (February 1993): 79-87.
104
The distinction between these two types of feedback has been discussed more thoroughly in
the first chapter of this paper.
 40

For the study, 24 non-disabled volunteers (eighteen male and six female), ranging in age

from nineteen to thirty-three years old, were divided into three groups with equal proportions of

male and female subjects in each group. Each group was randomly assigned a relative feedback

frequency condition of concurrent feedback, 100% feedback, or 50% feedback. Subjects in each

of the groups were then asked to perform an elbow extension task while the force exerted by

their muscles was recorded by an oscilloscope.105 The subjects were shown a graph of elbow-

extensor force and were asked to make their own elbow extensor force match that shown on the

graph. The concurrent feedback group was allowed to view a real-time illustration of their force

as it was graphed in relation to the desired force graph. This group was also allowed to view the

completed graph of each attempt for eight seconds following each attempt. The 100% feedback

group was not allowed to view the real time graphing but was allowed to see the completed

graph of each attempt after the completion of each attempt. Finally, the 50% feedback group

was only allowed to view the completed graph of each attempt after the completion of every

other attempt.

During the skill-acquisition phase of the research, each participant completed ten blocks

of ten attempts, each with a rest period of at least one minute between each block. Once all of

the skill-acquisition attempts were completed, an immediate retention test was administered,

wherein the subjects were required to perform the same task thirty additional times and all

groups received no feedback. An additional, delayed retention test of thirty attempts was

administered forty-eight hours after the immediate retention test.

105
An oscilloscope is an instrument which measures differences in electrical potentials. Often,
as was the case with this experiment, those differences are graphed on a cathode-ray screen with
the electrical potential differences plotted on the Y axis, as a function of time, which is plotted
on the X axis.
 41

The results showed that, during the skill-acquisition phase, subjects in the concurrent

feedback group exhibited significantly less deviation from the target force (error), than either the

100% or 50% feedback groups. However, in the immediate retention test the 50% feedback

group exhibited 58% less error than the concurrent feedback group and the 100% feedback group

exhibited 39% less error than the concurrent feedback group. The delayed retention test

produced similar results with the 50% feedback group exhibiting 52% less error than the

concurrent feedback group and the 100% feedback group exhibiting 26% less error than the

concurrent feedback group. Additionally, the 50% feedback group exhibited 31% less error than

the 100% feedback group on the immediate retention test, and 36% less error than the 100%

feedback group on the delayed retention test. Figure 2.1 below presents a graphic representation

of the results listed above. 106

Figure 2.1 Subject performance on acquisition-phase and retention trials

106
Vander Linden, 1993, 85.
 42

Root mean square error is represented on the Y axis and trials are represented on the X axis.

Results of the immediate and delayed retention tests are also represented directly to the right of

the trial results.

These results indicate that, while concurrent feedback may provide for immediate

improvements in performance ability during skill-acquisition, concurrent feedback is less

effective when relatively long-term changes in performance ability are desired. Additionally,

when comparing the frequency of feedback provided after the completion of the task, less

frequent feedback appears to be more effective at producing relatively long-term improvements

in performance ability. That is to say, less frequent feedback is more effective at facilitating

actual learning of a motor skill.

Feedback Frequency and the Acquisition of Voice Tasks

Effects of Practice With and Without Knowledge of Results on

Jitter and Shimmer Levels in Normally Speaking Women107

In her 1995 study, Carole Ferrand sought to determine if KR was in fact beneficial to the

acquisition of a vocal task, specifically the development of different levels of jitter (variations in

frequency) and shimmer (variations in amplitude).108 Ferrands study assigned 30, normally

107
Carole T. Ferrand, Effects of Practice With and Without Knowledge of Results on Jitter and
Shimmer Levels in Normally Speaking Women, Journal of Voice 9 (1995): 419-423.
108
In his book, Principles of Voice Production, Ingo R. Titze defines jitter and shimmer as short
term [meaning from one glottal pulse to the next] variability in fundamental frequency and
amplitude, respectively. In the 1960s, early investigators noticed that when they viewed speech
waveforms on an oscillograph, each period was slightly different from any other period. They
noted that the fundamental frequency appeared jittery, hence the term jitter. The term
shimmer was then coined as a partner term referring to fluctuations in amplitude (Iowa City,
 43

speaking women subjects into one of two groups. The first group received augmented aural and

visual feedback (referred to as the KR group) while the other group received no feedback during

practice sessions of a vocal task (referred to as the NKR group). The vocal task consisted of a

two-second prolongation of the vowel [a]. After a baseline reading, where both groups received

no feedback, all subjects underwent two practice sessions. Each practice session consisted of 15

trials of sustaining [a] for two seconds and the second practice session was held two days after

the first practice session. The KR group received feedback pertaining to their performance in

two ways. First they received visual feedback, which consisted of watching a real-time

waveform of their prolongation. Second, they received verbal feedback provided by the

instructor at the end of each prolongation. The NKR group received no feedback during or after

their prolongations. A final session, wherein no KR was provided, was held one week after the

subjects final practice session. The researchers refer to this as a transfer test. However, because

the subjects were performing the same test, delayed by some amount of time after the skill was

acquired, this test actually measured retention and not transfer.

The results of the study were split. Means of the shimmer levels did not change

significantly over the course of the study. Additionally, no significant difference between the

two groups occurred. Jitter values, on the other hand, were significantly lower (which is

desirable) in the KR group during practice sessions than in the NKR group. However, during the

transfer (retention) test, the KR groups means climbed to at or above baseline measures whereas

the NKR groups means climbed only to halfway between the measures from practice session

one and practice session two. These findings indicate that providing KR improved immediate

IA: NCVS, 2000), 313. Carole Ferrand indicates in her opening paragraph that jitter and
shimmer have become a widely used measure of the stability, and ostensibly health, of a
phonation.
 44

performance (indicated by the significantly lower practice session readings) but withholding KR

improved learning (indicated by the smaller increase in readings in the transfer [retention] test).

The Role of Knowledge of Results in Performance

and Learning of a Voice Motor Task109

In this study, researchers examined the effect of relative frequency of augmented external

feedback on the acquisition, retention, and transfer of a vowel nasalization task. 30 subjects, all

of whom were vocally untrained (meaning that they had received fewer than 5 formal singing or

speaking lessons) were randomly assigned to one of three groups receiving 100%, 50% or No

KR. KR was administered as a visual indication of percentage of nasalance as measured by a

Kay Elemetrics Nasometer. The skill acquisition task consisted of nasalizing a sustained [i]

vowel. Subjects were instructed to sustain the vowel [i] in a traditional oral phonation for two

seconds before switching to a nasalized [] for four seconds. To test transfer, subjects were

required to perform the same task using the vowel [a].

During the skill acquisition phase, subjects in the 100% KR group received immediate

feedback following each trial. Subjects in the 50% KR group received immediate feedback

following every other trial. Subjects in the No KR group were given no feedback regarding

accuracy. Each group performed five blocks of eight trials of the tasks outlined above.

Five minutes after completing the acquisition phase, retention was tested with the subject

performing two blocks of eight trials of the same task while none of the subjects received KR.

109
Kimberly Steinhauer and Judith Preston Grayhack, The Role of Knowledge of Results in
Performance and Learning of a Voice Motor Task, Journal of Voice 14, no. 2 (2000): 137-145.
 45

The transfer phase of the study consisted of two blocks of eight trials of the same task using the

vowel [a] and was conducted one day after the acquisition phase.

The results of the study showed a significant difference between the 100% KR groups

and the groups receiving 50% or no KR. The subjects in the less frequent KR groups performed

the nasalance tasks more accurately during the retention and transfer phases of the study meaning

that an increase in relative frequency of KR led to a decrease in both motor performance and

learning of a vowel nasalization task. This result is in keeping with similar research as noted

earlier. However, this study also found there to be no effect on immediate motor performance as

a result of increased relative frequency of KR. This finding is consistent with Winstein and

Schmidts110 findings in experiment one of their research reported earlier, but inconsistent with

Schmidts earlier hypotheses111 and Verdolini and Lees similar assertions112.

Steinhauer and Grayhack provide two possible explanations for the deviance of the

findings from the expected outcomes. One explanation is rooted in what they call the

dynamical theories of motor control as asserted by J. Kelso113 and J. Folkins.114 In the context

of these theories, the researchers contend that perhaps the nasalization of [i] was essentially too

110
Winstein and Schmidt, 1990.
111
Richard A. Schmidt, A Schema Theory of Discrete Motor Skill Learning, Psychological
Review 82 (1975): 225-260.
112
Katherine Verdolini and Timothy D. Lee. Optimizing Motor Learning in Speech
Interventions, In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine
M. Sapienza and Janina K Casper (Austin, TX: Pro-ed, 2004), 403-446.
113
JA Kelso and B. Tuller, A Dynamical Basis for Action Systems, in Handbook of Cognitive
Neuroscience, ed. M. Gazzaniza (New York: Plenum Press, 1984), 321-356.
114
JW Folkins, Issues in Speech Motor Control and Their Relation to the Speech of Individuals
with Cleft Palate, Cleft Palate Journal 22/2 (1985) 106-122.
 46

easy of a task, meaning that KR was useless. However, 100% KR was overwhelming to the

subjects, causing them to override their own, correct, sensory feedback.

The second explanation provided for the similar performances between the No KR and

the 50% KR groups comes in the form of several possible criticisms of the study design. They

contend that perhaps the sample size was too small, or possibly the subjects own auditory

feedback provided them with enough information to correctly perform the task. Finally, it is

possible that, even though the group assignments were randomized, more skilled subjects were

assigned to the No KR and 50% KR groups.

Effects of Feedback Frequency and Timing on Acquisition,

Retention, and Transfer of Speech Skills in

Acquired Apraxia of Speech115

Austermann Hula and her fellow researchers performed two separate studies examining

the acquisition and retention of speech skills in adult subjects with apraxia of speech116 in

relation to varied frequency or timing of feedback. The first study, using four adult subjects,

examined the effect of feedback frequency on retention and transfer of speech skills. A second

study, using two adult subjects, examined the effect of feedback timing on the same skills. The

research of both studies was intended to test the applicability of motor learning principles to the

115
Shannon N. Austermann Hula and others, Effects of Feedback Frequency and Timing on
Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech, Journal
of Speech, Language, and Hearing Research 51 (2008), 1088-1113.
116
Apraxia of speech is a motor speech disorder which causes the individual to have difficulty
physically producing the speech sounds that they have no difficulty planning consciously.
Individuals with apraxia of speech often present with slow, monotonic speech.
 47

practice of voice and speech therapy. Only the study regarding feedback frequency will be

reported here as it is the most applicable to the current research.

During this study, participants (three male and one female) were provided speech

therapy, in an attempt to aid the participants in regaining their abilities to produce non-word

syllables (such as [m] or [s]) as well as whole words. A single-subject, alternating-treatment

design was chosen for this study, meaning that the therapy for each subject alternated between a

high frequency feedback (HFF) condition, and a low frequency feedback (LFF) condition.

During the HFF condition, simple KR (correct or incorrect) was provided after 100% of the

subjects attempts to produce a syllable or word prescribed by the researcher. During the LFF

condition, the same simple KR was provided following only 60% of the subjects attempts. For

the attempts in which no feedback was provided, a two-second interval was observed between

each attempt.

HFF and LFF treatment phases were each four weeks long with a four-week maintenance

period in between. Following each HFF or LFF treatment phase, retention of the trained skills

was tested using ten repetitions of six trained non-word syllables. At the same period of time,

transfer of the trained skills to untrained word behaviors was tested using twelve related but

untrained non-word items, and six related but untrained real-word transfer items.

The study found that two of the four subjects responded with increased retention and

transfer of speech skills following the low frequency feedback phase, indicating that that

feedback condition had more effectively facilitated learning for those two subjects. The other

two subjects exhibited consistent retention and transfer of skills during both feedback phases,

indicating that feedback frequency had little effect on their learning. These findings lend some

support to the idea of applying principles of motor learning to the speech system but also raise
 48

more questions as to why feedback manipulation produced expected results in some subjects but

not in others. The authors account for this discrepancy by suggesting that because the individual

treatment plans for these two subjects required the practice and acquisition of relatively simpler

tasks than the tasks practiced by the other two subjects, their progress was expectedly more

consistent throughout the therapy process.

Conclusion

The three arm-positioning studies cited117 are only a small portion of a large body of

research that indicates a correlation between lower frequency of feedback and higher levels of

retention and/or transfer (i.e. learning), at least in relatively simple motor tasks. These studies

(particularly the Winstein and Schmidt research) are thorough in their scope and solid in their

design. The second set of studies reported118 sought to meet the difficult task of applying similar

uses of KR to the acquisition of the more complex motor tasks required for coordinated

phonation. While these studies did not produce as decisive results as the previous studies, their

results still indicate at least some positive interaction between less frequent KR and higher levels

of learning.

As noted at the onset of this chapter, the lack of research aimed at applying these same

principles to classical singing technique was the impetus for the current research. Ample

groundwork has been laid by the research in simple movement tasks. Similarly, that groundwork

has been adequately directed toward voice production tasks. It stands to reason that there is now

117
Carolee J. Winstein and Richard A. Schmidt, 1990; Timothy D. Lee, Margaret A. White and
Heather Carnahan, 1990; Darl W. Vander Linden, James H. Cauraugh and Tracy A. Greene,
1993.
118
Carole T. Ferrand, 1995; Kimberly Steinhauer and Judith Preston Grayhack, 2000; Shannon
N. Austermann Hula and others, 2008.
 49

justification to further this body of research by testing the application of reduced-frequency

feedback to realm of classical singing voice instruction.

 50

CHAPTER 3

METHODOLOGY

Among teachers of singing, there are as many different teaching styles as there are

teachers. However, it has been observed that there are several common threads in terms of

teaching techniques that are apparent in most voice studios. Frequency of feedback is one of

those common threads. Many teachers are anxious to give immediate feedback to the student

following every attempt to perform a given task. This occurs, understandably, for many reasons.

First, instructors mistakenly look to the immediate improvement of performance as evidence of

effective teaching practices. Second, students expect return for their investment. They are

happier after a lesson when they feel that their performance has made significant improvement.

Unfortunately, this elation wears off quickly when they are unable to repeat the performance a

few hours later in the practice room.

Hypothesis

The study outlined in the following pages tested the hypothesis that decreasing the

frequency of feedback in voice lessons, while possibly depressing immediate performance,

actually fosters learning, resulting in increased retention (ability to perform the task after a

period of time with no instruction) and transfer (ability to perform a similar, but previously

untrained task) of voice performance skills.

 51

Theoretical Framework

The hypothesis stated above stems directly from Schmidts schema theory119 and its

application to clinical voice therapy and studio voice by Verdolini and Lee, both of which were

referenced earlier. These references claim that augmented feedback (knowledge of results) is

necessary for motor learning. This feedback can vary in a number of ways including frequency

and timing (the current study focused only on variations in frequency, though a future study will

likely consider variation of timing). It is also noted in the literature that cognitive effort, in the

form of hypothesis generation and evaluation, is essential for learning, though conscious trying

may actually be detrimental to learning.120 If it is assumed that reducing frequency of feedback

can increase the cognitive effort of the student by requiring them to create and test their own

hypotheses, it can be interpolated that reducing feedback frequency will thereby increase

learning.

Subjects

8 subjects (5 females and 3 males), ranging in age from 18 to 25 years old, participated in

this study. Subjects were recruited from the researchers undergraduate applied voice studio at

Monmouth College (Monmouth, IL). All subjects were clearly informed of what would be

required of them with their participation in the study, including that increased cognitive attention

would be required throughout the study. They were also instructed to actively weigh their own

performance against the desired outcomes and to try new methods of achieving the desired

119
Richard A. Schmidt, A Schema Theory of Discrete Motor Skill Learning, Psychological
Review 82 (1975): 225-260.
120
Katherine Verdolini and Timothy D. Lee. Optimizing Motor Learning in Speech
Interventions, In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine
M. Sapienza and Janina K. Casper (Austin, TX: Pro-ed, 2004), 403-446.
 52

outcomes during times that they noticed lower amounts of feedback. However, subjects were not

informed concerning when feedback frequency levels would be changed. As evidence that the

participants understood this instruction, they were invited to sign an informed consent

document.121

Once informed consent was obtained, subjects were asked to complete a subject intake

form,122 which included information regarding previous singing experience and history of vocal

health. Subjects indicating a history of vocal pathologies or chronic dysphonia would have been

excused from the study. However, none of the subjects indicated a history of any such excluding

factors.

The design of the study (discussed further in the following section), which allowed each

subject to act as his or her own control, meant that students with widely varied levels of ability

could participate in the study without posing difficulties in measuring results. However, there

were a few exclusionary factors that were taken into account during the recruitment of subjects.

First, because of the difficulty and often slow progress of teaching a student to match pitches, it

was required that candidates already be able to match pitch before they were admitted to the

study. Second, because the study calls for a capella singing (singing without accompaniment),

subjects were required to have sufficient tonal memory as to be able to perform America the

Beautiful without piano accompaniment.

These abilities were judged by the researcher during the first meeting, immediately

following the consent process. In regard to matching pitch, the researcher played a pitch on the

piano and asked the subject to sing that same pitch. This process was repeated five times with

121
See appendix A for the IRB-approved informed consent document.
122
See appendix B for the subject intake form.
 53

various pitches across the subjects vocal pitch range. In regard to tonal memory, the subject

was asked to sing America the Beautiful a capella. Only subjects who were able to match

correctly 4 of the 5 pitches and who were able to perform America the Beautiful with fewer

than 5 wrong notes were admitted to the study.

A total of 8 students met the above criteria and were admitted to the study. Background

information from each of their subject intake forms is reported in figure 3-1.

Retention/
Years of Vocal Health
Subject # Age Gender Transfer
Experience Issues
test key
Frequent strep
1 19 Female 5 C
throat
2 19 Female 5 None Bb

3 19 Male 1 None Bb

4 20 Female 1 None Bb
Frequent
5 19 Female 1 C
laryngitis
6 21 Male 6 None Bb

7 18 Male .5 None C

8 25 Female 1 None G
Table 3.1 - Demographics of eight study participants

Instrumentation

Instrumentation for this study was relatively simple. A piano and music stand were

necessary for the lessons during the skill acquisition phases. Audio from the subjects

retention/transfer tests were recorded using a Tascam DR-1 digital audio recorder producing a
 54

256 Kbps mp3 file and sampling at 44.1 kHz. These audio files were recorded to compact disc

and distributed to the volunteer auditors.

Skill Acquisition Phase

A single-subject, alternating-treatment design, similar to that of the research done by

Austermann Hula,123 was chosen for this study because of the flexibility it allowed in adjusting

and accounting for individual differences.124 The duration of the study was 15 weeks with

subjects receiving 1 thirty-minute voice lesson per week. These 15 lessons were divided into 3

blocks of 4 lessons each and a final block of 3 lessons.125 With each block, instruction technique

alternated between high-frequency feedback (HFF) and low-frequency feedback (LFF).

For this study, each subject was quasi-randomly assigned126 to one of two feedback

groups: g-1 and g-2. Students in g-1 began with their first block being instructed with HFF.

123
Shannon N. Austermann Hula and others, Effects of Feedback Frequency and Timing on
Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech, Journal
of Speech, Language, and Hearing Research 51 (2008): 1088-1113.
124
Kevin P. Kearns, Flexibility of Single-Subject Experimental Designs Part II: Design
Selection and Arrangement of Experimental Phases, Journal of Speech and Hearing Disorders
51 (August 1986): 211.
125
A study duration of 15 weeks was chosen to fit within the 15 week semester at Monmouth
College, at which the experiment was conducted.
126
At the beginning of the semester, subjects were assigned a lesson time based on their
availability. Once the lesson schedule was determined, the first four lessons of the day were
assigned to g-1 and the last four students were assigned to g-2. In the strictest sense, this was not
a random assignment of groups. However, since the original lesson schedule was determined by
the subjects availability, assignment to g-1 or g-2 was mostly out of the control of the
researcher. This design was decided upon because it allowed the researcher to switch feedback
frequency only once during the day, ideally leading to a more consistent approach to feedback
frequency than if he were required to switch frequency with each subject. Throughout the course
of the semester, two subjects (subjects 4 and 8) required a change in lesson time. In these cases,
the subjects retained their group assignment, maintaining the same feedback frequency schedule.
 55

Alternately, students in g-2 began with their first block being instructed with LFF. During

lessons in the HFF blocks, instruction took place in a similar fashion to most voice studios; each

attempt by the subject to perform the given task was followed by some level of KR and/or KP

from the instructor, resulting in a 100% feedback condition. KR was provided in the form of a

report of the result of the attempt relative to the desired outcome. If the attempt matched the

desired outcome, positive KR was provided by the researcher saying yes or good. If the

attempt did not match the desired outcome, negative KR was occasionally provided by the

researcher saying, no or not quite. More frequently, however, if the attempt did not match

the desired outcome, more informative KP was provided by the researcher. KP included

guidance pertaining to how to improve performance of the task in order to improve the chances

of the subsequent attempts more closely approximating the desired outcome. Again, the same

type and amount of KP or KR was provided regardless of the skill that was being attempted.

During lessons in the LFF blocks, students were asked to perform the given task twice or

three times before further feedback/guidance was provided by the instructor, resulting in a 33%

to 66% feedback condition. This feedback was provided in the same combination of KR and/or

KP as it was provided during the HFF blocks.

In each lesson, both in HFF and LFF blocks, time was equally divided between working

on vocalization exercises and vocal literature in hopes of replicating the conditions in many

collegiate voice studios. Vocalization exercises were varied but all were aimed at improving the

subjects tone quality, breath management, and/or intonation.127 KP, therefore, was designed to

guide the subject toward meeting the goals of the individual exercises as they related to those

127
See appendix C for sample vocal exercises for each quality.
 56

qualities. When working on vocal literature, the same qualities were the focus of the KP.

Accordingly, auditors of the retention/transfer tests were asked to assess these same qualities.

It was expected that variations in amounts of time that each subject spent practicing

singing between lessons would have a significant impact on his or her ability to perform the

retention/transfer test. To help minimize the impact, subjects were instructed to perform at least

two hours of practice throughout the week between lessons. It was also expected that some

subjects would spend more than the required time, while others may still spend less time. To

help account for variations on performance ability that may have resulted from variations in

practice time, subjects were required to complete a weekly practice record. To minimize the risk

that their knowledge of my expectations may lead them to indicate falsely more or less practice

time on their practice log, actual logs were not viewed by the researcher until after the

completion of the study.

Retention/Transfer Tests

Subjects were required to perform an identical vocal task at the beginning of each lesson

to test retention from the previous weeks lesson. The vocal task chosen for this study was

singing the first verse of America the Beautiful without piano accompaniment. Subjects with

lower voices (i.e. mezzo-sopranos and baritones) performed the song in the key of Bb major,

while those with higher voices (i.e. sopranos and tenors) performed the song in the key of C

major. One subject (#8) presented with a particularly low singing voice and had considerable

difficulty transitioning into head register. For this reason, this subject performed the song in the

unusually low key of G major. A more thorough discussion of the results from this subject will

be discussed along with the rest of the subjects in chapters four and five.
 57

Because America the Beautiful was never trained during the skill acquisition phases,

this task measured skill transfer as well as retention. Thus, retention and transfer of vocal

technique, as demonstrated by a performance of the vocal task, temporally separated from

instruction, served as the measured variable, while frequency of feedback served as the

manipulated variable. Subjects were recorded performing America the Beautiful to serve as a

baseline of the subjects vocal qualities.

An audio recording of each performance (15 recordings for each of 8 subjects) of

America the Beautiful was collected and catalogued for later blinded analysis. At the

conclusion of the 15-week study period, audio files were recorded to a CD in a random order

(created using an online random sequence generator)128 such that recordings of individual

subjects were not placed consecutively on the playlist, and such that recordings of individual

subjects were not placed in chronological order on the playlist. Copies of this CD were then sent

to five college and university faculty members who served as voluntary auditors. These auditors

were asked to evaluate each audio file using the perceptual analysis evaluation scale included in

appendix D. This evaluation scale was developed using a small portion of a form developed by

Dr. Katherine Eberle-Fink for assessment of singing129 and it asked auditors to assess each

performance in terms of tone quality, breath management, and intonation. For the purposes of

this study tone quality was defined as the tonal and acoustic properties of the voice during the

performance (e.g. bright versus dark); breath management was defined as the ability of the

128
Random Sequence Generator, available at http://www.random.org/sequences/, accessed
12/6/2010.
129
Katherine Eberle-Fink, Perceptual Acoustic Assessment of Singing, Journal of Singing -
The Official Journal of the National Association of Teachers of Singing 63, no. 1 (September-
October 2006): 35-43.
 58

subject to perform the excerpt with sufficient air to maintain the quality of their tone throughout

individual phrases; and intonation was defined as the ability of the subject to perform the excerpt

in such a manner that the auditors perceived individual notes to match the correct pitches of the

excerpt. Auditors were selected by the researcher, with attention to their experience in

perceptually evaluating vocal performances.

A few key elements of the study design should be kept in mind while reading the results

of the current research. First, it should be noted that, while individual tests were only evaluating

performance, data from the entire study presented trends in retention and transfer levels that can

be interpreted as indications of overall learning. It is also important to bear in mind that, due to

the single-subject design of the proposed study, individual subjects were only evaluated against

their own previous performances and not against the performances of other subjects. Finally,

because of the limited number of subjects, statistically-significant trends were not determined.

Expected Results

Due to the fact that increased practice will, in all likelihood, lead to some increase in

ability, it was expected that performance levels would increase throughout the duration of the

study. However, if the hypothesis that lower frequency feedback will increase retention and

transfer of vocal skills were to hold true, it was expected that assessment values of the

retention/transfer tests would trend up more sharply during LFF blocks than in HFF blocks. The

actual results will be discussed in the following chapter.

 59

CHAPTER 4

REPORTING OF RESULTS

Upon completion of the data collection phase of this study, the audio recordings of each

subjects retention/transfer tests were sent to five volunteer auditors as explained in chapter 3.

Once the audio recordings had been evaluated by the volunteer auditors, the quantitative data

produced by those evaluations were compiled and the results are reported here. A more thorough

discussion of the interpretation of those results will be presented in the following chapter.

For each subject, the scores from the auditors assessments were plotted on three graphs:

one for each of the vocal characteristics that were assessed (tone quality, breath management,

and intonation. Each of these graphs plotted all auditors scores for each week as well as an

average of those scores.130 The average of these scores evens out the discrepancies between

auditors and presents a more constant view of the subjects progression throughout the course of

the study.

In order to more clearly see any correlation between a subjects scores and the varied

level of feedback, the line of average scores were divided into four segments, one for each phase

of the study in which feedback frequency was varied. Even though each phase was four weeks

in duration (with the exception of the final phase, which was only three weeks in duration), the

data for each phase consisted of five scores. This is because the scores represent the subjects

performance on retention tests delayed by one week. For example, a subject started the study by

producing a baseline performance of the test. That subject then received a lesson with either

high or low frequency feedback. At the beginning of the next meeting (week two), the subject

130
It appears that the five auditors were frequently in agreement as to whether a voice was
improving or declining, though the amplitude of that change varied from auditor to auditor.
 60

produced another performance of the test. This performance shows the level of retention of the

skills addressed at the previous lesson. Similarly, the tests from weeks three, four, and five (even

though instruction in week five will switch to the opposite feedback frequency condition) will

show the level of retention of the skills addressed in the lessons of weeks two, three, and four,

respectively. Thus, phase one consists of five data points.

The reporting of phase two will begin with the results of the test performed on week five,

which can been said to serve as a baseline for the subjects level of performance ability at the

beginning of the phase. A similar process will be followed for phases three and four. However,

there are only three data points for the results of phase four. This is due to the fact that, because

of the scheduling of the academic institution at which this research was conducted, it was not

possible to collect results of a final retention test one week following the final lesson.

The average auditors scores for each phase are plotted on separate graphs along with a

trend line (line of best fit) for those data points. That trend line serves to exhibit the overall

progression of the subjects performance ability throughout that phase. A comparison of the

slopes of the trend lines (negative or positive), which we can think of as learning trends, from

each phase can be used to determine any correlation between feedback frequency rate and an

individuals performance ability.

The results for each of the eight subjects will be reported in the following pages. Each

subject will have an introduction to the background of the subject, taken from the subject intake

form completed by each subject before the study began and including information from the

subjects practice record, completed throughout the duration of the study. Following the

introduction, results from each of the three voice characteristics (tone quality, breath
 61

management, and intonation) will be presented. For each characteristic, three figures will

summarize the findings for that subjects performance of that characteristic.

Subject One

Subject one was 19-year old female who had been studying college level voice with the

researcher for one year prior to the beginning of the study period. She also had studied privately

with her choir teacher throughout her four years of high school. In addition to her experience

studying private voice, she also reported having sung in choirs for five years. On her intake

form, she indicated that she had had strep throat more than twice in a six-month span. Though

she did contract a mild cold during the middle of the study period, the illness did not prevent her

from taking part in the study at any point.

Subject one was assigned to group g-1, meaning that she received low-frequency

feedback during weeks one through four, high-frequency feedback during weeks five through

eight, low-frequency feedback during weeks nine through twelve, and high-frequency feedback

during weeks thirteen through fifteen. She did not appear for her lesson on week nine, the

absence of data for which is exhibited in the results for this subject.

Over the course of the study, subject one practiced an average of one hour per week.

However, her levels of practice were higher during the beginning phases than in later phases.

Subject one averaged 1.4, and 1.1 hours/week of practice during phases one and two

respectively. That practice dropped to 0.8 and 0.7 hours/week of practice during phases three

and four respectively. These levels of practice may bear some weight on the amount of progress

made during the latter phases.

 62

Subject 1 Tone Quality

The overall results for the tone quality of subject one are reported below in table 4.1 and

those results are graphed in figure 4.1

Subject One Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 62 50 50 45 50 51.4
2 65 20 45 35 40 41
LFF

3 54 30 50 20 30 36.8
4 62 50 40 20 50 44.4
5 52 40 60 25 50 45.4
6 58 40 50 40 50 47.6
HFF

7 58 40 50 40 50 47.6
8 66 35 65 30 40 47.2
9
10 70 60 50 60 50 58
LFF

11 54 35 40 35 30 38.8
12 58 30 55 25 50 43.6
13 60 20 45 47 40 42.4
HFF

14 62 30 55 40 40 45.4
15 60 40 60 25 60 49
Table 4.1 Subject one tone quality results

Subject One Tone Quality

100
90
Auditors' Scores /100

80 Aud. 1
70
60 Aud. 2
50 Aud. 3
40
30 Aud. 4
20 Aud. 5
10
Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Weeks
Figure 4.1 Subject one tone quality results
 63

Figure 4.2 shows the average of the scores for tone quality from all five auditors divided

into low and high frequency feedback phases. As discussed above, a linear trend line was added

to each graph to show the overall trend throughout that phase. These trend lines were created

using linear regression of the points on the graph and the equation for each trend line is listed in

the legend in y = mx + b where m is the slope of the line and b is the point at which the line

intercepts the y axis.

S. 1 Phase 1 (LFF) S. 1 Phase 2 (HFF)

60 60

55 55
Auditors' scores / 100

Auditors' score /100

50
50
45
Average 45 Average
40
Linear Linear
40 (Average)
35 (Average)
y = -0.86x + 46.38 y = 0.54x + 45.6
30 35
1 2 3 4 5 5 6 7 8 9
Week Week

S. 1 Phase 3 (LFF) S. 1 Phase 4 (HFF)

65
60
60
Auditors' score /100

55
Auditors' score /100

55
50 50
45 Average
45 Average
40 Linear Linear
(Average) 40
35 (Average)
y = -4.2x + 60.4 y = 3.3x + 39
30 35
9 10 11 12 13 13 14 15
Week Week

Figure 4.2 Subject one average tone quality results by phase

 64

Subject One Breath Management

The overall results of the breath management for subject one are listed below in the chart

in table 4.2 and graphed in figure 4.3.

Subject One Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 60 40 40 65 25 46
2 74 20 45 40 40 43.8
HFF

3 70 25 50 75 50 54
4 60 30 30 60 40 44
5 60 35 50 80 40 53
6 68 35 50 70 25 49.6
LFF

7 58 30 45 50 25 41.6
8 62 30 75 60 25 50.4
9
10 60 40 35 70 20 45
LFF

11 64 20 35 70 25 42.8
12 50 5 50 80 35 44
13 58 20 35 60 40 42.6
HFF

14 65 20 45 50 25 41
15 50 20 60 75 40 49
Table 4.2 Subject one breath management results

Subject One Breath Management

100
90
80
Auditors' Scores /100

70 Aud. 1
60 Aud. 2
50
Aud. 3
40
30 Aud. 4
20 Aud. 5
10 Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Weeks
Figure 4.3 Subject one breath management results
 65

Figure 4.4 shows the average of the scores for breath management from all five auditors

including linear trend lines. The results for the breath management of subject one are mixed

with positive slopes phases one and four, but negative slopes in phases two and three.

S. 1 Phase 1 (LFF) S. 1 Phase 2 (HFF)

60
60
55
55

Auditors' score /100

Auditors' scor /100

50 50

45 45
Average Average
40 40
Linear Linear
35 (Average) 35 (Average)
y = 1.42x + 43.9 y = -1.58x + 52.6
30 30
1 2 3 4 5 5 6 7 8 9
Week Week

S.1 Phase 3 S. 1 Phase 4 (HFF)

60
60

55
Auditors' Score /100

55
Auditors' Score /100

50 50

45 Average Average
45
Linear Linear
40 40
(Average) (Average)
y = -0.6x + 45.7 y = 3.2x + 37.8
35 35
9 10 11 12 13 13 14 15
Week Week

Figure 4.4 Subject one average breath management results by phase

 66

Subject One Intonation

The overall results for intonation for subject one presented are in the chart in table 4.3

and are graphed in figure 4.5.

Subject One Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 45 5 10 25 25 22
2 35 0 10 40 25 22
LFF

3 36 0 25 60 25 29.2
4 40 0 15 20 25 20
5 28 5 15 10 10 13.6
6 30 0 30 30 15 21
HFF

7 42 10 30 60 50 38.4
8 60 10 70 75 60 55
9
10 38 5 20 40 40 28.6
LFF

11 36 0 15 60 25 27.2
12 40 0 35 40 25 28
13 46 0 20 55 10 26.2
HFF

14 40 0 25 40 10 23
15 42 5 35 60 50 38.4
Table 4.3 Subject one intonation results

Subject One Intonation

100
90
80
Auditors' Scores /100

Aud. 1
70
60 Aud. 2
50 Aud. 3
40
Aud. 4
30
20 Aud. 5
10 Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Weeks

Figure 4.5 Subject one intonation results

 67

Figure 4.6 shows the average of the scores for intonation from all five auditors including

trend lines. The results for intonation returned to the trends exhibited by the results for tone

quality, with the LFF instruction condition appearing to hinder intonation. During the LFF

phases (1 and 3) the scores for intonation declined with slopes of -1.8 and -0.64 respectively.

During the HFF phases (2 and 4) the scores increased with slopes of 14.16 and 6.1 respectively.

S. 1 Phase 1 (LFF) S. 1 Phase 2 (HFF)

65
35 60
55

Auditors' Scores /100

30
Auditors' Scores /100

50
25
45
20 40
Average 35
15 Average
30
10 25 Linear
Linear
20 (Average)
5 (Average)
15
y = -1.88x + 27 y = 14.16x - 3.4
0 10
1 2 3 4 5 5 6 7 8 9
Week Week

S. 1 Phase 3 (LFF) S. 1 Phase 4 (HFF)

35 45

40
Auditors' Score /100
Auditors' Score /100

30
35
Average
25 30
Average Linear
25 (Average)
20 Linear
(Average) 20
y = 6.1x + 17
y = -0.64x + 29.74 15
15 13 14 15
9 10 11 12 13
Week
Week

Figure 4.6 Subject one average intonation results by phase

 68

Subject Two

Subject two was a 19 year old female who had studied voice with the researcher for one

year prior to the start of the study. Additionally, she reported having studied private voice with

her high school choir teacher for four years prior to college. In addition to her private voice

instruction, subject two reported having participated in choirs for her four years in high school as

well as one year at college. Subject two reported having no record of vocal health issues. Nor

did she indicate having had chronic loss of voice. Health related issues did not present any

problems during her participation in the study.

Subject two was assigned to g-1, meaning that she received low-frequency feedback

during weeks one through four, high-frequency feedback during weeks five through eight, low-

frequency feedback during weeks nine through twelve, and high-frequency feedback during

weeks thirteen through fifteen. She did not appear for her lesson on week three, the absence of

data for which is exhibited in the results for this subject.

Over the course of the study, subject two reported having practiced an average of 3.5

hours/week. Unlike subject one, subject twos practice was more uniform throughout the term

with 3.75, and 3.25 hours/week during phases one and two respectively, and 3.75 and 3.0

hours/week during phases three and four respectively.

Subject Two Tone Quality

The aggregate scores for the tone quality of subject two from all five auditors are

reported in the chart in table 4.4 and in the graph in figure 4.7.
 69

Subject Two Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 76 40 60 75 50 60.2
2 62 30 55 75 40 52.4
HFF

3
4 62 20 75 75 50 56.4
5 60 25 50 40 40 43
6 60 35 80 35 40 50
LFF

7 70 60 55 60 25 54
8 64 65 55 25 40 49.8
9 64 50 55 60 50 55.8
10 58 50 65 40 50 52.6
HFF

11 58 25 50 40 40 42.6
12 60 30 50 25 60 45
13 58 50 45 60 40 50.6
LFF

14 62 35 65 50 60 54.4
15 52 20 55 40 30 39.4
Table 4.4 Subject two tone quality results

Subject Two Tone Quality

100
90
80 Aud. 1
Auditors' Score /100

70 Aud. 2
60
Aud. 3
50
40 Aud. 4
30 Aud. 5
20 Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.7 Subject two tone quality results
 70

Figure 4.8 below presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 2 Phase 1 (LFF) S. 2 Phase 2 (HFF)

70 65
60 60
Auditors' Score /100

Auditors' Score /100

50
55
40
Average 50 Average
30
45
20 Linear Linear
(Average) 40 (Average)
10
y = -3.04x + 62.12 y = 2.54x + 42.9
0 35
1 2 3 4 5 5 6 7 8 9
Week Week

S. 2 Phase 3 (LFF) S. 2 Phase 4 (HFF)

65 60

60 50
Auditors' Score /100
Auditors' Score /100

55 40

50 30
Average
Average
45 20
Linear
Linear
40 10 (Average)
(Average)
y = -5.6x + 59.333
y = -1.8x + 54.72 0
35
9 10 11 12 13 13 14 15
Week Week

Figure 4.8 Subject two average tone quality results by phase

 71

Subject Two Breath Management

The aggregate scores for the breath management of subject two from all five auditors are

presented in the chart in table 4.5 and in the graph in figure 4.9.

Subject Two Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 80 40 45 80 50 59
2 68 20 50 90 40 53.6
LFF

3
4 70 15 75 95 50 61
5 70 15 60 75 50 54
6 70 25 75 75 50 59
HFF

7 70 50 50 80 50 60
8 70 50 60 65 50 59
9 75 40 50 85 60 62
10 60 25 60 80 50 55
LFF

11 76 10 55 75 50 53.2
12 64 20 50 90 40 52.8
13 60 40 55 98 50 60.6
HFF

14 66 20 65 90 50 58.2
15 58 20 45 65 40 45.6
Table 4.5 Subject two breath management results

Subject Two Breath Management

100
90
80
Auditors' Score /100

Aud. 1
70
60 Aud. 2
50 Aud. 3
40 Aud. 4
30
Aud. 5
20
10 Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.9 Subject two breath management results

 72

Figure 4.10 presents the average of the scores for breath management from all five

auditors, divided into phases and including a linear trend line for each phase.

S. 2 Phase 1 (LFF) S. 2 Phase 2 (HFF)

70
70

65
Auditors' Score /100

65

Auditors' Score /100

60 60

55 Average Average
55
Linear Linear
50 (Average) 50 (Average)
y = -0.26x + 57.68 y = 1.6x + 54
45 45
1 2 3 4 5 5 6 7 8 9
Week Week

S. 2 Phase 3 (LFF) S. 2 Phase 4 (HFF)

70 70

65 65
Auditors' Score /100

Auditors' Score /100

60 60
Average Average
55 55
Linear Linear
50 (Average) 50 (Average)
y = -0.5x + 58.22
y = -7.5x + 69.8
45 45
9 10 11 12 13 13 14 15
Week Week

Figure 4.10 Subject two average breath management results by phase

 73

Subject Two Intonation

The aggregate scores for the intonation of subject two from all five auditors are presented

in the chart in table 4.6 and in the graph in figure 4.11.

Subject Two Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 82 40 70 95 85 74.4
2 78 40 75 85 75 70.6
LFF

3
4 75 15 85 90 80 69
5 78 10 80 90 75 66.6
6 62 30 75 95 80 68.4
HFF

7 70 40 45 90 85 66
8 66 60 85 90 80 76.2
9 78 40 75 90 75 71.6
10 78 15 90 80 60 64.6
LFF

11 80 10 72 80 75 63.4
12 78 20 75 98 85 71.2
13 70 45 75 90 85 73
HFF

14 75 40 90 80 85 74
15 68 20 80 90 85 68.6
Table 4.6 Subject two intonation results

Subject Two Intonation

100
90
80 Aud. 1
Auditors' Score /100

70
Aud. 2
60
Aud. 3
50
40 Aud. 4
30 Aud. 5
20
Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.11 Subject two intonation results
 74

Figure 4.12 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 2 Phase 1 (LFF) S. 2 Phase 2 (HFF)

80 80

75 75
Auditors' Score /100

Auditors' Score /100

70 70

65 Average 65 Average

Linear Linear
60 (Average) 60
(Average)
y = -1.72x + 75.31 y = 1.78x + 64.42
55 55
1 2 3 4 5 5 6 7 8 9
Week Week

S. 2 Phase 3 (LFF) S. 2 Phase 4 (HFF)

80
80
75
Auditors' Score /100

Auditors' Score /100

75
70
70 Average
65 Average

Linear 65 Linear
60 (Average) (Average)
y = 0.94x + 65.94 y = -2.2x + 76.267
55 60
9 10 11 12 13 13 14 15
Week Week

Figure 4.12 Subject two average intonation results by phase

 75

Subject Three

Subject three was a 19 year old male who had studied private voice with the researcher

for one year prior to the start of the study. He reported having never studied private voice prior

to studying with the researcher. However, subject three reported that he had participated in

several high school and community choirs over the previous six years. Subject three reported no

previous vocal health issues or chronic loss of voice. Health related issues did not present any

problems throughout his participation in the study.

Subject three was assigned to g-1, meaning that he received low-frequency feedback

during weeks one through four, high-frequency feedback during weeks five through eight, low-

frequency feedback during weeks nine through twelve, and high-frequency feedback during

weeks thirteen through fifteen. He appeared for all of his lessons throughout the term, leaving no

gaps in the data.

Throughout the course of the study, subject three reported having practiced an average of

1 hour/week. Similar to subject two, subject threes practice time was relatively consistent

throughout the duration of the study. He reported an average of 1.1 and 1.0 hours/week during

phases one and two respectively, and 1.2 and 0.75 hours/week during phases three and four

respectively.

Subject Three Tone Quality

The aggregate scores for the tone quality of subject three from all five auditors are

presented in the chart in table 4.7 and in the graph in figure 4.13.
 76

Subject Three Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 58 35 50 40 50 46.6
2 52 25 45 25 50 39.4
LFF

3 58 25 35 40 50 41.6
4 56 20 40 10 50 35.2
5 62 35 40 20 50 41.4
6 58 40 60 60 60 55.6
HFF

7 64 25 50 30 50 43.8
8 60 30 55 20 50 43
9 56 30 30 20 50 37.2
10 64 30 45 15 50 40.8
LFF

11 53 10 60 20 60 40.6
12 53 20 40 40 50 40.6
13 50 35 40 30 50 41
HFF

14 56 30 55 25 60 45.2
15 60 30 55 10 50 41
Table 4.7 Subject three tone quality results

Subject Three Tone Quality

70

60
Aud. 1
Auditors' Score /100

50
Aud. 2
40
Aud. 3
30 Aud. 4
20 Aud. 5

10 Average

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.13 Subject three tone quality results
 77

Figure 4.14 presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 3 Phase 1 (LFF) S. 3 Phase 2 (HFF)

60 60

55 55
Auditors' Score /100

Auditors' Score /100

50 50

45 45
Average Average
40 40
Linear Linear
35 (Average) 35 (Average)
y = -1.46x + 45.22 y = -2.1x + 50.5
30 30
1 2 3 4 5 5 6 7 8 9
Week Week

S. 3 Phase 3 (LFF) S. 3 Phase 4 (HFF)

55 55
Auditors' Score /100

Auditors' Score /100

50 50

45 45
Average Average

40 Linear 40 Linear
(Average) (Average)
y = 0.74x + 37.82 y = 0.0x + 42.4
35 35
9 10 11 12 13 13 14 15
Week Week

Figure 4.14 Subject three average tone quality results by phase

 78

Subject Three Breath Management

The aggregate scores for the breath management of subject three from all five auditors

are presented in the chart in table 4.8 and in the graph in figure 4.15.

Subject Three Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 62 25 30 75 25 43.4
2 48 10 50 90 25 44.6
LFF

3 52 5 45 60 40 40.4
4 54 15 35 40 20 32.8
5 65 20 40 60 25 42
6 56 15 60 80 40 50.2
HFF

7 50 20 40 40 25 35
8 52 10 60 60 25 41.4
9 48 20 45 75 40 45.6
10 64 10 40 50 25 37.8
LFF

11 65 10 55 50 25 41
12 64 20 30 65 20 39.8
13 60 25 20 65 25 39
HFF

14 62 20 50 75 50 51.4
15 52 20 40 75 40 45.4
Table 4.8 Subject three breath management results

Subject Three Breath Management

100
90
80 Aud. 1
Auditors' Score /100

70
Aud. 2
60
50 Aud. 3
40 Aud. 4
30 Aud. 5
20
Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.15 Subject three breath management results

 79

Figure 4.16 presents the average of the scores for breath management from all five

auditors, divided into phases and including a linear trend line for each phase.

S. 3 Phase 1 (LFF) S. 3 Phase 2 (HFF)

50 55

50
45

Auditors' Score /100

Auditors' Score /100

45
40
40
Average Average
35
35
Linear Linear
30 (Average) 30 (Average)
y = -1.46x + 45.02 y = -0.16x + 43.32
25 25
1 2 3 4 5 5 6 7 8 9
Week Week

S. 3 Phase 3 (LFF) S. 3 Phase 4 (HFF)

55 55

50 50
Auditors' Score /100

Auditors' Score /100

45 45

Average Average
40 40
Linear Linear
35 (Average) 35 (Average)
y = -1.12x + 44 y = 3.2x + 38.867
30 30
9 10 11 12 13 13 14 15
Week Week

Figure 4.16 Subject three average breath management results by phase

 80

Subject Three Intonation

The aggregate scores for the intonation of subject three from all five auditors are

presented in the chart in table 4.9 and in the graph in figure 4.17.

Subject Three Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 25 0 10 30 10 15
2 44 5 40 50 25 32.8
LFF

3 48 0 30 30 25 26.6
4 28 0 20 10 10 13.6
5 38 5 45 60 50 39.6
6 42 0 50 40 25 31.4
HFF

7 55 5 75 50 60 49
8 46 5 70 60 40 44.2
9 38 5 25 30 50 29.6
10 49 0 55 30 50 36.8
LFF

11 44 0 50 40 50 36.8
12 28 0 15 35 10 17.6
13 38 0 10 35 10 18.6
HFF

14 42 5 60 50 50 41.4
15 45 5 45 40 50 37
Table 4.9 Subject three intonation results

Subject Three Intonation

80
70
60 Aud. 1
Auditors' Score /100

50 Aud. 2

40 Aud. 3

30 Aud. 4

20 Aud. 5

10 Average

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.17 Subject three intonation results
 81

Figure 4.18 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 3 Phase 1 (LFF) S. 3 Phase 2 (HFF)

45 55

40 50

Auditors' Score /100

Auditors' Score /100

35
45
30
40
25 Average Average
35
20
Linear Linear
15 (Average) 30 (Average)
y = 3x + 16.52 y = -0.72x + 40.92
10 25
1 2 3 4 5 5 6 7 8 9
Week Week

S. 3 Phase 3 (LFF) S. 3 Phase 4 (HFF)

40 45

35 40
Auditors' Score /100

Auidtors' Score /100

30 35

25 30
Average
Average
20 25
Linear Linear
15 (Average) 20 (Average)
y = -4.12x + 40.24 y = 9.2x + 13.933
10 15
9 10 11 12 13 13 14 15
Week Week

Figure 4.18 Subject three avearge intonation results by phase

 82

Subject Four

Subject four was a 20 year old female who had studied private voice with the researcher

for one semester prior to the onset of the study. Prior to that point, she had not studied private

voice with any other instructors. However, subject four reported having participated in choir for

one year during high school and for one year at college. She reported having no vocal health

issues or chronic loss of voice. Health related issues did not present any problem for her

participation in the study.

Subject four was assigned to g-1, meaning that she received low-frequency feedback

during weeks one through four, high-frequency feedback during weeks five through eight, low-

frequency feedback during weeks nine through twelve, and high-frequency feedback during

weeks thirteen through fifteen. She did not appear for her lesson on week fourteen, the absence

of data for which is exhibited in the results for this subject.

Throughout the course of the study, subject four reported having practiced an average of

5.4 hours/week. While subject fours practice time was consistently higher than that of many

other subjects, it did vary from phase to phase. She practiced most frequently during phases one

and three (both LFF phases), averaging 5.8 and 7.8 hours/week respectively. Practice time

dropped during phases two and four (both HFF phases) to 5.3 and 3.7 hours/week respectively.

However, the significantly higher practice levels in phases one and three did not consistently

produce higher scores on the retention/transfer tests.

Subject Four Tone Quality

The aggregate scores for the tone quality of subject four from all five auditors are

presented in the chart in table 4.10 and in the graph in figure 4.19.
 83

Subject Four Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 60 35 50 45 40 46
2 60 30 50 25 30 39
LFF

3 54 40 55 40 50 47.8
4 54 25 55 40 60 46.8
5 54 25 75 45 30 45.8
6 60 40 40 60 65 53
HFF

7 62 25 55 40 40 44.4
8 58 30 50 30 70 47.6
9 58 25 75 50 50 51.6
10 54 40 45 45 50 46.8
LFF

11 62 60 70 30 60 56.4
12 60 50 60 40 40 50
13 48 40 60 35 25 41.6
HFF

14
15 54 40 65 40 25 44.8
Table 4.10 Subject four tone quality results

Subject Four Tone Quality

80
70
60 Aud. 1
Auditors' Score /100

50 Aud. 2
40 Aud. 3

30 Aud. 4

20 Aud. 5

10 Average

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.19 Subject four tone quality results
 84

Figure 4.20 presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 4 Phase 1 (LFF) S. 4 Phase 2 (HFF)

55 60

50
Auditors' Score /100

Auditors' Score /100

55

45
50
40 Average Average

Linear 45 Linear
35
(Average) (Average)
y = 0.74x + 42.86 y = 0.62x + 46.62
30 40
1 2 3 4 5 5 6 7 8 9
Week Week

S. 4 Phase 3 (LFF) S. 4 Phase 4 (HFF)

60 55

55
Auditors' Score /100

Auditors' Score /100

50
50
45
45 Average Average

Linear 40 Linear
40
(Average) (Average)
y = -1.68x + 54.32 y = 1.6x + 40
35 35
9 10 11 12 13 13 14 15
Week Week

Figure 4.20 Subject foure average tone quality results by phase

 85

Subject Four Breath Management

The aggregate scores for the tone quality of subject four from all five auditors are

presented in the chart in table 4.11 and in the graph in figure 4.21.

Subject Four Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 68 20 50 90 30 51.6
2 60 15 50 90 50 53
LFF

3 56 25 50 40 50 44.2
4 72 20 55 60 50 51.4
5 62 25 75 60 50 54.4
6 64 30 40 90 65 57.8
HFF

7 70 20 65 50 50 51
8 68 20 50 60 60 51.6
9 52 10 70 80 50 52.4
10 60 25 50 70 40 49
LFF

11 52 40 65 60 60 55.4
12 52 30 50 90 50 54.4
13 56 25 40 45 40 41.2
HFF

14
15 60 25 50 60 40 47
Table 4.11 Subject four breath management results

Subject Four Breath Management

100
90
80 Aud. 1
Auditors' Score /100

70
Aud. 2
60
50 Aud. 3
40 Aud. 4
30 Aud. 5
20
Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.21 Subject four breath management results
 86

Figure 4.22 presents the average of the scores for breath management from all five

auditors, divided into phases and including a linear trend line for each phase.

S. 4 Phase 1 (LFF) S. 4 Phase 2 (HFF)

60 60

55 55
Auditors' Score /100

Auditors' Score /100

50 50

Average Average
45 45
Linear Linear
40 (Average) 40 (Average)
y = 0.4x + 49.72 y = -1.02x + 56.5
35 35
1 2 3 4 5 5 6 7 8 9
Week Week

S. 4 Phase 3 (LFF) S. 4 Phase 4 (HFF)

60 60

55 55
Auditors' Score /100

Auditors' Score /100

50 50
Average Average
45 45
Linear Linear
40 (Average) 40 (Average)
y = -1.7x + 55.58 y = 2.9x + 38.3
35 35
9 10 11 12 13 13 14 15
Week Week

Figure 4.22 Subject four average breath management results by phase

 87

Subject Four Intonation

The aggregate scores for the intonation of subject four from all five auditors are presented

in the chart in table 4.12 and in the graph in figure 4.23.

Subject Four Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 78 20 80 60 90 65.6
2 75 20 80 90 85 70
LFF

3 75 25 80 90 85 71
4 70 5 80 75 80 62
5 75 20 85 75 80 67
6 72 40 70 90 80 70.4
HFF

7 76 10 75 60 80 60.2
8 76 30 80 70 80 67.2
9 74 15 90 90 85 70.8
10 70 40 60 90 50 62
LFF

11 74 15 85 80 85 67.8
12 70 10 70 80 75 61
13 75 50 80 90 80 75
HFF

14
15 80 30 75 75 80 68
Table 4.12 Subject four intonation results

Subject Four Intonation

100
90
80 Aud. 1
Auditors' Score /100

70
Aud. 2
60
50 Aud. 3
40 Aud. 4
30 Aud. 5
20
Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.23 Subject four intonation results

 88

Figure 4.24 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 4 Phase 1 (LFF) S. 4 Phase 2 (HFF)

75 80

75
Auditors' Score /100

Auditors' Score /100

70

70
65
Average Average
65
Linear Linear
60
(Average) 60 (Average)
y = -0.52x + 68.68 y = 0.44x + 65.8
55 55
1 2 3 4 5 5 6 7 8 9
Week Week

S. 4 Phase 3 (LFF) S. 4 Phase 4 (HFF)

80 80

75
Auditors' Score /100

Auditors' Score /100

75

70
70
Average Average
65
Linear 65 Linear
60 (Average) (Average)
y = 0.74x + 65.1 y = -3.5x + 78.5
55 60
9 10 11 12 13 13 14 15
Week Week

Figure 4.24 Subject four avearge intonation results by phase

 89

Subject Five

Subject five was a 19 year old female who had studied private voice with the researcher

for one year prior to the onset of the study. Prior to that point, she had not studied private voice

with any other instructor. Subject five did, however, report having participated in choir for four

years in high school and for one year at college. She reported having problems with chronic

laryngitis, having lost her voice more than once during the span of six months, which the

researcher had witnessed during his experience with this subject in the previous year. While

health problems had been a significant factor for this subject in the past, she was able to remain

healthy throughout the duration of the study and health related issues did not present any

problem for her participation in the study.

Subject five was assigned to group g-2 meaning that she received the reverse feedback

frequency phase progression as the previous subjects. She received high-frequency feedback

during phase one, low-frequency feedback during phase two, high-frequency feedback during

phase three, and low-frequency feedback during phase four. Subject five did not appear for her

lesson at week five, the absence of data for which is exhibited in the results for this subject.

Over the course of the study, subject five reported having practiced an average of 1.0

hours/week. However, subject fives practice was less consistent over the course of the study.

She reported having practiced an average of 1.3 and 1.1 hours/week during phases one and two

respectively, and 1.0 and 0.6 hours/week during phases three and four respectively.

Subject Five Tone Quality

The aggregate scores for the tone quality of subject five from all five auditors are

presented in the chart in table 4.13 and in the graph in figure 4.25.
 90

Subject Five Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 58 30 45 25 40 39.6
2 53 25 25 40 50 38.6
HFF

3 56 30 55 60 50 50.2
4 45 15 25 30 75 38
5
6 50 35 50 60 50 49
LFF

7 48 20 35 30 25 31.6
8 42 30 25 30 25 30.4
9 50 20 20 45 25 32
10 52 20 20 25 25 28.4
HFF

11 58 20 35 25 30 33.6
12 45 40 45 30 50 42
13 46 15 45 20 25 30.2
LFF

14 52 25 30 10 10 25.4
15 62 20 40 20 50 38.4
Table 4.13 Subject five tone quality results

Subject Five Tone Quality

80
70
Auditors' Score /100

60 Aud. 1
50 Aud. 2
40 Aud. 3
30 Aud. 4
20 Aud. 5
10 Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.25 Subject five tone quality results

 91

Figure 4.26 presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 5 Phase 1 (HFF) S. 5 Phase 2 (LFF)

60 55

55 50
Auditors' Score /100

Auditors' Score /100

50 45

45 40
Average Average
40 35
Linear Linear
35 (Average) 30 (Average)
y = 0.68x + 39.9 y = -5.22x + 54.02
30 25
1 2 3 4 5 5 6 7 8 9
Week Week

S. 5 Phase 3 (HFF) S. 5 Phase 4 (LFF)

50 45

45 40
Auditors' Score /100

Auditors' Score /100

40 35

35 Average 30 Average

Linear Linear
30 (Average) 25 (Average)
y = 1x + 30.24 y = 4.1x + 23.133
25 20
9 10 11 12 13 13 14 15
Week Week

Figure 4.26 Subject five avearge tone quality results by phase

 92

Subject Five Breath Management

The aggregate scores for the breath management of subject five from all five auditors are

presented in the chart in table 4.14 and in the graph in figure 4.27.

Subject Five Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 52 15 50 75 40 46.4
2 74 15 20 60 0 33.8
HFF

3 62 5 50 80 50 49.4
4 52 20 25 75 50 44.4
5
6 62 25 55 70 50 52.4
LFF

7 52 10 35 70 25 38.4
8 54 20 35 60 40 41.8
9 54 15 25 40 25 31.8
10 60 5 20 60 40 37
HFF

11 48 10 50 75 50 46.6
12 70 30 40 60 25 45
13 52 10 50 70 10 38.4
LFF

14 50 5 35 75 10 35
15 68 10 35 80 40 46.6
Table 4.14 Subject five breath management results

Subject Five Breath Management

90
80
70 Aud. 1
Auditors' Score /100

60 Aud. 2
50
Aud. 3
40
Aud. 4
30
20 Aud. 5
10 Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.27 Subject five breath management results

 93

Figure 4.28 presents the average of the scores for breath management from all five

auditors, divided into phases and including a linear trend line for each phase.

S. 5 Phase 1 (HFF) S. 5 Phase 2 (LFF)

60 60

55 55

Auditors' Score /100

Auditors' Score /100

50 50

45 45
Average Average
40 40
Linear Linear
35 (Average) 35 (Average)
y = 0.96x + 41.1 y = -5.84x + 61.54
30 30
1 2 3 4 5 5 6 7 8 9
Week Week

S. 5 Phase 3 (HFF) S. 5 Phase 4 (LFF)

50 50

45 45
Auditors' Score /100

Auditors' Score /100

40 40

35 Average 35 Average

Linear Linear
30 30 (Average)
(Average)
y = 2.12x + 33.4 y = 4.1x + 31.8
25 25
9 10 11 12 13 13 14 15
Week Week

Figure 4.28 Subject five avearge breath management results by phase

 94

Subject Five Intonation

The aggregate scores for the intonation of subject five from all five auditors are presented

in the chart in table 4.15 and in the graph in figure 4.29.

Subject Five Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 54 5 60 60 70 49.8
2 72 10 60 65 60 53.4
HFF

3 78 10 85 50 60 56.6
4 75 30 50 90 75 64
5
6 75 10 75 70 50 56
LFF

7 52 10 75 80 50 53.4
8 68 30 55 60 75 57.6
9 49 15 70 70 50 50.8
10 75 10 80 75 70 62
HFF

11 68 10 60 45 60 48.6
12 62 30 55 90 50 57.4
13 48 5 60 40 50 40.6
LFF

14 42 5 55 75 25 40.4
15 78 15 60 60 60 54.6
Table 4.15 Subject five intonation results

Subject Five Intonation

100
90
80 Aud. 1
Auditors' Score /100

70
Aud. 2
60
50 Aud. 3
40 Aud. 4
30 Aud. 5
20
Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Weeks
Figure 4.29 Subject five intonation results
 95

Figure 4.30 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 5 Phase 1 (HFF) S. 5 Phase 2 (LFF)

70 70

65 65
Auditors' Score /100

Auditors' Score /100

60 60
Average
Average
55 55
Linear
Linear
50 (Average) 50 (Average)
y = 4.58x + 44.5 y = -1.14x + 58.44
45 45
1 2 3 4 5 5 6 7 8 9
Week Week

S. 5 Phase 3 (HFF) S. 5 Phase 4 (LFF)

65 60

60 55
Auditors' Score /100

Auditors' Score /100

55 50

50 45
Average Average
45 40
Linear Linear
40 (Average) 35 (Average)
y = -2.5x + 59.38 y = 7x + 31.2
35 30
9 10 11 12 13 13 14 15
Week Week

Figure 4.30 Subject five average intonation results by phase

 96

Subject Six

Subject six was a 21 year old male who had studied private voice with the researcher for

one year prior to the onset of the study. Prior to that point, he had studied at a college level for

three years and for one year in high school. Additionally, subject six had participated in several

school and community choirs over the previous ten years. Subject six reported having no vocal

health issues or chronic loss of voice. Health related issues did not present any problems for his

participation in the study.

Subject six was assigned to group g-2 meaning that he received high-frequency

feedback during phase one, low-frequency feedback during phase two, high-frequency feedback

during phase three, and low-frequency feedback during phase four. Subject six failed to appear

for his lessons on weeks eight, eleven, and fifteen, the absence of data for which is exhibited in

the results for this subject. These absences pose some question to the validity of the results from

the phases three and four.

Subject six misplaced his practice record at some point during the study and was unable

to return it to the researcher at the conclusion of the study. Consequently he had no record of the

amount of time he had spent practicing.

Subject Six Tone Quality

The aggregate scores for the tone quality of subject six from all five auditors are

presented in the chart in table 4.16 and in the graph in figure 4.31.
 97

Subject Six Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 78 50 80 90 100 79.6
2 72 50 75 75 100 74.4
HFF

3 80 60 80 75 100 79
4 78 60 80 75 90 76.6
5 76 50 85 80 100 78.2
6 78 60 80 75 100 78.6
LFF

7 85 90 75 65 100 83
8
9 84 60 85 75 90 78.8
10 85 80 85 90 90 86
HFF

11
12 80 60 90 75 100 81
13 90 90 65 65 100 82
LFF

14 87 80 85 70 90 82.4
15
Table 4.16 Subject six tone quality results

Subject Six Tone Quality

100
90
80 Aud. 1
Auditors' Score /100

70
Aud. 2
60
50 Aud. 3
40 Aud. 4
30 Aud. 5
20
Average
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week
Figure 4.31 Subject six tone quality results
 98

Figure 4.32 presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 6 Phase 1 (HFF) S. 6 Phase 2 (LFF)

85 85

Auditors' Score /100

Auditors' Score /100

80 80

75 75
Average Average

70 Linear 70 Linear
(Average) (Average)
y = -0.06x + 77.74 y = 0.2571x + 78.943
65 65
1 2 3 4 5 5 6 7 8 9
Week Week

S. 6 Phase 3 (HFF) S. 6 Phase 4 (LFF)

90 90
Auditors' Score /100

Auditors Score /100

85 85

80 80
Average Average

Linear Linear
75 75
(Average) (Average)
y = 0.14x + 81.53 y = 0.4x + 81.6
70 70
9 10 11 12 13 13 14 15
Week Week

Figure 4.32 Subject six average tone quality results by phase

 99

Subject Six Breath Management

The aggregate scores for the breath management of subject six from all five auditors are

presented in the chart in table 4.17 and in the graph in figure 4.33.

Subject Six Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 82 40 90 90 100 80.4
2 74 40 65 90 80 69.8
HFF

3 70 50 80 90 85 75
4 75 50 80 80 85 74
5 72 40 80 90 100 76.4
6 62 50 75 85 85 71.4
LFF

7 80 85 60 80 60 73
8
9 82 50 85 80 75 74.4
10 78 60 85 95 85 80.6
HFF

11
12 82 40 90 80 90 76.4
13 85 80 60 70 50 69
LFF

14 84 55 90 90 80 79.8
15
Table 4.17 Subject six breath management results

Subject Six Breath Management

100
90
80
Auditors' Score /100

Aud. 1
70
60 Aud. 2
50 Aud. 3
40
Aud. 4
30
20 Aud. 5
10 Average
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.33 Subject six breath management results

 100

Figure 4.34 presents the average of the scores for breath management from all five

auditors, divided into phases and including a linear trend line for each phase.

S. 6 Phase 1 (HFF) S. 6 Phase 2 (LFF)

85 85
Auditors' Score /100

Auditors' Score /100

80 80

75 75
Average Average

70 Linear 70 Linear
(Average) (Average)
y = -0.38x + 76.26 y = -0.1829x + 74.303
65 65
1 2 3 4 5 5 6 7 8 9
Week Week

S. 6 Phase 3 (HFF) S. 6 Phase 4 (LFF)

85 90

85
Auditors' Score /100

Auditors' Score /100

80

80
75
Average Average
75

70 Linear Linear
(Average) 70 (Average)
y = -1.5x + 79.6 y = 10.8x + 58.2
65 65
9 10 11 12 13 13 14 15
Week Week

Figure 4.34 Subject six average breath management results by phase

 101

Subject Six Intonation

The aggregate scores for the intonation of subject six from all five auditors are presented

in the chart in table 4.18 and in the graph in figure 4.35.

Subject Six Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1 70 30 90 75 100 73
2 45 25 65 70 75 56
HFF

3 52 10 80 70 85 59.4
4 49 40 75 90 75 65.8
5 60 30 80 75 100 69
6 64 40 75 75 100 70.8
LFF

7 78 75 60 60 50 64.6
8
9 65 30 75 80 85 67
10 88 60 90 95 90 84.6
HFF

11
12 48 10 90 80 85 62.6
13 60 25 20 60 90 51
LFF

14 54 40 85 95 75 69.8
15
Table 4.18 Subject six intonation results

Subject Six Intonation

100
90
80
Auditors' Score /100

70 Aud. 1
60
Aud. 2
50
40 Aud. 3
30 Aud. 4
20 Aud. 5
10
0 Average
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.35 Subject six intonation results

 102

Figure 4.36 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 6. Phase 1 (HFF) S. 6 Phase 2 (LFF)

80 80

75
Auditors' Socre /100

Auditors' Score /100

75
70

65 70 Average
Average
60 Linear
Linear 65
(Average)
55 (Average)
y = 0.18x + 64.1 y = -0.7943x + 70.034
50 60
1 2 3 4 5 5 6 7 8 9
Week Week

S. 6 Phase 3 (HFF) S. 6 Phase 4 (LFF)

90 90
85
Auditors' Score /100
Auditors' Score /100

80 80
75
70 70
65 Average Average
60 Linear
Linear 60
55 (Average) (Average)
50 y = 18.8x + 32.2
y = -5.4x + 82.5 50
45
13 14 15
9 10 11 12 13
Week
Week

Figure 4.36 Subject six average intonation results by phase

 103

Subject Seven

Subject seven was an 18 year old male who had studied private voice for six months with

his high school choir teacher prior to the onset of the study. In addition to his private voice

instruction, he reported having participated in choirs for three years in high school. Subject

seven reported no vocal health issues or chronic loss of voice. Health related issues did not

present any problems for his participation in the study.

Subject seven was assigned to group g-2 meaning that he received high-frequency

feedback during phase one, low-frequency feedback during phase two, high-frequency feedback

during phase three, and low-frequency feedback during phase four. While subject seven did

appear at every lesson, he did not choose to participate in the study until week two. This meant

that his phase one was only three weeks in duration. For consistency with the other subjects,

however, week one is indicated in the reporting of his results as having no score.

Over the course of the study, subject seven reported having practiced an average of 1.8

hours/week. Adjusted for phases, he practiced an average of 2.1 and 1.9 hours/week during

phases one and two respectively, and 1.6 and 2.0 hours/week during phases three and four

respectively.

Subject Seven Tone Quality

The aggregate scores for the intonation of subject seven from all five auditors are

presented in the chart in table 4.19 and in the graph in figure 4.37.
 104

Subject Seven Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1
2 76 60 70 75 60 68.2
HFF

3 75 60 75 70 80 72
4 72 50 75 30 70 59.4
5 70 50 75 80 100 75
6 66 60 75 60 75 67.2
LFF

7 65 40 55 60 60 56
8 72 50 70 75 80 69.4
9 75 65 75 40 75 66
10 76 60 75 40 80 66.2
HFF

11 65 60 65 35 70 59
12 76 50 75 60 75 67.2
13 70 45 75 75 80 69
LFF

14 75 50 70 70 80 69
15 78 40 75 40 90 64.6
Table 4.19 Subject seven tone quality results

Subject Seven Tone Quality

100
90
80
Auditors' Score /100

70 Aud. 1
60 Aud. 2
50
40 Aud. 3
30 Aud. 4
20 Aud. 5
10
0 Average
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.37 Subject seven tone quality results

 105

Figure 4.38 presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 7 Phase 1 (HFF) S. 7 Phase 2 (LFF)

80 80

75 75
Auditors' Score /100

Auditors' Score /100

70 70

Average Average
65 65
Linear Linear
60 (Average) 60 (Average)
y = 0.78x + 65.92 y = -1.58x + 71.46
55 55
1 2 3 4 5 5 6 7 8 9
Week Week

S. 7 Phase 3 (HFF) S. 7 Phase 4 (LFF)

80 80

75 75
Auditors' Score /100

Auditors' Score /100

70 70

Average Average
65 65
Linear Linear
60 (Average) 60 (Average)
y = 0.7x + 63.38 y = -2.2x + 71.933
55 55
9 10 11 12 13 13 14 15
Week Week

Figure 4.38 Subject seven average tone quality results by phase

 106

Subject Seven Breath Management

The aggregate scores for the breath management of subject seven from all five auditors

are presented in the chart in table 4.20 and in the graph in figure 4.39.

Subject Seven Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1
2 60 40 80 95 70 69
HFF

3 60 30 65 90 90 67
4 68 50 80 65 50 62.6
5 40 10 50 60 70 46
6 56 15 70 80 85 61.2
LFF

7 52 15 25 70 60 44.4
8 62 50 60 75 50 59.4
9 70 50 70 80 75 69
10 62 50 70 50 75 61.4
HFF

11 75 40 50 80 60 61
12 50 5 60 85 60 52
13 70 10 65 90 90 65
LFF

14 60 25 60 90 80 63
15 58 30 60 45 90 56.6
Table 4.20 Subject seven breath management results

Subject Seven Breath Management

100
90
80
Auditors' Score /100

70 Aud. 1
60 Aud. 2
50
40 Aud. 3
30 Aud. 4
20 Aud. 5
10
0 Average
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.39 Subject seven breath management results

 107

Figure 4.40 presents the average of the scores for breath management from all five

auditors, divided into phases and including a linear trend line for each phase.

S. 7 Phase 1 (HFF) S. 7 Phase 2 (LFF)

75 75
70 70
Auditors' Score /100

Auditors' Score /100

65 65
60 60
55 Average 55 Average
50 Linear 50
Linear
(Average)
45 45 (Average)
y = -7.34x + 86.84 y = 4.42x + 42.74
40 40
1 2 3 4 5 5 6 7 8 9
Week Week

S. 7 Phase 3 (HFF) S. 7 Phase 4 (LFF)

75 75
70 70
Auditors' Score /100

Auditors' Score /100

65 65
60 60
55 Average 55 Average
50 50 Linear
Linear
45 (Average) 45 (Average)
y = -1.74x + 66.9 y = -4.2x + 69.933
40 40
9 10 11 12 13 13 14 15
Week Week

Figure 4.40 Subject seven average breath management results by phase

 108

Subject Seven Intonation

The aggregate scores for the breath management of subject seven from all five auditors

are presented in the chart in table 4.21 and in the graph in figure 4.41.

Subject Seven Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1
2 49 10 60 80 50 49.8
HFF

3 46 10 80 70 85 58.2
4 51 25 80 50 25 46.2
5 48 0 75 50 75 49.6
6 45 10 60 50 75 48
LFF

7 50 0 40 52 25 33.4
8 40 0 55 30 10 27
9 28 10 10 30 25 20.6
10 38 10 40 40 25 30.6
HFF

11 46 10 15 60 50 36.2
12 48 5 70 95 85 60.6
13 48 5 45 50 75 44.6
LFF

14 30 0 45 75 50 40
15 40 0 35 50 50 35
Table 4.21 Subject seven intonation results

Subject Seven Intonation

100
90
80
Auditors' Score /100

70 Aud. 1
60 Aud. 2
50
40 Aud. 3
30 Aud. 4
20 Aud. 5
10
0 Average
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Week

Figure 4.41 Subject seven intonation results

 109

Figure 4.42 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 7 Phase 1 (HFF) S. 7 Phase 2 (LFF)

70 55
50
65
Auditors' Score /100

Auditors' Score /100

45
60
40
55 35
Average
Average 30
50
Linear 25 Linear
45 (Average) (Average)
20
y = -1.26x + 55.36 y = -7.9x + 59.42
40 15
1 2 3 4 5 5 6 7 8 9
Week Week

S. 7 Phase 3 (HFF) S. 7 Phase 4 (LFF)

65 50
60
55
Auditors' Score /100
Auditors' Score /100

45
50
45
40 40
Average Average
35
30 Linear
Linear 35
25 (Average)
(Average)
20 y = -4.8x + 49.467
y = 7.8x + 15.12
15 30
9 10 11 12 13 13 14 15
Week Week

Figure 4.42 Subject seven average intonation results by phase

 110

Subject Eight

Subject eight was a 25 year old female who had studied private voice with another

college instructor for one year prior to the onset of the study. She reported having never

participated in choirs previous to the start of the study period, nor had she studied voice with the

researcher prior to the study.

While subject eight reported having no known vocal pathologies or a history of chronic

vocal health problems, she acquired an illness, causing her to withdraw from college, and

consequently the study, after week nine of the study period. Also, as mentioned briefly in

chapter three, subject eight presented with an unusually low speaking and singing voice.

Consequently, she performed the retention/transfer tests in the key of G Major, while all other

low voices perfromed the test in the key of Bb Major She reported that her voice had always

been low and that her previous voice instructor assigned vocal literature in a tenor range for her

to sing. Over the course of the first few meetings, the researcher found that subject eight was

failing to access her head register and was singing exclusively using her chest register. In that

same time period, subject eight was introduced to the head register and the bulk of the instruction

time during her lesson times was spent practicing moving in to and out of her head register. By

doing so, her range was extended to a relatively normal low alto range. However, for the sake of

continuity, she continued to perform her retention/transfer tests in the key of G Major.

Because of her sudden and unexpected departure from the study, subject eight did not

return her practice log. Consequently, there is no record of the amount of time she spent

practicing during her participation in the study.

 111

Subject Eight Tone Quality

The aggregate scores for the tone quality of subject eight from all five auditors are

presented in the chart in table 4.22 and in the graph in figure 4.43.

Subject Eight Tone Quality

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1
2 62 40 75 40 60 55.4
HFF

3 60 50 60 60 75 61
4 55 60 60 45 60 56
5 54 50 50 40 60 50.8
6 60 40 40 46 50 47.2
LFF

7 62 45 65 80 80 66.4
8 60 55 65 50 65 59
Table 4.22 Subject eight tone quality results

Subject Eight Tone Quality

100
90
80
Auditors' Score /100

70 Aud. 1
60
Aud. 2
50
40 Aud. 3
30 Aud. 4
20
10 Aud. 5
0 Average
1 2 3 4 5 6 7 8
Week

Figure 4.43 Subject eight tone quality results

 112

Figure 4.44 presents the average of the scores for tone quality from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 8 Phase 1 (HFF) S. 8 Phase 2 (LFF)

65 70

65
Auditors' Score /100

Auditors' Score /100

60
60

55 55
Average Average
50
50 Linear Linear
(Average) 45 (Average)
y = -1.88x + 62.38 y = 4.38x + 44.9
45 40
1 2 3 4 5 5 6 7 8
Week Week

Figure 4.44 Subject eight average tone quality results by phase

Subject Eight Breath Management

The aggregate scores for the breath management of subject eight from all five auditors

are presented in the chart in table 4.23 and in the graph in figure 4.45.

Subject Eight Breath Management

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1
2 75 25 75 75 60 62
HFF

3 60 20 45 90 50 53
4 62 50 35 80 50 55.4
5 60 20 40 80 50 50
6 75 60 40 68 40 56.6
LFF

7 62 25 60 95 80 64.4
8 62 40 70 75 60 61.4
Table 4.23 Subject eight breath mangement results
 113

Subject Eight Breath Management

100
90
80
Auditors' Score /100

70 Aud. 1
60
Aud. 2
50
40 Aud. 3
30 Aud. 4
20
10 Aud. 5
0 Average
1 2 3 4 5 6 7 8
Week

Figure 4.45 Subject eight breath management results

Figure 4.46 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 8 Phase 1 (HFF) S. 8 Phase 2 (LFF)

65 70

65
Auditors' Score /100

Auditors' Score /100

60
60
55 Average Average
55
Linear Linear
50
(Average) 50 (Average)
y = -3.36x + 66.86 y = 4.2x + 47.6
45 45
1 2 3 4 5 5 6 7 8
Week Week

Figure 4.46 Subject eight average breath management results by phase

 114

Subject Eight Intonation

The aggregate scores for the intonation of subject eight from all five auditors are

presented in the chart in table 4.24 and in the graph in figure 4.47.

Subject Eight Intonation

Week Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average
1
2 52 5 75 60 50 48.4
HFF

3 38 5 50 70 25 37.6
4 40 15 55 65 40 43
5 38 0 30 50 50 33.6
6 46 20 45 70 40 44.2
LFF

7 51 20 75 70 75 58.2
8 49 10 75 60 50 48.8
Table 4.24 Subject eight intonation results

Subject Eight Intonation

100
90
80
Auditors' Score /100

70 Aud. 1
60 Aud. 2
50
40 Aud. 3
30 Aud. 4
20 Aud. 5
10
0 Average
1 2 3 4 5 6 7 8
Week

Figure 4.47 Subject eight intonation results

 115

Figure 4.48 presents the average of the scores for intonation from all five auditors,

divided into phases and including a linear trend line for each phase.

S. 8 Phase 1 (HFF) S. 8 Phase 2 (LFF)

55 65
60
50
Auditors' Score /100

Auditors' Score /100

55
45 50

40 Average 45
Average
Linear 40
35 Linear
(Average) 35 (Average)
y = -3.9x + 54.3 y = 5.96x + 31.3
30 30
1 2 3 4 5 5 6 7 8
Week Week

Figure 4.48 Subject eight average intonation results by phase

The results reported above may be used as a reference in the more thorough discussion

contained in the following chapter. That chapter will interpret the results reported above and

attempt to draw conclusions from them regarding the effect of variations in relative feedback

frequency on the retention/transfer of classical singing techniques.

 116

CHAPTER 5

DISCUSSION OF RESULTS

The purpose of this study was to test the hypothesis that LFF instruction would be beneficial to

long-term improvements in the subjects ability to perform classical vocal techniques. As seen

in the results reported in chapter 4, the findings of the study do not on the surface support a

strong positive correlation between LFF instruction and the desired improvements in

performance ability in all subjects. However, some positive correlation may be seen in subjects

with higher levels of ability.

Results That Support the Hypothesis

If the hypothesis were to hold true, it would be expected that the learning trends during

LFF phases would be steeper than during HFF phases. Upon analysis of the scores from the

retention/transfer tests, it appeared that only five of the possible twenty-four scenarios (three

voice characteristics for each of eight subjects) presented those results. These scenarios will be

discussed below. The results for each scenario are represented in the graphs in chapter four. In

order to save space, those graphs were not recreated here.

Subject Six Tone Quality

The results for the tone quality for subject six presented steeper learning trends during

LFF phases (phases 2 and 3 in this case) than in HFF phases. During phase one (HFF), the

learning trend was actually slightly negative with a slope of -0.06. That trend reversed during

phase two, the first of the LFF phases, to a positive slope of 0.257. When instruction returned to

the HFF condition in phase three, the learning trend remained positive but with a more moderate
 117

slope of 0.14. Finally, with the last LFF instruction condition in phase four, the learning trend

again steepened to the most aggressive slope of the study at 0.4. With positive learning trend

slopes in the LFF phases, and at a steeper slope than in the HFF phases, these findings lend

support to the hypothesis.

Subject Six Breath Management

The results for the breath management of subject six are similar to those for that subjects

tone quality reported above. However, in this scenario the overall trend across the duration of

the study was only very slightly positive. The learning trend in phase one (HFF) was negative

with a slope of -0.38. Phase two (LFF) was still negative, but with a more gradual slope of -0.18.

Phase three, returning to the HFF condition, was once again more aggressively negative with a

slope of -1.5. Finally, phase four (LFF) was positive, with a quite steep slope of 10.8. Thus,

with the two LFF phases presenting either less aggressive negative or positive slopes, these

findings provide some support of the hypothesis.

Subject Eight Tone Quality, Breath Management, and Intonation

Because subject eight was forced to end her participation in the study due to health and

academic problems after only eight weeks, she was only present for one phase each of the LFF

and HFF instruction conditions. However, her results for each vocal characteristic presented

steeper learning trends during the LFF condition than in the HFF condition. In fact, the learning

trends for all three characteristics were negative in the HFF phases but positive in the LFF

phases. For tone quality, her learning trend in phase one (HFF) had a slope of -1.88, while her

learning trend in phase two (LFF) had a slope of 4.38. For breath management, her learning
 118

trend in phase one had a slope of -3.36, while her learning trend in phase two had a slope of 4.2.

Finally, for intonation, her learning trend in phase one had a slope of -3.9, while her learning

trend in phase two had a slope of 5.96. These results do provide some relatively weak support

for the hypothesis. However, because of subject eights early departure from the study, these

results may be explained by factors other than that she benefited from the LFF instruction

condition as there was no second round of HFF and LFF phases against which the results for

phases one and two could be compared.

Results That Do Not Support the Hypothesis

Of the remaining nineteen scenarios, four presented results that were in direct

contradiction to the hypothesis. In these three cases, learning trends were in fact steeper during

HFF phases than during LFF phases. These cases will be discussed below.

Subject One Tone Quality

The results for the tone quality of subject one were exactly opposite of those expected if

the hypothesis were to hold true. During phase one (LFF), subject ones learning trend was

negative with a slope of -0.86. In phase two (HFF) however, that trend reversed to a positive

slope of 0.54. Phase three (LFF) saw a return to a negative learning trend with a slope of -4.2.

Finally, during phase four (HFF) the learning trend returned to a positive slope of 3.3. These

findings would indicate that not only was LFF instruction condition not beneficial to the

improvement of tone quality in subject one, it was actually detrimental, with subject one actually

performing worse each week during those phases.

 119

Subject One Intonation

In regard to intonation, subject ones performance again appeared to hindered by the LFF

instruction condition. During phase one (LFF), subject ones learning trend was negative with a

slope of -1.88. With the change to the HFF instruction condition in phase two, her learning trend

turned sharply positive with a slope of 14.16. During phase three (LFF), the learning trend again

returned to a negative slope of -0.64 and the study finished with a positive slope of 6.1 during the

final HFF instruction of phase four.

Subject Three Breath Management

The results for the breath management of subject three presented three negative learning trends

and one positive trend. Phase one (LFF) began the study with a negative learning trend with a

slope of -1.46. That negative trend moderated somewhat during phase two (HFF) to a slope of -

.16, but increased again in phase three (LFF) to a slope of -1.12. Finally, phase four (HFF) saw

the first positive learning trend with a slope of 3.2. With these results both LFF phases were

more negative than either of the HFF phases, thus contradicting the expected correlation.

Subject Seven Tone Quality

Similar to the two previous scenarios, the results for the tone quality of subject seven

again appeared to indicate that the LFF instruction condition actually hindered performance on

the retention/transfer tests. During phase one (HFF), subject ones learning trended up with a

slope of 0.78. The introduction of the LFF instruction condition was accompanied by a negative

turn in subject sevens learning trend with a slope of -1.58. That trend was again reversed with

the HFF instruction condition of phase three, resulting in a learning trend with a positive slope of
 120

0.7. Finally, the return of the LFF instruction condition in phase four saw a return to a negative

learning trend with a slope of -2.2.

Results That Are Mixed

The remaining fifteen scenarios presented results that were mixed, meaning that no clear

correlation between feedback frequency and performance on retention/transfer tests is clearly

visible in these cases. Each of these scenarios will be discussed below, followed by a discussion

of how these results might be interpreted.

Subject One Breath Management

The results for breath management of subject one were mixed. During phase one (LFF),

subject ones learning trend was positive with a slope of 1.42. That trend was reversed with the

HFF instruction condition of phase two, presenting a learning trend with a negative slope of

-1.58. During the LFF instruction condition of phase three, the learning trend remained negative

with a slope of -0.6 before turning positive once again in phase four (HFF) with a slope of 3.2.

Subject Two Tone Quality

The results for the tone quality of subject two again failed to present a clear correlation.

During phase one (LFF), the subjects learning trend was negative with a slope of -3.04. That

trend was reversed during phase two (HFF), presenting a learning trend with a positive slope of

2.54. That positive slope could not be maintained however, turning negative again in phase three

(LFF) with a slope of -1.8, and staying negative throughout the conclusion of the study in phase

four (HFF) with a slope of -5.6.

 121

Subject Two Breath Management

Similar to the results for her tone quality, the results for the breath management of subject

two again had three phases with negatively-sloped learning trends and one phase with a positive

learning trend. Interestingly, these results even presented the same pattern as the tone quality

results. Phase one (LFF) presented a learning trend with a negative slope of -.26. Phase two

(HFF) presented the only positive learning trend with a slope of 1.6. Phases three (LFF) and four

(HFF) again presented negative learning trends with slopes of -0.5 and -7.5 respectively.

Subject Two Intonation

In regard to intonation, the learning trend for subject two in phase one (LFF) was once

again negative, with a slope of -1.72. Phase two (HFF) once again presented a positive learning

trend with a slope of 1.78. Unlike in the results for breath management and tone quality for the

same subject, the learning trend for subject two in phase three (LFF) remained positive with a

slope of 0.94. The learning trend turned negative again, however, in phase four (HFF) with a

slope of -2.2.

Subject Three Tone Quality

The results for the tone quality of subject three presented two phases with negative

learning trends, one phase with a positive learning trend, and one phase with a static learning

trend. During phase one (LFF), the learning trend was negative with a slope of -1.46. That

negative trend continued through phase two (HFF) with a slope of -2.1. In phase three (LFF) the

learning trend finally turned positive with a slope of 0.74, but the trend then held static through

phase four (HFF) with a slope of 0.

 122

Subject Three Intonation

The results for the intonation of subject three presented two phases with positive learning

trends (one LFF and the other HFF) and two phases with negative learning trends (again one LFF

and the other HFF). Phase one (LFF), had a positive learning trend with a slope of 3.0 while

phase two (HFF) had a negative learning trend with a slope of -0.72. The negative learning trend

continued into the LFF instruction condition of phase three which presented a slope of -4.12. In

the final phase (HFF), the learning trend turned sharply positive with a slope of 9.2.

Subject Four Tone Quality

The results for the tone quality of subject four presented three phases with positive

learning trends and one phase with a negative learning trend. The study began with phases one

(LFF) and two (HFF) revealing positive learning trends with slopes of 0.74 and 0.62

respectively. During phase three (LFF) the learning trend turned negative with a slope of -1.68.

Finally, in phase four (HFF), the learning trend returned to a positive ascent with a slope of 1.6.

Subject Four Breath Management

The results for the breath management of subject four had two phase with positive slopes

and two with negative slopes. However, these results fail support the hypothesis because the two

negative slopes are in contrasting feedback conditions. Phase one (LFF) had a positive learning

trend with a slope of 0.4. That trend reversed in phase two (HFF) to a negative slope of -1.02

and remained negative through phase three (LFF) with a slope of -1.7. In the final phase (HFF)

the learning trend once again turned positive with a slope of 2.9.
 123

Subject Four Intonation

The results for the intonation of subject four were similar to those for the breath

management of the same subject in that they again presented two phases with positive learning

trends and two phases with negative slopes, each consisting of results from contrasting feedback

conditions. Phase one (LFF) presented a negative learning trend with a slope of -0.52. The trend

then turned positive through phases two (HFF) and three (LFF) with slopes of 0.44 and 0.74

respectively. The study concluded with a positive learning trend in phase four (HFF) with a

slope of -3.5.

Subject Five Tone Quality

The tone quality of subject five provided results that presented three phases with positive

learning trends and one phase with a negative learning trend. Phase one (HFF) presented a

positive learning trend with a slope of 0.68. The learning trend turned sharply negative in phase

two (LFF) with a slope of -5.22. That negative trend was reversed in phase three (HFF) with a

slope of 1.0 and remained positive through the conclusion of the study in phase four (LFF) with a

slope of 4.1.

Subject Five Breath Management

The results for the breath management of subject five are very similar to the results for

tone quality for the same subject. Phase one (HFF) presented a positive learning trend with a

slope of 0.96. The learning trend again turned negative in phase two (LFF) with a slope of -5.84.

During phase three (HFF) the learning trend once again turned positive with a slope of 2.12 and

remained positive through phase four (LFF) with a slope of 4.1.

 124

Subject Five Intonation

The results for the intonation of subject five presented two phases with positive learning

trends and two phases with negative learning trends, each consisting of contrasting feedback

conditions. Phase one (HFF) presented a positive learning trend with a slope of 4.58. Phases

two (LFF) and three (HFF) reversed the positive trend and presented learning trends with slopes

of -1.14 and -2.5 respectively. The final phase of the study (LFF) returned to a positive learning

trend with a slope of 7.0.

Subject Six Intonation

The results for the intonation of subject followed the same pattern as the results for the

intonation of subject five reported above. Phase one (HFF) presented a slightly positive learning

trend with a slope of 0.18. That trend was reversed in phases two (LFF) and three (HFF) which

presented negative learning trends with slopes of -0.79 and -5.4 respectively. Phase four

presented a return to a positive learning trend with a slope of 18.8. However, because subject six

did not appear for his final lesson, there are only two data points, representing the retention over

a single week, in phase four. Consequently, it is likely that the slope for phase four is artificially

high.

Subject Seven Breath Management

The results for the breath management of subject seven presented three phases with

negative learning trends and one phase with a positive learning trend. Phase one (HFF)

presented a negative learning trend with a slope of -7.34. The only positive learning trend came

in phase two (LFF) with slope of 4.42. The learning trend turned negative again in phase three
 125

(HFF) with a slope of -1.74 and remained negative through phase four (LFF) with a slope of -4.2.

Subject Seven Intonation

The results for the intonation of subject seven again presented three negative learning

trends and one positive learning trend. In this case, phases one (HFF) and two (LFF) both

presented negative learning trends with slopes of -1.26 and -7.9 respectively. Phase three (HFF)

presented the only positive learning trend with a slope of 7.8 before the trend again turned

negative in phase four (LFF) with a slope of -4.8.

Interpreting the Mixed Results

Because of the large number of scenarios in which the results were mixed, it is desirable

to attempt to interpret those results in order to determine if there is some correlation that can be

drawn between the frequency of feedback and the subjects performances on the

retention/transfer tests. In order to clarify these results, the two learning trends for each feedback

condition were added together (i.e. LFF slope + LFF slope and HFF slope + HFF slope) to

determine an overall trend for that condition. An analysis of those sums revealed that six

scenarios supported the hypothesis, while the remaining nine scenarios contradicted the

hypothesis. A summary of these calculations are included in tables 5.1 (supporting hypothesis)

and 5.2 (contradicting hypothesis) on the following page.

 126

Scenarios Supporting Hypothesis

Subject/Characteristic Sum of LFF Learning Trend Sum of HFF Learning Trend
S. 2 Breath Management -0.76 -5.9
S. 3 Tone Quality -1.44 -2.1
S. 4 Intonation 0.22 -3.06
S. 5 Intonation 5.86 2.08
S. 6 Intonation 18.01 -5.22
S. 7 Breath Management 0.22 -9.08
Table 5.1 Scenarios supporting the hypothesis

Scenarios Contradicting Hypothesis

Subject/Characteristic Sum of LFF Learning Trend Sum of HFF Learning Trend
S. 1 Breath Management 1.36 1.62
S. 2 Tone Quality -4.84 -3.6
S. 2 Intonation -0.78 -0.42
S.3 Intonation -1.12 8.48
S. 4 Tone Quality -0.94 2.22
S. 4 Breath Management -1.3 1.88
S. 5 Tone Quality -1.1 1.68
S. 5 Breath Management -1.74 3.08
S. 7 Intonation -12.7 6.54
Table 5.2 Scenarios contradicting the hypothesis

Initial Summary of Results

Of the twenty-four possible scenarios, it appears that five clearly support the hypothesis.

An additional six scenarios support the hypothesis when a sum is taken of both the LFF and the

HFF phases, for a total of eleven scenarios supporting the hypothesis. Oppositely, three

scenarios clearly contradict the hypothesis. An additional nine scenarios contradict the

hypothesis when a sum is taken of both the LFF and the HFF phases for a total of twelve

scenarios contradicting the hypothesis. The fact that the results are evenly divided between those

that support the hypothesis and those that contradict it, indicates that a strong correlation
 127

between feedback frequency during skill acquisition and a subjects performance ability on a

retention/transfer test, either positive or negative, cannot be surmised from this study.

If, as the studies summarized in chapter 2 suggest, research has indicated that lowering

the frequency of feedback during instruction is beneficial to the retention of newly-acquired

motor skills, the question arises as to why those findings did not hold true in the current study.

Several possible responses to that question will be discussed in the following chapter. One

response, which may be addressed by new analysis, is discussed below.

Variation of Skills Being Trained

One possible cause for the unexpected results may stem from the nature of the classical

vocal training that was taking place in this study. In typical college-level voice instruction, the

vocal skills receiving the bulk of the students and the instructors attention would often vary

throughout the course of a semester. For instance, it is unlikely that breath management, tone

quality, and intonation would all garner attention at every lesson. It is much more likely, instead,

that attention during any given lesson would end up being focused more on one of those

characteristics than on the other two.

In attempt to compensate for this variation in focus during individual lessons, the scores

for all three characteristics were added together to form a total score for the overall quality of the

performance on each weeks retention/transfer test. This process was repeated for all five

auditors and an average of the scores from all five auditors was calculated. From there, the same

process used with the scores from the individual characteristics was instituted with this data to

create overall learning scores for each subject.

 128

This analysis of the results revealed that three of the subjects (S.2, S.6, and S.8)

responded more favorably to the LFF instruction condition, supporting the hypothesis. However,

the results from the remaining five subjects (S.1, S.3, S.4, S.5, and S. 7) suggest that they

responded more favorably to the HFF instruction condition, contradicting the hypothesis.

Graphs of these results are shown and discussed in the following pages.

Subjects That Support the Hypothesis

Figure 5.1 shows the average of the overall scores for subject two from all five auditors

including linear trend lines. The overall scores for subject two do not appear at first glance to

support the hypothesis as the only positive learning trend occurred during phase two. However,

an average of the learning trend slopes for the LFF and the HFF phases reveals that, because of

the severity of the decline in phase four, the LFF instruction condition may have been less

damaging to the subjects progress. These results provide very weak support of the hypothesis.

Figure 5.2 shows the average of the overall scores for subject six from all five auditors

including linear trend lines. The overall scores for subject six lend somewhat stronger support

for the hypothesis than did those from subject two. However, the dramatic positive learning

trend in phase four, coupled with the fact that subject six was only present for two of the three

weeks for that phase, diminishes the credibility of these results.

 129

S. 2 Phase 1 (LFF) S. 2 Phase 2 (HFF)

205 200

195
190
Auditors' Score /300

Auditors' Score /300

185
180
175
Average 170
165
160 Average
Linear
155 (Average)
Linear
y = -5.02x + 195.11 150
145 (Average)
1 2 3 4 5 y = 6.12x + 160.32
140
Week
5 6 7 8 9
Week

S. 2 Phase 3 (LFF) S. 2 Phase 4 (HFF)

200 200

190 190
Auditors' Score /300
Auditors' Score /300

180 180

170 170
Average Average
160 160
Linear Linear
150 (Average) 150 (Average)
y = -1.42x + 178.98 y = -15.3x + 205.4
140 140
9 10 11 12 13 13 14 15
Week Week

Figure 5.1 Subject two average overall results by phase

 130

S. 6 Phase 1 (HFF) S. 6 Phase 2 (LFF)

240 240

230
230
Auditors' Score /300

Auditors' Score /300

220
220
210 Average Average

Linear 210 Linear

200
(Average) (Average)
y = -0.26x + 218.1 y = -0.72x + 223.28
190 200
1 2 3 4 5 5 6 7 8 9
Week Week

S. 6 Phase 3 (HFF) S. 6 Phase 4 (LFF)

260 250

250 240
Auditors' Score /300

Auditors' Score /300

240
230
230
220
220
Average Average
210
210
Linear Linear
200 (Average) 200 (Average)
y = -6.76x + 243.63 y = 30x + 172
190 190
9 10 11 12 13 13 14 15
Week Week

Figure 5.2 Subject six average overall results by phase

 131

Figure 5.3 shows the average of the overall scores for subject eight from all five auditors

including linear trend lines. The overall scores for subject eight would appear to clearly support

the hypothesis, with a positive learning trend occurring in phase two but a negative learning

trend occurring in phase one. Again, however, the fact that subject eight was unable to complete

the study, resulting in scores from only eight weeks, means that it is impossible to know if these

results would have been sustained in the remainder of the study.

S. 8 Phase 1 (HFF) S. 8 Phase 2 (LFF)

180 190

170 180
Auditors' Score /300

Auditors' Score /300

160 170

150 160
Average Average
140 150
Linear Linear
130 (Average) 140 (Average)
y = -9.1x + 183.3 y = 14.54x + 123.8
120 130
1 2 3 4 5 5 6 7 8
Week Week

Figure 5.3 Subject eight average overall results by phase

The overall scores of only three of the eight subjects seem to support the hypothesis that

reduced frequency of feedback would result in higher levels of learning, as evidenced by higher

scores on performances of retention/transfer tests. Additionally, the results of each of these three

subjects provide only relatively weak support of the hypothesis.

 132

Subjects that Contradict the Hypothesis

Figure 5.4 shows the average of the overall scores for subject one from all five auditors

including linear trend lines. Subject ones overall scores clearly contradict the hypothesis with

positive learning trends occurring in the HFF phases, but negative learning trends occurring in

both of the LFF phases.

S. 1 Phase 1 (LFF) S. 1 Phase 2 (HFF)

125 170
160
120
Auditors' Score /300

Auditors' Score /300

150
115 140

110 Average 130 Average

Linear 120
105 Linear
(Average) 110 (Average)
y = -1.32x + 117.28 y = 13.12x + 94.8
100 100
1 2 3 4 5 5 6 7 8 9
Week Week

S. 1 Pahse 3 (LFF) S. 1 Phase 4 (HFF)

140 140

130
Auditors' Score /300

Auditors' Score /300

130

120
120
Average 110 Average

110 Linear Linear

(Average) 100
(Average)
y = -5.44x + 135.84 y = 12.6x + 93.8
100 90
9 10 11 12 13 13 14 15
Week Week

Figure 5.4 Subject one average overall results by phase

 133

Figure 5.5 shows the average of the overall scores for subject three from all five auditors

including linear trend lines. The overall scores for subject three provide less clear contradiction

of the hypothesis than did those of subject one. Similar to the results for subject six, the dramatic

positive learning trend in phase four overrules the less dramatic trends in the previous three

phases, lending mathematical support to the HFF feedback condition of that phase.

S.3 Phase 1 (LFF) S. 3 Phase 2 (HFF)

140 160
130 150
Auditors' Score /300

Auditors' Score /300

120
140
110
130
100 Average Average
120
90
Linear Linear
80 (Average) 110 (Average)
y = 0.08x + 106.76 y = -2.98x + 134.74
70 100
1 2 3 4 5 5 6 7 8 9
Week Week

S. 3 Phase 3 (LFF) S. 3 Phase 4 (HFF)

135 155
145
125
Auditors' Score /300

Auditors' Score /300

135
115 125

105 Average 115 Average

105
Linear Linear
95
(Average) 95 (Average)
y = -4.5x + 122.06 y = 12.4x + 95.2
85 85
9 10 11 12 13 13 14 15
Week Week

Figure 5.5 Subject three average overall results by phase

 134

Figure 5.6 shows the average of the overall scores for subject four from all five auditors

including linear trend lines. The overall scores for subject four are again, less clearly

contradictory to the hypothesis. However, the combined learning trends for the HFF and the

LFF phases reveal that learning trends were higher during phases two and four.

S. 4 Phase 1 (LFF) S. 4 Phase 2 (HFF)

170 190

180
Auditors' Score /300

Auditors' Score /300

165

170
160
Average 160 Average

155 Linear Linear

150
(Average) (Average)
y = 0.62x + 161.26 y = 0.04x + 168.92
150 140
1 2 3 4 5 5 6 7 8 9
Week Week

S. 4 Phase 3 (LFF) S. 4 Phase 4 (HFF)

190 170
180
Auditors' Score /300

Auditors' Score /300

170 165

160
160
150 Average Average
140 155
Linear Linear
130 (Average) (Average)
y = -6.36x + 178.72 y = 1x + 156.8
120 150
9 10 11 12 13 13 14 15
Week Week

Figure 5.6 Subject four average overall results by phase

 135

Figure 5.7 shows the average of the overall scores for subject five from all five auditors

including linear trend lines. The overall scores for subject five reveal dramatic shifts in learning

trends between the two LFF phases (phases two and four). While the learning trends in phases

one and three were less dramatic, they both had positive slopes, contradicting the hypothesis.

S. 5 Phase 1 (HFF) S. 5 Phase 2 (LFF)

160 160

150
150

Auditors' Score /300

Auditors' Score /300

140
140
130
130 Average Average
120
Linear Linear
120 110
(Average) (Average)
y = 6.22x + 125.5 y = -12.2x + 174
110 100
1 2 3 4 5 5 6 7 8 9
Week Week

S. 5 Phase 3 (HFF) S. 5 Phase 4 (LFF)

150 150

140 140
Auditors' Score /300

Auditors' Score /300

130
130
120
120 Average Average
110
110 Linear Linear
(Average) 100 (Average)
y = 0.62x + 123.02 y = 15.2x + 86.133
100 90
9 10 11 12 13 13 14 15
Week Week

Figure 5.7 Subect five average overall results by phase

 136

Figure 5.8 shows the average of the overall scores for subject seven from all five auditors

including linear trend lines. The overall scores for subject seven are weakly contradictory to the

hypothesis due to the relatively steep negative learning trends in both LFF phases. However, the

trend in phase one (HFF) is also steeply negative, weakening the conclusion that the HFF

instruction condition was more beneficial for this subject.

S. 7 Phase 1 (HFF) S. 7 Phase 2 (LFF)

210 190
180
200
Auditors' Score /300

Auditors' Score /300

170
190 160

180 Average 150

Average
140
170 Linear Linear
(Average) 130 (Average)
y = -7.82x + 208.12 y = -5.06x + 173.62
160 120
1 2 3 4 5 5 6 7 8 9
Week Week

S. 7 Phase 3 (HFF) S. 7 Phase 4 (LFF)

200 190

190 180
Auditors' Score /300

Auditors' Score /300

180
170
170
Average 160 Average
160
Linear Linear
150 150
(Average) (Average)
y = 6.76x + 145.4 y = -11.2x + 191.33
140 140
9 10 11 12 13 13 14 15
Week Week

Figure 5.8 Subject seven average overall results by phase

 137

The combined results of the five subjects reported above indicate that it is not possible

from this data to draw the conclusion that the LFF instruction condition is more beneficial to the

acquistion of classical singing techniques for all subjects.

Conclusions

The mixed nature of the results of this experiment make it difficult to draw a strong

conclusion regarding the effect of variations in relative feedback frequency on the

retention/transfer, and therefore learning, of classical singing skills in all subjects. However, one

relevant trend did appear to be supported by the results: subjects with higher levels of

performance ability responded more positively to lower relative feedback frequency instruction

conditions, while subjects with lower levels of performance appeared to respond negatively to

lower relative feedback frequency instruction conditions. Of the eleven scenarios whose results

appeared to support the hypothesis, all but two had average weekly scores above 50/100. These

eleven scenarios had an overall average weekly score of 59.89/100. Additionally, when viewing

the overall scores for each subject as reported above, all three subjects whose results supported

the hypothesis (S. 2, S. 6, and S. 8) achieved average weekly scores above 150/300.

Conversely, of the thirteen scenarios whose results appeared to contradict the hypothesis,

all but three had average weekly scores below 50/100. The overall average weekly score for

these scenarios was only 45.86/100. Similarly, three (S. 1, S. 3, and S. 5) of the five subjects

whose overall scores reported previously in the chapter appeared to contradict the hypothesis

achieved average weekly scores below 150/300. Subjects four and seven achieved average

weekly scores above 150/300, yet their learning trends still appear to contradict the hypothesis.
 138

That having been said, the support for these findings are effected by two key factors.

First, the scores from the auditors assessments of the vocal charateristics of most of the subjects,

while fluctuating from week to week and between phases, appeared not to have significant

improvement over the course of the entire semester. For example, the overall scores for subject

one showed a positive learning trend of only 0.46 points per week. This translates to only an

improvement of only 6.9 points (on a scale of 300 points) over the course of the semester. One

possible explanation for such low score improvement could be found in the fact that the tests

being assessed were at least in part transfer tests. It may be possible, had the tests been purely

retention tests using the songs the subjects were studying throughout the term, that their scores

would have shown more significant improvement.

A second factor that may have effected the findings of this study was the wide variablity

between the scores provided by individual auditors. An attempt was made to even out that

variability by averaging the scores and using those averages to determine the learning trends

reported in chapter 4 and previously in this chapter. However, averaging the scores does not

completely negate the descrepancies. In future research, it may be beneficial to provide the

auditors with examples of various levels of performance in each category, which the auditors

would then be able to refer back to in order to ensure that they were assessing the recordings

against a more consistent standard. These vocal characaterstics will be defined in detail to

ensure that each judge is clear about the aspects of the performance they are rating.

Additionally, it would be beneficial to intersperse repetitions of individual recordings into the

track list for each judge. Doing so would allow the researcher to calculate intra-judge reliability.

This experiments findings, though possibly influenced by the factors listed above, that

subjects with higher levels of performance ability respond more favorably to LFF instruction
 139

conditions lends additional support to Salmonis guidance hypothesis.131 As noted in chapter

two, this hypothesis posits that instructional feedback will be most useful early in practice. As

the learner progresses, however, that instructional feedback will become less beneficial, and

lowering the frequency with which that feedback is provided may prevent the learner from

becoming reliant on the feedback, and thereby increase his/her retention and/or transfer of the

desired skill.

In addition to lending support to the guidance hypothesis, the findings of this experiment

indicate that motor learning theory regarding feedback frequency may indeed be applicable to

the teaching and learning of classical singing technique. It would also stand to reason that

application of motor learning theory regarding other aspects of motor skill acquisition (e.g.

feedback timing and variability of practice) may be appropriate and further research in these

areas is warranted.

131
Alan W. Salmoni, Richard A. Schmidt and Charles B. Walter, Knowledge of Results and
Motor Learning: A Review and Critical Reappraisal, Psychological Bulletin 95, no. 3 (1984):
335-386.
 140

CHAPTER 6

IMPLICATIONS FOR TEACHERS

AND FUTURE RESEARCH

The previous chapters have laid the theoretical groundwork for an experiment studying

the effects of variations in feedback frequency on the retention of classical singing techniques,

presented the methodology for that experiment, and reported and discussed the results of that

experiment. In the following pages, the discussion will turn to how the findings of that research

may or may not impact current teaching trends in private voice studio instruction. Additionally,

the findings of this study carry some implications for what related future research may be

warranted. Those implications will be presented, including a discussion of possible causes for

unexpected outcomes in the current experiment.

Implications for Teachers of Singing

As the results reported and discussed in the previous chapters suggest, sweeping changes

to studio voice instruction techniques are not warranted by the findings of this experiment.

Instead, the findings of this experiment suggest that care must be taken within the studio to

ensure that the instruction techniques being utilized are meeting the needs of each individual

student. That having been said, the findings of this experiment did indicate that students of

differing levels of performance ability may react differently to variations in the frequency with

which feedback is provided during instruction. Therefore, it would be prudent for voice

instructors to consider the needs of their students and, as many instructors already do, vary their

teaching techniques (including feedback frequency) to meet those needs.

 141

Inexperienced Singers

The findings of this experiment indicate that lowering the frequency with which feedback

is provided during the instruction of relatively inexperienced singers may not be beneficial to

those singers. In fact, when working with beginning singers or with singers whose performance

abilities are less developed, these findings indicate that such lowered frequency feedback may

actually be detrimental to the retention and transfer of newly acquired vocal skills. In these

instances, providing more frequent feedback during the early stages of skill acquisition may be

more beneficial to the student.

Because of the often excruciatingly slow pace of acquiring new vocal skills, it is difficult

to say exactly what constitutes the early stages of skill acquisition for new singers. Instead of

associating this stage of learning to sing with a certain amount of time, it is more apt to associate

it with a level of performance ability. In the current study, it appeared that those subjects who

responded negatively to the LFF instruction conditions were those whose scores were less than

50/100. Referring to the instructions given to the auditors at the time of their instruction, a score

of 50 indicated that the performance varied moderately from their perceived standard for a well-

trained voice. Therefore, it may be helpful for instructors to regularly assess the performance of

their students against their standard. As the performances levels of less-skilled students begin to

approximate only moderate variations from the standard, the research here indicates that it would

be time to begin reducing the frequency with which feedback is provided to that student.

Experienced Singers

The findings of this experiment indicate that those subjects who responded most

favorably to the LFF instruction conditions were those whose scores were above 50/100. It
 142

would stand to reason, then, that when teaching students whose performance abilities are

relatively high, lowering the frequency with which feedback is provided would be beneficial to

those students. However, an important distinction should be made here between the overall

performance ability of the student, and his or her current ability to perform a new task. Even if

the overall performance ability of a student is relatively high, he or she may have difficulty

performing a new task given by the instructor. In this case, the important performance ability to

consider is that with which he or she is able to perform the task being trained during that

instruction period. As a new task is introduced, even to a relatively skilled performer, it appears

from these findings, and the guidance hypothesis that they support, that relative feedback

frequency should remain high. Once the students ability to perform the new task begins to

approximate moderate variations from the standard the instructor should begin to reduce the

frequency of feedback.

It may seem that constantly assessing a students ability to perform a task as compared to

a standard would be a cumbersome technique to implement in a studio teachers instruction.

However, most teachers likely already perform such assessments in their current teaching

techniques. As a teacher listens to his or her student sing a vocal exercise or a passage from a

song, he or she is assessing the performance and deciding what elements of the performance

need to be addressed once the performance is complete and how those elements would be most

effectively addressed. The findings of this experiment indicate that, in addition to deciding what

to address and how to address it, the instructor should also decide whether or not feedback

should be given at all, based on the proficiency with which the performance is executed.
 143

Implications for Further Research

The findings of this experiment indicate at least some level of applicability of the

theories pertaining to feedback frequency, developed and researched in other fields, to the field

of classical singing instruction. Further research examining the applicability of other elements of

motor-learning theory to the same field is therefore warranted. The first area in which further

research would seem most appropriate would be that of continuing research examining the effect

of variations in feedback frequency. This research could stem from, and attempt to address,

some of the unexpected outcomes of the current study.

Unexpected Outcomes of the Current Study

Care was taken in the design of this study to ensure that the research conditions closely

approximated those found in typical private voice studios. However, by matching these

conditions, many variables were introduced that may have impacted the outcomes. Indeed, the

results of the current study did exhibit several unexpected outcomes, which may have been

caused by these variables. Two of these outcomes will be discussed below, followed by a

discussion of some of the confounding variables, out of the control of the investigator, which

may have had an impact on the outcomes. Finally, several variations to the study design will be

suggested that may help address some of these variables, leading to results that may be more

widely applicable.

Negative Learning Trends

Perhaps the most unexpected outcome of the research was number of phases whose

learning trends were negative, meaning that the students performance abilities had actually
 144

decreased during that time period. From the design of the study, it was anticipated that there

were would be significant variations in the slopes of the learning trends between phases with

different feedback conditions. However, it was also expected that the majority of the phases

would exhibit learning trends with positive slopes, with phases from one feedback condition

exhibiting learning trends with more steeply positive slopes than the other condition. As it

turned out, over half of the phases exhibited negative learning trends. Additionally, four of the

eight subjects exhibited negative trends in the overall performance across the duration of the

study. These results are disturbing to the instructor who hopes that all students will improve

somewhat over the course of a semester of study. However, the amount of time spent practicing

by the student as well as the level of motivation to progress may have had an impact on these

results.

Lack of Homogeneity among Results

At the onset of the study, it was hoped that the results would be more homogenized in

their support or contradiction of the hypothesis. However, the results were nearly evenly split

between those that supported the hypothesis and those that contradicted it. One explanation

regarding which subjects supported the hypothesis and which did not has been discussed earlier

in terms of the performance ability of the subjects. However, other variables may have had an

impact on these results. Some of the variables that may have impacted the homogeneity of the

results, in addition to the variables listed in the previous paragraph, include the health of the

subjects at the time of the retention tests, the ability of the instructor to provide meaningful

feedback in both LFF and HFF instruction conditions, and the use of the piano by the instructor

throughout the instruction phases.

 145

Student practice time

The amount of time spent by individual subjects varied greatly between subjects.

Additionally, the amount of time spent practicing during a given week, or even during a given

phase, also varied within each subject. For example, the average amount of time spent practicing

varied from less than one hour per week in one subject to nearly four hours per week in another

subject. However, the time spent practicing by the first subject ranged from less than 45 minutes

per week during phase three to nearly an hour and a half per week during phase one. While this

variation in the amount of time spent practicing was out of the control of the researcher, it was

documented throughout the duration of the study, allowing for outside analysis of the results in

relation to practice time. Upon analysis, however, there did not appear to be a clear correlation

between the amount of time spent by a subject practicing during a given phase of the study and

the level of their performance on the retention/transfer tests for the same phase. Changes to the

study, as discussed later in this chapter, may help to minimize the impact of practice time on the

results of a future study.

Subject health at the time of retention tests

Any teacher of voice will likely agree that one of the most frustrating factors affecting the

progress of a student is that students ability (or inability) to stay healthy. The health of

participants may well have impacted the outcomes of the current study. Subjects frequently

reported to the researcher that they were not feeling 100%. However, this happened with such

high frequency that it was not possible to prevent them from performing a retention test every

time. Instead, subjects were only prevented from participating when their ailment was to such an

extent that a notable deterioration in vocal quality was evident to the researcher and these
 146

instances are indicated in the results reported previously. Again, this practice is in keeping with

the conditions found in many voice studios (especially those in which students are receiving

academic credit for participation), wherein students are encouraged to participate in their lessons

in as great an extent as it is healthy to do so. Variations to the study design, similar to those

indicated in the discussion of practice time, may again help to minimize the impact that student

health has on the results of a future study.

Student motivation

Motivation is widely accepted as a major factor in student performance and learning in

every discipline.132 Edward Deci and Richard Ryan note in their 1985 on intrinsic motivation

text that research has indicated that being intrinsically motivated to learn improves the quality

of learning.133 The implication, then, is that a lack of intrinsic motivation to learn will

negatively impact the quality of learning. The current research had no way of measuring the

level of motivation to learn possessed by each subject. Additionally, due to the longitudinal

nature of this research, variations in motivation, which were certain to occur across the duration

of the study, likely had an impact on the results. The single-subject design of this research,

allowing each subject to serve as his or her own control and to not be compared to other subjects,

meant that motivational variance between subjects was less of a concern than was the variance

within each subject as the study progressed. In future research in which subjects are being

compared to one another, it would be most desirable to make attempts to equalize the motivation

132
Richard A. Schmidt, Motor Learning and Performance: A Problem-based Approach, 3rd ed.
(Champaign, IL:Human Kinetics Books, 2004), 191.
133
Edward L. Deci and Richard M Ryan, Intrinsic Motivation and Self-Determination in Human
Behavior, (New York: Plenum Press, 1985), 256.
 147

of all subjects. However, because of the difficulty of measuring and/or influencing intrinsic

motivation, it may be more practical to make attempts to adjust the study design to minimize the

opportunity for the levels of intrinsic motivation to change throughout the course of the study.

Instructors ability to provide meaningful feedback in both feedback conditions

Because this study was carried out in a college voice studio where students were paying

tuition and receiving academic credit for participation, every attempt was made by the researcher

to meet the design parameters of the study while simultaneously providing a meaningful and

productive learning environment in keeping with the academic mission of the college. To this

end, the researcher did not follow a strict script in each lesson, opting instead to attempt to

design the feedback provided in order to meet the needs being presented by the subject at that

moment in time. While this attempt may have been made with the interest of the students

progress in mind, changing the quality or direction of the feedback throughout the lesson may

have had an impact on the results of the study. Some of the changes to the study design

discussed later may again help to mitigate these impacts.

Reliance on piano as form of concurrent feedback during instruction

As is the case in most voice studios, the researcher in the current experiment utilized a

piano during lessons to introduce new vocal exercises and to help guide subjects through difficult

passages in their vocal literature. Assuming (and this may be a larger assumption for some

students than other) that the students were capable of distinguishing when they were singing the

same pitches and rhythms that the piano was playing at that exact time, the piano could be

classified as a source of concurrent feedback, which Richard Schmidt and Gabriele Wulf define
 148

in their 1997 article titled Continuous Concurrent Feedback Degrades Skill Learning:

Implications for Training and Simulation as supplemental information presented to the learner

during the actual action.134 As the title of the article suggests, their study found that learners

who were provided with continuous concurrent feedback during skill acquisition performed less

well on retention tests than did those learners who were not provided with the concurrent

feedback.135 The negative impact of concurrent feedback on learning is likely due, as the

guidance hypothesis would suggest, to the fact that such feedback is overly guiding and produces

in learners a dependency upon feedback.

If the piano used in the current research did indeed act as a source of continuous

concurrent feedback, then it is possible that that feedback negatively impacted the learning of the

participants. Additionally, because the retention/transfer tests were performed without piano

accompaniment, there was a difference between the acquisition and the testing conditions.

While research by Winstein and Schmidt136 (reported in chapter two) indicated that such a

difference in conditions may not have been as big of a factor in performance as had previously

been suggested, attempting to minimize that difference in future research would likely still be

desirable. Controlling the amount and frequency of piano use in further research may help

minimize the unintended impact of this feedback on the results of the study that research.

134
Richard A. Schmidt and Gabriele Wulf, Continuous Concurrent Feedback Degrades Skill
Learning: Implications for Training and Simulation, Human Factors 39, no. 4 (1997): 509.
135
Schmidt and Wulf, 509.
136
Carolee J. Winstein, and Richard A. Schmidt, Reduced Frequency of Knowledge of Results
Enhances Motor Skill Learning, Journal of Experimental Psychology: Learning, Memory, and
Cognition 16, no. 4 (1990): 677-791.
 149

Variation in Auditors Scores

A final uncontrolled variable that may have impacted the results of the current study is

the variation in scores given by the volunteer auditors. For example, the overall (sum) scores for

subject one had a standard deviation of 39.36. In attempt to even out these scores, averages were

taken and reported in the results sections of this paper. However, the magnitude of the standard

deviations may have had in impact on the findings of this research.

Further Feedback Frequency Research

Given the variables listed above that may have impacted the findings of the current

experiment, a second experiment examining the effect of variations in relative frequency of

feedback on the acquisition of classical singing techniques is suggested here in which attempts

are made to control these variables or minimize their impact on the findings. While it was

intentional in the design of the current study to approximate the learning environment found in

many private voice studios, this aspect of the study may have led to a variety of difficulties in

producing clear and generalizable results. Consequently, the experiment proposed here steps

away significantly from the studio environment.

Hypothesis

The hypothesis for this experiment would remain the same as that of the current study

being discussed in this paper, namely that a reduced feedback frequency instruction condition

would be beneficial to the acquisition of a new vocal task, resulting in a more accurate

performance on retention tests of that task by those students who received that type of

instruction.
 150

Subjects

For the best generalizability of the results from this research, a representative sample of

college level voice students, including various ages, abilities, and experience levels would be

needed. Once a significant sample size of subjects was established, subjects would be divided

into one of two feedback condition groups: g-1, g-2.

Task

This experiment would examine the effects of variations in relative frequency of

feedback on the acquisition of a single, new, vocal exercise. The vocal exercise would consist of

a series of ten pitches to be sung in succession. The number of pitches may be varied according

the ability level of the subjects participating, though the same series must be used for all

participants. It seems that ten pitches may be an appropriate level to start with, as it is unlikely

that many of the subjects would be able to repeat that length of a sequence without some level of

practice. However, it is likely that all subjects would be able to make some level of progress in

their ability to perform the sequence given practice and guidance.

Procedure

A recording of the sequence would be played for the subject and then the experiment,

consisting of a skill acquisition phase and a retention test, would begin. During the skill

acquisition phase, subjects would complete 27 attempts to perform the task. Subjects in g-1

would receive feedback after every attempt (100% KR), while subjects in g-2 would receive
 151

feedback after only every third attempt (33% KR).137 Feedback would be in the form of an

additional playing of the desired series, played immediately following each attempt for which

feedback is to be provided. For each attempt, the number of pitches from the sequence that the

subject performed correctly would be recorded.

Following the completion of the 27th attempt, the subjects would be dismissed and

instructed to return in 24 hours for a retention test. Because of the nature of music, it may be

necessary to instruct the subjects to try not to hum, whistle, or otherwise perform the sequence

during those hours, as every time they did so, it would essentially serve as a no-KR attempt,

which could ostensibly skew the results of the retention test. When the subjects returned for the

retention test, they would simply be given the starting pitch of the sequence and asked to sing the

entire sequence. The number of pitches performed correctly would be recorded in the same

manner as during the skill acquisition phase.

Results

Once the scores for both groups have been collected, an analysis of variance (ANOVA)

could be calculated, revealing which group performed more accurately on the skill acquisition

phase attempts. A t-test could be used to calculate the results of the retention test attempts.

Discussion

Simply removing the longitudinal aspect of the research reported in this paper, and

replacing it with the design outlined above, would minimize the impact of many of the variables

discussed previously in this chapter. The effects of practice time and subject health would

137
33% KR was chosen because that level of reduced-frequency feedback has been shown in the
research by Winstein and Schmidt, reported earlier, to be most effective.
 152

immediately be almost completely mitigated (though it is plausible that one or more subjects

would become ill during the 24 hours between the skill acquisition phase and the retention test.

Motivation may not be entirely mitigated, though two aspects of the design would contribute to

lessening its impact on the results. First, because the time span of this study is markedly shorter,

variation in individual motivation levels would be less likely to occur. Second, by ensuring that

a statistically significant sample size was used, the effects of any one individual with low levels

of motivation would be decreased. Additionally, because feedback would only be provided after

an attempt was made and never as concurrent feedback, any reliance on the recording as a form

of concurrent feedback would be completely removed. Finally, because the results would consist

simply of a numerical count of the correct pitches, the need for aural assessment and evaluation,

and therefore any negative impact on findings that may result from it, would also be completely

removed.

Feedback Timing Research

In addition to the frequency with which feedback is provided, related research may be

conducted regarding the impact variations in the timing at which feedback is provided.

Verdolini and Lee discuss the timing in which feedback is provided in terms of concurrent versus

terminal feedback,138 indicating that concurrent feedback, as noted earlier, is detrimental to

learning. Schmidt and Lee take the discussion further to examine whether or not the amount of

138
Katherine Verdolini and Timothy D. Lee. Optimizing Motor Learning in Speech
Interventions, In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine
M. Sapienza and Janina K. Casper (Austin, TX: Pro-ed, 2004), 419.
 153

time by which feedback is delayed has an effect on learning.139 In this discussion, Schmidt and

Lee discuss research conducted in 1990 by Swinnen, Schmidt, and Nicholson,140 the results of

which indicated that subjects who received feedback following a 3.2 second delay after task

completion, scored better on retention tests than did those subjects who received feedback

immediately following task completion. These results were found using a relatively simple arm-

positioning task as the skill being acquired. It would be valuable to test whether or not the same

results could be replicated in the acquisition of a vocal task.

A simple way to design this research would be to conduct a variation of the experiment

outlined above. For this new experiment, the subjects and task could remain the same.

However, a change could be made to the procedure such that feedback timing, and not feedback

frequency, could be varied between the two groups.

Hypothesis

Given the research in other fields described above, it is expected that subjects receiving

delayed feedback will perform more accurately on retention tests, thereby indicating a higher

level of learning took place under the delayed feedback condition during the skill acquisition

phase.

139
Richard A. Schmidt and Timothy D. Lee, Motor Control and Learning: A Behavioral
Emphasis. (Champaign, IL: Human Kinetics, 2005), 347-348.
140
Stephen P. Swinnen and others, Information Feedback for Skill Acquisition: Instantaneous
Knowledge of Results Degrades Learning, Journal of Experimental Psychology: Learning,
Memory, and Cognitions 16, no. 4 (1990): 706-716.
 154

Procedure

For this experiment, feedback frequency would remain the same (i.e. all subjects would

receive feedback after each attempt, or after every third attempt, dependent on which condition

proves to maximize learning in the previous experiment). Similar to the previous experiment, all

subjects would make 27 attempts to perform the pitch series during the skill-acquisition phase

and feedback would consist of an additional playing of an audio recording of the desired series.

Subjects in g-1 would receive feedback immediately following each attempt (within one second)

for which feedback is to be provided. Alternatively, feedback provided to subjects in g-2 would

be delayed by five seconds.141

Results

Again, the number of pitches performed correctly for each attempt would be recorded and

the results would be calculated using the same statistical analyses to reveal which group

performed more accurately during the skill acquisition phase and in the retention test.

Discussion

The feedback timing research described above could have significant impacts on the

teaching techniques within the private voice studio. If the hypothesis is supported by the

research, it would suggest that teachers should consider allowing their students a few seconds

after an attempt is made at performing a vocal task before they begin to provide feedback. By

141
A delay of five seconds was chosen, as opposed to a longer or shorter delay, using the
research of Austerman-Hula, et al. reported in chapter 2 as a model: Shannon N. Austermann
Hula and others, Effects of Feedback Frequency and Timing on Acquisition, Retention, and
Transfer of Speech Skills in Acquired Apraxia of Speech, Journal of Speech, Language, and
Hearing Research 51 (2008), 1088-1113.
 155

doing so, students are allowed time to form their own hypotheses about their performance, as

compared to the desired performance, and mentally come up with ways to test those hypotheses

in their subsequent performance attempts. Studies suggest that the act of forming and testing

their own hypotheses is highly beneficial to student learning142 and the results of this study could

provide support for the generalizability of these findings.

Variability of Practice and Task Distribution Research143

In addition to the effects of various feedback conditions, motor-learning theory has

provided some questions regarding how much time should be spent training any one skill without

a variation being introduced during a given training session and the most effective sequence in

which skills should be trained. Verdolini and Lee refer to these two concerns as variability of

practice and task distribution respectively.144

Variability of Practice

Variability of practice refers to the question of whether it is most effective to train a

certain skill using only one task, which would be referred to as non-variable practice, or to use

more than one task to train the same skill, which would be termed variable practice.145 In their

discussion of variable versus non-variable practice, Verdolini and Lee cite several studies

142
Verdolini and Lee, 422.
143
In this context, the term practice refers not to time spent by the learner rehearsing a skill
outside of the training session, but rather to the time spent training any one task during the
training session.
144
Verdolini and Lee, 422-427.
145
Verdolini and Lee, 422.
 156

conducted by McCracken and Stelmach,146 Shea and Kohl,147 Cleave and Fey,148 and Fey,

Cleave, Long, and Hughes,149 all of which examined the effects of variable practice on the

acquisition of timing and force production tasks. However, Verdolini and Lee note that few, if

any, experiments have been conducted examining these same conditions in the training of vocal

tasks. Given the implications of motor-learning theory on the question of variability of practice,

it seems that a research study designed to examine the effects of variability of practice on the

acquisition of vocal tasks would be highly relevant to the current discussion.

Hypothesis

The findings of the studies listed above found that subjects who trained using non-

variable practice conditions performed more accurately during training, but those subjects who

trained in variable practice conditions consistently performed better on retention tests. Given

these findings, it is expected that non-variable practice conditions, while improving performance

during training, will degrade retention, and therefore learning, of a new vocal task.

146
Hugh D. McCracken and George E. Stelmach, A Test of the Schema Theory of Discrete
Motor Learning, Journal of Motor Behavior 9 (1977): 193-201.
147
Charles H. Shea and Robert M. Kohl, Specificity and Variability of Practice, Research
Quarterly for Exercise and Sport 61, no. 2 (1990): 169-177.
148
Patricia L. Cleave and Marc E. Fey, Two Approaches to Facilitation of Grammar in Children
with Language Impairments: Rationale and Description, American Journal of Speech-Language
Pathology 6 (1997): 22-32.
149
Marc E. Fey and others, Two Approaches to Facilitation of Grammar in Children with
Language Impairment: An Experimental Evaluation, Journal of Speech and Hearing Research
36 (1993): 141-157.
 157

Subjects

For this experiment, a sample of subjects similar to that of the previous two studies can

be used. Subjects will again be divided into two groups: g-1 and g-2.

Task

For this study, subjects will be trained to sing a relatively complex melody. The melody

should be at a level of complexity that ensures that few students will be able to perform it

flawlessly without some training.

Procedure

Similar to the two previous experiments, this study will consist of a skill acquisition

phase and a retention test to be performed 24 hours following the completion of the skill

acquisition phase. During the skill acquisition phase, all subjects will make twenty five attempts

to perform the task, with a recording of the melody played between each attempt. Subjects in

group g-1 will serve as the non-variable practice group and will perform the melody in the same

key for each of the twenty five attempts. Subjects in group g-2, on the other hand, will serve as

the variable practice group, and will perform the melody in five different keys (real imitation)

throughout the twenty five attempts. In other words, for each attempt, the subject will be

randomly assigned one of possible five keys (including the original). The melody will be played

in the new key before the attempt is made. The number of pitches and/or rhythms performed

correctly on each attempt will be recorded for each group. Following the twenty five attempts,

the subjects will be dismissed and asked to return 24 hours later. When the subjects return, the

starting pitch of the melody in the original key will be played and the subjects will be asked to
 158

perform the melody in that key. The number of pitches and/or rhythms performed correctly will

again be recorded.

Results

Again, the number of pitches performed correctly for each attempt would be recorded and

the results would be calculated using the same statistical analyses to reveal which group

performed more accurately during the skill acquisition phase and in the retention test.

Discussion

If the results of the above experiment support the hypothesis, it could again have

implications for studio voice instruction. If variable practice appears to be beneficial to the

learning of a new vocal task, then teachers of singing should evaluate the number of times they

encourage their students to repeat the same exercise or excerpt from vocal literature without

variation. In light of the results of the study, it may be desirable for teachers to limit the number

of times an exercise or excerpt should be repeated before a new task or a variation of that task is

introduced to the training.

Task Distribution

Assuming that a variable practice training condition is favorable, the next question to be

addressed is that of the most effective sequence in which to order that practice. This will be

referred to as task distribution here. The question of how best to order exercises aimed at

different skills throughout the duration of a lesson is of particular importance to the voice

instructor. Specifically, should the exercise aimed at individual skills be blocked together
 159

(blocked practice), or should exercises aimed at different skills be randomly interspersed with

one another (random practice)?

The results of some of the initial research in this area, performed by John Shea and

Robyn Morgan in 1979, required two groups of individuals to make three sequential arm

movements.150 A total of 54 trials of the tasks were completed in three sets of 18 trials (one set

for each of the arm movements. Subjects in the first group completed 18 trials of one movement

before moving to the next movement, representing a blocked practice environment. Subjects in

the other group practiced six trials of each of the three movements in each 18-trial set

representing a random practice environment.151 A retention test consisting of 18 trials (six trials

for each of the three movements) was conducted after a ten-minute delay following the

completion of the skill acquisition phase. The time it took to perform each movement was

recorded throughout the skill acquisition phase and during the retention tests. The results of this

research indicated that those subjects who practiced in a blocked practice environment

consistently performed the arm movements more quickly throughout the skill acquisition phase

of the experiment. However, on the retention test delayed by ten minutes, those subjects who

practiced in the random practice environment consistently performed the arm movements more

quickly.

The difficulty in designing research in this area to be applied to vocal skill acquisition is

that each task being trained will have its own set of appropriate associated skills that are

available to alternate with the desired task. The research study described immediately above is

150
John B. Shea and Robyn L. Morgan, Contextual Interference Effects on the Acquisition,
Retention, and Transfer of a Motor Skill, Journal of Experimental Psychology: Human
Learning and Memory 5, no. 2 (1979): 179-187.
151
Ibid., 181.
 160

of a very basic level of variance in practice related to a single task. However, one could easily

imagine the wide variety of skills, all aimed at improving the overall quality of a voice, that

could be addressed within a single voice lesson. The experiment described below will simply

choose three vocal exercises related to pitch, rhythm, and vocal onset. However, it should be

noted that the results may well be generalized to the variability in practice of all aspects of vocal

technique.

Hypothesis

Given that the results of Shea and Morgans research have been replicated several times

and applied to various skills and tasks,152 it is expected that random practice conditions will

benefit the retention of newly acquired vocal tasks.

Subjects

For this experiment, a sample of subjects similar to that of the previous two studies can

be used. Subjects would again be divided into two groups: g-1 and g-2.

Tasks

In this experiment, subjects would be asked to perform three vocal exercises. In the first

exercise, aimed at pitch control, the subjects would be required to repeat a series of pitches

similar to that used in the research proposed for the feedback experiments. In the second task,

aimed at improving rhythmic skills, students would be asked to repeat a rhythmic sequence.

152
See reviews of research in Schmidt and Lee, 2004, and Richard A. Magill and Kellie G. Hall,
A Review of the Contextual Interference Effect in Motor Skill Acquisition, Human Movement
Science 9 (1990): 241-289.
 161

Finally, a task aimed at controlling the onset of sound production would require the subjects to

produce four repetitions of an [a] vowel on specified pitch. With this task, subjects would be

instructed to produce a balanced onset, somewhere between breathy and glottal.

Procedure

Similar to the design of Shea and Morgans research, this experiment would consist of a

total of 54 trials during the skill acquisition phase: 18 trials of each task. Subjects in group g-1

would serve as the blocked practice group and perform all 18 trials of each task before moving

on to the next task. Subjects in group g-2 would serve as the random practice group and would

perform their trials in such an order that six trials of each task would be included in each 18-trial

set. A retention test, in which subjects would perform six trials of each task, would be

administered 24 hours following the completion of the skill acquisition phases. For the first two

tasks, a simple number of correct pitches or rhythms could be recorded to serve as the results for

each trial. A correct performance of the third task, however, is somewhat more subjective and

may require setting an acceptable threshold of perceived deviation from the desired outcome.

The number of trials performed below that threshold could then be recorded.

Results

The number of correct, or acceptable, trials would be recorded and analyzed using the

same statistical analyses as used in the experiments outlined earlier in this chapter.
 162

Discussion

If the results of this experiment support the hypothesis, indicating that a random practice

learning environment is beneficial to the retention of newly acquired vocal skills, teacher of

singing may again be encouraged to attempt the application of these principles to their own

teaching. For example, during a typical 30-minute vocal technique lesson, instructors could

replicate a random practice environment by altering the order in which exercises aimed at

various vocal skills are prescribed to the student. The application of this concept to a voice

lesson focused on the training of vocal literature, often referred to as a coaching session. In this

instance, a random practice environment could be achieved by spending a relatively small

amount of time on several songs rather than spending the bulk of the time on a single song. For

example, a 30-minute coaching session may be made more effective by spending ten minutes on

each of three songs than by spending the entire 30 minutes coaching a single song.

Perceptual Analysis and Evaluation Research

While removed from the realm of motor-learning theory, the current study has shed more

light on the need for further research into the effectiveness and accuracy of perceptual analysis of

a vocal performance by an auditor or group of auditors. This type of analysis and evaluation is

used regularly at vocal competitions and auditions throughout the industry. Katherine Eberle-

Fink has written on this matter and has suggested the use of a perceptual acoustic assessment

form,153 upon which the assessment form used by the auditors in the current study was loosely

based. With the completion of the auditory assessments required by this research, some data is

153
Katherine Eberle-Fink, Perceptual Acoustic Assessment of Singing, Journal of Singing -
The Official Journal of the National Association of Teachers of Singing 63 no. 1
(September/October 2006): 35-43.
 163

now available to examine the abilities of different auditors to assess the same vocal sample and

come to similar conclusions regarding that samples quality. A relevant bit of research would

now be to examine the five different scores for each vocal sample in terms of the range of scores,

the mean deviation (average deviation from the mean score), and the standard deviation. This

statistical analysis may provide some fodder to prompt further research into the area of

perceptual acoustic analysis and eventually lead to more standardized scoring by auditors and

judges of competitions and auditions.

Conclusion

The field of motor-learning theory has been evolving for nearly a century, yet it continues

to yield new and intriguing findings regarding the most effective practices for training muscles to

produce a desired movement consistently and accurately. The current study examined the effects

of variations in relative frequency of feedback on the retention of classical singing techniques in

a setting closely resembling that of a typical studio voice lesson. The results of this research

indicated weak support of the hypothesis that reducing the relative frequency of feedback would

improve retention. However, this support was only found among those subjects who consistently

had the highest scores from a panel of auditors. Among the remaining subjects with consistently

lower scores, a low-frequency feedback instruction condition appeared to actually be detrimental

to their retention. A number of variables may have contributed to a lack of homogeneity among

the results. In attempt to control for these variables, a revised version of this study has been

suggested. This new study, however, creates an environment which does not resemble a typical

studio voice lesson and therefore is much less likely and/or practical to be recreated by voice

teachers in their own studio instruction. The results of the current research, while not providing
 164

astoundingly strong support for the hypothesis, do indicate some relevant trends that warrant

further research into the application of motor-learning theory to studio voice instruction. It is the

hope of the researcher to carry out some of this research, including the experiments described in

this chapter in the near future.

 165

APPENDIX A

INFORMED CONSENT DOCUMENT

Project Title: Application of Principles from Motor-Learning Theory to the Studio Voice
Lesson: Effects of Feedback Frequency on Retention of Classical Singing Technique

Principal Investigator: Lynn Maxfield

Research Team Contact: Lynn Maxfield

208-406-7748
lmaxfield@monmouth.edu

This consent form describes the research study to help you decide if you want to participate. This
form provides important information about what you will be asked to do during the study, about
the risks and benefits of the study, and about your rights as a research subject.
If you have any questions about or do not understand something in this form, you should
ask the research team for more information.
You should discuss your participation with anyone you choose such as family or friends.
Do not agree to participate in this study unless the research team has answered your
questions and you decide that you want to be part of this study.

WHAT IS THE PURPOSE OF THIS STUDY?

This is a research study. We are inviting you to participate in this research study because you are
a student in Lynn Maxfields voice studio.

The purpose of this research study is to investigate the effectiveness of feedback given by an
instructor during a voice lesson. The hypothesis is that less frequent feedback from the instructor
will actually result in more efficient learning by the student. It is hoped that the information
from this study will lead to more effective and efficient vocal instruction.

HOW MANY PEOPLE WILL PARTICIPATE?

Approximately 12 people will take part in this study at the University of Iowa.

HOW LONG WILL I BE IN THIS STUDY?

If you agree to take part in this study, your involvement will last for approximately 15 weeks.
The study will require your attendance at 16, weekly voice lessons throughout the semester with
two lessons being given during the final week. Each lesson will be approximately 30 minutes.
 166

WHAT WILL HAPPEN DURING THIS STUDY?

If you agree to participate in this study you will first be asked to fill out a brief questionnaire
regarding your previous singing experience and your vocal health. Following the questionnaire,
the researcher will evaluate your ability to match pitch and your tonal memory. The researcher
will play a pitch on the piano and ask you to match that pitch with your voice. You will be asked
to do this five (5) times at pitches across your range. You will then be asked to sing America
the Beautiful a capella (without accompaniment) to evaluate your tonal memory and intonation.

Following this initial evaluation you will receive 16 30-minute voice lessons. These lessons will
be conducted in a similar fashion to most normal voice lessons except that in some of the
lessons, the instructor (Lynn Maxfield) will provide more feedback to the student and in some of
the lessons he will provide less feedback. At the beginning of each lesson, you will make an
audio recording of yourself singing America the Beautiful a capella. Throughout the week in
between lessons, you will be asked to complete at least 2 hours of voice practice.

Audio/Video Recording or Photographs

One aspect of this study involves making audio recordings of you. These recordings will serve
as a measure of your performance ability at individual moments throughout the duration of the
study. Following completion of the 15-week study, these recordings will be evaluated by five
(5) college voice faculty members. These faculty members will not know your name nor will
they know at what point during the study each recording was made. Once the recordings have
been evaluated, they will be destroyed.

[ ] Yes [ ] No I give you permission to make audio recordings of me during this study.

WHAT ARE THE RISKS OF THIS STUDY?

You may experience one or more of the risks indicated below from being in this study. In
addition to these, there may be other unknown risks, or risks that we did not anticipate,
associated with being in this study.

You may feel some discomfort or embarrassment in knowing that your voice will be
evaluated by other voice faculty.
There are no other foreseeable risks to participating.

WHAT ARE THE BENEFITS OF THIS STUDY?

We dont know if you will benefit from being in this study.

 167

However, we hope that, in the future, other people might benefit from this study because the
information gleaned from this study may lead to more effective and efficient techniques for vocal
instruction and training.

WILL IT COST ME ANYTHING TO BE IN THIS STUDY?

You will not have any additional costs for being in this research study.

WILL I BE PAID FOR PARTICIPATING?

You will not be paid for being in this research study.

WHO IS FUNDING THIS STUDY?

The University and the research team are receiving no payments from other agencies,
organizations, or companies to conduct this research study.

WHAT ABOUT CONFIDENTIALITY?

We will keep your participation in this research study confidential to the extent permitted by law.
However, it is possible that other people such as those indicated below may become aware of
your participation in this study and may inspect and copy records pertaining to this research.
Some of these records could contain information that personally identifies you.
federal government regulatory agencies,
auditing departments of the University of Iowa, and
the University of Iowa Institutional Review Board (a committee that reviews and
approves research studies)

To help protect your confidentiality, we will keep any documents containing your personal
information in a secure location and remove your name and any identifying information from
any recording or document before it is distributed to any of the recording evaluators. If we write
a report or article about this study or share the study data set with others, we will do so in such a
way that you cannot be directly identified.

IS BEING IN THIS STUDY VOLUNTARY?

Taking part in this research study is completely voluntary. You may choose not to take part at
all. If you decide to be in this study, you may stop participating at any time. If you decide not to
be in this study, or if you stop participating at any time, you wont be penalized or lose any
 168

benefits for which you otherwise qualify. If you choose not to participate or stop participating at
any time, this decision will in no way effect your final grade for the course, nor will it effect the
level of commitment on the part of the instructor.

Will I Receive New Information About the Study while Participating?

If we obtain any new information during this study that might affect your willingness to continue
participating in the study, well promptly provide you with that information.

Can Someone Else End my Participation in this Study?

Under certain circumstances, the researchers might decide to end your participation in this
research study earlier than planned. This might happen because you have missed two (2) or more
consecutive lessons or you have missed three (3) or more non-consecutive lessons throughout the
term of the study. Your participation in the study may also be ended if you fail to complete two
(2) hours of practice each week.

WHAT IF I HAVE QUESTIONS?

We encourage you to ask questions. If you have any questions about the research study itself,
please contact: Lynn Maxfield, 208-406-7748, lmaxfield@monmouth.edu. If you experience a
research-related injury, please contact Eileen Finnegan, 319-335-8717.

If you have questions, concerns, or complaints about your rights as a research subject or about
research related injury, please contact the Human Subjects Office, 340 College of Medicine
Administration Building, The University of Iowa, Iowa City, Iowa, 52242, (319) 335-6564, or e-
mail irb@uiowa.edu. General information about being a research subject can be found by
clicking Info for Public on the Human Subjects Office web site, http://research.uiowa.edu/hso.
To offer input about your experiences as a research subject or to speak to someone other than the
research staff, call the Human Subjects Office at the number above.

This Informed Consent Document is not a contract. It is a written explanation of what will
happen during the study if you decide to participate. You are not waiving any legal rights by
signing this Informed Consent Document. Your signature indicates that this research study has
 169

been explained to you, that your questions have been answered, and that you agree to take part in
this study. You will receive a copy of this form.

Subject's Name (printed): ________________________________________________________

Do not sign this form if todays date is on or after $STAMP_EXP_DT .

________________________________________ _____________________________
(Signature of Subject) (Date)

Statement of Person Who Obtained Consent

I have discussed the above points with the subject or, where appropriate, with the subjects
legally authorized representative. It is my opinion that the subject understands the risks,
benefits, and procedures involved with participation in this research study.

__________________________________________ _____________________________
(Signature of Person who Obtained Consent) (Date)
 170

APPENDIX B

SUBJECT INTAKE FORM

Subject Number: __________ Date: ____________

Age__________

Previous singing experience

Private Study
Duration and dates______________________________________________________________

Location(s)____________________________________________________________________
_____________________________________________________________________________

Instructor(s)____________________________________________________________________
______________________________________________________________________________

Solo:
Duration and dates______________________________________________________________

Location(s)____________________________________________________________________
_____________________________________________________________________________

Ensemble:
Duration and dates______________________________________________________________

Location(s)____________________________________________________________________
_____________________________________________________________________________

Director(s)_____________________________________________________________________

Vocal Health History

Do you now have, or have you ever had any vocal health issues that may affect your ability
to maintain active participation (i.e. singing daily) in a full semester of voice lessons (nodes,
nodules, polyps, cysts etc.)?
______________________________________________________________________________
______________________________________________________________________________
______________________________________________________________________________

Have you ever lost your voice more than once within the span of six (6) months? Y or N
If yes, please explain: ___________________________________________________________
______________________________________________________________________________
 171

APPENDIX C

SAMPLE VOCAL EXERCISES

Tone Quality Exercises154

and/or

Breath Management Exercises155

and/or

154
Richard Miller, Structure of Singing: System and Art in Vocal Technique, (Belmont, CA:
Wadsorth Group/Thomson Learning) 1996.
155
Ibid.
 172

Intonation Exercises

and/or
 173

APPENDIX D

PERCEPTUAL AUDITORY ASSESSMENT FORM

Assessors Name:________________________________________ Track Number:__________

Please assess the singers performance in each of the following areas on a scale of 1-100 with the
score of 100 indicating a performance completely meeting your perceived standard for a well-
trained voice. For example, a score of 50 would indicate that the performance deviated
moderately from your standard for that area. A score of 0 would indicate that the performance
deviated severely from your standard for that area. For your convenience, a line scale has been
provided for each area, on which you may easily make a tic mark indicating score as you listen.
Once you have finished listening, please indicate an exact numerical score in the space provided
to the right of the scale.

Tone Quality __________________________________________ ___/100

0 50 100

Breath Management __________________________________________ ___/100

0 50 100

Intonation __________________________________________ ___/100

0 50 100
 174

BIBLIOGRAPHY

Adams, Jack A. A Closed-Loop Theory of Motor Learning. Journal of Motor Behavior 3

(1971): 111-150.

Adams, Jack A. Theoretical Issues for Knowledge of Results. in Information Processing in

Motor Control and Learning. ed. George. E. Stelmach. New York: Academic Press.
1978.

Annett, John. Feedback and Human Behavior. Middlesex, England: Penguin Books. 1959.

Bilodeau, Edward.A., Ima McD Bilodea, and David A. Schumsky, Some Effects of Introducing
and Withdrawing Knowledge of Results Early and Late in Practice. Journal of
Experimental Psychology 58 (1959): 142-144.

Bilodeau, Ina M. Information Feedback. in The Acquisition of Skill. ed. Edward A. Bilodeau.
New York: Academic Press, 1966.

Cleave, Patricia L. and Marc E. Fey. Two Approaches to Facilitation of Grammar in Children
with Language Impairments: Rational and Description. American Journal of Speech-
Language Pathology 6 (1997): 22-32.

Deci, Edward L. and Richard M. Ryan. Intrinsic Motivation and Self-Determination in Human
Behavior. New York: Plenum Press, 1985.

Eberle-Fink, Katherine. Perceptual Acoustic Assessment of Singing. Journal of Singing The

Official Journal of the Nation Association of Teachers of Singing 63, no. 1
(September/October 2006): 35-43.

Ellwell, J.L. and G. C. Grindley. The Effect of Knowledge of Results on Learning and
Performance. British Journal of Psychology 29 (1938): 39-54.

Evans, S.H. A Brief Statement of Schema Theory. Psychonomic Science 8 (1967): 87-88.

Ferrand, Carole T. Effects of Practice With and Without Knowledge of Results on Jitter and
Shimmer Levels in Normally Speaking Women. Journal of Voice 9 (1995): 491-423.

Fey, Marc E., Patricia L. Cleave, Steven H. Long, and Diana L. Hughes Two Approaches to
Facilitation of Grammar in Children with Language Impairment: An Experimental
Evaluation. Journal of Speech and Hearing Research 36 (1993): 141-157.

Folkins, J. W. Issues in Speech Motor Control and Their Relation to the Speech of Individuals
with Cleft Palate. Cleft Palate Journal 22, no. 2 (1985): 106-122.
 175

Helding, Lynn. Voice Science and Vocal Art, Part Two: Motor Learning Theory. Journal of
Singing 64, no. 4 (March/April 2008): 417-428.

Ho, Linda, and John B. Shea. Effects of Relative Frequency of Knowledge of Results on
Retention of a Motor Skill. Perceptial Motor Skills 46 (1978): 859-866.

Hula, Shannon N. Austermann, Donald A. Robin, Edwin Maas, Kirrie J. Ballard, and Richard A.
Schmidt. Effects of Feedback Frequency and Timing on Acquisition, Retention, and
Transfer of Speech Skills in Acquired Apraxia of Speech. Journal of Speech, Language,
and Hearing Research 51 (2008): 1088-1113.

Kearns, Kevin P. Flexibility of Single-Subject Experimental Designs Part II: Design Selection
and Arrangement of Experimental Phases. Journal of Speech and Hearing Disorders 51
(August 1986): 204-214.

Kelso, J. A., and B. Tuller. A Cynamical Basis for Action Systems. in Handbook of Cognitive
Neuroscience. ed. M. Gazzaniza. New York: Plenum Press, 1984, 321-356.

Lavery, J.J. and F. H. Suddon. Retention of Simple Motor Skills as a Function of the Number of
Trials by Which KR is Delayed. Perceptual and Motor Skills 15 (1962): 231-237.

Lee, Timothy D., Margaret A. White and Heather Carnahan. On the Role of Knowledge of
Results in Motor Learning: Exploring the Guidance Hypothesis. Journal of Motor
Behavior 22, no. 2 (1990): 191-208.

Magill, Richard.A. and C.A. Wood. Knowledge of Results Precision as a Learning Variable in
Motor Skill Acquisition. Research Quarterly for Exercise and Sport 57 (1986): 170-173.

Magill, Richard A. and Kellie G. Hall. A Review of the Contextual Interference Effect in Motor
Skill Acquisition. Human Movement Science 9 (1990): 241-289.

McCracken, Hugh D. and George E. Stelmach. A Test of the Schema Theory of Discrete Motor
Learning. Journal of Motor Behavior 9 (1977): 193-201.

Miller, Richard R. Structure of Singing: System and Art in Vocal Technique Belmont, CA:
Wadsorth Group/Thomson Learning, 1996.

Salmoni, Alan W., Richard A Schmidt, and Charles B. Walter. Knowledge of Results and
Motor Learning: A Review and Critical Reappraisal. Psychological Bulletin 95, no.3
(1984): 335-386.

Schmidt, Richard A. Past and Future Issues in Motor Programming. Research Quarterly for
Exercise and Sport 51 (1980): 122-140.

Schmidt, Richard A. Schema Theory: Implications for Movement Education. Motor Skills:
Theory into Practice 2 (1977): 36-38.
 176

Schmidt, Richard A. A Schema Theory of Discrete Motor Skill Learning. Psychological

Review 82 (1975): 225-260.

Schmidt, Richard A. Control Processes in Motor Skills. Exercise and Sport Science Review 4
(1976): 229-261

Schmidt, Richard A. Movement Education and the Schema Theory. In Report of the 1976
Conference: National Association for Physical Education of College Women held in
Cedar Falls, IA 3-8 June 1976, ed. E. Crawford.

Schmidt, Richard A. and Gabriele Wulf. Continuous Concurrent Feedback Degrades Skill
Learning: Implications for Training and Simulation. Human Factor 39, no. 4 (1997):
509-525.

Schmidt, Richard A. and Timothy D. Lee. Motor Learning and Performance: From Principles to
Practice. Champaign IL: Human Kinetics Books, 1991.

Schmidt, Richard A. Motor Control and Learning: A Behavioral Emphasis 3rd Ed. Champaign,
IL: Human Kinetics Books, 1999.

Schmidt, Richard A. Motor Learning and Performance. 3rd ed. Champaign, IL: Human Kinetics
Books, 2004.

Schmidt, Richard A. Motor Learning and Performance: A Problem-Based Approach. 3rd Ed.
Champaign, IL: Human Kinetics Books, 2004.

Schmidt, Richard A. Motor Learning and Performance: From Principles to Practice.

Champaign, IL: Human Kinetics Books, 1991.

Schmidt, Richard A., Claudia Lang, and Douglas E. Young. Optimizing Summary Knowledge
of Results for Skill Learning. Human Movement Science 9 (1990): 325-348.

Shea, Charles H. and Robert M. Kohl. Specificity and Variability of Practice. Research
Quarterly for Exercise and Sport 61, no. 2 (1990): 169-177.

Shea, John B. and Robyn L. Morgan. Contextual Interference Effects on the Acquisition,
Retention, and Transfer of a Motor Skill. Journal of Experimental Psychology: Human
Learning and Memory 5, no.2 (1979): 179-187.

Sherwood, D. E. Effect of Bandwidth Knowledge of Results on Movement Consistency.

Perceptual and Motor Skills 66 (1988): 535-542.

Steinhauer, Kimberly, and Judith Preston Grayhack. The Role of Knowldege of Results in
Performacne and Learning of a Voice Motor Task. Journal of Voice 14, no. 2 (2000):
137-145.
 177

Swinnen, Stephan P., Richard A. Schmidt, Diane E. Nicholson, and Diane C. Shapiro.
Information Feedback for Skill Acquisition: Instantaneous Knowledge of Results
Degrades Learning. Journal of Experimental Psychology: Learning, Memory, and
Cognition 16, no. 4 (1990): 706-716.

Thorndike, Edward L. Education Psychology: Briefer Course. New York: Teacher College,
Columbia University. 1923.

Titze, Ingo R. Principles of Voice Production. Iowa City, IA: NCVS, 2000.

Tolman, Edward C. Purposive Behavior of Animals and Men. New York: Century, 2003.

Varder-Linden, Darl W., James H. Cauraugh, and Tracy A. Greene. The Effect of Frequency of
Kinetic Feedback on Learning an Isometric Force Production Task in Nondisabled
Subjects. Physical Therapy 73, no. 2 (1993): 79-87.

Verdolini, Katherine and Timothy D. Lee. Optimizing Motor Learning in Speech

Interventions. in Vocal Rehabilitation for Medical Speech-Language Patholody. ed.
Christine M. Sapienza and Janina K. Casper. Austin, TX: Pro-ed, 2004.

Verdolini, Katherine. Motor Learning Principles: How to Train. in Vocology by Ingo R. Titze
and Katherine Verdolini. Currently in press: used by permission.

Winstein, Carolee J., and Richard A. Schmidt. Reduced Frequency of Knowledge of Results
Enhances Motor Skill Learning. Journal of Experimental psychology: Learning,
Memory, and Cognition 16, no. 4 (1990): 677-691.