2011
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.
by
An Abstract
May 2011
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
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.
__________________________________________________________
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
May 2011
CERTIFICATE OF APPROVAL
___________________________
PH.D. THESIS
____________
_______________________________
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 3 METHODOLOGY.. 50
Hypothesis.. 50
Theoretical Framework...51
Subjects...51
Instrumentation... 53
Skill Acquisition Phase...54
Retention/Transfer Tests 56
Expected Results 58
Subject One 61
Subject Two 68
Subject Three.. 75
Subject Four72
Subject Five 89
Subject Six.. 96
Subject Seven... 103
Subject Eight.110
iv
Interpreting the Mixed Results. 125
Initial Summary of Results... 126
Variation of Skills Being Trained.128
Conclusions...137
BIBLIOGRAPHY 174
v
LIST OF TABLES
Table
vi
4.21 Subject seven intonation results. 108
vii
LIST OF FIGURES
Figure
viii
4.20 Subject four average tone quality results by phase.. 84
ix
4.43 Subject eight tone quality results... 111
x
1
CHAPTER 1
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.
skills were fleeting in nature, he had not actually learned the behavior. Rather he had seen a
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.
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
In the 1960s and 70s, the majority of theories pertaining to motor-learning were what
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
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
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
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
5. Human performers can be aware of whether or not their motor movement was correct, often
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
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
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
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
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
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
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
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
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
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
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
24
Adams, 126.
25
Ibid., 126.
26
Ibid., 125.
27
Ibid., 124.
9
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
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
Proprioceptive Exteroceptive KR KP
Muscles Vision
Audition
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.
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
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
56
Schmidt, 2004, 283-306.
17
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
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
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
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
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
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
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
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
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
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
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
instrumentation.84 For the purpose of the following study, knowledge of results was
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
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
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
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
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.
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
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
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
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
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
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
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
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
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)
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
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
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
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
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
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)
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.
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
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
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
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
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
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
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
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
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
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
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
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
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
These results indicate that, while concurrent feedback may provide for immediate
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
in performance ability. That is to say, less frequent feedback is more effective at facilitating
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
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).
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
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
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
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
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
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
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
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
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
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.
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
Hypothesis
The study outlined in the following pages tested the hypothesis that decreasing the
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
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
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
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
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
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
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
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
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
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
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
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
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,
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
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
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
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
The overall results for the tone quality of subject one are reported below in table 4.1 and
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
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
55 55
Auditors' scores / 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
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
The overall results of the breath management for subject one are listed below in the chart
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
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.
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
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
The overall results for intonation for subject one presented are in the chart in table 4.3
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
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 4.6 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.
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
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
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
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
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
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
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
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.
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
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.
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
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 4.10 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.
65
Auditors' Score /100
65
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
65 65
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
The aggregate scores for the intonation of subject two from all five auditors are presented
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
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.
75 75
Auditors' Score /100
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
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
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
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
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.
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
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
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.
55 55
Auditors' Score /100
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
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
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.
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
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 4.16 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.
50
45
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
50 50
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
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.
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
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.
40 50
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
35 40
Auditors' 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
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
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
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
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
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
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.
50
Auditors' Score /100
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
55
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
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.
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
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.
55 55
Auditors' Score /100
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
55 55
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
The aggregate scores for the intonation of subject four from all five auditors are presented
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
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 4.24 presents the average of the scores for intonation from all five auditors,
divided into phases and including a linear trend line for each phase.
75
Auditors' Score /100
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
75
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
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
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.
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
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
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 4.26 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.
55 50
Auditors' Score /100
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
45 40
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
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.
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
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 4.28 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.
55 55
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
45 45
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
The aggregate scores for the intonation of subject five from all five auditors are presented
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
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.
65 65
Auditors' Score /100
60 55
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
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
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
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
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
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
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.
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
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
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.
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
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 4.34 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.
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
85
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
The aggregate scores for the intonation of subject six from all five auditors are presented
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
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 4.36 presents the average of the scores for intonation from all five auditors,
divided into phases and including a linear trend line for each phase.
75
Auditors' Socre /100
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
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
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
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.
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
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
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 4.38 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.
75 75
Auditors' Score /100
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
75 75
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
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.
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
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 4.40 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.
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
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.
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
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 4.42 presents the average of the scores for intonation from all five auditors,
divided into phases and including a linear trend line for each phase.
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
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
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
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.
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
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 4.44 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.
65
Auditors' Score /100
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
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.
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
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 4.46 presents the average of the scores for intonation from all five auditors,
divided into phases and including a linear trend line for each phase.
65
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
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.
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
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 4.48 presents the average of the scores for intonation from all five auditors,
divided into phases and including a linear trend line for each phase.
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
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
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
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
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
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
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
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.
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
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
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.
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
The remaining fifteen scenarios presented results that were mixed, meaning that no clear
visible in these cases. Each of these scenarios will be discussed below, followed by a discussion
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.
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
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.
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.
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
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.
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.
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
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.
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.
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
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
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.
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.
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
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)
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
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
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
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
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.
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
195
190
Auditors' Score /300
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
230
230
Auditors' Score /300
260 250
250 240
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 5.3 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
170 180
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
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
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
130
Auditors' Score /300
130
120
120
Average 110 Average
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.
135
115 125
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.
180
Auditors' Score /300
170
160
Average 160 Average
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 5.7 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.
150
150
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
140 140
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 5.8 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
190 180
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
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
Conclusions
The mixed nature of the results of this experiment make it difficult to draw a strong
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
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.
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
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
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
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
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
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
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
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,
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.
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.
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
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
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
Student motivation
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.
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
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
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
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
Given the variables listed above that may have impacted the findings of the current
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
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
Task
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
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
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
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.
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
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
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
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
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
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
Variability of Practice
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
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
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
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
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
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
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
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.
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
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
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
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
APPENDIX A
Project Title: Application of Principles from Motor-Learning Theory to the Studio Voice
Lesson: Effects of Feedback Frequency on Retention of Classical Singing Technique
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.
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.
Approximately 12 people will take part in this study at the University of Iowa.
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
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.
[ ] Yes [ ] No I give you permission to make audio recordings of me during 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.
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.
You will not have any additional costs for being in this research study.
The University and the research team are receiving no payments from other agencies,
organizations, or companies to conduct this research study.
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.
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.
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.
________________________________________ _____________________________
(Signature of Subject) (Date)
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
Age__________
Location(s)____________________________________________________________________
_____________________________________________________________________________
Instructor(s)____________________________________________________________________
______________________________________________________________________________
Solo:
Duration and dates______________________________________________________________
Location(s)____________________________________________________________________
_____________________________________________________________________________
Ensemble:
Duration and dates______________________________________________________________
Location(s)____________________________________________________________________
_____________________________________________________________________________
Director(s)_____________________________________________________________________
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
and/or
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
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.
BIBLIOGRAPHY
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.
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. 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., 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.
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. 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.