Application of principles from motor-learning theory to ...

University of Iowa

Iowa Research Online Theses and Dissertations

2011

Application of principles from motor-learning theory to the studio voice lesson: effects of feedback frequency on retention of classical singing technique Lynn Milo Maxfield University of Iowa

Follow this and additional works at: http://ir.uiowa.edu/etd Part of the Music Commons 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." doctoral dissertation, University of Iowa, 2011. http://ir.uiowa.edu/etd/1021.

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

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

by Lynn Milo Maxfield

An Abstract Of a thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Music in the Graduate College of The University of Iowa

May 2011

Thesis Supervisor: Associate Professor John Muriello

1 ABSTRACT

Over the past several decades, cognitive and behavioral scientists have been researching the most effective practices for training muscles to produce specific movements consistently and accurately. That research has led to relatively wide acceptance of several best practices for the training of motor skills. One such practice is the reduction in the frequency with which augmented (external) feedback is provided by the instructor/trainer during skill-acquisition. This theory of low-frequency feedback has been examined by research in a wide variety of fields ranging from exercise and sport to voice therapy and rehabilitation. Prior to the study reported here, however, this theory had not been applied to the acquisition of vocal skills associated with classical singing techniques. The current research consisted of an alternating treatment singlesubject 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 highfrequency feedback (HFF) instruction condition and a low-frequency feedback (LFF) instruction condition. At the beginning of each session, a vocal sample was recorded to test the retention of the skills trained in the previous lesson. Those recordings were evaluated by a panel of five college voice instructors who provided a numerical score (out of a possible 100 pts.) for each sample on the basis of tone quality, breath management, and intonation. The results of this study indicated that three of the eight subjects retained more vocal skill ability during the LFF phases of the study, while the remaining five subjects retained less vocal skill ability during the LFF phases of the study. It was also seen that the three subjects who responded favorably to the LFF instruction condition were also those whose scores were higher throughout the duration of the

2 study. These findings would appear to indicate that an LFF instruction condition may be more beneficial to more experienced or more skilled singers, while an HFF instruction condition may be more beneficial to more novice singers. In the final chapter of this report, several modifications to this study are suggested along with suggestions for future research regarding the application of other principles from motor-learning theory to the acquisition of new vocal skills.

Abstract Approved:

__________________________________________________________ Thesis Supervisor __________________________________________________________ Title and Department __________________________________________________________ Date

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

by Lynn Milo Maxfield

A thesis submitted in partial fulfillment of the requirements for the Doctor of Philosophy degree in Music in the Graduate College of The University of Iowa

May 2011

Thesis Supervisor: Associate Professor John Muriello

Graduate College The University of Iowa Iowa City, Iowa CERTIFICATE OF APPROVAL ___________________________ PH.D. THESIS ____________ This is to certify that the Ph. D. thesis of Lynn Milo Maxfield has been approved by the Examining Committee for the thesis requirement for the Doctor of Philosophy degree in Music at the May 2011 graduation Thesis Committee: _______________________________ John Muriello, Thesis Supervisor _______________________________ Eileen Finnegan _______________________________ Katherine Eberle-Fink _______________________________ David Puderbaugh _______________________________ Stephen Swanson

To Ellen

ii

ACKNOWLEDGMENTS

I would like to thank Dr. John Muriello, Dr. Eileen Finnegan, and Dr. Katherine EberleFink for their time, patience, and expert advice in the preparation of this thesis. Special thanks are extended to Dr. Muriello for his keen eye in reading and proofing chapter drafts. I would also like to thank Dr. Katherine Verdolini and Dr. Ingo Titze for their willingness to allow me to read and cite material from their as yet unpublished text on vocology, and the music faculty and administration at Monmouth College for allowing me to use their facilities to conduct this research. Additionally, I would like to thank my entire family for their seemingly endless encouragement. Finally, I must thank my wife, Ellen, for her love and supporting encouragement throughout my academic career and especially through the preparation of this thesis.

iii

TABLE OF CONTENTS LIST OF TABLES……………………………………………………………………….. vi LIST OF FIGURES……………………………………………………………………...viii CHAPTER 1 INTRODUCTION TO PRINCIPLES OF MOTOR LEARNING………… 1 Jack A. Adams‟ Closed-Loop Theory……………………………………………….. 2 Richard A. Schmidt‟s Schema Theory………………………………………………. 9 Motor Learning Defined and the Role of Feedback………………………………... 12 Application to Voice Science and Need for the Current Study……………………. 23 CHAPTER 2 LITERATURE REVIEW………………………………………………… 27 Feedback Frequency and Simple Arm-Positioning/Timing Motor Tasks…………..28 Feedback Frequency and the Acquisition of Voice Tasks…………………………. 42 Conclusion………………………………………………………………………….. 48 CHAPTER 3 METHODOLOGY……………………………………………………….. 50 Hypothesis………………………………………………………………………….. 50 Theoretical Framework……………………………………………………………...51 Subjects……………………………………………………………………………...51 Instrumentation……………………………………………………………………... 53 Skill Acquisition Phase……………………………………………………………...54 Retention/Transfer Tests…………………………………………………………… 56 Expected Results…………………………………………………………………… 58 CHAPTER 4 REPORTING OF RESULTS……………………………………………...59 Subject One………………………………………………………………………… 61 Subject Two………………………………………………………………………… 68 Subject Three……………………………………………………………………….. 75 Subject Four…………………………………………………………………………72 Subject Five………………………………………………………………………… 89 Subject Six………………………………………………………………………….. 96 Subject Seven……………………………………………………………………... 103 Subject Eight……………………………………………………………………….110 CHAPTER 5 DISCUSSION OF RESULTS……………………………………………116 Results That Support the Hypothesis……………………………………………... 116 Results That Do Not Support the Hypothesis……………………………………...118 Results That Are Mixed……………………………………………………………120 iv

Interpreting the Mixed Results……………………………………………………. 125 Initial Summary of Results………………………………………………………... 126 Variation of Skills Being Trained………………………………………………….128 Conclusions………………………………………………………………………...137 CHAPTER 6 IMPLICATIONS FOR TEACHERS AND FUTURE RESEARCH…….140 Implications for Teachers of Singing……………………………………………... 140 Implications for Future Research…………………………………………………. 143 Conclusion………………………………………………………………………… 163 APPENDIX A INFORMED CONSENT DOCUMENT………………………………. 165 APPENDIX B SUBJECT INTAKE FORM…………………………………………… 170 APPENDIX C SAMPLE VOCAL EXERCISES……………………………………… 171 APPENDIX D PERCEPTUAL AUDITORY ASSESSMENT FORM………………... 173 BIBLIOGRAPHY……………………………………………………………………… 174

v

LIST OF TABLES

Table 3.1 – Demographics of eight study participants………………………………………..53 4.1 – Subject one tone quality results…………………………………………………. 62 4.2 – Subject one breath management results…………………………………………. 64 4.3 – Subject one intonation results…………………………………………………… 66 4.4 – Subject two tone quality results…………………………………………………. 69 4.5 – Subject two breath management results…………………………………………. 71 4.6 – Subject two intonation results…………………………………………………… 73 4.7 – Subject three tone quality results………………………………………………... 76 4.8 – Subject three breath management results………………………………………... 78 4.9 – Subject three intonation results………………………………………………… 80 4.10 – Subject four tone quality results……………………………………………….. 83 4.11 – Subject four breath management results……………………………………….. 85 4.12 – Subject four intonation results…………………………………………………. 87 4.13 – Subject five tone quality results………………………………………………... 90 4.14 – Subject five breath management results……………………………………….. 92 4.15 – Subject five intonation results………………………………………………….. 94 4.16 – Subject six tone quality results………………………………………………… 97 4.17 – Subject six breath management results………………………………………… 99 4.18 – Subject six intonation results…………………………………………………..101 4.19 – Subject seven tone quality results…………………………………………….. 104 4.20 – Subject seven breath management results……………………………………. 106 vi

4.21 – Subject seven intonation results………………………………………………. 108 4.22 – Subject eight tone quality results……………………………………………... 111 4.23 – Subject eight breath management results……………………………………... 112 4.24 – Subject eight intonation results……………………………………………….. 114 5.1 – Scenarios supporting the hypothesis…………………………………………… 126 5.2 – Scenarios contradicting the hypothesis………………………………………… 126

vii

LIST OF FIGURES

Figure 1.1 – Categories of feedback receive by the learner…………………………………... 15 2.1 – Subject performance on acquisition-phase and retention trials…………………. 41 4.1 – Subject one tone quality results…………………………………………………. 62 4.2 – Subject one average tone quality results by phase………………………………. 63 4.3 – Subject one breath management results…………………………………………. 64 4.4 – Subject one average breath management results by phase……………………… 65 4.5 – Subject one intonation results…………………………………………………… 66 4.6 – Subject one average intonation results by phase…………………………………67 4.7 – Subject two tone quality results…………………………………………………. 69 4.8 – Subject two average tone quality results by phase……………………………….70 4.9 – Subject two breath management results…………………………………………. 71 4.10 – Subject two average breath management results by phase…………………….. 71 4.11 – Subject two intonation results………………………………………………….. 73 4.12 – Subject two average intonation results by phase………………………………. 74 4.13 – Subject three tone quality results………………………………………………. 76 4.14 – Subject three average tone quality results by phase…………………………….77 4.15 – Subject three breath management results………………………………………. 78 4.16 – Subject three average breath management results by phase…………………… 79 4.17 – Subject three intonation results………………………………………………… 80 4.18 – Subject three average intonation results……………………………………….. 81 4.19 – Subject four tone quality results……………………………………………….. 83 viii

4.20 – Subject four average tone quality results by phase…………………………….. 84 4.21 – Subject four breath management results……………………………………….. 85 4.22 – Subject four average breath management results by phase……………………..86 4.23 – Subject four intonation results…………………………………………………. 87 4.24 – Subject four average intonation results by phase………………………………. 88 4.25 – Subject five tone quality results………………………………………………... 90 4.26 – Subject five average tone quality results by phase……………………………...91 4.27 – Subject five breath management results………………………………………...92 4.28 – Subject five average breath management results by phase…………………….. 93 4.29 – Subject five intonation results………………………………………………….. 94 4.30 – Subject five average intonation results………………………………………… 95 4.31 – Subject six tone quality results………………………………………………… 97 4.32 – Subject six average tone quality results by phase……………………………… 98 4.33 – Subject six breath management results………………………………………… 99 4.34 – Subject six average breath management results by phase…………………… 100 4.35 – Subject six intonation results…………………………………………………..101 4.36 – Subject six average intonation results by phase………………………………. 102 4.37 – Subject seven tone quality results…………………………………………….. 104 4.38 – Subject seven average tone quality results by phase…………………………..105 4.39 – Subject seven breath management results……………………………………..106 4.40 – Subject seven average breath management results by phase…………………. 107 4.41 – Subject seven intonation results………………………………………………. 108 4.42 – Subject seven average intonation results by phase…………………………… 109

ix

4.43 – Subject eight tone quality results……………………………………………... 111 4.44 – Subject eight average tone quality results by phase…………………………...112 4.45 – Subject eight breath management results……………………………………... 113 4.46 – Subject eight average breath management results by phase………………….. 113 4.47 – Subject eight intonation results……………………………………………….. 114 4.48 – Subject eight average intonation results by phase……………………………. 115 5.1 – Subject two overall average results by phase…………………………………...129 5.2 – Subject six overall average results by phase…………………………………… 130 5.3 – Subject eight overall average results by phase………………………………….131 5.4 – Subject one overall average results by phase…………………………………... 132 5.5 – Subject three overall average results by phase………………………………….133 5.6 – Subject four overall average results by phase………………………………….. 134 5.7 – Subject five overall overage results by phase…………………………………...135 5.8 – Subject seven overall average results by phase………………………………... 136

x

1 CHAPTER 1 INTRODUCTION TO PRINCIPLES OF MOTOR LEARNING

Frank is a 30 year old tenor who has been studying classical singing at a post-secondary level for the past 12 years. He has a capable voice, but has experienced difficulty in maintaining consistency in his range and tone quality. He has studied with four different voice teachers and performed in several master classes. Frequently, he has experienced elation in a lesson or master class as a tidbit of instruction from the teacher has allowed him to perform exercises or passages that he had previously thought were above his abilities. Just as frequently, however, that elation was quickly washed away as he returned to the practice room only to find that his new-found ability was just as unattainable as it had previously been. In Frank‟s case, it is reasonable to determine that, because his recently-acquired vocal skills were fleeting in nature, he had not actually “learned” the behavior. Rather he had seen a momentary improvement in his performance ability. Frank‟s experience is certainly not uncommon, and teachers of singing and students alike can undoubtedly readily recall multiple examples of similar situations. The task for instructors, then, is to determine how they can best structure the learning environment in their studio to facilitate the development of relatively stable, long-term improvements in the abilities of their students, which are the very definition of learning. Theories pertaining to how humans process information in order to produce long-term changes in skilled behavior have been developing for centuries and continue to develop as new information becomes available. In the 20th century, educational theorists and psychologists such as John Dewey, Robert Gagne, Jean Piaget, Burrhus Skinner, Lev Vygotsky, and Edward

2 Thorndike1 all contributed to a mountainous volume of research and theory regarding human learning and understanding, much of which is still referenced today in teacher education programs around the world. Since learning to sing involves the coordination of hundreds of muscles throughout the body, those theories that pertain to the learning of motor behavior are of particular interest to the studio voice instructor. The research study reported in the following chapters has examined two particularly paramount theories, those put forth by Jack A. Adams in 19712 and by Richard Schmidt in 1975.3 Additionally, subsequent research and publications furthering the understanding of motor-learning theory are included in attempt to bring further clarity to the question of how frequently instruction (feedback) should be provided to students during the learning process. Adams‟ and Schmidt‟s theories are summarized below, followed by a more thorough discussion of the qualities of feedback.

Jack A. Adams’ Closed-Loop Theory In the 1960s and 70s, the majority of theories pertaining to motor-learning were what Adams considered to be “open-loop” theories4 based primarily on understandings developed

1

Edward L. Thorndike, Education Psychology: Briefer Course, (New York: Teacher College, Columbia University 1923). 2

Jack A. Adams, “A Closed-Loop Theory of Motor Learning.” Journal of Motor Behavior 3 (1971): 111-150. 3

Richard A. Schmidt, “A Schema Theory of Discrete Motor Skill Learning,” Psychological Review 82 (1975): 225-260. 4

Adams, 117.

3 early in the 19th century.5 Chief among these were Thorndike‟s “Empirical Law of Effect,” which states, “To the situation, „a modifiable connection being made by him between [a stimulus] and [a response] and being accompanied or followed by a satisfying state of affairs,‟ man responds, other things being equal, by an increase in the strength of that connection.”6 In other words, following a response with a reward will increase the likelihood of repeating that response and, inversely, following a response with punishment will decrease the likelihood of repeating that response.7 Adams took issue with applying this simplified stimulus-response approach to an overarching understanding of all skilled-behavior acquisition. While most of the research completed by Thorndike and his associates in support of his theory was completed using animal subjects, Adams points out that in many key ways animal behavior is dissimilar to human behavior. The difference between animal and human behavior form the backbone of Adams‟ concerns, which he articulated with five shortcomings of the Empirical Law of Effect.8 1. Delaying the reward (positive feedback) does not elicit the same response from animals and humans. Delaying the reward for animals does in fact depress performance, while doing the same for humans produces little or no effect on performance at all. 2. When the reward or reinforcement is withdrawn from animal subjects, performance of the desired behavior declines. However, the same is not true for humans.

5

Certainly, additional research was being conducted throughout the century. However, Adams cites primarily works from the early 19th c., indicating that they significantly influenced the contemporary theories, and in response, his own theory. 6

Thorndike, 71.

7

Adams, 113.

8

Ibid., 114-115.

4 3. Human subjects do not simply repeat behaviors just because those behaviors are rewarded. Instead, Adams claims, humans attempt to correct their error in an effort to produce a more accurate response in relation to the desired behavior. Adams points to research by Elwell and Grindley as support for this claim.9 4. “Humans covertly guide their motor behavior with verbal responses, at least in the early stages of learning.”10 This verbal guidance does not fit into the stimulus-response model that follows from Thorndike‟s law. 5. Human performers can be aware of whether or not their motor movement was correct, often without an outside reward provided by an external source. Adams responded to the perceived shortfalls of Thorndike‟s theories by developing his own theory of motor learning. As noted earlier, nearly all of the theories of learning at the time of Adams‟ research were open-loop theories. This meant that, in order for a motor response to occur, some sort of outside stimulus response was needed. Furthermore, as long as the necessary motivational and habitual states were achieved along with said stimulus, that same response would always be produced.11 Adams noted that this explanation did not take into account the effect that feedback, and specifically feedback regarding error, had on guiding the completion of the motor task, both during and after an attempt was made to complete that task. He therefore determined that his theory would need to be closed-loop. More precisely, his closed-loop theory

9

J. L. Ellwell and G. C. Grindley, “The Effect of Knowledge of Results on Learning and Performance,” British Journal of Psychology 29 (1938): 39-54. 10

Adams, 115.

11

Ibid., 117.

5 “must be error-centered, with a reference mechanism against which feedback from the response is compared for the detection and correction of error.”12 Adams began the exposition of his closed-loop theory with a discussion of the nature of knowledge of results (KR),13 which he initially defines simply as information provided to the subject regarding his or her performance of the desired task that eventually leads the subject to a correct performance attempt.14 However, he does further distinguish between qualitative KR and quantitative KR. According to his definition, qualitative KR is “dichotomous.” This type of KR would only provide the learner with information regarding which direction away from the desired task his or her attempt varied (e.g. too long or too short).15 In the case of Frank the tenor from the beginning of the chapter, if his instructor simply told him, “You are flat,” the instructor would be providing qualitative KR. Alternatively, Adams defines quantitative KR as providing the same information as qualitative KR, but with some scaled numerical information regarding the degree to which the error was made.16 According to this definition, Frank‟s instructor could provide quantitative KR by saying, “You were a quarter-tone flat.” Of special importance to Adams‟ theory is that KR differs from feedback. Adams seems to refer to KR only in regard to

12

Adams, 120.

13

It should be noted that Adams‟ definition of KR varies rather significantly from Schmidt‟s definition, which will be discussed later in the chapter. 14

Adams, 122.

15

Ibid.

16

Ibid.

6 information received from outside the subject‟s body. Information received from and perceived by the performer‟s body is what he refers to as feedback.17 After his discussion of KR, Adams made the case for two states of memory, referred to as the perceptual trace and the memory trace, both of which he claimed to have a determining role in the performance of any motor task. The memory trace would be responsible for initiating the movement, choosing its initial direction and determining early portions of the movement.18 The perceptual trace, on the other hand, would be responsible for guiding the movement to the correct location along the prescribed pathway.19 For clarity in the distinction between the two traces, Adams‟ discusses the latter first. In attempt to retain the same clarity, the traces will be discussed in the same order here. Adams‟ concept of the perceptual trace is rooted in “the notion of a trace or image which perception has used for a long time to account for the recognition of exteroceptive stimuli.”20 In this sense, when a stimulus is experienced for the first time, whether that stimulus is seeing an image or performing a motor task, a perceptual trace is created and imprinted. When that stimulus is encountered again, it is recognized when it matches the imprint of the original perceptual trace.21 For example, the first time that Frank sang an A4, an imprint of the motor skills necessary to produce that pitch, referred to by Adams‟ as a perceptual trace, was created. Subsequently, every time that he attempted to reproduce that pitch again, he would access that

17

Adams, 123.

18

Ibid., 125.

19

Ibid., 124.

20

Ibid.

21

Ibid.

7 perceptual trace and weigh it against the current feedback stimuli he was receiving in order to determine the correctness or error of that attempt. What if, however, Frank had never successfully negotiated the A4? How would he ever develop a correct perceptual trace for producing that pitch? Adams accounts for this in his theory by distinguishing between two stages in perceptual trace development: the verbal-motor stage, and the motor stage. Early on in a subject‟s training, when mistakes are frequently being made, the subject cannot simply repeat movements that the perceptual trace would recognize as having made before or the subject would repeat the errors from past attempts. In this stage (the verbal-motor stage) the subject must instead rely on KR to make each attempt different from the previous ones.22 Once the KR has been reporting relatively small errors for “some time,” it is safe to assume that, since the correct movement has been made fairly frequently, the perceptual trace is sufficiently strong that the subject has moved into the motor stage. During the motor stage, KR can be dropped out completely and “conscious behavior eventually becomes automatic.”23 With his theory relying on the perceptual trace for guidance of motor tasks, Adams admits that it would be tempting to stay with a single-trace theory with the perceptual trace also being responsible for initiating the task. However, Adams notes three strong reasons that a second trace, which he terms the memory trace, is necessary for a complete understanding of motor skill acquisition. First, if one trace were responsible for both initiating a response and testing its correctness, the response would always be judged against itself, meaning that it would always necessarily be judged as having occurred with zero error. Second, the perceptual trace 22

Adams, 124

23

Ibid., 125.

8 requires feedback, which can only occur after a response begins. Consequently, something else must be responsible for initiating the response in the first place.24 The third justification for the memory trace requires a distinction between recall and recognition. Adams defines the terms: “Recall is a response production and recognition is identification of a stimulus or a response.”25 Accordingly, the initiation of a motor response is an example of motor recall, whereas judging the correctness of that motor response is an example of motor recognition. If no one trace can be responsible for both recall and recognition, then there must be a need for a second trace. The memory trace, therefore, is responsible for initiating a response, while the perceptual trace, as discussed earlier, is responsible for judging the correctness of that response. It is important to note that, according to Adams‟ theory, the perceptual trace can not be formed without experiencing the correct location at least once.26 Additionally, Adams‟ postulated that, because of its closed-loop nature, learning could take place, once the perceptual trace had been established, without the subject being aware of the accuracy of his or her performance (i.e without KR).27 These were key points with which Richard R. Schmidt took issue, ultimately bringing him to develop his own theory.

24

Adams, 126.

25

Ibid., 126.

26

Ibid., 125.

27

Ibid., 124.

9 Richard A. Schmidt’s Schema Theory While Schmidt appreciated much of the work that Adams had done (most importantly, that Adams‟ theory attempted to test relatively long-term learning, whereas previous studies had focused primarily on improvements in performance28), he took issue with several points of Adams‟ closed-loop theory. In particular, Schmidt disagreed that the subject needed to experience the correct location in order to learn to be able to move accurately to that position.29 Additionally, Schmidt took issue with the idea that a subject would be able to continue learning without subsequent KR.30 It was these issues that Schmidt hoped to address in his own theory.31 The primary focus of Schmidt‟s theory concerns how the development of a schema32 (a theory long accepted in verbal memory studies) applies to motor-learning. Schmidt uses the process of learning to recognize a dog as an example of forming a schema. The idea… is that in order to perceive a set of visual stimuli (e.g., a dog) and to classify these stimuli correctly in the category “dog,” we need not have previously received the particular set of stimuli in question. Through our past experiences with seeing dogs, we store these stimuli in recognition memory and also abstract these stimuli into a concept related to dogs for additional storage. This concept forms the basis of a schema or rule of determining whether a new set of visual stimuli should be classified into the category “dog‟ or not. Thus to recognize an animal as a dog, we need not have ever seen that

28

Schmidt, 227.

29

Ibid., 228.

30

Ibid.

31

Ibid., 230.

32

S. H. Evans defines a schema as “a characteristic of some population of objects [which] consists of a set of rules serving as instructions for producing a population prototype.” Evans, S. H. “A Brief Statement of Schema Theory,” Psychonomic Science 8 (1967): 87-88.

10 particular animal before, and with the use of the schema for dogs, we correctly identify the animal‟s category.33 Before a schema for a set of motor stimuli can be developed, a goal-oriented movement must be produced several times, and sensory information from that movement must be stored. Schmidt outlines four sequential steps of storing information included in the production of a goal-oriented movement:34 1. Storing the initial conditions - consists of information received from receptors prior to a response such as proprioceptive information about the positions of the limbs and body in space. 2. Storing the response specifications for the motor program - consists of information specifying alterations to the general motor program. These alterations include changes in speed, force, etc. made to the motor program before it is run off. 3. Storing the sensory consequences of the response produced – consists of actual feedback stimuli received from the sensory organs while the motor program is being run. As a result, these consequences would be a copy of the afferent information received by the brain during completion of the task (movement). 4. Storing the outcome information – consists of the success of the response in relation to the outcome originally intended. This information consists of information received after the movement. Therefore, absence of feedback would result in no outcome information (a major difference from Adams‟ theory).

33

Schmidt, 233.

34

Ibid., 235.

11 Information from these four steps are stored together after the movement and, after a number of these movements (Schmidt does not specify how many), the subject begins to abstract the information about the relationship among these four sources of information, forming a schema of that particular movement. Once the schema is formed, Schmidt outlines how that schema operates to produce a movement in the order in which those steps occur:35 1. A desired outcome is specified/chosen. 2. Initial conditions that may affect the production of that outcome are determined. 3. New specifications for motor program are selected, which accommodate variances in the initial conditions and, using relationships between outcomes and sensory responses stored from previous attempts, expected feedback is selected. 4. Movement is initiated by running off the motor program. 5. Within 200 milliseconds, sensory receptors begin providing information about the movements occurring. 6. Feedback information is conveyed to feedback states selected earlier and compared to the expected outcomes. From this comparison, a response error is created. 7. Information regarding the response error is fed back to the schema for additional adjustments. At the same time, error information is fed to the labeling system, resulting in subjective reinforcement and fed back to the schema as subjective information. 8. Final error information is provided by the experimenter or teacher in the form of Knowledge of Results (KR). This information stems from the measured outcome of the subject‟s response (movement).

35

Schmidt, 239-240.

12 Motor Learning Defined and the Role of Feedback After publishing several studies and articles seeming to affirm his theories,36,37,38,39 Schmidt went on to write a textbook along with Timothy D. Lee, the fourth edition of which is referenced here.40 In this text, Schmidt finally amalgamates all of the main points of his theory into a clear definition of motor learning: Motor learning is a set of processes associated with practice or experience leading to relatively permanent changes in the capability for movement.41 An important element of this definition is the inclusion of the words “relatively permanent changes,” which set motor learning apart from fleeting and inconsistent improvements in performance. To highlight this point, in a chapter of a book currently in press, Katherine Verdolini dicusses three core issues providing further distinction between learning and performance. First, the nature of learning is dynamic. Acquiring new abilities does not result in accumulation of “bits of information in some discrete location in the head.”42 Instead, learning is a change in the

36

Richard A. Schmidt, “Control Processes in Motor Skills,” Exercise and Sport Sciences Reviews 4 (1976): 229-261. 37

Richard A. Schmidt, “Movement Education and the Schema Theory,” in Report of the 1976 Conference: National Association for Physical Education of College Women held in Cedar Falls, IA 3-8 June 1976, ed. E. Crawford. 38

Richard A. Schmidt, “Schema Theory: Implications for movement education,” Motor Skills: Theory into Practice 2 (1977): 36-38. 39

Richard A. Schmidt, “Past and Future Issues in Motor Programming,” Research Quarterly for Exercise and Sport 51 (1980): 122-140. 40

Richard A. Schmidt and Timothy D. Lee, Motor Learning and Performance: From Principles to Practice, (Champaign, IL: Human Kinetics Books 1991). 41

42

Ibid., 264-265.

Katherine Verdolini, “Motor Learning Principles: How to Train,” in Vocolgy, by Ingo R. Titze and Katherine Verdolini, (currently in press; used by permission), 3.

13 probability that a certain response will be given in the future as a result of practice.43 Second, learning cannot be observed directly. It can only be inferred by observing changes in individual performances. The third issue is what Verdolini may argue to be the most important: a temporary improvement in performance ability does not indicate that learning has taken place. Only after a change in performance ability has proven to be relatively stable over time can it be said that the new skill has been learned.44 Once a clear definition of motor learning had been developed, Schmidt turned his efforts to determining what factors most greatly influenced the achievement of motor learning. Among other factors including attention and motivation on the part of the learner, Schmidt realized that feedback from the instructor greatly influenced motor learning. More specifically, extrinsic feedback regarding errors “is one of the more important sources of information.”45 In terms of motor learning theory, as defined by Schmidt, feedback refers to information received by the learner before, during, and after a task has been attempted.46 This feedback can be divided into feedback that is received from either intrinsic or extrinsic sources.47 Feedback from intrinsic sources would include sensory information that arises as a natural consequence of attempting to perform a task.48 This intrinsic feedback can be further divided into categories of either proprioceptive or exteroceptive feedback. Proprioceptive feedback refers to sensory 43

Verdolini, 4.

44

Ibid.

45

Richard A. Schmidt, Motor Learning and Performance: A Problem-based Approach, 3rd ed., (Champaign, IL:Human Kinetics Books, 2004), 305. 46

Ibid., 276.

47

Ibid., 277.

48

Ibid.

14 information received from sources within the learner‟s 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 learner‟s own body, such as a mechanical device or an instructor.50 Ideally, this feedback should provide information that the learner cannot receive on his/her own without the aid of the outside information source, and should supplement the intrinsic feedback that the learner has already received.51 Extrinsic feedback can provide information regarding the learner‟s performance in one of two ways. First, it can provide information regarding the result of the performance attempt, such as how close the attempt was to the target behavior. To use Schmidt‟s terminology, this type of feedback can be referred to as Knowledge of Results (KR).52 Often, KR is redundant, simply restating the information that the learner has already gleaned through intrinsic feedback. In this case, the feedback is of little value to the learner. However, KR can be beneficial, even necessary, when the learner‟s 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 Schmidt‟s terminology, this type of feedback can be referred to as Knowledge of Performance. In most cases, Knowledge of Performance (KP) is more beneficial to learners than KR, as it provides information that the learner would be unable to receive through any other means.54 Figure 1.1 is an adaptation of a figure in Schmidt‟s text depicting the breakdown of all information received by the learner into the different categories of feedback discussed here.55

All Information Received by Learner

Related to the Action

Available Before Movement

Not Related

Available After Movement

Intrinsic Feedback

Proprioceptive Muscles

Extrinsic Feedback

Exteroceptive Vision Audition

Figure 1.1 - Categories of feedback received by the learner

54

Ibid., 281.

55

Ibid., 277.

KR

KP

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 one‟s 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 Frank‟s attempt. Additionally, information provided by a computer display using voice analysis software, electronic tuning equipment, and sound pressure metering devices all would fall under the category of extrinsic feedback. Furthermore, if the instructor told Frank, “You were flat on the high A,” that feedback would be classified as KR. Assuming that Frank was able to discern intonation during his/her performance, this KR would be redundant and of little value. On the other hand, if the instructor said something to the effect of “The high A had a pressed tone quality,” that information would be classified as KP. Since Frank‟s own aural perception of his own sound can be distorted, this KP could be useful information. While it may be interesting and useful to investigate how to improve a learner‟s ability to perceive and interpret his/her own intrinsic feedback, this study is aimed at informing training practices from the instructor‟s point of view. Consequently, the use and quality of extrinsic feedback is more pertinent. When considering extrinsic feedback, Schmidt suggests several questions to ask regarding what its properties are, when to provide it, how much information to provide, and how precise to be when giving it.56

56

Schmidt, 2004, 283-306.

17 Properties of Extrinsic Feedback In terms of the properties of extrinsic feedback, Schmidt states that it can serve to motivate a learner, reinforce a behavior, inform the learner, and/or produce a dependency on the feedback.57 The first of these properties, motivating a learner, may not be the primary objective for the instructor when providing feedback. However, the motivational properties of that feedback often provide an added benefit to the learner by encouraging them to continue to give their best effort, even when faced with multiple repetitions and monotonous training sessions.58 Reinforcing feedback will, as its name implies, reinforce a certain behavior in one of two ways. First, positively reinforcing feedback will provide the learner with an experience, which due to its pleasant nature, would increase the likelihood that the desired behavior will be repeated (similar to Thorndike‟s stimulus-response relationship).59 The second possibility, negatively reinforcing feedback, will provide the learner with an experience consisting of the removal of an unpleasant stimulus, thereby increasing the likelihood that the desired behavior will be repeated.60 To clarify, return again to Frank. If Frank performs a certain difficult passage correctly for the first time and the instructor says, “That was it. Good job!” this would be an example of positively reinforcing feedback. Alternately, after the previous five wrong attempts the instructor could have said with increasing impatience, “NO! THAT IS NOT CORRECT. DO IT AGAIN!” If, then, after the first correct attempt the instructor said nothing but let Frank finally continue to the next phrase, this would be an example of negatively reinforcing feedback,

57

Schmidt, 2004, 282.

58

Ibid.

59

Ibid., 284.

60

Ibid.

18 because the unpleasant experience of being berated by the instructor was removed. In his 1978 study, Adams found that positively reinforcing feedback is more effective than negatively reinforcing feedback. Incidentally, the same study also found that both positively and negatively reinforcing feedback were more effective motivators than was the use of punishment.61 When instructors think about feedback to their students, providing information is often their primary objective. Indeed, the very definitions of feedback listed by both Adams and Schmidt state that feedback is inherently information. Schmidt, however, indicates a slightly more pointed definition of informative feedback as feedback that provides the learner with the “direction they need to correct their errors and to modify their future performance.”62 This definition should be kept in mind in the following discussion of how much and what information to provide during feedback. Throughout the learning process, the instructor must be wary of providing feedback in such a manner that the learner becomes dependent on that feedback in order to produce the desired behavior. It has recently become evident that producing frequent augmented, verbal feedback to a learner can guide the learner to the correct movement in much the same way that physical guidance aids do. However, Schmidt points63 to a 1959 study by J. Annett where it was found that learners who practiced with a physical guidance aid could not perform the task once the guidance aid was taken away.64 This was a result of the fact that the learners had become

61

Jack A. Adams, “Theoretical Issues for Knowledge of Results,” in Information Processing in Motor Control and Learning, ed. G. E. Stelmach (New York: Academic Press, 1978), 229-240. 62

Schmidt, 2004, 286.

63

Ibid., 225.

64

J. Annett, Feedback and Human Behavior (Middlesex, England: Penguin Books, 1959).

19 dependent on the guidance aid. Schmidt maintains that learners can develop the same dependency on verbal feedback when it is provided too frequently.65 Thus, the instructor must ask the question of when to provide feedback.

When to Provide Extrinsic Feedback The question of when to provide feedback actually consists of at least two questions: a) whether or not to provide feedback at all, and b) if it is determined that feedback is necessary, how frequently should it be provided? When deciding whether or not to provide feedback, the first consideration should be whether or not feedback is necessary.66 While this consideration may seem intuitive, instructors frequently provide feedback almost as an instinctual response to the completion of a learner‟s attempt. Rather than providing this type of reactionary feedback, the instructor would be better off to consider the complexity of the task weighed against the ability and experience of the learner.67 The question of how frequently to provide feedback is the very subject of the study discussed in the following chapters of this paper. Schmidt states that it is important to make a distinction between absolute feedback frequency and relative feedback frequency. Absolute feedback frequency is simply a statement of how many times feedback was provided during a training or instruction session. Alternatively, relative feedback frequency refers to the number of times feedback was provided relative to the total number of attempts made during the session.68

65

Schmidt, 2004, 287.

66

Ibid., 289.

67

Ibid.

20 For example, if in a lesson Frank attempted to sing a phrase 15 times and the instructor provided feedback 5 times, then the absolute feedback frequency for the session would be 5 and the relative feedback frequency would be 33%. Schmidt claims that increasing absolute feedback frequency will result in increased learning and cites several studies to support that claim. However, he also states that decreasing the relative feedback frequency will also result in increased learning. If both of these assertions are true, then the ideal situation for learning would be a large number of attempts made by the learner, with feedback being provided by the instructor only after every few attempts. Several studies have examined these assertions and they will be discussed in further detail in chapter 2.

How Much and What Information to Provide Since the memory capabilities of humans are limited, instructors must give careful consideration to how much information should be provided to the learner during each instance of feedback. Schmidt recommends focusing on one feature of the task that is most fundamental to its successful completion.69 He also suggests a method of feedback called summary feedback, which is again directly related to the question of frequency of feedback. In summary feedback, the instructor would withhold feedback for a number of attempts, and then provide the feedback in a summary form. For support of summary feedback, Schmidt points to studies conducted by J.J. Lavery70 and by Schmidt, Lange, and Young.71 In both studies, it was found that subjects

68

Schmidt, 2004, 303.

69

Ibid., 295.

70

J. J. Lavery and F. H. Suddon, “Retention of Simple Motor Skills as a Function of the Number of Trials by Which KR is Delayed,” Perceptual and Motor Skills, 15, 1962, 231-237.

21 who received summary feedback performed better on a retention test of a practiced task than those subjects who received immediate feedback following each attempt. However, Schmidt also notes that, as indicated by the study that he and his colleagues conducted, as the complexity of the task being learned increases, the number of attempts being summarized should decrease in order to maximize learning.72 A variation of summary feedback is to average all of the attempts being summarized to reveal a single trend that can then be addressed by feedback. This practice is known as average feedback.73 When deciding what information to provide, of primary importance to the instructor is to ensure that the information he or she is providing is addressing elements of the task that are under the learner‟s control.74 In order to ensure that an element being addressed by an instructor‟s 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 learner‟s 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 Schmidt‟s distinction between KR and KP.

How Precise to Be When Providing Feedback It may seem that determining how precise to be with feedback may be the same as deciding what information to provide. However, there is a distinction, which can be clarified by returning once again to the example of Frank. When Frank sings a flat note, his instructor has two questions he or she must decide to pose. First, will he or she tell Frank that he was flat? This would be an example of deciding what information to provide. Second, will he or she tell Frank just how far flat he was? This is an example of how precise to be. Schmidt notes that, in general, very precise feedback is not necessarily more effective than less precise feedback.76 Furthermore, Magill and Wood found that learners at an early stage will not benefit from highly precise feedback because the magnitude of their errors is so great.77 Instead, it may be more beneficial to provide information pertaining only to the direction (sharp or flat) and magnitude (quarter-tone) of the error. Even then, information regarding the direction of the error is more

75

Schmidt, 2004, 294.

76

Ibid., 300.

77

R.A. Magill and C.A. Wood, “Knowledge of Results Precision as a Learning Variable in Motor Skill Acquisition,” Research Quarterly for exercise and Sport 57 (1986): 170-173.

23 important than information regarding the magnitude.78 One method of feedback that deals with the matter of precision is the bandwidth feedback method. In this practice, the instructor would withhold feedback unless the attempt falls outside the realm of some acceptable result (i.e. outside an acceptable bandwidth). This practice is especially well-suited to tasks in which there is a very clearly defined desired outcome.79

Application to Voice Science and Need for the Current Study While Schmidt has clearly outlined his theory in terms that prove its basis in cognitive science, the application of that theory to practical training in singing and voice therapy has not yet been widely practiced. To this end, Katherine Verdolini and Timothy Lee published a chapter in Vocal Rehabilitation for Medical Speech-Language Pathology aimed at applying schema theory to the speech clinic or voice studio.80 Through their discussion, along with review of relevant research, Verdolini and Lee conclude that augmented feedback is necessary for motor learning, but that less frequent feedback may enhance learning more.81 Verdolini and Lee define augmented feedback as “feedback that is received from sources that have been augmented by an instructor or therapist"82 as opposed to “feedback that a person normally receives as a natural, sensory attribute of an action.”83 In other words, augmented feedback

78

Schmidt, 2004, 301.

79

Ibid.

80

Katherine Verdolini and Timothy D. Lee., “Optimizing Motor Learning in Speech Interventions,” In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine M. Sapienza and Janina K. Casper (Austin, TX: Pro-ed, 2004), 403-446. 81

Ibid., 435.

82

Verdolini and Lee, 417.

24 simply refers to feedback that is provided by an outside source. Augmented feedback may include verbal and/or graphic reports of the results of a previous performance attempt (knowledge of results) and can be provided either by an instructor or by various types of instrumentation.84 For the purpose of the following study, knowledge of results was supplemented by additional instruction during augmented feedback. Verdolini and Lee also spend some time clarifying the distinction between learning and performance. They use an old, clear definition from early research in experimental psychology that emphasizes differences between performance, skill, and skill acquisition (i.e. learning).85 In this definition, performance refers to an action that results in a measurable outcome. Skill refers to an individual‟s capability to perform. Finally, skill acquisition (learning) is defined as a process, which leads to relatively long-term changes in an individual‟s skill, or ability to perform. Simply put, learning results in permanent or semi-permanent changes in one‟s ability to perform. Verdolini and Lee, then, mandate that “any claims about learning must be based on transfer or retention tests conducted some period after a training session has been concluded.”86 Without a transfer or retention test, it would be impossible to test the permanence of the changes in performance. It is important to understand clearly the difference between performance and learning as studies87,88,89 suggest that, especially in regard to augmented feedback from the

83

Ibid.

84

Ibid.

85

Edward Tolman, Purposive Behavior of Animals and Men, (New York: Century, 1932).

86

Verdolini and Lee, 411.

87

E. A. Bilodeau, Ima McD Bilodeau, and Donald A. Schumsky, ”Some Effects of Introducing and Withdrawing Knowledge of Results Early and Late in Practice,” Journal of Experimental Psychology 58 (1959): 142-144.

25 experimenter or teacher, those practices that most enhance learning may actually depress immediate performance in lessons or therapy sessions. While Verdolini and Lee, and subsequently Lynn Helding,90 have all indicated that these principles of motor learning theory should be applicable to studio voice instruction, very little research has been conducted to provide quantitative evidence that the principles gleaned from years of research in other fields can be effectively applied in the voice studio. The current study, as outlined in Chapter 3, has been designed to test just such an application, especially in regard to the frequency at which feedback is provided by the voice instructor. Accordingly, this chapter has provided an overview of the principles of motor learning theory and has focused most heavily on the theory regarding feedback. For more thorough review of other elements effecting motor skill acquisition, such as individual differences, motivation, and practice variability, the reader is encouraged to investigate several textbooks written by Schmidt, each of which examine the issues of motor skill acquisition from different perspectives.91

88

Timothy Lee, Margaret A. White and Heather Carnahan, “On the Role of Knowledge of Results in Motor Learning: Exploring the Guidance Hypothesis,” Journal of Motor Behavior 22 (1990): 191-208. 89

Darl W. Vander-linden, James H. Cauraugh and Tracy A. Greene, “The Effect of Frequency of Kinetic Feedback on Learning an Isometric Force Production Task in Nondisabled Subjects,” Physical Therapy 73, no. 2 (1993): 79-87. 90

Lynn Helding, “Voice Science and Vocal Art, Part Two: Motor Learning Theory,” Journal of Singing 64, no. 4 (March/April 2008): 417-428. 91

Richard A. Schmidt, Motor Learning and Performance: From Principles to Practice (Champaign, IL: Human Kinetics Books, 1991); Richard A. Schmidt, Motor Control and Learning: A Behavioral Emphasis 3rd ed. (Champaign, IL: Human Kinetics Books, 1999; Richard A. Schmidt, Motor Learning and Performance: A Problem-Based Learning Approach, 4th ed. (Champaign, IL: Human Kinetics Books, 2004).

26 The amalgamation of the research discussed in this chapter has led the author to believe that there are several truths regarding feedback and learning in the classical voice studio. 1. Extrinsic feedback (referred to as augmented feedback by Verdolini and Lee), including instruction from the voice instructor, visual guidance from devices such as a hand mirror, and even aural guidance from a piano or other pitch-producing device, is necessary for the acquisition of the motor skills required by classical singing technique. 2. Extrinsic feedback can help reinforce desirable motor skills while inhibiting undesirable skills. 3. Extrinsic feedback will often guide a student singer to an immediate improvement in the ability to perform a vocal task. 4. Excessive extrinsic feedback, however, will interrupt a student singer‟s cognitive processes as he or she attempts to compare the outcome of an attempt at a vocal task with the expected or desired outcome of that task. 5. True acquisition (learning) of the skill required to perform a vocal task will only be evidenced by the ability of the student singer to perform that task consistently after a period during which there was an absence of instruction and/or the ability to transfer the requisite skills to a related, but unpracticed vocal task. If these five statements are indeed true, as a great deal of research suggests that they are, then it stands to reason that reducing the frequency with which extrinsic feedback is provided to a student singer, while not removing the feedback altogether, will lead to more effective acquisition of the motor skills required by classical singing technique. This statement provides the hypothesis that the current study attempts to test.

27 CHAPTER 2 LITERATURE REVIEW

Over the past several decades, many studies have been conducted that were aimed at applying the principles of motor learning theory, as outlined in chapter one of this paper, to various disciplines. Of particular relevance to the current research are those studies regarding feedback frequency. Six such studies have been summarized below. These studies have been categorized as those dealing with feedback frequency and simple arm-positioning/timing motor tasks, and those dealing with feedback frequency and vocal motor tasks. In deciding which and how many studies to include in this chapter, the researcher continued to gather articles and reports until the citations started to become circular (i.e. they all started citing the same research studies). The most frequently cited articles are those reported below. The study outlined in chapter three, then, was based on two criteria: how close it could be to the types of studies that were most frequently reported, and how well it could be adapted to the voice studio with minimal disruption to the learning environment. Of course, more studies have been conducted and published that focus on application of motor learning theory to various motor tasks. Each of the studies reported below, however, has provided guidance in the development of the current research design, which is presented in chapter three. While the studies in the second category are directly concerned with the role of various elements of feedback on the training of voice skills, none of them are centered in the realm of classical singing technique. The current study is designed to begin to fill this gap in the research.

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

Reduced Frequency of Knowledge of Results Enhances Motor Skill Learning92 In their paper, Carolee J. Winstein and Richard A. Schmidt began by noting that in previous research regarding the efficacy of reducing the relative frequency of feedback, absolute frequency of feedback was allowed to co-vary along with the relative frequency of feedback.93 Additionally, they note that previous research had been conducted using relatively simple tasks. The researchers claim that perhaps “enhanced learning [resulting from decreased relative frequency of feedback] cannot be demonstrated with such simple tasks.”94 This claim, according to the researchers, poses some doubt as to whether or not the results from the previous studies they cite95 can be generalized. In an attempt to more thoroughly and accurately test their hypothesis that less-frequent feedback would result in greater long-term improvement in the ability to perform a motor task, Winstein and Schmidt conducted three experiments, all of which used a more complex motor task. In each of these experiments, subjects were required to learn and replicate a four-stage elbow-extension/flexion movement, performed to a set degree of 92

Carolee J. Winstein, and Richard A. Schmidt, “Reduced Frequency of Knowledge of Results Enhances Motor Skill Learning,” Journal of Experimental Psychology: Learning, Memory, and Cognition 16, no. 4 (1990): 677-791. 93

You may recall from chapter one (page 17) that absolute frequency refers to the total number of times that feedback was provided, whereas relative frequency refers to the number of times feedback was provided in relation to the number of attempts made. 94

95

Winstein and Schimdt, 1990, 679.

e.g. Linda Ho and John B. Shea, “Effects of Relative Frequency of Knowledge of Results on Retention of a Motor Skill,” Perceptual Motor Skills 46 (1978): 859-866.

29 extension/flexion and within a set time parameter. Subjects were required to extend their elbow to a prescribed position, reverse the movement and flex the elbow to another prescribed position (different from the starting position), reverse again and extend the elbow to a different prescribed position, and finally flex the elbow once again to a final prescribed position, all within 800 milliseconds. A plywood sheet was placed such that subjects were not allowed to directly view their arm during their movement attempts. Instead, a video monitor was provided on which subjects could view their movements remotely. In this way, the researchers could manipulate how frequently the subjects received feedback (visualization of their movement attempts) regarding the accuracy of their movements.

Experiment One The authors note that many previous experiments regarding feedback frequency used retention tests where no KR was provided during the testing trials. Because of this, any evidence in these experiments that relatively low feedback frequency resulted in better performance on retention tests could be interpreted to mean that subjects who had been training with relatively low feedback frequency performed better simply because the testing environment was more similar to their training environment. This interpretation of the results was referred to as the specificity hypothesis. In order to test the specificity hypothesis, experiment one utilized four different retention-test conditions where subjects received feedback on 0%, 33%, 66%, or 100% of their retention test attempts to perform the extension-flexion task described above. During the skill acquisition phase of the experiment, however, subjects received feedback only on either 100% or 33% of their attempts. To this end, 136 undergraduate student volunteers (74 female

30 and 62 male) were divided into eight groups (two acquisition-phase feedback conditions for each of the four retention-test feedback conditions) of seventeen students each. The acquisition phase of the experiment consisted of two thirty-minute sessions, during which each subject made 99 attempts to perform the task. Each subject was allowed 15 to 20 seconds between each attempt, regardless of feedback condition. Those subjects in the 100% feedback groups received feedback after every trial. Subjects in the 33% feedback groups received feedback following only two of every six attempts, though not predictably after every third attempt. An immediate retention test of 27 attempts was conducted ten minutes following the completion of the second acquisition session. Accuracy data was collected for each attempt during both the acquisition and retention test phases. Analysis of the skill-acquisition phase data revealed that subjects in both groups of feedback frequency reduced their error (i.e. performed the task more accurately) throughout the phase. Predictably, the 33% feedback group exhibited higher levels of error throughout the acquisition phase. However, the difference in error between the two groups was not statistically reliable, indicating that less frequent feedback may not inhibit performance during skill acquisition quite as much as had been previously thought. As noted earlier, the 27-trial retention test was administered providing 0%, 33%, 66%, or 100% feedback. Consistently, the subjects who had received 33% feedback during the skillacquisition phase tended to perform slightly better than those subjects who received 100% feedback during skill-acquisition (though the difference was not statistically significant) regardless of the retention test condition. Therefore, the experiment provided no evidence in support of the specificity hypothesis.

31 Experiment Two As experiment one was designed to test the specificity hypothesis, experiment two was designed to test what the researchers termed the guidance hypothesis, which “predicts that error information should be most useful early in practice to drive the performer toward the goal response.”96 In other words, this hypothesis indicates that in the early stages of skill acquisition, more frequent feedback will be beneficial, while less frequent feedback will be more beneficial during the later stages of skill acquisition. In order to test the guidance hypothesis, Winstein and Schmidt designed an experiment in which feedback was provided more frequently early on in the skill acquisition phase and was reduced gradually in later stages. In experiment two, 58 undergraduate student volunteers (42 female and 16 male) were divided into two groups: a 100% feedback condition group, and a 50% feedback condition group. Subjects from both groups were required to perform the same arm flexion/extension task as was used in experiment one. Similar to experiment one, experiment two consisted of a skillacquisition phase, consisting of two practice sessions of 96 practice attempts each, followed by two no-feedback retention tests. One retention test was performed approximately five minutes following completion of the skill-acquisition phase, while the second retention test was performed approximately 24 hours later. During the skill acquisition phase, the 100% feedback group received feedback following every practice attempt, while the 50% feedback group received feedback following only half of the practice attempts. Differing from experiment one, however, feedback for this group was not evenly distributed throughout the phase. Instead, feedback was provided following each of the first 22 attempts in both of the two practice sessions. Following these attempts, which were essentially 100% feedback conditions, feedback

96

Winstein and Schmidt, 1990, 682.

32 was systematically reduced throughout the remaining attempts to such a point that subjects in this group received feedback following only 4 of the final 12 attempts. As with experiment one, the absolute feedback frequency was allowed to co-vary with the relative feedback frequency, meaning that as the relative frequency of feedback decreased, so did the absolute feedback frequency. Error data from the skill-acquisition phase indicated that both groups reduced their mean error throughout the duration of the phase. While not statistically significant in amplitude, the 50% feedback group tended to have slightly lower error scores during the second practice session. Similarly, the 50% feedback group maintained a lower level of error during the delay between the two practice sessions than did the 100% feedback group. This indicated that the 50% feedback group did not slide back in ability to perform the task nearly as much as the 100% feedback group. Error data from the immediate retention test, conducted five minutes following the completion of the acquisition phase, showed that the 50% feedback group performed with less error than the 100% feedback group, though again to a statistically insignificant degree. During the delayed retention test, however, the 50% feedback group performed with a statistically significant 35% lower rate of error than did the 100% feedback group. These findings indicate that providing feedback on 50% of attempts during skill acquisition led to a higher level of retention of the ability to perform the motor task when compared with providing feedback on 100% of the attempts. The researchers do note that supporters of the specificity hypothesis would argue that, because the subjects in the 100% feedback group had never practiced without feedback, it would come as no surprise that they would perform less accurately on a no-feedback retention test. While this hypothesis cannot be completely ruled out, the more accurate performance by the 50% feedback group on the delayed

33 retention test combined with the higher level of retention during the delay between the two practice sessions would imply that lower frequency feedback resulted in enhanced learning.

Experiment Three While the results of experiment two did imply that lower frequency feedback resulted in enhanced learning, it was unclear whether those positive effects were due to the similarity between the testing and training conditions of the 50% feedback group or if the lowered feedback frequency was truly beneficial to learning. In an attempt to address this question, the researchers designed experiment three wherein the feedback frequency schedules were kept the same during the skill acquisition phase as in experiment two. However, in experiment three KR was provided during the delayed retention test, whereas KR was not provided during the delayed retention test in experiment two. 46 undergraduate student subjects (26 female and 20 male) participated in experiment three, practicing the same elbow flexion/extension task as in experiments one and two. Similar to the results of experiment two, subjects in the 50% feedback group performed the task with less error in the initial trials of the second practice session, though the results were not statistically reliable (meaning simply that the variance in performance between the two groups was not of sufficient consistence to be considered reliable). The authors note that the these initial trials of the second practice session can be viewed as mid-acquisition phase retention test with KR provided since they were performed after a period of time during which no feedback was provided and no practice trials were performed (please see Chapter 1 pp. 18 and 22 for more discussion on retention tests). Additionally, the 50% feedback group performed the task with significantly less error on the delayed retention test performed 24 hours following the

34 completion of the skill acquisition phase, and in which 100% KR was provided. Since the 50% feedback subjects were able to perform the task with less error even though they had not practiced as in the 100% KR retention test conditions, these results provide some evidence that reduced relative frequency of feedback during skill acquisition is indeed beneficial to learning. Overall, the findings of all three studies combined seem to lend merit to the guidance hypothesis of KR as formalized by Alan Salmoni and Richard Schmidt in their 1984 review of KR in motor learning.97 This guidance hypothesis suggests that KR plays two roles in motor learning.98 First, it benefits the learner by guiding the learner toward the desired outcome (i.e. movement or motor task). Second, however, it is detrimental to the learner as it encourages, or at least allows, the learner to rely on and even become dependent on KR in order to achieve the desired outcome. In terms of reducing the relative frequency of KR, the guidance hypothesis would seem to suggest that any negative effects from reducing the amount of guidance received by the learner from KR would be offset by the beneficial effects of reducing the dependency-creating effects of KR. The findings of experiment one, which found little positive effect from reducing the relative frequency of KR, would support this view of the guidance hypothesis. The guidance hypothesis is further examined by the researchers of the following study.

97

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

Ibid., 380.

35 On the Role of Knowledge of Results in Motor Learning: Exploring the Guidance Hypothesis99 Later in 1990, after the previous research was published, Timothy D. Lee, Margaret A. White, and Heather Carnahan published this research in which the guidance hypothesis‟ claims of the detrimental effects of KR were examined. The authors first note that early research conducted by Jack A. Adams100 and Edward A. Bilodeau101 claim that learning takes place most effectively when KR is provided frequently, which is in direct contradiction to the guidance hypothesis. In order to examine the discrepancy between these two claims, the researchers designed three experiments in which a reciprocal movement task was either guided using a metronome (experiment one) or using augmented auditory KR (experiments two and three). All three experiments employed the same methodology throughout, with the exception of the manner in which the augmented feedback was provided. In each of the studies, 30 undergraduate students (15 male and 15 female) were asked to use a stylus to tap one of two metal plates that were connected to a computer to calculate the timing of the subjects‟ taps. The subjects were instructed to perform seven consecutive taps, alternating between the two plates with 500 milliseconds (ms) between each tap. All subjects performed this seven-tap trial at least 50 times with an intertrial interval of 5 seconds.

99

Timothy D. Lee, Margaret A. White and Heather Carnahan, “On the Role of Knowledge of Results in Motor Learning: Exploring the Guidance Hypothesis,” Journal of Motor Behavior 22, no. 2 (1990): 191-208. 100

Jack A. Adams, “A Closed-loop Theory of Motor Learning,” Journal of Motor Behavior 3 (1971): 111-150. 101

Ina M. Bilodeau, “Information Feedback,” in The Acquisition of Skill, ed. Edward A. Bilodeau (New York: Academic Press, 1966) 255-296.

36 For all three experiments, each of the 30 subjects was randomly assigned to one of three feedback conditions. The first group, referred to as the 100% group, received feedback following each of their 50 trials. The second group, referred to as the 50% -50T group, received feedback following only half of the trials. This group still performed only 50 trials, meaning that they received feedback in 25 instances. Consequently, for this group absolute feedback frequency was allowed to co-vary with relative feedback frequency. The third group, referred to as the 50% - 100T group, again received feedback following only half of the trials. However, this group performed 100 trials, allowing the absolute feedback frequency to remain consistent while the relative feedback frequency was reduced to 50%. Following the completion of the acquisition phase consisting of either 50 or 100 trials, subjects were allowed a 5-minute rest period before completing a retention test consisting of 20 trials, during which no feedback was provided. Statistical analyses were conducted on how close the actual inter-tap time gap was to the desired 500 ms goal, providing an absolute constant error for each group‟s performance of the task.

Experiment One For experiment one, subjects were provided with feedback in the form of constant guidance. In this experiment, subjects were guided to the correct inter-tap timing by a metronome set at 120 beats per minute (bpm). During the acquisition phase, all three groups appeared to perform at an equal level of timing accuracy, though the 100% group did present a statistically insignificant trend to perform at a slightly higher level of accuracy during the first ten guided trials. At the onset of the retention phase, all three groups again performed with similar accuracy. However, as the retention test progressed, accuracy for the 100% group

37 declined, while the 50% - 50T group maintained its level of accuracy and the 50% - 100T group increased in accuracy. Again, the difference in performance accuracy was not significant.

Experiment Two In experiment two, subjects were provided aural KR feedback following each tap. If the inter-tap time gap was too large (525 ms or more), the computer immediately provided 200 Hz tone, 80 ms in duration. If the inter-tap time gap was too small (475 ms or less), the computer immediately provided a 680 Hz tone, again 80 ms in duration. If the inter-tap time gap was between 476 and 524 ms, then no tone was provided. This absence of aural feedback was in fact KR in itself, as the subjects were able to interpret the absence as an accurate performance. During the acquisition phase, subjects were between three and six times more likely to produce error than in experiment one. Similar to experiment one, however, there appeared to be no significant effect from the relative frequency with which KR was provided. In the retention test, error was reduced considerably across all feedback groups. However, a statistically significant difference between any of the three feedback groups failed to emerge. One possible explanation provided by the authors for the lack of significant difference between the groups was in terms of the bandwidth of error allowed before a tone was produced. In experiment two, subjects were allowed to produce a timing variance of was produced, resulting in

25 ms before a tone

5% around the goal, or a 5% bandwidth. Previous research by D. E.

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

102

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

38 Experiment Three Citing the research by Sherwood, the authors designed experiment three to replicate experiment two in every aspect except that the acceptable error bandwidth was decreased from 5% to 1%. Thereby, a low frequency tone was produced if inter-tap time gap exceeded 505 ms and a high frequency tone was produced if the inter-tap time gap was less than 495 ms. During the acquisition phase of experiment three, there again appeared to be no difference between the three feedback groups in performance error of trials on which KR was provided. In trials on which KR was not provided (for the two 50% feedback groups only), performance appeared to be more consistent (not more accurate, but with a lesser degree of fluctuation) than in trials with KR. This increase in consistency was significant for both 50% feedback groups. Additionally, during the retention test, both 50% feedback groups performed with lower percentages of error than did the 100% feedback group, indicating as experiment one did, that lower frequency of feedback was better for retention, and therefore learning. The findings of experiment one, where subjects were guided to the correct timing using a metronome, and experiment three, where subjects were provided with KR within a low bandwidth of acceptable error, both lead to the conclusion that 50% feedback was more effective for learning. These results indicate that performance guidance and KR are both used in a similar manner by learners during skill acquisition, namely that KR is used to “guide” the learner to the desired outcome. The similarity between performance guidance and KR lends credibility to the guidance hypothesis.

39 The Effect of Frequency of Kinetic Feedback on Learning an Isometric Force Production Task in Nondisabled Subject103

In this study, Vander Linden, Cauraugh, and Greene sought to apply the concept of lessfrequent 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 elbowextensor force and were asked to make their own elbow extensor force match that shown on the graph. The concurrent feedback group was allowed to view a real-time illustration of their force as it was graphed in relation to the desired force graph. This group was also allowed to view the completed graph of each attempt for eight seconds following each attempt. The 100% feedback group was not allowed to view the real time graphing but was allowed to see the completed graph of each attempt after the completion of each attempt. Finally, the 50% feedback group was only allowed to view the completed graph of each attempt after the completion of every other attempt. During the skill-acquisition phase of the research, each participant completed ten blocks of ten attempts, each with a rest period of at least one minute between each block. Once all of the skill-acquisition attempts were completed, an immediate retention test was administered, wherein the subjects were required to perform the same task thirty additional times and all groups received no feedback. An additional, delayed retention test of thirty attempts was administered forty-eight hours after the immediate retention test.

105

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

41 The results showed that, during the skill-acquisition phase, subjects in the concurrent feedback group exhibited significantly less deviation from the target force (error), than either the 100% or 50% feedback groups. However, in the immediate retention test the 50% feedback group exhibited 58% less error than the concurrent feedback group and the 100% feedback group exhibited 39% less error than the concurrent feedback group. The delayed retention test produced similar results with the 50% feedback group exhibiting 52% less error than the concurrent feedback group and the 100% feedback group exhibiting 26% less error than the concurrent feedback group. Additionally, the 50% feedback group exhibited 31% less error than the 100% feedback group on the immediate retention test, and 36% less error than the 100% feedback group on the delayed retention test. Figure 2.1 below presents a graphic representation of the results listed above. 106

Figure 2.1 – Subject performance on acquisition-phase and retention trials

106

Vander Linden, 1993, 85.

42 Root mean square error is represented on the Y axis and trials are represented on the X axis. Results of the immediate and delayed retention tests are also represented directly to the right of the trial results. These results indicate that, while concurrent feedback may provide for immediate improvements in performance ability during skill-acquisition, concurrent feedback is less effective when relatively long-term changes in performance ability are desired. Additionally, when comparing the frequency of feedback provided after the completion of the task, less frequent feedback appears to be more effective at producing relatively long-term improvements in performance ability. That is to say, less frequent feedback is more effective at facilitating actual learning of a motor skill.

Feedback Frequency and the Acquisition of Voice Tasks

Effects of Practice With and Without Knowledge of Results on Jitter and Shimmer Levels in Normally Speaking Women107 In her 1995 study, Carole Ferrand sought to determine if KR was in fact beneficial to the acquisition of a vocal task, specifically the development of different levels of jitter (variations in frequency) and shimmer (variations in amplitude).108 Ferrand‟s study assigned 30, normally

107

Carole T. Ferrand, “Effects of Practice With and Without Knowledge of Results on Jitter and Shimmer Levels in Normally Speaking Women,” Journal of Voice 9 (1995): 419-423. 108

In his book, Principles of Voice Production, Ingo R. Titze defines jitter and shimmer as “short term [meaning from one glottal pulse to the next] variability in fundamental frequency and amplitude, respectively.” In the 1960s, early investigators noticed that when they viewed speech waveforms on an oscillograph, each period was slightly different from any other period. They noted that the fundamental frequency appeared jittery, hence the term “jitter”. The term “shimmer” was then coined as a partner term referring to fluctuations in amplitude (Iowa City,

43 speaking women subjects into one of two groups. The first group received augmented aural and visual feedback (referred to as the KR group) while the other group received no feedback during practice sessions of a vocal task (referred to as the NKR group). The vocal task consisted of a two-second prolongation of the vowel [a]. After a baseline reading, where both groups received no feedback, all subjects underwent two practice sessions. Each practice session consisted of 15 trials of sustaining [a] for two seconds and the second practice session was held two days after the first practice session. The KR group received feedback pertaining to their performance in two ways. First they received visual feedback, which consisted of watching a real-time waveform of their prolongation. Second, they received verbal feedback provided by the instructor at the end of each prolongation. The NKR group received no feedback during or after their prolongations. A final session, wherein no KR was provided, was held one week after the subjects‟ final practice session. The researchers refer to this as a transfer test. However, because the subjects were performing the same test, delayed by some amount of time after the skill was acquired, this test actually measured retention and not transfer. The results of the study were split. Means of the shimmer levels did not change significantly over the course of the study. Additionally, no significant difference between the two groups occurred. Jitter values, on the other hand, were significantly lower (which is desirable) in the KR group during practice sessions than in the NKR group. However, during the transfer (retention) test, the KR groups‟ means climbed to at or above baseline measures whereas the NKR groups‟ means climbed only to halfway between the measures from practice session one and practice session two. These findings indicate that providing KR improved immediate

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

44 performance (indicated by the significantly lower practice session readings) but withholding KR improved learning (indicated by the smaller increase in readings in the transfer [retention] test).

The Role of Knowledge of Results in Performance and Learning of a Voice Motor Task109 In this study, researchers examined the effect of relative frequency of augmented external feedback on the acquisition, retention, and transfer of a vowel nasalization task. 30 subjects, all of whom were vocally untrained (meaning that they had received fewer than 5 formal singing or speaking lessons) were randomly assigned to one of three groups receiving 100%, 50% or No KR. KR was administered as a visual indication of percentage of nasalance as measured by a Kay Elemetrics Nasometer. The skill acquisition task consisted of nasalizing a sustained [i] vowel. Subjects were instructed to sustain the vowel [i] in a traditional oral phonation for two seconds before switching to a nasalized [ĩ] for four seconds. To test transfer, subjects were required to perform the same task using the vowel [a]. During the skill acquisition phase, subjects in the 100% KR group received immediate feedback following each trial. Subjects in the 50% KR group received immediate feedback following every other trial. Subjects in the No KR group were given no feedback regarding accuracy. Each group performed five blocks of eight trials of the tasks outlined above. Five minutes after completing the acquisition phase, retention was tested with the subject performing two blocks of eight trials of the same task while none of the subjects received KR.

109

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

45 The transfer phase of the study consisted of two blocks of eight trials of the same task using the vowel [a] and was conducted one day after the acquisition phase. The results of the study showed a significant difference between the 100% KR groups and the groups receiving 50% or no KR. The subjects in the less frequent KR groups performed the nasalance tasks more accurately during the retention and transfer phases of the study meaning that “an increase in relative frequency of KR led to a decrease in both motor performance and learning of a vowel nasalization task.” This result is in keeping with similar research as noted earlier. However, this study also found there to be no effect on immediate motor performance as a result of increased relative frequency of KR. This finding is consistent with Winstein and Schmidt‟s110 findings in experiment one of their research reported earlier, but inconsistent with Schmidt‟s earlier hypotheses111 and Verdolini and Lee‟s similar assertions112. Steinhauer and Grayhack provide two possible explanations for the deviance of the findings from the expected outcomes. One explanation is rooted in what they call the “dynamical theories of motor control” as asserted by J. Kelso113 and J. Folkins.114 In the context of these theories, the researchers contend that perhaps the nasalization of [i] was essentially too

110

Winstein and Schmidt, 1990.

111

Richard A. Schmidt, “A Schema Theory of Discrete Motor Skill Learning,” Psychological Review 82 (1975): 225-260. 112

Katherine Verdolini and Timothy D. Lee. “Optimizing Motor Learning in Speech Interventions,” In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine M. Sapienza and Janina K Casper (Austin, TX: Pro-ed, 2004), 403-446. 113

JA Kelso and B. Tuller, “A Dynamical Basis for Action Systems,” in Handbook of Cognitive Neuroscience, ed. M. Gazzaniza (New York: Plenum Press, 1984), 321-356. 114

JW Folkins, “Issues in Speech Motor Control and Their Relation to the Speech of Individuals with Cleft Palate,” Cleft Palate Journal 22/2 (1985) 106-122.

46 easy of a task, meaning that KR was useless. However, 100% KR was overwhelming to the subjects, causing them to override their own, correct, sensory feedback. The second explanation provided for the similar performances between the No KR and the 50% KR groups comes in the form of several possible criticisms of the study design. They contend that perhaps the sample size was too small, or possibly the subjects‟ own auditory feedback provided them with enough information to correctly perform the task. Finally, it is possible that, even though the group assignments were randomized, more skilled subjects were assigned to the No KR and 50% KR groups.

Effects of Feedback Frequency and Timing on Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech115 Austermann Hula and her fellow researchers performed two separate studies examining the acquisition and retention of speech skills in adult subjects with apraxia of speech116 in relation to varied frequency or timing of feedback. The first study, using four adult subjects, examined the effect of feedback frequency on retention and transfer of speech skills. A second study, using two adult subjects, examined the effect of feedback timing on the same skills. The research of both studies was intended to test the applicability of motor learning principles to the

115

Shannon N. Austermann Hula and others, “Effects of Feedback Frequency and Timing on Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech,” Journal of Speech, Language, and Hearing Research 51 (2008), 1088-1113. 116

Apraxia of speech is a motor speech disorder which causes the individual to have difficulty physically producing the speech sounds that they have no difficulty planning consciously. Individuals with apraxia of speech often present with slow, monotonic speech.

47 practice of voice and speech therapy. Only the study regarding feedback frequency will be reported here as it is the most applicable to the current research. During this study, participants (three male and one female) were provided speech therapy, in an attempt to aid the participants in regaining their abilities to produce non-word syllables (such as [mə] or [sə]) as well as whole words. A single-subject, alternating-treatment design was chosen for this study, meaning that the therapy for each subject alternated between a high frequency feedback (HFF) condition, and a low frequency feedback (LFF) condition. During the HFF condition, simple KR (“correct” or “incorrect”) was provided after 100% of the subjects‟ attempts to produce a syllable or word prescribed by the researcher. During the LFF condition, the same simple KR was provided following only 60% of the subjects‟ attempts. For the attempts in which no feedback was provided, a two-second interval was observed between each attempt. HFF and LFF treatment phases were each four weeks long with a four-week maintenance period in between. Following each HFF or LFF treatment phase, retention of the trained skills was tested using ten repetitions of six trained non-word syllables. At the same period of time, transfer of the trained skills to untrained word behaviors was tested using twelve related but untrained non-word items, and six related but untrained real-word transfer items. The study found that two of the four subjects responded with increased retention and transfer of speech skills following the low frequency feedback phase, indicating that that feedback condition had more effectively facilitated learning for those two subjects. The other two subjects exhibited consistent retention and transfer of skills during both feedback phases, indicating that feedback frequency had little effect on their learning. These findings lend some support to the idea of applying principles of motor learning to the speech system but also raise

48 more questions as to why feedback manipulation produced expected results in some subjects but not in others. The authors account for this discrepancy by suggesting that because the individual treatment plans for these two subjects required the practice and acquisition of relatively simpler tasks than the tasks practiced by the other two subjects, their progress was expectedly more consistent throughout the therapy process.

Conclusion The three arm-positioning studies cited117 are only a small portion of a large body of research that indicates a correlation between lower frequency of feedback and higher levels of retention and/or transfer (i.e. learning), at least in relatively simple motor tasks. These studies (particularly the Winstein and Schmidt research) are thorough in their scope and solid in their design. The second set of studies reported118 sought to meet the difficult task of applying similar uses of KR to the acquisition of the more complex motor tasks required for coordinated phonation. While these studies did not produce as decisive results as the previous studies, their results still indicate at least some positive interaction between less frequent KR and higher levels of learning. As noted at the onset of this chapter, the lack of research aimed at applying these same principles to classical singing technique was the impetus for the current research. Ample groundwork has been laid by the research in simple movement tasks. Similarly, that groundwork has been adequately directed toward voice production tasks. It stands to reason that there is now 117

Carolee J. Winstein and Richard A. Schmidt, 1990; Timothy D. Lee, Margaret A. White and Heather Carnahan, 1990; Darl W. Vander Linden, James H. Cauraugh and Tracy A. Greene, 1993. 118

Carole T. Ferrand, 1995; Kimberly Steinhauer and Judith Preston Grayhack, 2000; Shannon N. Austermann Hula and others, 2008.

49 justification to further this body of research by testing the application of reduced-frequency feedback to realm of classical singing voice instruction.

50 CHAPTER 3 METHODOLOGY

Among teachers of singing, there are as many different teaching styles as there are teachers. However, it has been observed that there are several common threads in terms of teaching techniques that are apparent in most voice studios. Frequency of feedback is one of those common threads. Many teachers are anxious to give immediate feedback to the student following every attempt to perform a given task. This occurs, understandably, for many reasons. First, instructors mistakenly look to the immediate improvement of performance as evidence of effective teaching practices. Second, students expect return for their investment. They are happier after a lesson when they feel that their performance has made significant improvement. Unfortunately, this elation wears off quickly when they are unable to repeat the performance a few hours later in the practice room.

Hypothesis The study outlined in the following pages tested the hypothesis that decreasing the frequency of feedback in voice lessons, while possibly depressing immediate performance, actually fosters learning, resulting in increased retention (ability to perform the task after a period of time with no instruction) and transfer (ability to perform a similar, but previously untrained task) of voice performance skills.

51 Theoretical Framework The hypothesis stated above stems directly from Schmidt‟s schema theory119 and its application to clinical voice therapy and studio voice by Verdolini and Lee, both of which were referenced earlier. These references claim that augmented feedback (knowledge of results) is necessary for motor learning. This feedback can vary in a number of ways including frequency and timing (the current study focused only on variations in frequency, though a future study will likely consider variation of timing). It is also noted in the literature that cognitive effort, in the form of hypothesis generation and evaluation, is essential for learning, though conscious “trying” may actually be detrimental to learning.120 If it is assumed that reducing frequency of feedback can increase the cognitive effort of the student by requiring them to create and test their own hypotheses, it can be interpolated that reducing feedback frequency will thereby increase learning.

Subjects 8 subjects (5 females and 3 males), ranging in age from 18 to 25 years old, participated in this study. Subjects were recruited from the researcher‟s undergraduate applied voice studio at Monmouth College (Monmouth, IL). All subjects were clearly informed of what would be required of them with their participation in the study, including that increased cognitive attention would be required throughout the study. They were also instructed to actively weigh their own performance against the desired outcomes and to try new methods of achieving the desired 119

Richard A. Schmidt, “A Schema Theory of Discrete Motor Skill Learning,” Psychological Review 82 (1975): 225-260. 120

Katherine Verdolini and Timothy D. Lee. “Optimizing Motor Learning in Speech Interventions,” In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine M. Sapienza and Janina K. Casper (Austin, TX: Pro-ed, 2004), 403-446.

52 outcomes during times that they noticed lower amounts of feedback. However, subjects were not informed concerning when feedback frequency levels would be changed. As evidence that the participants understood this instruction, they were invited to sign an informed consent document.121 Once informed consent was obtained, subjects were asked to complete a subject intake form,122 which included information regarding previous singing experience and history of vocal health. Subjects indicating a history of vocal pathologies or chronic dysphonia would have been excused from the study. However, none of the subjects indicated a history of any such excluding factors. The design of the study (discussed further in the following section), which allowed each subject to act as his or her own control, meant that students with widely varied levels of ability could participate in the study without posing difficulties in measuring results. However, there were a few exclusionary factors that were taken into account during the recruitment of subjects. First, because of the difficulty and often slow progress of teaching a student to match pitches, it was required that candidates already be able to match pitch before they were admitted to the study. Second, because the study calls for a capella singing (singing without accompaniment), subjects were required to have sufficient tonal memory as to be able to perform “America the Beautiful” without piano accompaniment. These abilities were judged by the researcher during the first meeting, immediately following the consent process. In regard to matching pitch, the researcher played a pitch on the piano and asked the subject to sing that same pitch. This process was repeated five times with

121

See appendix A for the IRB-approved informed consent document.

122

See appendix B for the subject intake form.

53 various pitches across the subject‟s vocal pitch range. In regard to tonal memory, the subject was asked to sing “America the Beautiful” a capella. Only subjects who were able to match correctly 4 of the 5 pitches and who were able to perform “America the Beautiful” with fewer than 5 wrong notes were admitted to the study. A total of 8 students met the above criteria and were admitted to the study. Background information from each of their subject intake forms is reported in figure 3-1.

Subject #

Age

Gender

Years of Experience

Vocal Health Issues

Retention/ Transfer test key

1

19

Female

5

Frequent strep throat

C

2

19

Female

5

None

Bb

3

19

Male

1

None

Bb

4

20

Female

1

None

Bb

5

19

Female

1

Frequent laryngitis

C

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 subject‟s retention/transfer tests were recorded using a Tascam DR-1 digital audio recorder producing a

54 256 Kbps mp3 file and sampling at 44.1 kHz. These audio files were recorded to compact disc and distributed to the volunteer auditors.

Skill Acquisition Phase A single-subject, alternating-treatment design, similar to that of the research done by Austermann Hula,123 was chosen for this study because of the flexibility it allowed in adjusting and accounting for individual differences.124 The duration of the study was 15 weeks with subjects receiving 1 thirty-minute voice lesson per week. These 15 lessons were divided into 3 blocks of 4 lessons each and a final block of 3 lessons.125 With each block, instruction technique alternated between high-frequency feedback (HFF) and low-frequency feedback (LFF). For this study, each subject was quasi-randomly assigned126 to one of two feedback groups: g-1 and g-2. Students in g-1 began with their first block being instructed with HFF.

123

Shannon N. Austermann Hula and others, “Effects of Feedback Frequency and Timing on Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech,” Journal of Speech, Language, and Hearing Research 51 (2008): 1088-1113. 124

Kevin P. Kearns, “Flexibility of Single-Subject Experimental Designs Part II: Design Selection and Arrangement of Experimental Phases,” Journal of Speech and Hearing Disorders 51 (August 1986): 211. 125

A study duration of 15 weeks was chosen to fit within the 15 week semester at Monmouth College, at which the experiment was conducted. 126

At the beginning of the semester, subjects were assigned a lesson time based on their availability. Once the lesson schedule was determined, the first four lessons of the day were assigned to g-1 and the last four students were assigned to g-2. In the strictest sense, this was not a random assignment of groups. However, since the original lesson schedule was determined by the subjects‟ availability, assignment to g-1 or g-2 was mostly out of the control of the researcher. This design was decided upon because it allowed the researcher to switch feedback frequency only once during the day, ideally leading to a more consistent approach to feedback frequency than if he were required to switch frequency with each subject. Throughout the course of the semester, two subjects (subjects 4 and 8) required a change in lesson time. In these cases, the subjects retained their group assignment, maintaining the same feedback frequency schedule.

55 Alternately, students in g-2 began with their first block being instructed with LFF. During lessons in the HFF blocks, instruction took place in a similar fashion to most voice studios; each attempt by the subject to perform the given task was followed by some level of KR and/or KP from the instructor, resulting in a 100% feedback condition. KR was provided in the form of a report of the result of the attempt relative to the desired outcome. If the attempt matched the desired outcome, positive KR was provided by the researcher saying “yes” or “good”. If the attempt did not match the desired outcome, negative KR was occasionally provided by the researcher saying, “no” or “not quite.” More frequently, however, if the attempt did not match the desired outcome, more informative KP was provided by the researcher. KP included guidance pertaining to how to improve performance of the task in order to improve the chances of the subsequent attempts more closely approximating the desired outcome. Again, the same type and amount of KP or KR was provided regardless of the skill that was being attempted. During lessons in the LFF blocks, students were asked to perform the given task twice or three times before further feedback/guidance was provided by the instructor, resulting in a 33% to 66% feedback condition. This feedback was provided in the same combination of KR and/or KP as it was provided during the HFF blocks. In each lesson, both in HFF and LFF blocks, time was equally divided between working on vocalization exercises and vocal literature in hopes of replicating the conditions in many collegiate voice studios. Vocalization exercises were varied but all were aimed at improving the subjects‟ tone quality, breath management, and/or intonation.127 KP, therefore, was designed to guide the subject toward meeting the goals of the individual exercises as they related to those

127

See appendix C for sample vocal exercises for each quality.

56 qualities. When working on vocal literature, the same qualities were the focus of the KP. Accordingly, auditors of the retention/transfer tests were asked to assess these same qualities. It was expected that variations in amounts of time that each subject spent practicing singing between lessons would have a significant impact on his or her ability to perform the retention/transfer test. To help minimize the impact, subjects were instructed to perform at least two hours of practice throughout the week between lessons. It was also expected that some subjects would spend more than the required time, while others may still spend less time. To help account for variations on performance ability that may have resulted from variations in practice time, subjects were required to complete a weekly practice record. To minimize the risk that their knowledge of my expectations may lead them to indicate falsely more or less practice time on their practice log, actual logs were not viewed by the researcher until after the completion of the study.

Retention/Transfer Tests Subjects were required to perform an identical vocal task at the beginning of each lesson to test retention from the previous week‟s lesson. The vocal task chosen for this study was singing the first verse of “America the Beautiful” without piano accompaniment. Subjects with lower voices (i.e. mezzo-sopranos and baritones) performed the song in the key of Bb major, while those with higher voices (i.e. sopranos and tenors) performed the song in the key of C major. One subject (#8) presented with a particularly low singing voice and had considerable difficulty transitioning into head register. For this reason, this subject performed the song in the unusually low key of G major. A more thorough discussion of the results from this subject will be discussed along with the rest of the subjects in chapters four and five.

57 Because “America the Beautiful” was never trained during the skill acquisition phases, this task measured skill transfer as well as retention. Thus, retention and transfer of vocal technique, as demonstrated by a performance of the vocal task, temporally separated from instruction, served as the measured variable, while frequency of feedback served as the manipulated variable. Subjects were recorded performing “America the Beautiful” to serve as a baseline of the subjects‟ vocal qualities. An audio recording of each performance (15 recordings for each of 8 subjects) of “America the Beautiful” was collected and catalogued for later blinded analysis. At the conclusion of the 15-week study period, audio files were recorded to a CD in a random order (created using an online random sequence generator)128 such that recordings of individual subjects were not placed consecutively on the playlist, and such that recordings of individual subjects were not placed in chronological order on the playlist. Copies of this CD were then sent to five college and university faculty members who served as voluntary auditors. These auditors were asked to evaluate each audio file using the perceptual analysis evaluation scale included in appendix D. This evaluation scale was developed using a small portion of a form developed by Dr. Katherine Eberle-Fink for assessment of singing129 and it asked auditors to assess each performance in terms of tone quality, breath management, and intonation. For the purposes of this study tone quality was defined as the tonal and acoustic properties of the voice during the performance (e.g. bright versus dark); breath management was defined as the ability of the

128

Random Sequence Generator, available at http://www.random.org/sequences/, accessed 12/6/2010. 129

Katherine Eberle-Fink, “Perceptual Acoustic Assessment of Singing,” Journal of Singing The Official Journal of the National Association of Teachers of Singing 63, no. 1 (SeptemberOctober 2006): 35-43.

58 subject to perform the excerpt with sufficient air to maintain the quality of their tone throughout individual phrases; and intonation was defined as the ability of the subject to perform the excerpt in such a manner that the auditors perceived individual notes to match the correct pitches of the excerpt. Auditors were selected by the researcher, with attention to their experience in perceptually evaluating vocal performances. A few key elements of the study design should be kept in mind while reading the results of the current research. First, it should be noted that, while individual tests were only evaluating performance, data from the entire study presented trends in retention and transfer levels that can be interpreted as indications of overall learning. It is also important to bear in mind that, due to the single-subject design of the proposed study, individual subjects were only evaluated against their own previous performances and not against the performances of other subjects. Finally, because of the limited number of subjects, statistically-significant trends were not determined.

Expected Results Due to the fact that increased practice will, in all likelihood, lead to some increase in ability, it was expected that performance levels would increase throughout the duration of the study. However, if the hypothesis that lower frequency feedback will increase retention and transfer of vocal skills were to hold true, it was expected that assessment values of the retention/transfer tests would trend up more sharply during LFF blocks than in HFF blocks. The actual results will be discussed in the following chapter.

59 CHAPTER 4 REPORTING OF RESULTS

Upon completion of the data collection phase of this study, the audio recordings of each subject‟s 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 subject‟s scores and the varied level of feedback, the line of average scores were divided into four segments, one for each phase of the study in which feedback frequency was varied. Even though each phase was four weeks in duration (with the exception of the final phase, which was only three weeks in duration), the data for each phase consisted of five scores. This is because the scores represent the subjects‟ performance on retention tests delayed by one week. For example, a subject started the study by producing a baseline performance of the test. That subject then received a lesson with either high or low frequency feedback. At the beginning of the next meeting (week two), the subject

130

It appears that the five auditors were frequently in agreement as to whether a voice was improving or declining, though the amplitude of that change varied from auditor to auditor.

60 produced another performance of the test. This performance shows the level of retention of the skills addressed at the previous lesson. Similarly, the tests from weeks three, four, and five (even though instruction in week five will switch to the opposite feedback frequency condition) will show the level of retention of the skills addressed in the lessons of weeks two, three, and four, respectively. Thus, phase one consists of five data points. The reporting of phase two will begin with the results of the test performed on week five, which can been said to serve as a baseline for the subject‟s 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 subject‟s 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 individual‟s performance ability. The results for each of the eight subjects will be reported in the following pages. Each subject will have an introduction to the background of the subject, taken from the subject intake form completed by each subject before the study began and including information from the subject‟s practice record, completed throughout the duration of the study. Following the introduction, results from each of the three voice characteristics (tone quality, breath

61 management, and intonation) will be presented. For each characteristic, three figures will summarize the findings for that subject‟s performance of that characteristic.

Subject One Subject one was 19-year old female who had been studying college level voice with the researcher for one year prior to the beginning of the study period. She also had studied privately with her choir teacher throughout her four years of high school. In addition to her experience studying private voice, she also reported having sung in choirs for five years. On her intake form, she indicated that she had had strep throat more than twice in a six-month span. Though she did contract a mild cold during the middle of the study period, the illness did not prevent her from taking part in the study at any point. Subject one was assigned to group g-1, meaning that she received low-frequency feedback during weeks one through four, high-frequency feedback during weeks five through eight, low-frequency feedback during weeks nine through twelve, and high-frequency feedback during weeks thirteen through fifteen. She did not appear for her lesson on week nine, the absence of data for which is exhibited in the results for this subject. Over the course of the study, subject one practiced an average of one hour per week. However, her levels of practice were higher during the beginning phases than in later phases. Subject one averaged 1.4, and 1.1 hours/week of practice during phases one and two respectively. That practice dropped to 0.8 and 0.7 hours/week of practice during phases three and four respectively. These levels of practice may bear some weight on the amount of progress made during the latter phases.

62 Subject 1 Tone Quality The overall results for the tone quality of subject one are reported below in table 4.1 and those results are graphed in figure 4.1

Week

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

62 65 54 62 52 58 58 66

Subject One Tone Quality Aud. 2 Aud. 3 Aud. 4 50 50 20 45 30 50 50 40 40 60 40 50 40 50 35 65

70 54 58 60 62 60

60 35 30 20 30 40

50 40 55 45 55 60

Aud. 5 45 35 20 20 25 40 40 30

50 40 30 50 50 50 50 40

Average 51.4 41 36.8 44.4 45.4 47.6 47.6 47.2

60 35 25 47 40 25

50 30 50 40 40 60

58 38.8 43.6 42.4 45.4 49

Table 4.1 – Subject one tone quality results

Auditors' Scores /100

Subject One Tone Quality 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

9

Weeks

Figure 4.1 – Subject one tone quality results

10

11

12

13

14

15

63 Figure 4.2 shows the average of the scores for tone quality from all five auditors divided into low and high frequency feedback phases. As discussed above, a linear trend line was added to each graph to show the overall trend throughout that phase. These trend lines were created using linear regression of the points on the graph and the equation for each trend line is listed in the legend in y = mx + b where m is the slope of the line and b is the point at which the line intercepts the y axis.

S. 1 Phase 2 (HFF)

60

60

55

55

50 45

Average

40

Linear (Average) y = -0.86x + 46.38

35 30 1

2

3

4

Auditors' score /100

Auditors' scores / 100

S. 1 Phase 1 (LFF)

50 Average

45

Linear (Average) y = 0.54x + 45.6

40 35 5

5

6

7

8

9

Week

Week

S. 1 Phase 3 (LFF)

S. 1 Phase 4 (HFF)

65

60

55 50 45

Average

40

Linear (Average)

35

y = -4.2x + 60.4

30 9

10

11

12

13

Auditors' score /100

Auditors' score /100

60 55 50 Average

45

Linear (Average) y = 3.3x + 39

40 35 13

Week

Figure 4.2 – Subject one average tone quality results by phase

14 Week

15

64 Subject One Breath Management The overall results of the breath management for subject one are listed below in the chart in table 4.2 and graphed in figure 4.3.

Week

Aud. 1

HFF

LFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

60 74 70 60 60 68 58 62

Subject One Breath Management Aud. 2 Aud. 3 Aud. 4 40 40 20 45 25 50 30 30 35 50 35 50 30 45 30 75

60 64 50 58 65 50

40 20 5 20 20 20

35 35 50 35 45 60

Aud. 5

Average

65 40 75 60 80 70 50 60

25 40 50 40 40 25 25 25

46 43.8 54 44 53 49.6 41.6 50.4

70 70 80 60 50 75

20 25 35 40 25 40

45 42.8 44 42.6 41 49

Table 4.2 – Subject one breath management results

Auditors' Scores /100

Subject One Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

9

Weeks

Figure 4.3 – Subject one breath management results

10

11

12

13

14

15

65 Figure 4.4 shows the average of the scores for breath management from all five auditors including linear trend lines. The results for the breath management of subject one are mixed with positive slopes phases one and four, but negative slopes in phases two and three.

S. 1 Phase 1 (LFF)

S. 1 Phase 2 (HFF) 60

55 50 45

Average

40 Linear (Average) y = 1.42x + 43.9

35 30 1

2

3

4

Auditors' score /100

Auditors' scor /100

60

55 50 45

Average

40 Linear (Average) y = -1.58x + 52.6

35 30

5

5

6

Week

S.1 Phase 3

8

9

S. 1 Phase 4 (HFF)

60

60

55 50 45

Average

40

Linear (Average) y = -0.6x + 45.7

35 9

10

11 Week

12

13

Auditors' Score /100

Auditors' Score /100

7 Week

55 50 Average

45

Linear (Average) y = 3.2x + 37.8

40 35 13

14 Week

Figure 4.4 – Subject one average breath management results by phase

15

66 Subject One Intonation The overall results for intonation for subject one presented are in the chart in table 4.3 and are graphed in figure 4.5.

Week

Subject One Intonation Aud. 2 Aud. 3 Aud. 4 5 10 0 10 0 25 0 15 5 15 0 30 10 30 10 70

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

45 35 36 40 28 30 42 60 38 36 40 46 40 42

5 0 0 0 0 5

20 15 35 20 25 35

Aud. 5

Average

25 40 60 20 10 30 60 75

25 25 25 25 10 15 50 60

22 22 29.2 20 13.6 21 38.4 55

40 60 40 55 40 60

40 25 25 10 10 50

28.6 27.2 28 26.2 23 38.4

Auditors' Scores /100

Table 4.3 – Subject one intonation results

Subject One Intonation

100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7 Weeks

Figure 4.5 – Subject one intonation results

8

9

10

11

12

13

14

15

67 Figure 4.6 shows the average of the scores for intonation from all five auditors including trend lines. The results for intonation returned to the trends exhibited by the results for tone quality, with the LFF instruction condition appearing to hinder intonation. During the LFF phases (1 and 3) the scores for intonation declined with slopes of -1.8 and -0.64 respectively. During the HFF phases (2 and 4) the scores increased with slopes of 14.16 and 6.1 respectively.

S. 1 Phase 1 (LFF) 30 25 20 Average

15 10

Linear (Average) y = -1.88x + 27

5 0 1

2

3

4

Auditors' Scores /100

Auditors' Scores /100

35

S. 1 Phase 2 (HFF)

65 60 55 50 45 40 35 30 25 20 15 10

Average Linear (Average)

y = 14.16x - 3.4 5

5

6

8

9

Week

Week

S. 1 Phase 3 (LFF)

S. 1 Phase 4 (HFF)

35

45

30 25

Average Linear (Average) y = -0.64x + 29.74

20 15 9

10

11

12

Auditors' Score /100

Auditors' Score /100

7

40 35

Average

30

Linear (Average)

25 20

y = 6.1x + 17

15

13

Week

Figure 4.6 – Subject one average intonation results by phase

13

14 Week

15

68 Subject Two Subject two was a 19 year old female who had studied voice with the researcher for one year prior to the start of the study. Additionally, she reported having studied private voice with her high school choir teacher for four years prior to college. In addition to her private voice instruction, subject two reported having participated in choirs for her four years in high school as well as one year at college. Subject two reported having no record of vocal health issues. Nor did she indicate having had chronic loss of voice. Health related issues did not present any problems during her participation in the study. Subject two was assigned to g-1, meaning that she received low-frequency feedback during weeks one through four, high-frequency feedback during weeks five through eight, lowfrequency feedback during weeks nine through twelve, and high-frequency feedback during weeks thirteen through fifteen. She did not appear for her lesson on week three, the absence of data for which is exhibited in the results for this subject. Over the course of the study, subject two reported having practiced an average of 3.5 hours/week. Unlike subject one, subject two‟s practice was more uniform throughout the term with 3.75, and 3.25 hours/week during phases one and two respectively, and 3.75 and 3.0 hours/week during phases three and four respectively.

Subject Two Tone Quality The aggregate scores for the tone quality of subject two from all five auditors are reported in the chart in table 4.4 and in the graph in figure 4.7.

69

Week

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Subject Two Tone Quality Aud. 2 Aud. 3 Aud. 4 76 40 60 62 30 55

75 75

50 40

Average 60.2 52.4

62 60 60 70 64 64 58 58 60 58 62 52

75 40 35 60 25 60 40 40 25 60 50 40

50 40 40 25 40 50 50 40 60 40 60 30

56.4 43 50 54 49.8 55.8 52.6 42.6 45 50.6 54.4 39.4

20 25 35 60 65 50 50 25 30 50 35 20

75 50 80 55 55 55 65 50 50 45 65 55

Aud. 5

Table 4.4 – Subject two tone quality results

Auditors' Score /100

Subject Two Tone Quality 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average

1

2

3

4

5

6

7

8 9 Week

Figure 4.7 – Subject two tone quality results

10

11

12

13

14

15

70 Figure 4.8 below presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 2 Phase 2 (HFF) 65

60

60

50 40 Average

30 20

Linear (Average) y = -3.04x + 62.12

10

Auditors' Score /100

Auditors' Score /100

S. 2 Phase 1 (LFF) 70

0

55 50

Average

45

Linear (Average) y = 2.54x + 42.9

40 35

1

2

3

4

5

5

Week

60

50

55 50 Average

45

Linear (Average) y = -1.8x + 54.72

40 35 12

13

Auditors' Score /100

Auditors' Score /100

60

11

8

9

S. 2 Phase 4 (HFF)

65

10

7 Week

S. 2 Phase 3 (LFF)

9

6

40 30

Average

20

Linear (Average) y = -5.6x + 59.333

10 0 13

Week

Figure 4.8 – Subject two average tone quality results by phase

14 Week

15

71 Subject Two Breath Management The aggregate scores for the breath management of subject two from all five auditors are presented in the chart in table 4.5 and in the graph in figure 4.9.

Week

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Subject Two Breath Management Aud. 2 Aud. 3 Aud. 4 80 40 45 68 20 50

80 90

50 40

59 53.6

70 70 70 70 70 75 60 76 64 60 66 58

95 75 75 80 65 85 80 75 90 98 90 65

50 50 50 50 50 60 50 50 40 50 50 40

61 54 59 60 59 62 55 53.2 52.8 60.6 58.2 45.6

15 15 25 50 50 40 25 10 20 40 20 20

Aud. 5

75 60 75 50 60 50 60 55 50 55 65 45

Average

Table 4.5 – Subject two breath management results

Auditors' Score /100

Subject Two Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

9

Week

Figure 4.9 – Subject two breath management results

10

11

12

13

14

15

72 Figure 4.10 presents the average of the scores for breath management from all five auditors, divided into phases and including a linear trend line for each phase.

S. 2 Phase 1 (LFF)

S. 2 Phase 2 (HFF) 70

65 60 Average

55

Linear (Average) y = -0.26x + 57.68

50 45 1

2

3

4

Auditors' Score /100

Auditors' Score /100

70

65 60 Average

55

Linear (Average) y = 1.6x + 54

50 45

5

5

6

Week

8

9

Week

S. 2 Phase 3 (LFF)

S. 2 Phase 4 (HFF)

70

70

65

65

60 Average

55

Linear (Average) y = -0.5x + 58.22

50 45

Auditors' Score /100

Auditors' Score /100

7

60 Average

55

Linear (Average) y = -7.5x + 69.8

50 45

9

10

11 Week

12

13

13

14 Week

Figure 4.10 – Subject two average breath management results by phase

15

73 Subject Two Intonation The aggregate scores for the intonation of subject two from all five auditors are presented in the chart in table 4.6 and in the graph in figure 4.11.

Week

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Subject Two Intonation Aud. 2 Aud. 3 Aud. 4 82 40 70 78 40 75

95 85

Average 85 74.4 75 70.6

75 78 62 70 66 78 78 80 78 70 75 68

90 90 95 90 90 90 80 80 98 90 80 90

80 75 80 85 80 75 60 75 85 85 85 85

15 10 30 40 60 40 15 10 20 45 40 20

85 80 75 45 85 75 90 72 75 75 90 80

Aud. 5

69 66.6 68.4 66 76.2 71.6 64.6 63.4 71.2 73 74 68.6

Auditors' Score /100

Table 4.6 – Subject two intonation results

Subject Two Intonation

100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8 9 Week

Figure 4.11 – Subject two intonation results

10

11

12

13

14

15

74 Figure 4.12 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 2 Phase 2 (HFF) 80

75

75

70 65

Average

60

Linear (Average)

Auditors' Score /100

Auditors' Score /100

S. 2 Phase 1 (LFF) 80

y = -1.72x + 75.31

55 1

2

3

4

70 65

Average

60

Linear (Average) y = 1.78x + 64.42

55

5

5

6

Week

S. 2 Phase 3 (LFF)

8

9

S. 2 Phase 4 (HFF)

80

80

75 70 Average

65

Linear (Average) y = 0.94x + 65.94

60 55 9

10

11

12

13

Auditors' Score /100

Auditors' Score /100

7 Week

75 70

Average Linear (Average) y = -2.2x + 76.267

65 60 13

Week

Figure 4.12 – Subject two average intonation results by phase

14 Week

15

75 Subject Three Subject three was a 19 year old male who had studied private voice with the researcher for one year prior to the start of the study. He reported having never studied private voice prior to studying with the researcher. However, subject three reported that he had participated in several high school and community choirs over the previous six years. Subject three reported no previous vocal health issues or chronic loss of voice. Health related issues did not present any problems throughout his participation in the study. Subject three was assigned to g-1, meaning that he received low-frequency feedback during weeks one through four, high-frequency feedback during weeks five through eight, lowfrequency feedback during weeks nine through twelve, and high-frequency feedback during weeks thirteen through fifteen. He appeared for all of his lessons throughout the term, leaving no gaps in the data. Throughout the course of the study, subject three reported having practiced an average of 1 hour/week. Similar to subject two, subject three‟s practice time was relatively consistent throughout the duration of the study. He reported an average of 1.1 and 1.0 hours/week during phases one and two respectively, and 1.2 and 0.75 hours/week during phases three and four respectively.

Subject Three Tone Quality The aggregate scores for the tone quality of subject three from all five auditors are presented in the chart in table 4.7 and in the graph in figure 4.13.

76

Week

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

58 52 58 56 62 58 64 60 56 64 53 53 50 56 60

Subject Three Intonation Aud. 2 Aud. 3 Aud. 4 35 50 25 45 25 35 20 40 35 40 40 60 25 50 30 55 30 30 30 45 10 60 20 40 35 40 30 55 30 55

Aud. 5 40 25 40 10 20 60 30 20 20 15 20 40 30 25 10

50 50 50 50 50 60 50 50 50 50 60 50 50 60 50

Average 46.6 39.4 41.6 35.2 41.4 55.6 43.8 43 37.2 40.8 40.6 40.6 41 45.2 41

Table 4.7 – Subject three tone quality results

Subject Three Tone Quality 70

Auditors' Score /100

60 Aud. 1

50

Aud. 2

40

Aud. 3

30

Aud. 4

20

Aud. 5

10

Average

0 1

2

3

4

5

6

7

8 Week

Figure 4.13 – Subject three tone quality results

9

10

11

12

13

14

15

77 Figure 4.14 presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 3 Phase 2 (HFF)

60

60

55

55

50 45 Average

40

Linear (Average) y = -1.46x + 45.22

35 30 1

2

3

4

Auditors' Score /100

Auditors' Score /100

S. 3 Phase 1 (LFF)

50 45 Average 40 Linear (Average) y = -2.1x + 50.5

35 30

5

5

6

Week

S. 3 Phase 3 (LFF)

8

9

S. 3 Phase 4 (HFF)

55

55

50 45

Average Linear (Average)

40

y = 0.74x + 37.82

35 9

10

11

12

13

Auditors' Score /100

Auditors' Score /100

7 Week

50 45

Average Linear (Average)

40

y = 0.0x + 42.4

35 13

Week

Figure 4.14 – Subject three average tone quality results by phase

14 Week

15

78 Subject Three Breath Management The aggregate scores for the breath management of subject three from all five auditors are presented in the chart in table 4.8 and in the graph in figure 4.15.

Week

Aud. 1

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

HFF

Subject Three Breath Management Aud. 2 Aud. 3 Aud. 4 25 30 10 50 5 45 15 35 20 40 15 60 20 40 10 60 20 45 10 40 10 55 20 30 25 20 20 50 20 40

62 48 52 54 65 56 50 52 48 64 65 64 60 62 52

Aud. 5 75 90 60 40 60 80 40 60 75 50 50 65 65 75 75

25 25 40 20 25 40 25 25 40 25 25 20 25 50 40

Average 43.4 44.6 40.4 32.8 42 50.2 35 41.4 45.6 37.8 41 39.8 39 51.4 45.4

Table 4.8 – Subject three breath management results

Auditors' Score /100

Subject Three Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

9

10

Week

Figure 4.15 – Subject three breath management results

11

12

13

14

15

79 Figure 4.16 presents the average of the scores for breath management from all five auditors, divided into phases and including a linear trend line for each phase.

S. 3 Phase 1 (LFF)

S. 3 Phase 2 (HFF) 55

45 40 Average

35

Linear (Average)

30

Auditors' Score /100

Auditors' Score /100

50

y = -1.46x + 45.02

25 1

2

3

4

50 45 40 Average

35

Linear (Average) y = -0.16x + 43.32

30 25

5

5

6

Week

55

55

50

50

45 Average

40

Linear (Average)

35

y = -1.12x + 44

30 10

11 Week

8

9

S. 3 Phase 4 (HFF)

12

13

Auditors' Score /100

Auditors' Score /100

S. 3 Phase 3 (LFF)

9

7 Week

45 40

Average

35

Linear (Average) y = 3.2x + 38.867

30 13

14 Week

Figure 4.16 – Subject three average breath management results by phase

15

80 Subject Three Intonation The aggregate scores for the intonation of subject three from all five auditors are presented in the chart in table 4.9 and in the graph in figure 4.17.

Week

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

25 44 48 28 38 42 55 46 38 49 44 28 38 42 45

Subject Three Intonation Aud. 2 Aud. 3 Aud. 4 0 10 5 40 0 30 0 20 5 45 0 50 5 75 5 70 5 25 0 55 0 50 0 15 0 10 5 60 5 45

Aud. 5 30 50 30 10 60 40 50 60 30 30 40 35 35 50 40

Average 10 25 25 10 50 25 60 40 50 50 50 10 10 50 50

15 32.8 26.6 13.6 39.6 31.4 49 44.2 29.6 36.8 36.8 17.6 18.6 41.4 37

Table 4.9 – Subject three intonation results

Subject Three Intonation 80

Auditors' Score /100

70 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 Week

Figure 4.17 – Subject three intonation results

9

10

11

12

13

14

15

81 Figure 4.18 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 3 Phase 2 (HFF)

45

55

40

50

35 30 25

Average

20

Linear (Average) y = 3x + 16.52

15 10 1

2

3

4

Auditors' Score /100

Auditors' Score /100

S. 3 Phase 1 (LFF)

45 40 Average

35

Linear (Average) y = -0.72x + 40.92

30 25

5

5

6

Week

40

45

35

40

30 25 Average 20 Linear (Average) y = -4.12x + 40.24

15 10 10

11

8

9

S. 3 Phase 4 (HFF)

12

13

Auidtors' Score /100

Auditors' Score /100

S. 3 Phase 3 (LFF)

9

7 Week

35 30

Average

25

Linear (Average) y = 9.2x + 13.933

20 15 13

Week

Figure 4.18 – Subject three avearge intonation results by phase

14 Week

15

82 Subject Four Subject four was a 20 year old female who had studied private voice with the researcher for one semester prior to the onset of the study. Prior to that point, she had not studied private voice with any other instructors. However, subject four reported having participated in choir for one year during high school and for one year at college. She reported having no vocal health issues or chronic loss of voice. Health related issues did not present any problem for her participation in the study. Subject four was assigned to g-1, meaning that she received low-frequency feedback during weeks one through four, high-frequency feedback during weeks five through eight, lowfrequency feedback during weeks nine through twelve, and high-frequency feedback during weeks thirteen through fifteen. She did not appear for her lesson on week fourteen, the absence of data for which is exhibited in the results for this subject. Throughout the course of the study, subject four reported having practiced an average of 5.4 hours/week. While subject four‟s practice time was consistently higher than that of many other subjects, it did vary from phase to phase. She practiced most frequently during phases one and three (both LFF phases), averaging 5.8 and 7.8 hours/week respectively. Practice time dropped during phases two and four (both HFF phases) to 5.3 and 3.7 hours/week respectively. However, the significantly higher practice levels in phases one and three did not consistently produce higher scores on the retention/transfer tests.

Subject Four Tone Quality The aggregate scores for the tone quality of subject four from all five auditors are presented in the chart in table 4.10 and in the graph in figure 4.19.

83

Week

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

60 60 54 54 54 60 62 58 58 54 62 60 48

Subject Four Tone Quality Aud. 2 Aud. 3 Aud. 4 35 50 30 50 40 55 25 55 25 75 40 40 25 55 30 50 25 75 40 45 60 70 50 60 40 60

54

40

65

Aud. 5

Average

45 25 40 40 45 60 40 30 50 45 30 40 35

40 30 50 60 30 65 40 70 50 50 60 40 25

46 39 47.8 46.8 45.8 53 44.4 47.6 51.6 46.8 56.4 50 41.6

40

25

44.8

Table 4.10 – Subject four tone quality results

Subject Four Tone Quality 80

Auditors' Score /100

70 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 Week

Figure 4.19 – Subject four tone quality results

9

10

11

12

13

14

15

84 Figure 4.20 presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 4 Phase 1 (LFF)

S. 4 Phase 2 (HFF) 60

50 45 Average

40

Linear (Average) y = 0.74x + 42.86

35 30 1

2

3

4

Auditors' Score /100

Auditors' Score /100

55

55 50 Average Linear (Average) y = 0.62x + 46.62

45 40

5

5

6

Week

S. 4 Phase 3 (LFF)

8

9

S. 4 Phase 4 (HFF)

60

55

55 50 45

Average Linear (Average) y = -1.68x + 54.32

40 35 9

10

11

12

13

Auditors' Score /100

Auditors' Score /100

7 Week

50 45 Average 40

Linear (Average) y = 1.6x + 40

35 13

Week

Figure 4.20 – Subject foure average tone quality results by phase

14 Week

15

85 Subject Four Breath Management The aggregate scores for the tone quality of subject four from all five auditors are presented in the chart in table 4.11 and in the graph in figure 4.21.

Week

Subject Four Breath Management Aud. 2 Aud. 3 Aud. 4 20 50 15 50 25 50 20 55 25 75 30 40 20 65 20 50 10 70 25 50 40 65 30 50 25 40

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

68 60 56 72 62 64 70 68 52 60 52 52 56 60

25

50

Aud. 5 90 90 40 60 60 90 50 60 80 70 60 90 45

30 50 50 50 50 65 50 60 50 40 60 50 40

Average 51.6 53 44.2 51.4 54.4 57.8 51 51.6 52.4 49 55.4 54.4 41.2

60

40

47

Table 4.11 – Subject four breath management results

Auditors' Score /100

Subject Four Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average

1

2

3

4

5

6

7

8

9

10

Week

Figure 4.21 – Subject four breath management results

11

12

13

14

15

86 Figure 4.22 presents the average of the scores for breath management from all five auditors, divided into phases and including a linear trend line for each phase.

S. 4 Phase 2 (HFF)

60

60

55

55

50 Average

45

Linear (Average)

40

Auditors' Score /100

Auditors' Score /100

S. 4 Phase 1 (LFF)

y = 0.4x + 49.72

35 1

2

3

4

50 Average

45

Linear (Average) y = -1.02x + 56.5

40 35

5

5

Week

55

55

50 Average

45

Linear (Average)

40

y = -1.7x + 55.58

35 11

12

13

Auditors' Score /100

Auditors' Score /100

60

Week

8

9

S. 4 Phase 4 (HFF)

60

10

7 Week

S. 4 Phase 3 (LFF)

9

6

50 Average

45

Linear (Average)

40

y = 2.9x + 38.3

35 13

14 Week

Figure 4.22 – Subject four average breath management results by phase

15

87 Subject Four Intonation The aggregate scores for the intonation of subject four from all five auditors are presented in the chart in table 4.12 and in the graph in figure 4.23.

Week

Subject Four Intonation Aud. 2 Aud. 3 Aud. 4 20 80 20 80 25 80 5 80 20 85 40 70 10 75 30 80 15 90 40 60 15 85 10 70 50 80

Aud. 1

HFF

LFF

HFF

LFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

78 75 75 70 75 72 76 76 74 70 74 70 75 80

30

75

Aud. 5 60 90 90 75 75 90 60 70 90 90 80 80 90

90 85 85 80 80 80 80 80 85 50 85 75 80

Average 65.6 70 71 62 67 70.4 60.2 67.2 70.8 62 67.8 61 75

75

80

68

Table 4.12 – Subject four intonation results

Auditors' Score /100

Subject Four Intonation 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8 Week

Figure 4.23 – Subject four intonation results

9

10

11

12

13

14

15

88 Figure 4.24 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 4 Phase 1 (LFF)

S. 4 Phase 2 (HFF) 80

70 65

Average Linear (Average)

60

Auditors' Score /100

Auditors' Score /100

75

75 70 Average

65

Linear (Average)

60

y = -0.52x + 68.68 55

y = 0.44x + 65.8

55 1

2

3

4

5

5

6

Week

S. 4 Phase 3 (LFF)

8

9

S. 4 Phase 4 (HFF) 80

75 70 Average

65

Linear (Average)

60

Auditors' Score /100

80 Auditors' Score /100

7 Week

75 70 Average Linear (Average) y = -3.5x + 78.5

65

y = 0.74x + 65.1 55

60 9

10

11

12

13

13

Week

Figure 4.24 – Subject four avearge intonation results by phase

14 Week

15

89 Subject Five Subject five was a 19 year old female who had studied private voice with the researcher for one year prior to the onset of the study. Prior to that point, she had not studied private voice with any other instructor. Subject five did, however, report having participated in choir for four years in high school and for one year at college. She reported having problems with chronic laryngitis, having lost her voice more than once during the span of six months, which the researcher had witnessed during his experience with this subject in the previous year. While health problems had been a significant factor for this subject in the past, she was able to remain healthy throughout the duration of the study and health related issues did not present any problem for her participation in the study. Subject five was assigned to group g-2 meaning that she received the reverse feedback frequency phase progression as the previous subjects. She received high-frequency feedback during phase one, low-frequency feedback during phase two, high-frequency feedback during phase three, and low-frequency feedback during phase four. Subject five did not appear for her lesson at week five, the absence of data for which is exhibited in the results for this subject. Over the course of the study, subject five reported having practiced an average of 1.0 hours/week. However, subject five‟s practice was less consistent over the course of the study. She reported having practiced an average of 1.3 and 1.1 hours/week during phases one and two respectively, and 1.0 and 0.6 hours/week during phases three and four respectively.

Subject Five Tone Quality The aggregate scores for the tone quality of subject five from all five auditors are presented in the chart in table 4.13 and in the graph in figure 4.25.

90

Week

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

58 53 56 45

Subject Five Tone Quality Aud. 2 Aud. 3 Aud. 4 30 45 25 25 30 55 15 25

50 48 42 50 52 58 45 46 52 62

35 20 30 20 20 20 40 15 25 20

50 35 25 20 20 35 45 45 30 40

25 40 60 30

Aud. 5 40 50 50 75

Average 39.6 38.6 50.2 38

60 30 30 45 25 25 30 20 10 20

50 25 25 25 25 30 50 25 10 50

49 31.6 30.4 32 28.4 33.6 42 30.2 25.4 38.4

Table 4.13 – Subject five tone quality results

Subject Five Tone Quality 80 Auditors' Score /100

70 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 Week

Figure 4.25 – Subject five tone quality results

9

10

11

12

13

14

15

91 Figure 4.26 presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 5 Phase 2 (LFF)

60

55

55

50

50 45 Average 40 Linear (Average) y = 0.68x + 39.9

35 30 1

2

3

4

Auditors' Score /100

Auditors' Score /100

S. 5 Phase 1 (HFF)

45 40 Average 35 Linear (Average) y = -5.22x + 54.02

30 25

5

5

6

Week

50

45

45

40

40 Average

35

Linear (Average) y = 1x + 30.24

30 25 10

11

8

9

S. 5 Phase 4 (LFF)

12

13

Auditors' Score /100

Auditors' Score /100

S. 5 Phase 3 (HFF)

9

7 Week

35 Average

30

Linear (Average) y = 4.1x + 23.133

25 20 13

Week

Figure 4.26 – Subject five avearge tone quality results by phase

14 Week

15

92 Subject Five Breath Management The aggregate scores for the breath management of subject five from all five auditors are presented in the chart in table 4.14 and in the graph in figure 4.27.

Week

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

LFF

Subject Five Breath Management Aud. 2 Aud. 3 Aud. 4 15 50 15 20 5 50 20 25

Aud. 1 52 74 62 52 62 52 54 54 60 48 70 52 50 68

25 10 20 15 5 10 30 10 5 10

55 35 35 25 20 50 40 50 35 35

Aud. 5 75 60 80 75

40 0 50 50

Average 46.4 33.8 49.4 44.4

70 70 60 40 60 75 60 70 75 80

50 25 40 25 40 50 25 10 10 40

52.4 38.4 41.8 31.8 37 46.6 45 38.4 35 46.6

Table 4.14 – Subject five breath management results

Subject Five Breath Management 90 Auditors' Score /100

80 70

Aud. 1

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

Week

Figure 4.27 – Subject five breath management results

10

11

12

13

14

15

93 Figure 4.28 presents the average of the scores for breath management from all five auditors, divided into phases and including a linear trend line for each phase.

S. 5 Phase 2 (LFF)

60

60

55

55

50 45

Average

40 Linear (Average) y = 0.96x + 41.1

35 30 1

2

3

4

Auditors' Score /100

Auditors' Score /100

S. 5 Phase 1 (HFF)

50 45 Average 40 Linear (Average) y = -5.84x + 61.54

35 30

5

5

6

Week

50

50

45

45

40 Average

35

Linear (Average) y = 2.12x + 33.4

30 25 10

11 Week

8

9

S. 5 Phase 4 (LFF)

12

13

Auditors' Score /100

Auditors' Score /100

S. 5 Phase 3 (HFF)

9

7 Week

40 Average

35

Linear (Average) y = 4.1x + 31.8

30 25 13

14 Week

Figure 4.28 – Subject five avearge breath management results by phase

15

94 Subject Five Intonation The aggregate scores for the intonation of subject five from all five auditors are presented in the chart in table 4.15 and in the graph in figure 4.29.

Week

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

54 72 78 75

Subject Five Intonation Aud. 2 Aud. 3 Aud. 4 5 60 10 60 10 85 30 50

75 52 68 49 75 68 62 48 42 78

10 10 30 15 10 10 30 5 5 15

Aud. 5

75 75 55 70 80 60 55 60 55 60

60 65 50 90

70 60 60 75

Average 49.8 53.4 56.6 64

70 80 60 70 75 45 90 40 75 60

50 50 75 50 70 60 50 50 25 60

56 53.4 57.6 50.8 62 48.6 57.4 40.6 40.4 54.6

Table 4.15 – Subject five intonation results

Auditors' Score /100

Subject Five Intonation 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average

1

2

3

4

5

6

7

8 Weeks

Figure 4.29 – Subject five intonation results

9

10

11

12

13

14

15

95 Figure 4.30 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 5 Phase 2 (LFF)

70

70

65

65

60 Average 55

Linear (Average)

50

Auditors' Score /100

Auditors' Score /100

S. 5 Phase 1 (HFF)

60 Average

55

Linear (Average)

50

y = 4.58x + 44.5 45

y = -1.14x + 58.44

45 1

2

3

4

5

5

6

Week

65

60

60

55

55 50

Average

45

Linear (Average)

40

y = -2.5x + 59.38

35 10

11

8

9

S. 5 Phase 4 (LFF)

12

13

Auditors' Score /100

Auditors' Score /100

S. 5 Phase 3 (HFF)

9

7 Week

50 45 Average 40 Linear (Average)

35

y = 7x + 31.2

30 13

Week

Figure 4.30 – Subject five average intonation results by phase

14 Week

15

96 Subject Six Subject six was a 21 year old male who had studied private voice with the researcher for one year prior to the onset of the study. Prior to that point, he had studied at a college level for three years and for one year in high school. Additionally, subject six had participated in several school and community choirs over the previous ten years. Subject six reported having no vocal health issues or chronic loss of voice. Health related issues did not present any problems for his participation in the study. Subject six was assigned to group g-2 meaning that he received high-frequency feedback during phase one, low-frequency feedback during phase two, high-frequency feedback during phase three, and low-frequency feedback during phase four. Subject six failed to appear for his lessons on weeks eight, eleven, and fifteen, the absence of data for which is exhibited in the results for this subject. These absences pose some question to the validity of the results from the phases three and four. Subject six misplaced his practice record at some point during the study and was unable to return it to the researcher at the conclusion of the study. Consequently he had no record of the amount of time he had spent practicing.

Subject Six Tone Quality The aggregate scores for the tone quality of subject six from all five auditors are presented in the chart in table 4.16 and in the graph in figure 4.31.

97

Week

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

78 72 80 78 76 78 85

Subject Six Tone Quality Aud. 2 Aud. 3 Aud. 4 50 80 50 75 60 80 60 80 50 85 60 80 90 75

90 75 75 75 80 75 65

Aud. 5 100 100 100 90 100 100 100

Average 79.6 74.4 79 76.6 78.2 78.6 83

84 85

60 80

85 85

75 90

90 90

78.8 86

80 90 87

60 90 80

90 65 85

75 65 70

100 100 90

81 82 82.4

Table 4.16 – Subject six tone quality results

Auditors' Score /100

Subject Six Tone Quality 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average

1

2

3

4

5

6

7

8 Week

Figure 4.31 – Subject six tone quality results

9

10

11

12

13

14

15

98 Figure 4.32 presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 6 Phase 1 (HFF)

S. 6 Phase 2 (LFF) 85

80 75

Average Linear (Average)

70

Auditors' Score /100

Auditors' Score /100

85

y = -0.06x + 77.74

65 1

2

3

4

80 75

Average Linear (Average) y = 0.2571x + 78.943

70 65 5

5

6

8

9

Week

Week

S. 6 Phase 3 (HFF)

S. 6 Phase 4 (LFF) 90

85 80

Average Linear (Average)

75

y = 0.14x + 81.53

70 9

10

11

12

13

Auditors Score /100

90 Auditors' Score /100

7

85 80

Average Linear (Average)

75

y = 0.4x + 81.6

70 13

Week

Figure 4.32 – Subject six average tone quality results by phase

14 Week

15

99 Subject Six Breath Management The aggregate scores for the breath management of subject six from all five auditors are presented in the chart in table 4.17 and in the graph in figure 4.33.

Week

Subject Six Breath Management Aud. 2 Aud. 3 Aud. 4 40 90 40 65 50 80 50 80 40 80 50 75 85 60

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

82 74 70 75 72 62 80

Aud. 5 90 90 90 80 90 85 80

100 80 85 85 100 85 60

Average 80.4 69.8 75 74 76.4 71.4 73

82 78

50 60

85 85

80 95

75 85

74.4 80.6

82 85 84

40 80 55

90 60 90

80 70 90

90 50 80

76.4 69 79.8

Table 4.17 – Subject six breath management results

Auditors' Score /100

Subject Six Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

9

Week

Figure 4.33 – Subject six breath management results

10

11

12

13

14

15

100 Figure 4.34 presents the average of the scores for breath management from all five auditors, divided into phases and including a linear trend line for each phase.

S. 6 Phase 1 (HFF)

S. 6 Phase 2 (LFF) 85

80 75

Average Linear (Average) y = -0.38x + 76.26

70

Auditors' Score /100

Auditors' Score /100

85

65

80 75

Average Linear (Average)

70

y = -0.1829x + 74.303

65 1

2

3

4

5

5

6

Week

S. 6 Phase 3 (HFF)

8

9

S. 6 Phase 4 (LFF) 90

80 75

Average Linear (Average)

70

y = -1.5x + 79.6

65 9

10

11 Week

12

13

Auditors' Score /100

85 Auditors' Score /100

7 Week

85 80 Average

75

Linear (Average) y = 10.8x + 58.2

70 65 13

14 Week

Figure 4.34 – Subject six average breath management results by phase

15

101 Subject Six Intonation The aggregate scores for the intonation of subject six from all five auditors are presented in the chart in table 4.18 and in the graph in figure 4.35.

Week

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

70 45 52 49 60 64 78

Subject Six Intonation Aud. 2 Aud. 3 Aud. 4 30 90 25 65 10 80 40 75 30 80 40 75 75 60

Aud. 5

Average

75 70 70 90 75 75 60

100 75 85 75 100 100 50

73 56 59.4 65.8 69 70.8 64.6

65 88

30 60

75 90

80 95

85 90

67 84.6

48 60 54

10 25 40

90 20 85

80 60 95

85 90 75

62.6 51 69.8

Table 4.18 – Subject six intonation results

Auditors' Score /100

Subject Six Intonation 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8 Week

Figure 4.35 – Subject six intonation results

9

10

11

12

13

14

15

102 Figure 4.36 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 6. Phase 1 (HFF)

S. 6 Phase 2 (LFF) 80

75 70 65

Average

60

Linear (Average) y = 0.18x + 64.1

55

Auditors' Score /100

Auditors' Socre /100

80

50

75 70

Average

65

Linear (Average) y = -0.7943x + 70.034

60 1

2

3

4

5

5

6

Week

S. 6 Phase 3 (HFF)

8

9

S. 6 Phase 4 (LFF) 90

85 80 75 70 Average

65 60

Linear (Average)

55 50

y = -5.4x + 82.5

45 9

10

11

12

13

Auditors' Score /100

90 Auditors' Score /100

7 Week

80 70

Average Linear (Average) y = 18.8x + 32.2

60 50 13

Week

Figure 4.36 – Subject six average intonation results by phase

14 Week

15

103 Subject Seven Subject seven was an 18 year old male who had studied private voice for six months with his high school choir teacher prior to the onset of the study. In addition to his private voice instruction, he reported having participated in choirs for three years in high school. Subject seven reported no vocal health issues or chronic loss of voice. Health related issues did not present any problems for his participation in the study. Subject seven was assigned to group g-2 meaning that he received high-frequency feedback during phase one, low-frequency feedback during phase two, high-frequency feedback during phase three, and low-frequency feedback during phase four. While subject seven did appear at every lesson, he did not choose to participate in the study until week two. This meant that his phase one was only three weeks in duration. For consistency with the other subjects, however, week one is indicated in the reporting of his results as having no score. Over the course of the study, subject seven reported having practiced an average of 1.8 hours/week. Adjusted for phases, he practiced an average of 2.1 and 1.9 hours/week during phases one and two respectively, and 1.6 and 2.0 hours/week during phases three and four respectively.

Subject Seven Tone Quality The aggregate scores for the intonation of subject seven from all five auditors are presented in the chart in table 4.19 and in the graph in figure 4.37.

104

Week

Subject Seven Tone Quality Aud. 2 Aud. 3 Aud. 4

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

76 75 72 70 66 65 72 75 76 65 76 70 75 78

60 60 50 50 60 40 50 65 60 60 50 45 50 40

70 75 75 75 75 55 70 75 75 65 75 75 70 75

Aud. 5 75 70 30 80 60 60 75 40 40 35 60 75 70 40

Average 60 80 70 100 75 60 80 75 80 70 75 80 80 90

68.2 72 59.4 75 67.2 56 69.4 66 66.2 59 67.2 69 69 64.6

Table 4.19 – Subject seven tone quality results

Auditors' Score /100

Subject Seven Tone Quality 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8 Week

Figure 4.37 – Subject seven tone quality results

9

10

11

12

13

14

15

105 Figure 4.38 presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 7 Phase 2 (LFF)

80

80

75

75

70 Average

65

Linear (Average) y = 0.78x + 65.92

60 55 1

2

3

4

Auditors' Score /100

Auditors' Score /100

S. 7 Phase 1 (HFF)

70 Average

65

Linear (Average) y = -1.58x + 71.46

60 55

5

5

6

Week

75

75

70 Average

65

Linear (Average) y = 0.7x + 63.38

60 55 12

13

Auditors' Score /100

Auditors' Score /100

80

11

9

S. 7 Phase 4 (LFF)

80

10

8

Week

S. 7 Phase 3 (HFF)

9

7

70 Average

65

Linear (Average) y = -2.2x + 71.933

60 55 13

Week

Figure 4.38 – Subject seven average tone quality results by phase

14 Week

15

106 Subject Seven Breath Management The aggregate scores for the breath management of subject seven from all five auditors are presented in the chart in table 4.20 and in the graph in figure 4.39.

Week

Subject Seven Breath Management Aud. 2 Aud. 3 Aud. 4

Aud. 1

LFF

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

60 60 68 40 56 52 62 70 62 75 50 70 60 58

40 30 50 10 15 15 50 50 50 40 5 10 25 30

80 65 80 50 70 25 60 70 70 50 60 65 60 60

Aud. 5 95 90 65 60 80 70 75 80 50 80 85 90 90 45

Average 70 90 50 70 85 60 50 75 75 60 60 90 80 90

69 67 62.6 46 61.2 44.4 59.4 69 61.4 61 52 65 63 56.6

Table 4.20 – Subject seven breath management results

Auditors' Score /100

Subject Seven Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

9

10

Week

Figure 4.39 – Subject seven breath management results

11

12

13

14

15

107 Figure 4.40 presents the average of the scores for breath management from all five auditors, divided into phases and including a linear trend line for each phase.

S. 7 Phase 2 (LFF)

75

75

70

70

65 60 Average

55 50

Linear (Average) y = -7.34x + 86.84

45

Auditors' Score /100

Auditors' Score /100

S. 7 Phase 1 (HFF)

65 60 55

Average

50

Linear (Average) y = 4.42x + 42.74

45

40

40 1

2

3

4

5

5

6

Week

75

75

70

70

65 60 55

Average

50

Linear (Average) y = -1.74x + 66.9

45 40 10

11 Week

8

9

S. 7 Phase 4 (LFF)

12

13

Auditors' Score /100

Auditors' Score /100

S. 7 Phase 3 (HFF)

9

7 Week

65 60 55

Average

50

Linear (Average) y = -4.2x + 69.933

45 40 13

14 Week

Figure 4.40 – Subject seven average breath management results by phase

15

108 Subject Seven Intonation The aggregate scores for the breath management of subject seven from all five auditors are presented in the chart in table 4.21 and in the graph in figure 4.41.

Week

HFF

LFF

HFF

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

LFF

Subject Seven Intonation Aud. 2 Aud. 3 Aud. 4

Aud. 1 49 46 51 48 45 50 40 28 38 46 48 48 30 40

10 10 25 0 10 0 0 10 10 10 5 5 0 0

60 80 80 75 60 40 55 10 40 15 70 45 45 35

Aud. 5 80 70 50 50 50 52 30 30 40 60 95 50 75 50

Average 50 85 25 75 75 25 10 25 25 50 85 75 50 50

49.8 58.2 46.2 49.6 48 33.4 27 20.6 30.6 36.2 60.6 44.6 40 35

Table 4.21 – Subject seven intonation results

Auditors' Score /100

Subject Seven Intonation 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8 Week

Figure 4.41 – Subject seven intonation results

9

10

11

12

13

14

15

109 Figure 4.42 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 7 Phase 1 (HFF)

S. 7 Phase 2 (LFF) 55 50

65 60 55 Average 50 Linear (Average) y = -1.26x + 55.36

45 40 1

2

3

4

Auditors' Score /100

Auditors' Score /100

70

45 40 35

Average

30 Linear (Average)

25 20

y = -7.9x + 59.42

15

5

5

6

Week

S. 7 Phase 3 (HFF)

8

9

65 60 55 50 45 40 35 30 25 20 15

S. 7 Phase 4 (LFF) 50

Average Linear (Average) y = 7.8x + 15.12 9

10

11

12

13

Auditors' Score /100

Auditors' Score /100

7 Week

45 40

Average Linear (Average) y = -4.8x + 49.467

35 30 13

Week

Figure 4.42 – Subject seven average intonation results by phase

14 Week

15

110 Subject Eight Subject eight was a 25 year old female who had studied private voice with another college instructor for one year prior to the onset of the study. She reported having never participated in choirs previous to the start of the study period, nor had she studied voice with the researcher prior to the study. While subject eight reported having no known vocal pathologies or a history of chronic vocal health problems, she acquired an illness, causing her to withdraw from college, and consequently the study, after week nine of the study period. Also, as mentioned briefly in chapter three, subject eight presented with an unusually low speaking and singing voice. Consequently, she performed the retention/transfer tests in the key of G Major, while all other low voices perfromed the test in the key of Bb Major She reported that her voice had always been low and that her previous voice instructor assigned vocal literature in a tenor range for her to sing. Over the course of the first few meetings, the researcher found that subject eight was failing to access her head register and was singing exclusively using her chest register. In that same time period, subject eight was introduced to the head register and the bulk of the instruction time during her lesson times was spent practicing moving in to and out of her head register. By doing so, her range was extended to a relatively normal low alto range. However, for the sake of continuity, she continued to perform her retention/transfer tests in the key of G Major. Because of her sudden and unexpected departure from the study, subject eight did not return her practice log. Consequently, there is no record of the amount of time she spent practicing during her participation in the study.

111 Subject Eight Tone Quality The aggregate scores for the tone quality of subject eight from all five auditors are presented in the chart in table 4.22 and in the graph in figure 4.43.

Week

Subject Eight Tone Quality Aud. 2 Aud. 3 Aud. 4

Aud. 1

LFF

HFF

1 2 3 4 5 6 7 8

62 60 55 54 60 62 60

40 50 60 50 40 45 55

75 60 60 50 40 65 65

Aud. 5 40 60 45 40 46 80 50

Average 60 75 60 60 50 80 65

55.4 61 56 50.8 47.2 66.4 59

Table 4.22 – Subject eight tone quality results

Auditors' Score /100

Subject Eight Tone Quality 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5 Week

Figure 4.43 – Subject eight tone quality results

6

7

8

112 Figure 4.44 presents the average of the scores for tone quality from all five auditors, divided into phases and including a linear trend line for each phase.

S. 8 Phase 1 (HFF)

S. 8 Phase 2 (LFF) 70

60 55

Average Linear (Average) y = -1.88x + 62.38

50

Auditors' Score /100

Auditors' Score /100

65

45

65 60 55

Average

50

Linear (Average)

45

y = 4.38x + 44.9

40 1

2

3

4

5

5

Week

6

7

8

Week

Figure 4.44 – Subject eight average tone quality results by phase

Subject Eight Breath Management The aggregate scores for the breath management of subject eight from all five auditors are presented in the chart in table 4.23 and in the graph in figure 4.45.

Subject Eight Breath Management

LFF

HFF

Week

Aud. 1 1 2 3 4 5 6 7 8

Aud. 2 75 60 62 60 75 62 62

Aud. 3 25 20 50 20 60 25 40

Table 4.23 – Subject eight breath mangement results

Aud. 4 75 45 35 40 40 60 70

Aud. 5 75 90 80 80 68 95 75

Average 60 50 50 50 40 80 60

62 53 55.4 50 56.6 64.4 61.4

113

Auditors' Score /100

Subject Eight Breath Management 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5

6

7

8

Week

Figure 4.45 – Subject eight breath management results

Figure 4.46 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 8 Phase 1 (HFF)

S. 8 Phase 2 (LFF) 70

60 55

Average Linear (Average)

50

Auditors' Score /100

Auditors' Score /100

65

65 60 Average

55

Linear (Average)

50

y = -3.36x + 66.86 45

y = 4.2x + 47.6

45 1

2

3 Week

4

5

5

6

7 Week

Figure 4.46 – Subject eight average breath management results by phase

8

114 Subject Eight Intonation The aggregate scores for the intonation of subject eight from all five auditors are presented in the chart in table 4.24 and in the graph in figure 4.47.

Subject Eight Intonation Week

Aud. 1

LFF

HFF

1 2 3 4 5 6 7 8

Aud. 2 52 38 40 38 46 51 49

Aud. 3 5 5 15 0 20 20 10

Aud. 4 75 50 55 30 45 75 75

Aud. 5 60 70 65 50 70 70 60

Average 50 25 40 50 40 75 50

48.4 37.6 43 33.6 44.2 58.2 48.8

Table 4.24 – Subject eight intonation results

Auditors' Score /100

Subject Eight Intonation 100 90 80 70 60 50 40 30 20 10 0

Aud. 1 Aud. 2 Aud. 3 Aud. 4 Aud. 5 Average 1

2

3

4

5 Week

Figure 4.47 – Subject eight intonation results

6

7

8

115 Figure 4.48 presents the average of the scores for intonation from all five auditors, divided into phases and including a linear trend line for each phase.

S. 8 Phase 1 (HFF)

S. 8 Phase 2 (LFF) 65 60

50 45 Average

40

Linear (Average) y = -3.9x + 54.3

35 30 1

2

3

4

5

Auditors' Score /100

Auditors' Score /100

55

55 50 45

Average

40

Linear (Average) y = 5.96x + 31.3

35 30 5

Week

6

7

8

Week

Figure 4.48 – Subject eight average intonation results by phase

The results reported above may be used as a reference in the more thorough discussion contained in the following chapter. That chapter will interpret the results reported above and attempt to draw conclusions from them regarding the effect of variations in relative feedback frequency on the retention/transfer of classical singing techniques.

116 CHAPTER 5 DISCUSSION OF RESULTS

The purpose of this study was to test the hypothesis that LFF instruction would be beneficial to long-term improvements in the subjects‟ ability to perform classical vocal techniques. As seen in the results reported in chapter 4, the findings of the study do not on the surface support a strong positive correlation between LFF instruction and the desired improvements in performance ability in all subjects. However, some positive correlation may be seen in subjects with higher levels of ability.

Results That Support the Hypothesis If the hypothesis were to hold true, it would be expected that the learning trends during LFF phases would be steeper than during HFF phases. Upon analysis of the scores from the retention/transfer tests, it appeared that only five of the possible twenty-four scenarios (three voice characteristics for each of eight subjects) presented those results. These scenarios will be discussed below. The results for each scenario are represented in the graphs in chapter four. In order to save space, those graphs were not recreated here.

Subject Six Tone Quality The results for the tone quality for subject six presented steeper learning trends during LFF phases (phases 2 and 3 in this case) than in HFF phases. During phase one (HFF), the learning trend was actually slightly negative with a slope of -0.06. That trend reversed during phase two, the first of the LFF phases, to a positive slope of 0.257. When instruction returned to the HFF condition in phase three, the learning trend remained positive but with a more moderate

117 slope of 0.14. Finally, with the last LFF instruction condition in phase four, the learning trend again steepened to the most aggressive slope of the study at 0.4. With positive learning trend slopes in the LFF phases, and at a steeper slope than in the HFF phases, these findings lend support to the hypothesis.

Subject Six Breath Management The results for the breath management of subject six are similar to those for that subject‟s tone quality reported above. However, in this scenario the overall trend across the duration of the study was only very slightly positive. The learning trend in phase one (HFF) was negative with a slope of -0.38. Phase two (LFF) was still negative, but with a more gradual slope of -0.18. Phase three, returning to the HFF condition, was once again more aggressively negative with a slope of -1.5. Finally, phase four (LFF) was positive, with a quite steep slope of 10.8. Thus, with the two LFF phases presenting either less aggressive negative or positive slopes, these findings provide some support of the hypothesis.

Subject Eight Tone Quality, Breath Management, and Intonation Because subject eight was forced to end her participation in the study due to health and academic problems after only eight weeks, she was only present for one phase each of the LFF and HFF instruction conditions. However, her results for each vocal characteristic presented steeper learning trends during the LFF condition than in the HFF condition. In fact, the learning trends for all three characteristics were negative in the HFF phases but positive in the LFF phases. For tone quality, her learning trend in phase one (HFF) had a slope of -1.88, while her learning trend in phase two (LFF) had a slope of 4.38. For breath management, her learning

118 trend in phase one had a slope of -3.36, while her learning trend in phase two had a slope of 4.2. Finally, for intonation, her learning trend in phase one had a slope of -3.9, while her learning trend in phase two had a slope of 5.96. These results do provide some relatively weak support for the hypothesis. However, because of subject eight‟s early departure from the study, these results may be explained by factors other than that she benefited from the LFF instruction condition as there was no second round of HFF and LFF phases against which the results for phases one and two could be compared.

Results That Do Not Support the Hypothesis Of the remaining nineteen scenarios, four presented results that were in direct contradiction to the hypothesis. In these three cases, learning trends were in fact steeper during HFF phases than during LFF phases. These cases will be discussed below.

Subject One Tone Quality The results for the tone quality of subject one were exactly opposite of those expected if the hypothesis were to hold true. During phase one (LFF), subject one‟s learning trend was negative with a slope of -0.86. In phase two (HFF) however, that trend reversed to a positive slope of 0.54. Phase three (LFF) saw a return to a negative learning trend with a slope of -4.2. Finally, during phase four (HFF) the learning trend returned to a positive slope of 3.3. These findings would indicate that not only was LFF instruction condition not beneficial to the improvement of tone quality in subject one, it was actually detrimental, with subject one actually performing worse each week during those phases.

119 Subject One Intonation In regard to intonation, subject one‟s performance again appeared to hindered by the LFF instruction condition. During phase one (LFF), subject one‟s learning trend was negative with a slope of -1.88. With the change to the HFF instruction condition in phase two, her learning trend turned sharply positive with a slope of 14.16. During phase three (LFF), the learning trend again returned to a negative slope of -0.64 and the study finished with a positive slope of 6.1 during the final HFF instruction of phase four.

Subject Three Breath Management The results for the breath management of subject three presented three negative learning trends and one positive trend. Phase one (LFF) began the study with a negative learning trend with a slope of -1.46. That negative trend moderated somewhat during phase two (HFF) to a slope of .16, but increased again in phase three (LFF) to a slope of -1.12. Finally, phase four (HFF) saw the first positive learning trend with a slope of 3.2. With these results both LFF phases were more negative than either of the HFF phases, thus contradicting the expected correlation.

Subject Seven Tone Quality Similar to the two previous scenarios, the results for the tone quality of subject seven again appeared to indicate that the LFF instruction condition actually hindered performance on the retention/transfer tests. During phase one (HFF), subject one‟s learning trended up with a slope of 0.78. The introduction of the LFF instruction condition was accompanied by a negative turn in subject seven‟s learning trend with a slope of -1.58. That trend was again reversed with the HFF instruction condition of phase three, resulting in a learning trend with a positive slope of

120 0.7. Finally, the return of the LFF instruction condition in phase four saw a return to a negative learning trend with a slope of -2.2.

Results That Are Mixed The remaining fifteen scenarios presented results that were mixed, meaning that no clear correlation between feedback frequency and performance on retention/transfer tests is clearly visible in these cases. Each of these scenarios will be discussed below, followed by a discussion of how these results might be interpreted.

Subject One Breath Management The results for breath management of subject one were mixed. During phase one (LFF), subject one‟s learning trend was positive with a slope of 1.42. That trend was reversed with the HFF instruction condition of phase two, presenting a learning trend with a negative slope of -1.58. During the LFF instruction condition of phase three, the learning trend remained negative with a slope of -0.6 before turning positive once again in phase four (HFF) with a slope of 3.2.

Subject Two Tone Quality The results for the tone quality of subject two again failed to present a clear correlation. During phase one (LFF), the subject‟s learning trend was negative with a slope of -3.04. That trend was reversed during phase two (HFF), presenting a learning trend with a positive slope of 2.54. That positive slope could not be maintained however, turning negative again in phase three (LFF) with a slope of -1.8, and staying negative throughout the conclusion of the study in phase four (HFF) with a slope of -5.6.

121 Subject Two Breath Management Similar to the results for her tone quality, the results for the breath management of subject two again had three phases with negatively-sloped learning trends and one phase with a positive learning trend. Interestingly, these results even presented the same pattern as the tone quality results. Phase one (LFF) presented a learning trend with a negative slope of -.26. Phase two (HFF) presented the only positive learning trend with a slope of 1.6. Phases three (LFF) and four (HFF) again presented negative learning trends with slopes of -0.5 and -7.5 respectively.

Subject Two Intonation In regard to intonation, the learning trend for subject two in phase one (LFF) was once again negative, with a slope of -1.72. Phase two (HFF) once again presented a positive learning trend with a slope of 1.78. Unlike in the results for breath management and tone quality for the same subject, the learning trend for subject two in phase three (LFF) remained positive with a slope of 0.94. The learning trend turned negative again, however, in phase four (HFF) with a slope of -2.2.

Subject Three Tone Quality The results for the tone quality of subject three presented two phases with negative learning trends, one phase with a positive learning trend, and one phase with a static learning trend. During phase one (LFF), the learning trend was negative with a slope of -1.46. That negative trend continued through phase two (HFF) with a slope of -2.1. In phase three (LFF) the learning trend finally turned positive with a slope of 0.74, but the trend then held static through phase four (HFF) with a slope of 0.

122 Subject Three Intonation The results for the intonation of subject three presented two phases with positive learning trends (one LFF and the other HFF) and two phases with negative learning trends (again one LFF and the other HFF). Phase one (LFF), had a positive learning trend with a slope of 3.0 while phase two (HFF) had a negative learning trend with a slope of -0.72. The negative learning trend continued into the LFF instruction condition of phase three which presented a slope of -4.12. In the final phase (HFF), the learning trend turned sharply positive with a slope of 9.2.

Subject Four Tone Quality The results for the tone quality of subject four presented three phases with positive learning trends and one phase with a negative learning trend. The study began with phases one (LFF) and two (HFF) revealing positive learning trends with slopes of 0.74 and 0.62 respectively. During phase three (LFF) the learning trend turned negative with a slope of -1.68. Finally, in phase four (HFF), the learning trend returned to a positive ascent with a slope of 1.6.

Subject Four Breath Management The results for the breath management of subject four had two phase with positive slopes and two with negative slopes. However, these results fail support the hypothesis because the two negative slopes are in contrasting feedback conditions. Phase one (LFF) had a positive learning trend with a slope of 0.4. That trend reversed in phase two (HFF) to a negative slope of -1.02 and remained negative through phase three (LFF) with a slope of -1.7. In the final phase (HFF) the learning trend once again turned positive with a slope of 2.9.

123 Subject Four Intonation The results for the intonation of subject four were similar to those for the breath management of the same subject in that they again presented two phases with positive learning trends and two phases with negative slopes, each consisting of results from contrasting feedback conditions. Phase one (LFF) presented a negative learning trend with a slope of -0.52. The trend then turned positive through phases two (HFF) and three (LFF) with slopes of 0.44 and 0.74 respectively. The study concluded with a positive learning trend in phase four (HFF) with a slope of -3.5.

Subject Five Tone Quality The tone quality of subject five provided results that presented three phases with positive learning trends and one phase with a negative learning trend. Phase one (HFF) presented a positive learning trend with a slope of 0.68. The learning trend turned sharply negative in phase two (LFF) with a slope of -5.22. That negative trend was reversed in phase three (HFF) with a slope of 1.0 and remained positive through the conclusion of the study in phase four (LFF) with a slope of 4.1.

Subject Five Breath Management The results for the breath management of subject five are very similar to the results for tone quality for the same subject. Phase one (HFF) presented a positive learning trend with a slope of 0.96. The learning trend again turned negative in phase two (LFF) with a slope of -5.84. During phase three (HFF) the learning trend once again turned positive with a slope of 2.12 and remained positive through phase four (LFF) with a slope of 4.1.

124 Subject Five Intonation The results for the intonation of subject five presented two phases with positive learning trends and two phases with negative learning trends, each consisting of contrasting feedback conditions. Phase one (HFF) presented a positive learning trend with a slope of 4.58. Phases two (LFF) and three (HFF) reversed the positive trend and presented learning trends with slopes of -1.14 and -2.5 respectively. The final phase of the study (LFF) returned to a positive learning trend with a slope of 7.0.

Subject Six Intonation The results for the intonation of subject followed the same pattern as the results for the intonation of subject five reported above. Phase one (HFF) presented a slightly positive learning trend with a slope of 0.18. That trend was reversed in phases two (LFF) and three (HFF) which presented negative learning trends with slopes of -0.79 and -5.4 respectively. Phase four presented a return to a positive learning trend with a slope of 18.8. However, because subject six did not appear for his final lesson, there are only two data points, representing the retention over a single week, in phase four. Consequently, it is likely that the slope for phase four is artificially high.

Subject Seven Breath Management The results for the breath management of subject seven presented three phases with negative learning trends and one phase with a positive learning trend. Phase one (HFF) presented a negative learning trend with a slope of -7.34. The only positive learning trend came in phase two (LFF) with slope of 4.42. The learning trend turned negative again in phase three

125 (HFF) with a slope of -1.74 and remained negative through phase four (LFF) with a slope of -4.2.

Subject Seven Intonation The results for the intonation of subject seven again presented three negative learning trends and one positive learning trend. In this case, phases one (HFF) and two (LFF) both presented negative learning trends with slopes of -1.26 and -7.9 respectively. Phase three (HFF) presented the only positive learning trend with a slope of 7.8 before the trend again turned negative in phase four (LFF) with a slope of -4.8.

Interpreting the Mixed Results Because of the large number of scenarios in which the results were mixed, it is desirable to attempt to interpret those results in order to determine if there is some correlation that can be drawn between the frequency of feedback and the subjects‟ performances on the retention/transfer tests. In order to clarify these results, the two learning trends for each feedback condition were added together (i.e. LFF slope + LFF slope and HFF slope + HFF slope) to determine an overall trend for that condition. An analysis of those sums revealed that six scenarios supported the hypothesis, while the remaining nine scenarios contradicted the hypothesis. A summary of these calculations are included in tables 5.1 (supporting hypothesis) and 5.2 (contradicting hypothesis) on the following page.

126

Scenarios Supporting Hypothesis Subject/Characteristic Sum of LFF Learning Trend S. 2 Breath Management -0.76 S. 3 Tone Quality -1.44 S. 4 Intonation 0.22 S. 5 Intonation 5.86 S. 6 Intonation 18.01 S. 7 Breath Management 0.22 Table 5.1 – Scenarios supporting the hypothesis

Sum of HFF Learning Trend -5.9 -2.1 -3.06 2.08 -5.22 -9.08

Scenarios Contradicting Hypothesis Subject/Characteristic Sum of LFF Learning Trend S. 1 Breath Management 1.36 S. 2 Tone Quality -4.84 S. 2 Intonation -0.78 S.3 Intonation -1.12 S. 4 Tone Quality -0.94 S. 4 Breath Management -1.3 S. 5 Tone Quality -1.1 S. 5 Breath Management -1.74 S. 7 Intonation -12.7 Table 5.2 – Scenarios contradicting the hypothesis

Sum of HFF Learning Trend 1.62 -3.6 -0.42 8.48 2.22 1.88 1.68 3.08 6.54

Initial Summary of Results Of the twenty-four possible scenarios, it appears that five clearly support the hypothesis. An additional six scenarios support the hypothesis when a sum is taken of both the LFF and the HFF phases, for a total of eleven scenarios supporting the hypothesis. Oppositely, three scenarios clearly contradict the hypothesis. An additional nine scenarios contradict the hypothesis when a sum is taken of both the LFF and the HFF phases for a total of twelve scenarios contradicting the hypothesis. The fact that the results are evenly divided between those that support the hypothesis and those that contradict it, indicates that a strong correlation

127 between feedback frequency during skill acquisition and a subject‟s performance ability on a retention/transfer test, either positive or negative, cannot be surmised from this study. If, as the studies summarized in chapter 2 suggest, research has indicated that lowering the frequency of feedback during instruction is beneficial to the retention of newly-acquired motor skills, the question arises as to why those findings did not hold true in the current study. Several possible responses to that question will be discussed in the following chapter. One response, which may be addressed by new analysis, is discussed below.

Variation of Skills Being Trained One possible cause for the unexpected results may stem from the nature of the classical vocal training that was taking place in this study. In typical college-level voice instruction, the vocal skills receiving the bulk of the student‟s and the instructor‟s attention would often vary throughout the course of a semester. For instance, it is unlikely that breath management, tone quality, and intonation would all garner attention at every lesson. It is much more likely, instead, that attention during any given lesson would end up being focused more on one of those characteristics than on the other two. In attempt to compensate for this variation in focus during individual lessons, the scores for all three characteristics were added together to form a total score for the overall quality of the performance on each week‟s retention/transfer test. This process was repeated for all five auditors and an average of the scores from all five auditors was calculated. From there, the same process used with the scores from the individual characteristics was instituted with this data to create overall learning scores for each subject.

128 This analysis of the results revealed that three of the subjects (S.2, S.6, and S.8) responded more favorably to the LFF instruction condition, supporting the hypothesis. However, the results from the remaining five subjects (S.1, S.3, S.4, S.5, and S. 7) suggest that they responded more favorably to the HFF instruction condition, contradicting the hypothesis. Graphs of these results are shown and discussed in the following pages.

Subjects That Support the Hypothesis Figure 5.1 shows the average of the overall scores for subject two from all five auditors including linear trend lines. The overall scores for subject two do not appear at first glance to support the hypothesis as the only positive learning trend occurred during phase two. However, an average of the learning trend slopes for the LFF and the HFF phases reveals that, because of the severity of the decline in phase four, the LFF instruction condition may have been less damaging to the subject‟s progress. These results provide very weak support of the hypothesis. Figure 5.2 shows the average of the overall scores for subject six from all five auditors including linear trend lines. The overall scores for subject six lend somewhat stronger support for the hypothesis than did those from subject two. However, the dramatic positive learning trend in phase four, coupled with the fact that subject six was only present for two of the three weeks for that phase, diminishes the credibility of these results.

129

S. 2 Phase 1 (LFF)

S. 2 Phase 2 (HFF)

205

200 190

185 175 Average

165

Linear (Average) y = -5.02x + 195.11

155 145 1

2

3

4

Auditors' Score /300

Auditors' Score /300

195

180 170 Average

160

Linear (Average) y = 6.12x + 160.32

150

5

140

Week

5

200

200

190

190

180 170 Average 160 Linear (Average)

150

7 8 Week

9

S. 2 Phase 4 (HFF)

Auditors' Score /300

Auditors' Score /300

S. 2 Phase 3 (LFF)

6

180 170 Average 160 Linear (Average)

150

y = -15.3x + 205.4

y = -1.42x + 178.98 140

140 9

10

11

12

13

Week

Figure 5.1 – Subject two average overall results by phase

13

14 Week

15

130

S. 6 Phase 1 (HFF)

S. 6 Phase 2 (LFF) 240

230 220 210

Average Linear (Average) y = -0.26x + 218.1

200 190 1

2

3

4

Auditors' Score /300

Auditors' Score /300

240

230

220 Average 210

Linear (Average) y = -0.72x + 223.28

200

5

5

Week

7

8

9

Week

S. 6 Phase 3 (HFF)

S. 6 Phase 4 (LFF)

260

250

250

240

240 230 220

Average

210 Linear (Average) y = -6.76x + 243.63

200 190

Auditors' Score /300

Auditors' Score /300

6

230 220 Average

210

Linear (Average) y = 30x + 172

200 190

9

10

11

12

13

Week

Figure 5.2 – Subject six average overall results by phase

13

14 Week

15

131 Figure 5.3 shows the average of the overall scores for subject eight from all five auditors including linear trend lines. The overall scores for subject eight would appear to clearly support the hypothesis, with a positive learning trend occurring in phase two but a negative learning trend occurring in phase one. Again, however, the fact that subject eight was unable to complete the study, resulting in scores from only eight weeks, means that it is impossible to know if these results would have been sustained in the remainder of the study.

S. 8 Phase 2 (LFF)

180

190

170

180

160 150 Average

140

Linear (Average) y = -9.1x + 183.3

130 120 1

2

3

4

Auditors' Score /300

Auditors' Score /300

S. 8 Phase 1 (HFF)

170 160 Average

150

Linear (Average) y = 14.54x + 123.8

140 130

5

5

Week

6

7

8

Week

Figure 5.3 – Subject eight average overall results by phase

The overall scores of only three of the eight subjects seem to support the hypothesis that reduced frequency of feedback would result in higher levels of learning, as evidenced by higher scores on performances of retention/transfer tests. Additionally, the results of each of these three subjects provide only relatively weak support of the hypothesis.

132 Subjects that Contradict the Hypothesis Figure 5.4 shows the average of the overall scores for subject one from all five auditors including linear trend lines. Subject one‟s overall scores clearly contradict the hypothesis with positive learning trends occurring in the HFF phases, but negative learning trends occurring in both of the LFF phases.

S. 1 Phase 1 (LFF)

S. 1 Phase 2 (HFF) 170 160

120 115 Average

110

Linear (Average) y = -1.32x + 117.28

105 100 1

2

3

4

Auditors' Score /300

Auditors' Score /300

125

150 140 130

Average

120

Linear (Average) y = 13.12x + 94.8

110 100

5

5

6

Week

8

9

Week

S. 1 Pahse 3 (LFF)

S. 1 Phase 4 (HFF) 140

130 120

Average Linear (Average) y = -5.44x + 135.84

110 100

Auditors' Score /300

140 Auditors' Score /300

7

130 120 Average

110

Linear (Average) y = 12.6x + 93.8

100 90

9

10

11

12

13

Week

Figure 5.4 – Subject one average overall results by phase

13

14 Week

15

133 Figure 5.5 shows the average of the overall scores for subject three from all five auditors including linear trend lines. The overall scores for subject three provide less clear contradiction of the hypothesis than did those of subject one. Similar to the results for subject six, the dramatic positive learning trend in phase four overrules the less dramatic trends in the previous three phases, lending mathematical support to the HFF feedback condition of that phase.

S. 3 Phase 2 (HFF)

140

160

130

150

120 110 100

Average

90

Linear (Average) y = 0.08x + 106.76

80 70 1

2

3

4

Auditors' Score /300

Auditors' Score /300

S.3 Phase 1 (LFF)

140 130 Average 120 Linear (Average) y = -2.98x + 134.74

110 100

5

5

6

Week

8

9

Week

S. 3 Phase 3 (LFF)

S. 3 Phase 4 (HFF)

135

155 145

125 115 Average

105

Linear (Average) y = -4.5x + 122.06

95 85 9

10

11

12

Auditors' Score /300

Auditors' Score /300

7

135 125 115

Average

105

13

Week

Figure 5.5 – Subject three average overall results by phase

Linear (Average) y = 12.4x + 95.2

95 85 13

14 Week

15

134 Figure 5.6 shows the average of the overall scores for subject four from all five auditors including linear trend lines. The overall scores for subject four are again, less clearly contradictory to the hypothesis. However, the combined learning trends for the HFF and the LFF phases reveal that learning trends were higher during phases two and four.

S. 4 Phase 1 (LFF)

S. 4 Phase 2 (HFF) 190

165 160 Average 155

Linear (Average) y = 0.62x + 161.26

150 1

2

3

4

Auditors' Score /300

Auditors' Score /300

170

180 170 Average

160

Linear (Average) y = 0.04x + 168.92

150 140

5

5

6

Week

7

8

9

Week

S. 4 Phase 3 (LFF)

S. 4 Phase 4 (HFF)

190

170

170 160 150

Average

140

Linear (Average) y = -6.36x + 178.72

130 120 9

10

11

12

Auditors' Score /300

Auditors' Score /300

180 165 160 Average 155

Linear (Average) y = 1x + 156.8

150

13

Week

Figure 5.6 – Subject four average overall results by phase

13

14 Week

15

135 Figure 5.7 shows the average of the overall scores for subject five from all five auditors including linear trend lines. The overall scores for subject five reveal dramatic shifts in learning trends between the two LFF phases (phases two and four). While the learning trends in phases one and three were less dramatic, they both had positive slopes, contradicting the hypothesis.

S. 5 Phase 1 (HFF)

S. 5 Phase 2 (LFF) 160

150 140 130

Average Linear (Average) y = 6.22x + 125.5

120 110 1

2

3

4

Auditors' Score /300

Auditors' Score /300

160

150 140 130 Average

120

Linear (Average) y = -12.2x + 174

110 100 5

5

150

150

140

140

130 Average

120

Linear (Average) y = 0.62x + 123.02

110 100 11

8

9

S. 5 Phase 4 (LFF)

12

Auditors' Score /300

Auditors' Score /300

S. 5 Phase 3 (HFF)

10

7 Week

Week

9

6

130 120 Average

110

Linear (Average) y = 15.2x + 86.133

100 90

13

Week

Figure 5.7 – Subect five average overall results by phase

13

14 Week

15

136 Figure 5.8 shows the average of the overall scores for subject seven from all five auditors including linear trend lines. The overall scores for subject seven are weakly contradictory to the hypothesis due to the relatively steep negative learning trends in both LFF phases. However, the trend in phase one (HFF) is also steeply negative, weakening the conclusion that the HFF instruction condition was more beneficial for this subject.

S. 7 Phase 1 (HFF)

S. 7 Phase 2 (LFF) 190 180

200 190 180

Average Linear (Average) y = -7.82x + 208.12

170 160 1

2

3

4

Auditors' Score /300

Auditors' Score /300

210

170 160 150

Average

140

Linear (Average) y = -5.06x + 173.62

130 120

5

5

Week

190

180

180 170 Average

160

Linear (Average) y = 6.76x + 145.4

150 140 12

Auditors' Score /300

Auditors' Score /300

190

11

8

9

S. 7 Phase 4 (LFF)

200

10

7 Week

S. 7 Phase 3 (HFF)

9

6

170 160

Average Linear (Average) y = -11.2x + 191.33

150 140

13

Week

Figure 5.8 – Subject seven average overall results by phase

13

14 Week

15

137 The combined results of the five subjects reported above indicate that it is not possible from this data to draw the conclusion that the LFF instruction condition is more beneficial to the acquistion of classical singing techniques for all subjects.

Conclusions The mixed nature of the results of this experiment make it difficult to draw a strong conclusion regarding the effect of variations in relative feedback frequency on the retention/transfer, and therefore learning, of classical singing skills in all subjects. However, one relevant trend did appear to be supported by the results: subjects with higher levels of performance ability responded more positively to lower relative feedback frequency instruction conditions, while subjects with lower levels of performance appeared to respond negatively to lower relative feedback frequency instruction conditions. Of the eleven scenarios whose results appeared to support the hypothesis, all but two had average weekly scores above 50/100. These eleven scenarios had an overall average weekly score of 59.89/100. Additionally, when viewing the overall scores for each subject as reported above, all three subjects whose results supported the hypothesis (S. 2, S. 6, and S. 8) achieved average weekly scores above 150/300. Conversely, of the thirteen scenarios whose results appeared to contradict the hypothesis, all but three had average weekly scores below 50/100. The overall average weekly score for these scenarios was only 45.86/100. Similarly, three (S. 1, S. 3, and S. 5) of the five subjects whose overall scores reported previously in the chapter appeared to contradict the hypothesis achieved average weekly scores below 150/300. Subjects four and seven achieved average weekly scores above 150/300, yet their learning trends still appear to contradict the hypothesis.

138 That having been said, the support for these findings are effected by two key factors. First, the scores from the auditors‟ assessments of the vocal charateristics of most of the subjects, while fluctuating from week to week and between phases, appeared not to have significant improvement over the course of the entire semester. For example, the overall scores for subject one showed a positive learning trend of only 0.46 points per week. This translates to only an improvement of only 6.9 points (on a scale of 300 points) over the course of the semester. One possible explanation for such low score improvement could be found in the fact that the tests being assessed were at least in part transfer tests. It may be possible, had the tests been purely retention tests using the songs the subjects were studying throughout the term, that their scores would have shown more significant improvement. A second factor that may have effected the findings of this study was the wide variablity between the scores provided by individual auditors. An attempt was made to even out that variability by averaging the scores and using those averages to determine the learning trends reported in chapter 4 and previously in this chapter. However, averaging the scores does not completely negate the descrepancies. In future research, it may be beneficial to provide the auditors with examples of various levels of performance in each category, which the auditors would then be able to refer back to in order to ensure that they were assessing the recordings against a more consistent standard. These vocal characaterstics will be defined in detail to ensure that each judge is clear about the aspects of the performance they are rating. Additionally, it would be beneficial to intersperse repetitions of individual recordings into the track list for each judge. Doing so would allow the researcher to calculate intra-judge reliability. This experiment‟s findings, though possibly influenced by the factors listed above, that subjects with higher levels of performance ability respond more favorably to LFF instruction

139 conditions lends additional support to Salmoni‟s guidance hypothesis.131 As noted in chapter two, this hypothesis posits that instructional feedback will be most useful early in practice. As the learner progresses, however, that instructional feedback will become less beneficial, and lowering the frequency with which that feedback is provided may prevent the learner from becoming reliant on the feedback, and thereby increase his/her retention and/or transfer of the desired skill. In addition to lending support to the guidance hypothesis, the findings of this experiment indicate that motor learning theory regarding feedback frequency may indeed be applicable to the teaching and learning of classical singing technique. It would also stand to reason that application of motor learning theory regarding other aspects of motor skill acquisition (e.g. feedback timing and variability of practice) may be appropriate and further research in these areas is warranted.

131

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

140 CHAPTER 6 IMPLICATIONS FOR TEACHERS AND FUTURE RESEARCH

The previous chapters have laid the theoretical groundwork for an experiment studying the effects of variations in feedback frequency on the retention of classical singing techniques, presented the methodology for that experiment, and reported and discussed the results of that experiment. In the following pages, the discussion will turn to how the findings of that research may or may not impact current teaching trends in private voice studio instruction. Additionally, the findings of this study carry some implications for what related future research may be warranted. Those implications will be presented, including a discussion of possible causes for unexpected outcomes in the current experiment.

Implications for Teachers of Singing As the results reported and discussed in the previous chapters suggest, sweeping changes to studio voice instruction techniques are not warranted by the findings of this experiment. Instead, the findings of this experiment suggest that care must be taken within the studio to ensure that the instruction techniques being utilized are meeting the needs of each individual student. That having been said, the findings of this experiment did indicate that students of differing levels of performance ability may react differently to variations in the frequency with which feedback is provided during instruction. Therefore, it would be prudent for voice instructors to consider the needs of their students and, as many instructors already do, vary their teaching techniques (including feedback frequency) to meet those needs.

141 Inexperienced Singers The findings of this experiment indicate that lowering the frequency with which feedback is provided during the instruction of relatively inexperienced singers may not be beneficial to those singers. In fact, when working with beginning singers or with singers whose performance abilities are less developed, these findings indicate that such lowered frequency feedback may actually be detrimental to the retention and transfer of newly acquired vocal skills. In these instances, providing more frequent feedback during the early stages of skill acquisition may be more beneficial to the student. Because of the often excruciatingly slow pace of acquiring new vocal skills, it is difficult to say exactly what constitutes the “early stages” of skill acquisition for new singers. Instead of associating this stage of learning to sing with a certain amount of time, it is more apt to associate it with a level of performance ability. In the current study, it appeared that those subjects who responded negatively to the LFF instruction conditions were those whose scores were less than 50/100. Referring to the instructions given to the auditors at the time of their instruction, a score of 50 indicated that the performance varied moderately from their perceived standard for a welltrained 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 student‟s 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 student‟s ability to perform a task as compared to a standard would be a cumbersome technique to implement in a studio teacher‟s instruction. However, most teachers likely already perform such assessments in their current teaching techniques. As a teacher listens to his or her student sing a vocal exercise or a passage from a song, he or she is assessing the performance and deciding what elements of the performance need to be addressed once the performance is complete and how those elements would be most effectively addressed. The findings of this experiment indicate that, in addition to deciding what to address and how to address it, the instructor should also decide whether or not feedback should be given at all, based on the proficiency with which the performance is executed.

143 Implications for Further Research The findings of this experiment indicate at least some level of applicability of the theories pertaining to feedback frequency, developed and researched in other fields, to the field of classical singing instruction. Further research examining the applicability of other elements of motor-learning theory to the same field is therefore warranted. The first area in which further research would seem most appropriate would be that of continuing research examining the effect of variations in feedback frequency. This research could stem from, and attempt to address, some of the unexpected outcomes of the current study.

Unexpected Outcomes of the Current Study Care was taken in the design of this study to ensure that the research conditions closely approximated those found in typical private voice studios. However, by matching these conditions, many variables were introduced that may have impacted the outcomes. Indeed, the results of the current study did exhibit several unexpected outcomes, which may have been caused by these variables. Two of these outcomes will be discussed below, followed by a discussion of some of the confounding variables, out of the control of the investigator, which may have had an impact on the outcomes. Finally, several variations to the study design will be suggested that may help address some of these variables, leading to results that may be more widely applicable.

Negative Learning Trends Perhaps the most unexpected outcome of the research was number of phases whose learning trends were negative, meaning that the students‟ performance abilities had actually

144 decreased during that time period. From the design of the study, it was anticipated that there were would be significant variations in the slopes of the learning trends between phases with different feedback conditions. However, it was also expected that the majority of the phases would exhibit learning trends with positive slopes, with phases from one feedback condition exhibiting learning trends with more steeply positive slopes than the other condition. As it turned out, over half of the phases exhibited negative learning trends. Additionally, four of the eight subjects exhibited negative trends in the overall performance across the duration of the study. These results are disturbing to the instructor who hopes that all students will improve somewhat over the course of a semester of study. However, the amount of time spent practicing by the student as well as the level of motivation to progress may have had an impact on these results.

Lack of Homogeneity among Results At the onset of the study, it was hoped that the results would be more homogenized in their support or contradiction of the hypothesis. However, the results were nearly evenly split between those that supported the hypothesis and those that contradicted it. One explanation regarding which subjects supported the hypothesis and which did not has been discussed earlier in terms of the performance ability of the subjects. However, other variables may have had an impact on these results. Some of the variables that may have impacted the homogeneity of the results, in addition to the variables listed in the previous paragraph, include the health of the subjects at the time of the retention tests, the ability of the instructor to provide meaningful feedback in both LFF and HFF instruction conditions, and the use of the piano by the instructor throughout the instruction phases.

145 Student practice time The amount of time spent by individual subjects varied greatly between subjects. Additionally, the amount of time spent practicing during a given week, or even during a given phase, also varied within each subject. For example, the average amount of time spent practicing varied from less than one hour per week in one subject to nearly four hours per week in another subject. However, the time spent practicing by the first subject ranged from less than 45 minutes per week during phase three to nearly an hour and a half per week during phase one. While this variation in the amount of time spent practicing was out of the control of the researcher, it was documented throughout the duration of the study, allowing for outside analysis of the results in relation to practice time. Upon analysis, however, there did not appear to be a clear correlation between the amount of time spent by a subject practicing during a given phase of the study and the level of their performance on the retention/transfer tests for the same phase. Changes to the study, as discussed later in this chapter, may help to minimize the impact of practice time on the results of a future study.

Subject health at the time of retention tests Any teacher of voice will likely agree that one of the most frustrating factors affecting the progress of a student is that student‟s ability (or inability) to stay healthy. The health of participants may well have impacted the outcomes of the current study. Subjects frequently reported to the researcher that they were “not feeling 100%.” However, this happened with such high frequency that it was not possible to prevent them from performing a retention test every time. Instead, subjects were only prevented from participating when their ailment was to such an extent that a notable deterioration in vocal quality was evident to the researcher and these

146 instances are indicated in the results reported previously. Again, this practice is in keeping with the conditions found in many voice studios (especially those in which students are receiving academic credit for participation), wherein students are encouraged to participate in their lessons in as great an extent as it is healthy to do so. Variations to the study design, similar to those indicated in the discussion of practice time, may again help to minimize the impact that student health has on the results of a future study.

Student motivation Motivation is widely accepted as a major factor in student performance and learning in every discipline.132 Edward Deci and Richard Ryan note in their 1985 on intrinsic motivation text that research has indicated that “being intrinsically motivated to learn improves the quality of learning.”133 The implication, then, is that a lack of intrinsic motivation to learn will negatively impact the quality of learning. The current research had no way of measuring the level of motivation to learn possessed by each subject. Additionally, due to the longitudinal nature of this research, variations in motivation, which were certain to occur across the duration of the study, likely had an impact on the results. The single-subject design of this research, allowing each subject to serve as his or her own control and to not be compared to other subjects, meant that motivational variance between subjects was less of a concern than was the variance within each subject as the study progressed. In future research in which subjects are being compared to one another, it would be most desirable to make attempts to equalize the motivation

132

Richard A. Schmidt, Motor Learning and Performance: A Problem-based Approach, 3rd ed. (Champaign, IL:Human Kinetics Books, 2004), 191. 133

Edward L. Deci and Richard M Ryan, Intrinsic Motivation and Self-Determination in Human Behavior, (New York: Plenum Press, 1985), 256.

147 of all subjects. However, because of the difficulty of measuring and/or influencing intrinsic motivation, it may be more practical to make attempts to adjust the study design to minimize the opportunity for the levels of intrinsic motivation to change throughout the course of the study.

Instructor‟s ability to provide meaningful feedback in both feedback conditions Because this study was carried out in a college voice studio where students were paying tuition and receiving academic credit for participation, every attempt was made by the researcher to meet the design parameters of the study while simultaneously providing a meaningful and productive learning environment in keeping with the academic mission of the college. To this end, the researcher did not follow a strict script in each lesson, opting instead to attempt to design the feedback provided in order to meet the needs being presented by the subject at that moment in time. While this attempt may have been made with the interest of the students‟ progress in mind, changing the quality or direction of the feedback throughout the lesson may have had an impact on the results of the study. Some of the changes to the study design discussed later may again help to mitigate these impacts.

Reliance on piano as form of concurrent feedback during instruction As is the case in most voice studios, the researcher in the current experiment utilized a piano during lessons to introduce new vocal exercises and to help guide subjects through difficult passages in their vocal literature. Assuming (and this may be a larger assumption for some students than other) that the students were capable of distinguishing when they were singing the same pitches and rhythms that the piano was playing at that exact time, the piano could be classified as a source of concurrent feedback, which Richard Schmidt and Gabriele Wulf define

148 in their 1997 article titled “Continuous Concurrent Feedback Degrades Skill Learning: Implications for Training and Simulation” as “supplemental information presented to the learner during the actual action.”134 As the title of the article suggests, their study found that learners who were provided with continuous concurrent feedback during skill acquisition performed less well on retention tests than did those learners who were not provided with the concurrent feedback.135 The negative impact of concurrent feedback on learning is likely due, as the guidance hypothesis would suggest, to the fact that such feedback is overly guiding and produces in learners a dependency upon feedback. If the piano used in the current research did indeed act as a source of continuous concurrent feedback, then it is possible that that feedback negatively impacted the learning of the participants. Additionally, because the retention/transfer tests were performed without piano accompaniment, there was a difference between the acquisition and the testing conditions. While research by Winstein and Schmidt136 (reported in chapter two) indicated that such a difference in conditions may not have been as big of a factor in performance as had previously been suggested, attempting to minimize that difference in future research would likely still be desirable. Controlling the amount and frequency of piano use in further research may help minimize the unintended impact of this feedback on the results of the study that research.

134

Richard A. Schmidt and Gabriele Wulf, “Continuous Concurrent Feedback Degrades Skill Learning: Implications for Training and Simulation,” Human Factors 39, no. 4 (1997): 509. 135

136

Schmidt and Wulf, 509.

Carolee J. Winstein, and Richard A. Schmidt, “Reduced Frequency of Knowledge of Results Enhances Motor Skill Learning,” Journal of Experimental Psychology: Learning, Memory, and Cognition 16, no. 4 (1990): 677-791.

149 Variation in Auditors‟ Scores A final uncontrolled variable that may have impacted the results of the current study is the variation in scores given by the volunteer auditors. For example, the overall (sum) scores for subject one had a standard deviation of 39.36. In attempt to even out these scores, averages were taken and reported in the results sections of this paper. However, the magnitude of the standard deviations may have had in impact on the findings of this research.

Further Feedback Frequency Research Given the variables listed above that may have impacted the findings of the current experiment, a second experiment examining the effect of variations in relative frequency of feedback on the acquisition of classical singing techniques is suggested here in which attempts are made to control these variables or minimize their impact on the findings. While it was intentional in the design of the current study to approximate the learning environment found in many private voice studios, this aspect of the study may have led to a variety of difficulties in producing clear and generalizable results. Consequently, the experiment proposed here steps away significantly from the studio environment.

Hypothesis The hypothesis for this experiment would remain the same as that of the current study being discussed in this paper, namely that a reduced feedback frequency instruction condition would be beneficial to the acquisition of a new vocal task, resulting in a more accurate performance on retention tests of that task by those students who received that type of instruction.

150 Subjects For the best generalizability of the results from this research, a representative sample of college level voice students, including various ages, abilities, and experience levels would be needed. Once a significant sample size of subjects was established, subjects would be divided into one of two feedback condition groups: g-1, g-2.

Task This experiment would examine the effects of variations in relative frequency of feedback on the acquisition of a single, new, vocal exercise. The vocal exercise would consist of a series of ten pitches to be sung in succession. The number of pitches may be varied according the ability level of the subjects participating, though the same series must be used for all participants. It seems that ten pitches may be an appropriate level to start with, as it is unlikely that many of the subjects would be able to repeat that length of a sequence without some level of practice. However, it is likely that all subjects would be able to make some level of progress in their ability to perform the sequence given practice and guidance.

Procedure A recording of the sequence would be played for the subject and then the experiment, consisting of a skill acquisition phase and a retention test, would begin. During the skill acquisition phase, subjects would complete 27 attempts to perform the task. Subjects in g-1 would receive feedback after every attempt (100% KR), while subjects in g-2 would receive

151 feedback after only every third attempt (33% KR).137 Feedback would be in the form of an additional playing of the desired series, played immediately following each attempt for which feedback is to be provided. For each attempt, the number of pitches from the sequence that the subject performed correctly would be recorded. Following the completion of the 27th attempt, the subjects would be dismissed and instructed to return in 24 hours for a retention test. Because of the nature of music, it may be necessary to instruct the subjects to try not to hum, whistle, or otherwise perform the sequence during those hours, as every time they did so, it would essentially serve as a no-KR attempt, which could ostensibly skew the results of the retention test. When the subjects returned for the retention test, they would simply be given the starting pitch of the sequence and asked to sing the entire sequence. The number of pitches performed correctly would be recorded in the same manner as during the skill acquisition phase.

Results Once the scores for both groups have been collected, an analysis of variance (ANOVA) could be calculated, revealing which group performed more accurately on the skill acquisition phase attempts. A t-test could be used to calculate the results of the retention test attempts.

Discussion Simply removing the longitudinal aspect of the research reported in this paper, and replacing it with the design outlined above, would minimize the impact of many of the variables discussed previously in this chapter. The effects of practice time and subject health would 137

33% KR was chosen because that level of reduced-frequency feedback has been shown in the research by Winstein and Schmidt, reported earlier, to be most effective.

152 immediately be almost completely mitigated (though it is plausible that one or more subjects would become ill during the 24 hours between the skill acquisition phase and the retention test. Motivation may not be entirely mitigated, though two aspects of the design would contribute to lessening its impact on the results. First, because the time span of this study is markedly shorter, variation in individual motivation levels would be less likely to occur. Second, by ensuring that a statistically significant sample size was used, the effects of any one individual with low levels of motivation would be decreased. Additionally, because feedback would only be provided after an attempt was made and never as concurrent feedback, any reliance on the recording as a form of concurrent feedback would be completely removed. Finally, because the results would consist simply of a numerical count of the correct pitches, the need for aural assessment and evaluation, and therefore any negative impact on findings that may result from it, would also be completely removed.

Feedback Timing Research In addition to the frequency with which feedback is provided, related research may be conducted regarding the impact variations in the timing at which feedback is provided. Verdolini and Lee discuss the timing in which feedback is provided in terms of concurrent versus terminal feedback,138 indicating that concurrent feedback, as noted earlier, is detrimental to learning. Schmidt and Lee take the discussion further to examine whether or not the amount of

138

Katherine Verdolini and Timothy D. Lee. “Optimizing Motor Learning in Speech Interventions,” In Vocal Rehabilitation for Medical Speech-Language Pathology, ed. Christine M. Sapienza and Janina K. Casper (Austin, TX: Pro-ed, 2004), 419.

153 time by which feedback is delayed has an effect on learning.139 In this discussion, Schmidt and Lee discuss research conducted in 1990 by Swinnen, Schmidt, and Nicholson,140 the results of which indicated that subjects who received feedback following a 3.2 second delay after task completion, scored better on retention tests than did those subjects who received feedback immediately following task completion. These results were found using a relatively simple armpositioning task as the skill being acquired. It would be valuable to test whether or not the same results could be replicated in the acquisition of a vocal task. A simple way to design this research would be to conduct a variation of the experiment outlined above. For this new experiment, the subjects and task could remain the same. However, a change could be made to the procedure such that feedback timing, and not feedback frequency, could be varied between the two groups.

Hypothesis Given the research in other fields described above, it is expected that subjects receiving delayed feedback will perform more accurately on retention tests, thereby indicating a higher level of learning took place under the delayed feedback condition during the skill acquisition phase.

139

Richard A. Schmidt and Timothy D. Lee, Motor Control and Learning: A Behavioral Emphasis. (Champaign, IL: Human Kinetics, 2005), 347-348. 140

Stephen P. Swinnen and others, “Information Feedback for Skill Acquisition: Instantaneous Knowledge of Results Degrades Learning,” Journal of Experimental Psychology: Learning, Memory, and Cognitions 16, no. 4 (1990): 706-716.

154 Procedure For this experiment, feedback frequency would remain the same (i.e. all subjects would receive feedback after each attempt, or after every third attempt, dependent on which condition proves to maximize learning in the previous experiment). Similar to the previous experiment, all subjects would make 27 attempts to perform the pitch series during the skill-acquisition phase and feedback would consist of an additional playing of an audio recording of the desired series. Subjects in g-1 would receive feedback immediately following each attempt (within one second) for which feedback is to be provided. Alternatively, feedback provided to subjects in g-2 would be delayed by five seconds.141

Results Again, the number of pitches performed correctly for each attempt would be recorded and the results would be calculated using the same statistical analyses to reveal which group performed more accurately during the skill acquisition phase and in the retention test.

Discussion The feedback timing research described above could have significant impacts on the teaching techniques within the private voice studio. If the hypothesis is supported by the research, it would suggest that teachers should consider allowing their students a few seconds after an attempt is made at performing a vocal task before they begin to provide feedback. By

141

A delay of five seconds was chosen, as opposed to a longer or shorter delay, using the research of Austerman-Hula, et al. reported in chapter 2 as a model: Shannon N. Austermann Hula and others, “Effects of Feedback Frequency and Timing on Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech,” Journal of Speech, Language, and Hearing Research 51 (2008), 1088-1113.

155 doing so, students are allowed time to form their own hypotheses about their performance, as compared to the desired performance, and mentally come up with ways to test those hypotheses in their subsequent performance attempts. Studies suggest that the act of forming and testing their own hypotheses is highly beneficial to student learning142 and the results of this study could provide support for the generalizability of these findings.

Variability of Practice and Task Distribution Research143 In addition to the effects of various feedback conditions, motor-learning theory has provided some questions regarding how much time should be spent training any one skill without a variation being introduced during a given training session and the most effective sequence in which skills should be trained. Verdolini and Lee refer to these two concerns as variability of practice and task distribution respectively.144

Variability of Practice Variability of practice refers to the question of whether it is most effective to train a certain skill using only one task, which would be referred to as non-variable practice, or to use more than one task to train the same skill, which would be termed variable practice.145 In their discussion of variable versus non-variable practice, Verdolini and Lee cite several studies 142

Verdolini and Lee, 422.

143

In this context, the term practice refers not to time spent by the learner rehearsing a skill outside of the training session, but rather to the time spent training any one task during the training session. 144

Verdolini and Lee, 422-427.

145

Verdolini and Lee, 422.

156 conducted by McCracken and Stelmach,146 Shea and Kohl,147 Cleave and Fey,148 and Fey, Cleave, Long, and Hughes,149 all of which examined the effects of variable practice on the acquisition of timing and force production tasks. However, Verdolini and Lee note that few, if any, experiments have been conducted examining these same conditions in the training of vocal tasks. Given the implications of motor-learning theory on the question of variability of practice, it seems that a research study designed to examine the effects of variability of practice on the acquisition of vocal tasks would be highly relevant to the current discussion.

Hypothesis The findings of the studies listed above found that subjects who trained using nonvariable practice conditions performed more accurately during training, but those subjects who trained in variable practice conditions consistently performed better on retention tests. Given these findings, it is expected that non-variable practice conditions, while improving performance during training, will degrade retention, and therefore learning, of a new vocal task.

146

Hugh D. McCracken and George E. Stelmach, “A Test of the Schema Theory of Discrete Motor Learning,” Journal of Motor Behavior 9 (1977): 193-201. 147

Charles H. Shea and Robert M. Kohl, “Specificity and Variability of Practice,” Research Quarterly for Exercise and Sport 61, no. 2 (1990): 169-177. 148

Patricia L. Cleave and Marc E. Fey, “Two Approaches to Facilitation of Grammar in Children with Language Impairments: Rationale and Description,” American Journal of Speech-Language Pathology 6 (1997): 22-32. 149

Marc E. Fey and others, “Two Approaches to Facilitation of Grammar in Children with Language Impairment: An Experimental Evaluation,” Journal of Speech and Hearing Research 36 (1993): 141-157.

157 Subjects For this experiment, a sample of subjects similar to that of the previous two studies can be used. Subjects will again be divided into two groups: g-1 and g-2.

Task For this study, subjects will be trained to sing a relatively complex melody. The melody should be at a level of complexity that ensures that few students will be able to perform it flawlessly without some training.

Procedure Similar to the two previous experiments, this study will consist of a skill acquisition phase and a retention test to be performed 24 hours following the completion of the skill acquisition phase. During the skill acquisition phase, all subjects will make twenty five attempts to perform the task, with a recording of the melody played between each attempt. Subjects in group g-1 will serve as the non-variable practice group and will perform the melody in the same key for each of the twenty five attempts. Subjects in group g-2, on the other hand, will serve as the variable practice group, and will perform the melody in five different keys (real imitation) throughout the twenty five attempts. In other words, for each attempt, the subject will be randomly assigned one of possible five keys (including the original). The melody will be played in the new key before the attempt is made. The number of pitches and/or rhythms performed correctly on each attempt will be recorded for each group. Following the twenty five attempts, the subjects will be dismissed and asked to return 24 hours later. When the subjects return, the starting pitch of the melody in the original key will be played and the subjects will be asked to

158 perform the melody in that key. The number of pitches and/or rhythms performed correctly will again be recorded.

Results Again, the number of pitches performed correctly for each attempt would be recorded and the results would be calculated using the same statistical analyses to reveal which group performed more accurately during the skill acquisition phase and in the retention test.

Discussion If the results of the above experiment support the hypothesis, it could again have implications for studio voice instruction. If variable practice appears to be beneficial to the learning of a new vocal task, then teachers of singing should evaluate the number of times they encourage their students to repeat the same exercise or excerpt from vocal literature without variation. In light of the results of the study, it may be desirable for teachers to limit the number of times an exercise or excerpt should be repeated before a new task or a variation of that task is introduced to the training.

Task Distribution Assuming that a variable practice training condition is favorable, the next question to be addressed is that of the most effective sequence in which to order that practice. This will be referred to as task distribution here. The question of how best to order exercises aimed at different skills throughout the duration of a lesson is of particular importance to the voice instructor. Specifically, should the exercise aimed at individual skills be blocked together

159 (blocked practice), or should exercises aimed at different skills be randomly interspersed with one another (random practice)? The results of some of the initial research in this area, performed by John Shea and Robyn Morgan in 1979, required two groups of individuals to make three sequential arm movements.150 A total of 54 trials of the tasks were completed in three sets of 18 trials (one set for each of the arm movements. Subjects in the first group completed 18 trials of one movement before moving to the next movement, representing a blocked practice environment. Subjects in the other group practiced six trials of each of the three movements in each 18-trial set representing a random practice environment.151 A retention test consisting of 18 trials (six trials for each of the three movements) was conducted after a ten-minute delay following the completion of the skill acquisition phase. The time it took to perform each movement was recorded throughout the skill acquisition phase and during the retention tests. The results of this research indicated that those subjects who practiced in a blocked practice environment consistently performed the arm movements more quickly throughout the skill acquisition phase of the experiment. However, on the retention test delayed by ten minutes, those subjects who practiced in the random practice environment consistently performed the arm movements more quickly. The difficulty in designing research in this area to be applied to vocal skill acquisition is that each task being trained will have its own set of appropriate associated skills that are available to alternate with the desired task. The research study described immediately above is

150

John B. Shea and Robyn L. Morgan, “Contextual Interference Effects on the Acquisition, Retention, and Transfer of a Motor Skill,” Journal of Experimental Psychology: Human Learning and Memory 5, no. 2 (1979): 179-187. 151

Ibid., 181.

160 of a very basic level of variance in practice related to a single task. However, one could easily imagine the wide variety of skills, all aimed at improving the overall quality of a voice, that could be addressed within a single voice lesson. The experiment described below will simply choose three vocal exercises related to pitch, rhythm, and vocal onset. However, it should be noted that the results may well be generalized to the variability in practice of all aspects of vocal technique.

Hypothesis Given that the results of Shea and Morgan‟s research have been replicated several times and applied to various skills and tasks,152 it is expected that random practice conditions will benefit the retention of newly acquired vocal tasks.

Subjects For this experiment, a sample of subjects similar to that of the previous two studies can be used. Subjects would again be divided into two groups: g-1 and g-2.

Tasks In this experiment, subjects would be asked to perform three vocal exercises. In the first exercise, aimed at pitch control, the subjects would be required to repeat a series of pitches similar to that used in the research proposed for the feedback experiments. In the second task, aimed at improving rhythmic skills, students would be asked to repeat a rhythmic sequence.

152

See reviews of research in Schmidt and Lee, 2004, and Richard A. Magill and Kellie G. Hall, “A Review of the Contextual Interference Effect in Motor Skill Acquisition,” Human Movement Science 9 (1990): 241-289.

161 Finally, a task aimed at controlling the onset of sound production would require the subjects to produce four repetitions of an [a] vowel on specified pitch. With this task, subjects would be instructed to produce a balanced onset, somewhere between breathy and glottal.

Procedure Similar to the design of Shea and Morgan‟s research, this experiment would consist of a total of 54 trials during the skill acquisition phase: 18 trials of each task. Subjects in group g-1 would serve as the blocked practice group and perform all 18 trials of each task before moving on to the next task. Subjects in group g-2 would serve as the random practice group and would perform their trials in such an order that six trials of each task would be included in each 18-trial set. A retention test, in which subjects would perform six trials of each task, would be administered 24 hours following the completion of the skill acquisition phases. For the first two tasks, a simple number of correct pitches or rhythms could be recorded to serve as the results for each trial. A correct performance of the third task, however, is somewhat more subjective and may require setting an acceptable threshold of perceived deviation from the desired outcome. The number of trials performed below that threshold could then be recorded.

Results The number of correct, or acceptable, trials would be recorded and analyzed using the same statistical analyses as used in the experiments outlined earlier in this chapter.

162 Discussion If the results of this experiment support the hypothesis, indicating that a random practice learning environment is beneficial to the retention of newly acquired vocal skills, teacher of singing may again be encouraged to attempt the application of these principles to their own teaching. For example, during a typical 30-minute vocal technique lesson, instructors could replicate a random practice environment by altering the order in which exercises aimed at various vocal skills are prescribed to the student. The application of this concept to a voice lesson focused on the training of vocal literature, often referred to as a coaching session. In this instance, a random practice environment could be achieved by spending a relatively small amount of time on several songs rather than spending the bulk of the time on a single song. For example, a 30-minute coaching session may be made more effective by spending ten minutes on each of three songs than by spending the entire 30 minutes coaching a single song.

Perceptual Analysis and Evaluation Research While removed from the realm of motor-learning theory, the current study has shed more light on the need for further research into the effectiveness and accuracy of perceptual analysis of a vocal performance by an auditor or group of auditors. This type of analysis and evaluation is used regularly at vocal competitions and auditions throughout the industry. Katherine EberleFink 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 sample‟s quality. A relevant bit of research would now be to examine the five different scores for each vocal sample in terms of the range of scores, the mean deviation (average deviation from the mean score), and the standard deviation. This statistical analysis may provide some fodder to prompt further research into the area of perceptual acoustic analysis and eventually lead to more standardized scoring by auditors and judges of competitions and auditions.

Conclusion The field of motor-learning theory has been evolving for nearly a century, yet it continues to yield new and intriguing findings regarding the most effective practices for training muscles to produce a desired movement consistently and accurately. The current study examined the effects of variations in relative frequency of feedback on the retention of classical singing techniques in a setting closely resembling that of a typical studio voice lesson. The results of this research indicated weak support of the hypothesis that reducing the relative frequency of feedback would improve retention. However, this support was only found among those subjects who consistently had the highest scores from a panel of auditors. Among the remaining subjects with consistently lower scores, a low-frequency feedback instruction condition appeared to actually be detrimental to their retention. A number of variables may have contributed to a lack of homogeneity among the results. In attempt to control for these variables, a revised version of this study has been suggested. This new study, however, creates an environment which does not resemble a typical studio voice lesson and therefore is much less likely and/or practical to be recreated by voice teachers in their own studio instruction. The results of the current research, while not providing

164 astoundingly strong support for the hypothesis, do indicate some relevant trends that warrant further research into the application of motor-learning theory to studio voice instruction. It is the hope of the researcher to carry out some of this research, including the experiments described in this chapter in the near future.

165 APPENDIX A INFORMED CONSENT DOCUMENT

Project Title: Application of Principles from Motor-Learning Theory to the Studio Voice Lesson: Effects of Feedback Frequency on Retention of Classical Singing Technique Principal Investigator:

Lynn Maxfield

Research Team Contact:

Lynn Maxfield 208-406-7748 [email protected]

This consent form describes the research study to help you decide if you want to participate. This form provides important information about what you will be asked to do during the study, about the risks and benefits of the study, and about your rights as a research subject.  If you have any questions about or do not understand something in this form, you should ask the research team for more information.  You should discuss your participation with anyone you choose such as family or friends.  Do not agree to participate in this study unless the research team has answered your questions and you decide that you want to be part of this study. WHAT IS THE PURPOSE OF THIS STUDY? This is a research study. We are inviting you to participate in this research study because you are a student in Lynn Maxfield‟s voice studio. The purpose of this research study is to investigate the effectiveness of feedback given by an instructor during a voice lesson. The hypothesis is that less frequent feedback from the instructor will actually result in more efficient learning by the student. It is hoped that the information from this study will lead to more effective and efficient vocal instruction. HOW MANY PEOPLE WILL PARTICIPATE? Approximately 12 people will take part in this study at the University of Iowa.

HOW LONG WILL I BE IN THIS STUDY? If you agree to take part in this study, your involvement will last for approximately 15 weeks. The study will require your attendance at 16, weekly voice lessons throughout the semester with two lessons being given during the final week. Each lesson will be approximately 30 minutes.

166

WHAT WILL HAPPEN DURING THIS STUDY? If you agree to participate in this study you will first be asked to fill out a brief questionnaire regarding your previous singing experience and your vocal health. Following the questionnaire, the researcher will evaluate your ability to match pitch and your tonal memory. The researcher will play a pitch on the piano and ask you to match that pitch with your voice. You will be asked to do this five (5) times at pitches across your range. You will then be asked to sing “America the Beautiful” a capella (without accompaniment) to evaluate your tonal memory and intonation. Following this initial evaluation you will receive 16 30-minute voice lessons. These lessons will be conducted in a similar fashion to most normal voice lessons except that in some of the lessons, the instructor (Lynn Maxfield) will provide more feedback to the student and in some of the lessons he will provide less feedback. At the beginning of each lesson, you will make an audio recording of yourself singing “America the Beautiful” a capella. Throughout the week in between lessons, you will be asked to complete at least 2 hours of voice practice.

Audio/Video Recording or Photographs One aspect of this study involves making audio recordings of you. These recordings will serve as a measure of your performance ability at individual moments throughout the duration of the study. Following completion of the 15-week study, these recordings will be evaluated by five (5) college voice faculty members. These faculty members will not know your name nor will they know at what point during the study each recording was made. Once the recordings have been evaluated, they will be destroyed. [ ] Yes

[ ] No

I give you permission to make audio recordings of me during this study.

WHAT ARE THE RISKS OF THIS STUDY? You may experience one or more of the risks indicated below from being in this study. In addition to these, there may be other unknown risks, or risks that we did not anticipate, associated with being in this study.  

You may feel some discomfort or embarrassment in knowing that your voice will be evaluated by other voice faculty. There are no other foreseeable risks to participating.

WHAT ARE THE BENEFITS OF THIS STUDY? We don‟t know if you will benefit from being in this study.

167

However, we hope that, in the future, other people might benefit from this study because the information gleaned from this study may lead to more effective and efficient techniques for vocal instruction and training.

WILL IT COST ME ANYTHING TO BE IN THIS STUDY? You will not have any additional costs for being in this research study.

WILL I BE PAID FOR PARTICIPATING? You will not be paid for being in this research study.

WHO IS FUNDING THIS STUDY? The University and the research team are receiving no payments from other agencies, organizations, or companies to conduct this research study.

WHAT ABOUT CONFIDENTIALITY? We will keep your participation in this research study confidential to the extent permitted by law. However, it is possible that other people such as those indicated below may become aware of your participation in this study and may inspect and copy records pertaining to this research. Some of these records could contain information that personally identifies you.  federal government regulatory agencies,  auditing departments of the University of Iowa, and  the University of Iowa Institutional Review Board (a committee that reviews and approves research studies) To help protect your confidentiality, we will keep any documents containing your personal information in a secure location and remove your name and any identifying information from any recording or document before it is distributed to any of the recording evaluators. If we write a report or article about this study or share the study data set with others, we will do so in such a way that you cannot be directly identified.

IS BEING IN THIS STUDY VOLUNTARY? Taking part in this research study is completely voluntary. You may choose not to take part at all. If you decide to be in this study, you may stop participating at any time. If you decide not to be in this study, or if you stop participating at any time, you won‟t be penalized or lose any

168 benefits for which you otherwise qualify. If you choose not to participate or stop participating at any time, this decision will in no way effect your final grade for the course, nor will it effect the level of commitment on the part of the instructor. Will I Receive New Information About the Study while Participating? If we obtain any new information during this study that might affect your willingness to continue participating in the study, we‟ll promptly provide you with that information. Can Someone Else End my Participation in this Study? Under certain circumstances, the researchers might decide to end your participation in this research study earlier than planned. This might happen because you have missed two (2) or more consecutive lessons or you have missed three (3) or more non-consecutive lessons throughout the term of the study. Your participation in the study may also be ended if you fail to complete two (2) hours of practice each week. WHAT IF I HAVE QUESTIONS? We encourage you to ask questions. If you have any questions about the research study itself, please contact: Lynn Maxfield, 208-406-7748, [email protected]. 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 email [email protected]. General information about being a research subject can be found by clicking “Info for Public” on the Human Subjects Office web site, http://research.uiowa.edu/hso. To offer input about your experiences as a research subject or to speak to someone other than the research staff, call the Human Subjects Office at the number above.

This Informed Consent Document is not a contract. It is a written explanation of what will happen during the study if you decide to participate. You are not waiving any legal rights by signing this Informed Consent Document. Your signature indicates that this research study has

169 been explained to you, that your questions have been answered, and that you agree to take part in this study. You will receive a copy of this form. Subject's Name (printed): ________________________________________________________ Do not sign this form if today’s date is on or after $STAMP_EXP_DT .

________________________________________ (Signature of Subject)

_____________________________ (Date)

Statement of Person Who Obtained Consent I have discussed the above points with the subject or, where appropriate, with the subject‟s legally authorized representative. It is my opinion that the subject understands the risks, benefits, and procedures involved with participation in this research study.

__________________________________________ (Signature of Person who Obtained Consent)

_____________________________ (Date)

170 APPENDIX B SUBJECT INTAKE FORM

Subject Number: __________

Date: ____________

Age__________ Previous singing experience Private Study Duration and dates______________________________________________________________ Location(s)____________________________________________________________________ _____________________________________________________________________________ Instructor(s)____________________________________________________________________ ______________________________________________________________________________ Solo: Duration and dates______________________________________________________________ Location(s)____________________________________________________________________ _____________________________________________________________________________ Ensemble: Duration and dates______________________________________________________________ Location(s)____________________________________________________________________ _____________________________________________________________________________ Director(s)_____________________________________________________________________

Vocal Health History Do you now have, or have you ever had any vocal health issues that may affect your ability to maintain active participation (i.e. singing daily) in a full semester of voice lessons (nodes, nodules, polyps, cysts etc.)? ______________________________________________________________________________ ______________________________________________________________________________ ______________________________________________________________________________ Have you ever lost your voice more than once within the span of six (6) months? Y or N If yes, please explain: ___________________________________________________________ ______________________________________________________________________________

171 APPENDIX C SAMPLE VOCAL EXERCISES Tone Quality Exercises154

and/or

Breath Management Exercises155

and/or

154

Richard Miller, “Structure of Singing: System and Art in Vocal Technique,” (Belmont, CA: Wadsorth Group/Thomson Learning) 1996. 155 Ibid.

172 Intonation Exercises

and/or

173 APPENDIX D PERCEPTUAL AUDITORY ASSESSMENT FORM Assessor‟s Name:________________________________________ Track Number:__________ Please assess the singer‟s 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 welltrained voice. For example, a score of 50 would indicate that the performance deviated moderately from your standard for that area. A score of 0 would indicate that the performance deviated severely from your standard for that area. For your convenience, a line scale has been provided for each area, on which you may easily make a tic mark indicating score as you listen. Once you have finished listening, please indicate an exact numerical score in the space provided to the right of the scale.

Tone Quality

__________________________________________ 0 50 100

___/100

Breath Management __________________________________________ 0 50 100

___/100

Intonation

___/100

__________________________________________ 0 50 100

174 BIBLIOGRAPHY Adams, Jack A. “A Closed-Loop Theory of Motor Learning.” Journal of Motor Behavior 3 (1971): 111-150. Adams, Jack A. “Theoretical Issues for Knowledge of Results.” in Information Processing in Motor Control and Learning. ed. George. E. Stelmach. New York: Academic Press. 1978. Annett, John. Feedback and Human Behavior. Middlesex, England: Penguin Books. 1959. Bilodeau, Edward.A., Ima McD Bilodea, and David A. Schumsky, “Some Effects of Introducing and Withdrawing Knowledge of Results Early and Late in Practice.” Journal of Experimental Psychology 58 (1959): 142-144. Bilodeau, Ina M. “Information Feedback.” in The Acquisition of Skill. ed. Edward A. Bilodeau. New York: Academic Press, 1966. Cleave, Patricia L. and Marc E. Fey. “Two Approaches to Facilitation of Grammar in Children with Language Impairments: Rational and Description.” American Journal of SpeechLanguage Pathology 6 (1997): 22-32. Deci, Edward L. and Richard M. Ryan. Intrinsic Motivation and Self-Determination in Human Behavior. New York: Plenum Press, 1985. Eberle-Fink, Katherine. “Perceptual Acoustic Assessment of Singing.” Journal of Singing – The Official Journal of the Nation Association of Teachers of Singing 63, no. 1 (September/October 2006): 35-43. Ellwell, J.L. and G. C. Grindley. “The Effect of Knowledge of Results on Learning and Performance.” British Journal of Psychology 29 (1938): 39-54. Evans, S.H. “A Brief Statement of Schema Theory.” Psychonomic Science 8 (1967): 87-88. Ferrand, Carole T. “Effects of Practice With and Without Knowledge of Results on Jitter and Shimmer Levels in Normally Speaking Women.” Journal of Voice 9 (1995): 491-423. Fey, Marc E., Patricia L. Cleave, Steven H. Long, and Diana L. Hughes “Two Approaches to Facilitation of Grammar in Children with Language Impairment: An Experimental Evaluation.” Journal of Speech and Hearing Research 36 (1993): 141-157. Folkins, J. W. “Issues in Speech Motor Control and Their Relation to the Speech of Individuals with Cleft Palate.” Cleft Palate Journal 22, no. 2 (1985): 106-122.

175 Helding, Lynn. “Voice Science and Vocal Art, Part Two: Motor Learning Theory.” Journal of Singing 64, no. 4 (March/April 2008): 417-428. Ho, Linda, and John B. Shea. “Effects of Relative Frequency of Knowledge of Results on Retention of a Motor Skill.” Perceptial Motor Skills 46 (1978): 859-866. Hula, Shannon N. Austermann, Donald A. Robin, Edwin Maas, Kirrie J. Ballard, and Richard A. Schmidt. “Effects of Feedback Frequency and Timing on Acquisition, Retention, and Transfer of Speech Skills in Acquired Apraxia of Speech.” Journal of Speech, Language, and Hearing Research 51 (2008): 1088-1113. Kearns, Kevin P. “Flexibility of Single-Subject Experimental Designs Part II: Design Selection and Arrangement of Experimental Phases.” Journal of Speech and Hearing Disorders 51 (August 1986): 204-214. Kelso, J. A., and B. Tuller. “A Cynamical Basis for Action Systems.” in Handbook of Cognitive Neuroscience. ed. M. Gazzaniza. New York: Plenum Press, 1984, 321-356. Lavery, J.J. and F. H. Suddon. “Retention of Simple Motor Skills as a Function of the Number of Trials by Which KR is Delayed.” Perceptual and Motor Skills 15 (1962): 231-237. Lee, Timothy D., Margaret A. White and Heather Carnahan. “On the Role of Knowledge of Results in Motor Learning: Exploring the Guidance Hypothesis.” Journal of Motor Behavior 22, no. 2 (1990): 191-208. Magill, Richard.A. and C.A. Wood. “Knowledge of Results Precision as a Learning Variable in Motor Skill Acquisition.” Research Quarterly for Exercise and Sport 57 (1986): 170-173. Magill, Richard A. and Kellie G. Hall. “A Review of the Contextual Interference Effect in Motor Skill Acquisition.” Human Movement Science 9 (1990): 241-289. McCracken, Hugh D. and George E. Stelmach. “A Test of the Schema Theory of Discrete Motor Learning.” Journal of Motor Behavior 9 (1977): 193-201. Miller, Richard R. Structure of Singing: System and Art in Vocal Technique Belmont, CA: Wadsorth Group/Thomson Learning, 1996. Salmoni, Alan W., Richard A Schmidt, and Charles B. Walter. “Knowledge of Results and Motor Learning: A Review and Critical Reappraisal.” Psychological Bulletin 95, no.3 (1984): 335-386. Schmidt, Richard A. “Past and Future Issues in Motor Programming.” Research Quarterly for Exercise and Sport 51 (1980): 122-140. Schmidt, Richard A. “Schema Theory: Implications for Movement Education.” Motor Skills: Theory into Practice 2 (1977): 36-38.

176 Schmidt, Richard A. “A Schema Theory of Discrete Motor Skill Learning.” Psychological Review 82 (1975): 225-260. Schmidt, Richard A. “Control Processes in Motor Skills.” Exercise and Sport Science Review 4 (1976): 229-261 Schmidt, Richard A. “Movement Education and the Schema Theory.” In Report of the 1976 Conference: National Association for Physical Education of College Women held in Cedar Falls, IA 3-8 June 1976, ed. E. Crawford. Schmidt, Richard A. and Gabriele Wulf. “Continuous Concurrent Feedback Degrades Skill Learning: Implications for Training and Simulation.” Human Factor 39, no. 4 (1997): 509-525. Schmidt, Richard A. and Timothy D. Lee. Motor Learning and Performance: From Principles to Practice. Champaign IL: Human Kinetics Books, 1991. Schmidt, Richard A. Motor Control and Learning: A Behavioral Emphasis 3rd Ed. Champaign, IL: Human Kinetics Books, 1999. Schmidt, Richard A. Motor Learning and Performance. 3rd ed. Champaign, IL: Human Kinetics Books, 2004. Schmidt, Richard A. Motor Learning and Performance: A Problem-Based Approach. 3rd Ed. Champaign, IL: Human Kinetics Books, 2004. Schmidt, Richard A. Motor Learning and Performance: From Principles to Practice. Champaign, IL: Human Kinetics Books, 1991. Schmidt, Richard A., Claudia Lang, and Douglas E. Young. “Optimizing Summary Knowledge of Results for Skill Learning.” Human Movement Science 9 (1990): 325-348. Shea, Charles H. and Robert M. Kohl. “Specificity and Variability of Practice. Research Quarterly for Exercise and Sport 61, no. 2 (1990): 169-177. Shea, John B. and Robyn L. Morgan. “Contextual Interference Effects on the Acquisition, Retention, and Transfer of a Motor Skill.” Journal of Experimental Psychology: Human Learning and Memory 5, no.2 (1979): 179-187. Sherwood, D. E. “Effect of Bandwidth Knowledge of Results on Movement Consistency.” Perceptual and Motor Skills 66 (1988): 535-542. Steinhauer, Kimberly, and Judith Preston Grayhack. “The Role of Knowldege of Results in Performacne and Learning of a Voice Motor Task.” Journal of Voice 14, no. 2 (2000): 137-145.

177 Swinnen, Stephan P., Richard A. Schmidt, Diane E. Nicholson, and Diane C. Shapiro. “Information Feedback for Skill Acquisition: Instantaneous Knowledge of Results Degrades Learning.” Journal of Experimental Psychology: Learning, Memory, and Cognition 16, no. 4 (1990): 706-716. Thorndike, Edward L. Education Psychology: Briefer Course. New York: Teacher College, Columbia University. 1923. Titze, Ingo R. Principles of Voice Production. Iowa City, IA: NCVS, 2000. Tolman, Edward C. Purposive Behavior of Animals and Men. New York: Century, 2003. Varder-Linden, Darl W., James H. Cauraugh, and Tracy A. Greene. “The Effect of Frequency of Kinetic Feedback on Learning an Isometric Force Production Task in Nondisabled Subjects.” Physical Therapy 73, no. 2 (1993): 79-87. Verdolini, Katherine and Timothy D. Lee. “Optimizing Motor Learning in Speech Interventions.” in Vocal Rehabilitation for Medical Speech-Language Patholody. ed. Christine M. Sapienza and Janina K. Casper. Austin, TX: Pro-ed, 2004. Verdolini, Katherine. “Motor Learning Principles: How to Train.” in Vocology by Ingo R. Titze and Katherine Verdolini. Currently in press: used by permission. Winstein, Carolee J., and Richard A. Schmidt. “Reduced Frequency of Knowledge of Results Enhances Motor Skill Learning.” Journal of Experimental psychology: Learning, Memory, and Cognition 16, no. 4 (1990): 677-691.