1.3..

5:...

.
33.. ..

gauzbﬁfzmr

4.... .334...

: :thun_.._.5.

.
. ..Hu.=.u..
s. z
5.1.... 3.5.... .
‘1): vi;
31..
.33..

.. t". 3.
P... 3...:
Ir

,3...
:31

”MW... 5...

ins

. I. hut
:9!
9.2.1
11"? :1!
tﬁvst..uuvg.nﬁx
.1
5. 3:. 3:...
:3.qu 13.1.
.. .

an

ll- . M
an“... ”33.3.4.3...
5.4.3... 2: 1....

.9... ,

 

 

 

 

pl. .

.rax...(

.. imwazi.

.4. n.

8.3.3.: It... I... .3311...

. in...“ an... . . a...”
.. .. I l .

l

I. 14.25,)“. L?

Q. ‘1‘.

:11

a:

:2.
.4"). V2 51....
..1rtyel..sa...1i

1..s:l.{:.. 32:25
1: s... .i v

x 1

(Iris... 2. .2

It;
. .5331 .n,.vn.cv!!
. XIII-.511.
.17.»!!! .0,

x. . 1|.)-
35. v: n... .. 4!
01! viz-2.1....

Manx OIPHK'IOJJ.
b.

x
In... .1 ,9} .7

..&.§

. y 3..l L}

LJLJI if“ I l
maszrs - » -
Mrcmgan State
.2 Us .a‘.’ers;t
:30; ’3

This is to certify that the
dissertation entitled

INDIVIDUAL FEELING STATES AND PERFORMANCE DURING
TENNIS MATCHES: AN INTEGRATION OF THE IZOF AND IAPZ
MODELS

presented by

M. RYAN F LETT

has been accepted towards fulﬁllment
of the requirements for the

DOCTORAL degree in KlNESlOLOGY

 

 

0W2 Xiw

 

Major Professor’s Signature

5//0//O

Date

MSU is an Afﬁrmative Action/Equal Opportunity Employer

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5/08 K:IProjIAoc&Prele|RCfDateDue.indd

 

INDIVIDUAL FEELING STATES AND PERFORMANCE DURING TENNIS
MATCHES: AN INTEGRATION OF THE IZOF AND IAPZ MODELS

By

M. Ryan Flett

A DISSERTATION

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

DOCTOR OF PHILOSOPHY
Kinesiology

2010

ABSTRACT

INDIVIDUAL FEELING STATES AND PERFORMANCE DURING TENNIS
MATCHES: AN INTEGRATION OF THE IZOF AND IAPZ MODELS

By
M. Ryan Flett

For those researchers and practitioners interested in optimal performance in sport,
feeling states are one of the most important predictors of performance. Feeling states are
a broad category that encompasses affective (e. g., emotions, moods), somatic (e. g.,
physical sensations), and other forms of feeling. All of these terms are difficult to deﬁne,
and there is no consensus for any of them. Among these feeling states, emotions play the
most prominent role in predicting performance; and are also the most researched, best-
deﬁned, and most theoretically mature construct in the feeling / affect literature.

Feeling states are, however, only half of the feeling-performance relationship. The
most proliﬁc theory in the ﬁeld is the idiosyncratic Individual Zones of Optimal
Functioning theory (IZOF: Hanin, 1978, 1997, 2000a). More recently, Tenenbaum and
colleagues have developed the Individual Affect Performance Zone (IAPZ; Kamata,
Tenenbaum, & Hanin, 2002; Tenenbaum, Edmonds, & Eccles, 2008). In this study, both
idiographic models were combined in the following ways. (1) Rather than describing
feelings as continuous two-dimensional constructs (wherein players rate the pleasantness
and arousal-level of their feeling state), discrete items (e.g., conﬁdence, worry, calmness)
were selected to create each proﬁle. (2) Ordinal Logistic Regressions were conducted to
develop probabilistic estimates for four feeling categories (based on positive-/negative-

valence and functional/dysfunctional groupings).

Ten male college tennis players volunteered to participate in this study. Participants
developed individualized proﬁles, and then monitored the intensity of each item in the
proﬁle during competitive intra-squad matches. In total, 918 observations were recorded.
Ordinal logistic regressions were conducted to identify probabilistic performance curves.

Visual comparisons of individual zone proﬁles, qualitative comparison of the discrete
items selected by each participant, and statistical analyses provided support for the
validity of the combined IZOF-IAPZ method used in this dissertation as well as for the
idiosyncratic nature of feeling-performance relationships. Statistical analysis identiﬁed
76.3% interindividual differences based on comparisons of the location and width of each
feeling-performance zone. Intraindividual differences did not emerge across context: only
12.5% of the zone comparisons for serving and returning serve displayed signiﬁcant
differences. Ordinal Logistic Regression (OLR)-based models were generally accurate.
The complete four-category (function-valence) model predicted performance correctly
63.5% of the time (more than twice that of chance); and the four “trigger” items chosen
by each player as the most important for his performance were 66.4% accurate—the top
three trigger items were accurate 70.0% of the time.

The dissertation offers important data in the ﬁeld of IZOF research because a large
sample was collected during performances, and it was shown that OLR may be used to
develop probabilistic feeling-performance proﬁles using discrete feeling items. As such,
the interaction of valence and function can be studied along with simple proﬁles with

only three self-selected performance predictors.

I have learned a great deal about research and the science of sport psychology from my
graduate training. I am grateful and appreciative for the excellent schools and professors
that I have learned from. With a truly humble appreciation for everything that is bigger
than me—where I am from—I dedicate this to all of the coaches, athletes, parents,
friends, and family (in particular, my uncle, Dan Marisi), who have provided me with a
special understanding of, and appreciation for, applied sport psychology. There are great

teachers available wherever you are in life—you just have to keep your eyes and ears

open.

iv

ACKNOWLEDGEMENTS

First and foremost, I want to thank my participants in this study for their exceptional
focus and patience. They did not have any complaints, or express concerns with how their
participation in this study might affect their personal performance. They are good guys.
My time with them and with the coaches of the team has been a highlight of the past three
years. I wish them all the best. Every one of them has an incredible future ahead of him!

I am indebted to my committee for their assistant with this project. Dr. Feltz is a
fantastic leader in our ﬁeld and in our building. I cannot thank her enough for the
resources that she has provided me with. Dr. Ewing has been so generous with her time
over the years. I appreciate her open door and her practical insight. She has always given
me good, plain advice. Dr. Wong taught the best class I have taken at MSU and one of
the two of three best I have ever taken (I don’t have an ofﬁcial ranking system worked
out, but I am conﬁdent in that ranking). His class is what I expected every graduate class
would be: a carnival of ideas without mere regurgitation of text.

In 2004, I decided to return to school in Fall 2005. I talked to some people I trusted.
Asked them who the best person in sport psychology was. I was told that the best man—
the most quality person—in our ﬁeld was Dan Gould. I laughed and pointed out that he
was the biggest name in applied. 1 did not see how he could be at the top of the ﬁeld and
also be a person who I could see eye-to-eye with. But I trusted the advice I had asked for
and applied to MSU. The story does not end at my application—my ﬁrst application to
MSUébut for the sake of interjecting some brevity, I will end the story there. Dan is one
of the best people I have ever known. I can’t imagine how much crap he has taken for my

disregard for stupid things that only exist to help insecure people who are scared to leave

the ﬂock. Dan has supported me to be myself, regardless of where that takes me, and that
is where it will take me. For everything that Dan has done for me personally, I am also
indebted to Larry Lauer. He has been a friend and a good person to talk shop with.

The most important acknowledgement goes to my family. Some are here, some are
there, some are all over. I have had no greater inspiration in my life than from my
grandpa. It is too bad there is not a heaven. I would like to make ravioli with him; or have
lunch with him and grandma again——chicken sloup and ham sandwiches. Don’t know
why grandma pronounced it sloup. My mom and dad did good. I am a pretty impossible
person. They could have traded me in for one of the simple kids I think. My mom said
she was put here just to look out for me, and that is a full time job. I wish, too, that my
dad could read and understand the stuff that I have been doing over the past four years—
he was diagnosed with Alzheimer’s and Dementia in the ﬁrst year of my doctorate.

I am grateful for the assistance and friendship that Tracey, Darcie, Jo Ann, Verna, and
Jan have provided me. I would also like to thank my friends here at MSU for their
support and fun—especially thank RJ, Francois, Jenn, and Amanda. When I moved to
Gainseville, I lost touch with many friends from Saskatchewan. I regret this. Having
learned my lesson, I am thankful for my friends in Regina and Saskatoon who made so
much time for me during the holidays, and even went out of their way to visit me here.
Thanks to Kelly, Kyla, Connor, Blair, Jac, Ty, Scott, Steve, Mark, Scott, and Brian.

Music really makes a difference in my life. I am thankful for all the artists who share
their world-view, their creativity, and their liberation. I ﬁnd peace in music and in the
Dharma, and I am looking forward to ﬁnding more time for those things once again. I am

thankful for my experiences—good and bad.

vi

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. ix
LIST OF FIGURES ........................................................................................................... xi
CHAPTER ONE
INTRODUCTION ............................................................................................................... 1
Feelings and Performance in Sport ................................................................................ 1
Affect-Related Terminologies ................................................................................. 2
The Emotion-Performance Relationship ........................................................................ 8
The Individual Zones of Optimal Functioning Model ........................................... 10
The Individual Affect Performance Zones Model ................................................. 13
Issues with the IZOF and IAPZ Models Examined in this Study ................................ 15
Purposes and Hypotheses ............................................................................................. l6
Purposes ................................................................................................................. 16
Hypotheses ............................................................................................................. 16
Deﬁnition of Terminologies ........................................................................................ 17
Affect-Related Terms ............................................................................................. 17
Feeling-Performance Models ................................................................................. 19
CHAPTER TWO
REVIEW OF LITERATURE ............................................................................................ 22
Why and How Feelings Affect Functioning ................................................................ 22
Emotion and Adaptation ........................................................................................ 23
Characteristics of Emotion ..................................................................................... 27
Appraisal Models ................................................................................................... 30
Summary of Relationships between Feelings and Performance ............................ 35
Hanin’s Individual Zones of Optimal Functioning ...................................................... 36
Evolution of the “Zones” Model ............................................................................ 37
The Modern Individual Zones of Optimal Functioning Model ............................. 40
Theoretical Predictions of the IZOF Model ........................................................... 45
Measurement Issues ............................................................................................... 53
Critique of IZOF Theory and Literature ................................................................ 59
Other Issues Addressed in this Study ..................................................................... 63
Extending the IZOF Model: Introduction to Individual Affect-Related Performance
Zones (IAPZ) ......................................................................................................... 64
An Introduction to Individual Affect Performance Zones (IAPZ) ........................ 64
How IAPZ Improves the IZOF Model ................................................................... 67
Critique and Shortcomings of IAPZ ...................................................................... 72
Conclusion: Is IAPZ a method or a model? ........................................................... 76
Summary: Integrating the IZOF and IAPZ Frameworks ............................................. 78

vii

CHAPTER THREE

METHOD .......................................................................................................................... 80
Participants ............................................................................................................. 81
Measures of Feeling ............................................................................................... 82
Procedure ............................................................................................................... 84
Ethical Considerations ........................................................................................... 88

CHAPTER FOUR ‘

RESULTS .......................................................................................................................... 90
Treatment of Raw Data: The basics ....................................................................... 90
Descriptive Data ..................................................................................................... 92
Treatment of the Raw Data: Ordinal Logistic Regressions ................................... 96
Interindividual Differences .................................................................................. 100
Intraindividual Context Differences .................................................................... 123
Accuracy of Discrete Categories versus Trigger Items ....................................... 129

CHAPTER FIVE

DISCUSSION .................................................................................................................. 133
Integrating Elements of IZOF with IAPZ ............................................................ 133
Intraindividual Context Differences .................................................................... 136
Simpliﬁed F eeling-Perfonnance Proﬁles ............................................................ 138
Strengths and Limitations .................................................................................... 140
Future Research Directions .................................................................................. 144
Applied Implications ............................................................................................ 147

APPENDICES ................................................................................................................. 152

REFERENCES ................................................................................................................ 171

viii

LIST OF TABLES

Table 1. Summary of Empirical Articles in the area of Individual Affect Performance
Zone ................................................................................................................................... 65

Table 2. Distribution [and Percentage Proportion] of Performances for Each
Participant .......................................................................................................................... 91

Table 3. Individual Means and Standard Deviations for Positive-Functional Feeling
Intensities across Performance Levels ............................................................................... 94

Table 4. Individual Means and Standard Deviations for Negative-Functional Feeling
Intensities across Performance Levels ............................................................................... 95

Table 5. Individual Means and Standard Deviations for Negative-Dysfunctional Feeling
Intensities across Performance Levels ............................................................................... 95

Table 6. Individual Means and Standard Deviations for Positive-Dysfunctional Feeling
Intensities across Performance Levels ............................................................................... 96

Table 7. Player Support for each Row of (Synonymous) Positive and Negative Content
Item Descriptors ............................................................................................................... 102

Table 8. Interindividual Comparisons of Mean Difference [and Variance] for Positive-
Functional Feeling Intensities during Poor Performances ............................................... 159

Table 9. Interindividual Comparisons of Mean Difference [and Variance] for Positive-
Functional Feeling Intensities during Average Performances ......................................... 160

Table 10. Interindividual Comparisons of Mean Difference [and Variance] for Positive-
Functional Feeling Intensities during Good Performances .............................................. 161

Table 11. Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Functional Feeling Intensities during Poor Performances ............................................... 162

Table 12. Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Functional Feeling Intensities during Average Performances ......................................... 163

Table 13. Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Functional Feeling Intensities during Good Performances .............................................. 164

Table 14. Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Dysfunctional Feeling Intensities during Poor Performances ......................................... 165

ix

Table 15. Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Dysfunctional Feeling Intensities during Average Performances ................................... 166

Table 16. Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Dysfunctional Feeling Intensities during Good Performances ........................................ 167

Table 17. Interindividual Comparisons of Mean Difference [and Variance] for Positive-
Dysfunctional Feeling Intensities during Poor Performances ......................................... 168

Table 18. Interindividual Comparisons of Mean Difference [and Variance] for Positive-
Dysfunctional Feeling Intensities during Average Performances ................................... 169

Table 19. Interindividual Comparisons of Mean Difference [and Variance] for Positive-
Dysfunctional Feeling Intensities during Good Performances ........................................ 170

Table 20. Summary of Individual Differences across Performance and Feeling
Category ........................................................................................................................... 122

Table 21. Intraindividual Comparisons of Central Tendency and Variance F-Ratios across
Return and Serve Contexts for the Positive-Functional Feeling Category ...................... 125

Table 22. Intraindividual Comparisons of Central Tendency and Variance F -Ratios across
Return and Serve Contexts for the Negative-Functional Feeling Category .................... 126

Table 23. Intraindividual Comparisons of Central Tendency and Variance F -Ratios across
Return and Serve Contexts for the Negative-Dysfunctional Feeling Category ............... 127

Table 24. Intraindividual Comparisons of Central Tendency and Variance F-Ratios across
Return and Serve Contexts for the Positive-Dysﬁmctional Feeling Category ................ 128

Table 25. Accuracy of Performance Predictions for Feeling Categories and Trigger
Items ................................................................................................................................. 131

LIST OF FIGURES

Figure 1. Example of an Individual Proﬁle Graph for a College Distance Runner ........... 13
Figure 2. Example of the First Form from the Stepwise Individual Emotion Proﬁling
(IEP) System .................................................................................................................... 152
Figure 3. The Positive-Negative Affect (PNA-77) Scale: The second IEP form ............ 153
Figure 4. Example of the Monitoring Form used by Participants .................................... 154
Figure 5. Example of a Typical IAPZ Graph of Probabilistic Curves ............................... 68
Figure 6. Frequency of Combined Positive-Functional Intensity Ratings for All
Participants ....................................................................................................................... 155
Figure 7. Frequency of Combined Negative-Functional Intensity Ratings for All
Participants ....................................................................................................................... 156
Figure 8. Frequency of Combined Negative-Dysfunctional Intensity Ratings for All
Participants ....................................................................................................................... 157
Figure 9. Frequency of Combined Positive-Dysfunctional Intensity Ratings for All
Participants ....................................................................................................................... 158
Figure 10. Example of an Excel Spreadsheet used to Generate OLR Performance

Curves ................................................................................................................................ 99
Figure 11. Positive-Functional IAPZ Curves for Player 1 ............................................... 104
Figure l2. Positive-Functional IAPZ Curves for Player 2 ............................................... 105
Figure 13. Positive-Functional IAPZ Curves for Player 6 ............................................... 106
Figure 14. Positive-Functional IAPZ Curves for Player 4 ............................................... 107
Figure 15. Negative-Dysfunctional IAPZ Curves for Player 3 ........................................ 108
Figure 16. Negative-Dysfunctional IAPZ Curves for Player 2 ........................................ 109
Figure 17. Negative-Functional IAPZ Curves for Player 3 ............................................. 110
Figure 18. Positive-Dysfunctional IAPZ Curves for Player 1 ......................................... 111

xi

Figure 19. Positive-Functional Affect Performance Zones for Each Participant ............ 112
Figure 20. Negative-Functional Affect Performance Zones for Each Participant ........... 113
Figure 21. Negative-Dysfunctional Affect Performance Zones for Each Participant ..... 114

Figure 22. Positive-Dysfunctional Affect Performance Zones for Each Participant ....... 115

xii

CHAPTER ONE
INTRODUCTION

Have you ever been involved in, or seen, a sporting event that did not involve
emotion? Tears, screams, laughter, and dejected silence are as much a part of sport as
scoring and rules. Sport and emotion are inseparable. As the forthcoming review of
literature in emotion will explain, it is impossible to remove emotions from sport
experiences. Motivation, personal challenge, physical and psychological threat, reward
opportunities, and the need for strong coping skills—all of these things are implicit to
both sport and emotion (Lazarus, 1991 , 2000b). So long as athletes have an interest in the
outcome of their training or competition, they will experience emotions.

Some coaches, athletes, and even sport scientists incorrectly view emotions as bad
things that undermine performance (Botterill & Brown, 2002). In addition to this bias
against emotions (and feelings in general), some in sport believe that emotions should be
avoided or ignored because of their complexity. However, these perspectives are more
likely to harm performance than to facilitate it. Feelings are adaptive mediators of person-
goal relationships, and they are unavoidable. So athletes and practitioners must accept
them and learn to control them. Emotions and other feelings can enhance performance
by: (a) improving metacognitive understanding of the effects that emotions have on
functioning; (b) improving emotional self-awareness; and (c) by developing
psychological skills and routines that meet the unique emotional needs of athletes.

Feelings and Performance in Sport
The objective of this study is to examine the relationship between feeling states and

performance in sport. More speciﬁcally, the effects of feelings on performance during

competitive tennis matches will be assessed for the purpose of integrating the Individual
Zones of Optimal Functioning (IZOF) and Individual Affect-Related Performance Zones
(IAPZ) models. The design of the study was informed by appraisal theories of emotion,
and reviews of Individual Zones of Optimal Functioning (IZOF; Hanin, 1997, 2000a),
and Individual Affect Performance Zones (IAPZ; Kamata, et al., 2002; Tenenbaum, etal.,
2008) literature. In addition to scientiﬁc foundations, the design and procedures for this
study were inﬂuenced by the investigator’s applied experiences working with athletes.
The study of feelings in sport is important because athletics provide an ideal context for
the study of emotional appraisal and expression. As such, this research provides an
opportunity for sport psychology to contribute to the knowledge of general psychology.
Affect-Related Terminologies

One problem in sport psychology research and in practice is the failure to recognize
distinctions between similar affect-related terminologies. While these deﬁnitions are
somewhat contentious, researchers must nevertheless establish common meanings for the
words they use (Ekkikakis & Petruzzello, 2000). As much as possible, it is important for
these technical deﬁnitions to connect with non-academic practitioners such as coaches
and athletes. While this dissertation is interested in more general feeling states (e. g.,
feeling strong, conﬁdent, worried, satisﬁed, etc.), emotions are of particular interest
because they comprise the majority of items in individuals’ proﬁles, and there is more
advanced research and theory in the area of emotion than for any of the other affect-
related terminology. In addition to feeling and emotion, the terms affect, mood, anxiety,

arousal, and activation are described in the following section. In the second chapter,

theories are explored that describe how and why emotions function. These theories
provide a foundation for feeling-performance models such as Hanin’s (2000) IZOF.
Feeling states. Feeling states are perhaps the broadest category and are synonymous
with both affective conditions and physiological sensations such as pain. According to
Lazarus (1991, 2000b), sensations are reﬂexes, rather than emotions. Reﬂexes are
considered “pre-emotions” because they can promote certain emotional states and occur
without thought (Lazarus, Averill, & Opton, 1970). For instance, if George were to burn
his hand on a pot, the immediate reaction to pull his hand away from the pot is considered
a reﬂex. Because of his pain, the reﬂex is likely to precipitate an emotion such as anger
or fear. The onslaught of profanity uttered by George is a behavioral aspect of emotion
(e.g., coping), not of the reﬂex. Based on the connection with somatic sensations, feelings
like light, strong, quick, and ﬁerce may be considered feeling states.
Lazarus (1991, p. 57) clariﬁed the distinction between feeling and emotion:
The distinction between reﬂex and emotion is much like one that has been made
between feeling and emotion. These terms are often used interchangeably, but I think
it would be more precise to speak of feelings as sensory perception, as in feeling pain,
pleasure, and distaste, rather than as emotion. Although we speak constantly about
feelings when we mean emotions. . .it is more precise to restrict the word feeling to the
awareness of bodily sensations and to reserve the word emotion for occasions on
which there has been an appraisal of harm or beneﬁt.
Clifford Saron, psychologist at the Albert Einstein Medical School in New York City

elaborated in Goleman (1997, pp. 82-83):

Our common sense vocabulary is a central problem in talking about emotions. . .doubt
is not considered a basic emotion, like anger, happiness, or sadness. However, it is
certainly a feeling state that we can identify. When you say ‘doubt,’ I know what you
mean. . .Certainty, conﬁdence, and pride would be considered emotions.

The current study assumes a pragmatic point of view and includes the broadest
deﬁnition of possible affect-related states. This is because feeling conﬁdent, feeling
strong, or feeling sluggish, (etc.) may be very important for athletes. These items
certainly describe the athlete’s feeling state, but they are not true emotions in the view of
most researchers. As such, these descriptive items should be included within idiographic
performance proﬁles. Also, several items on the positive-negative affect scale (PNA,
Hanin, 1993b; Syrja & Hanin, 1997) are not considered to be basic emotions (e.g., fast,
inspired, aggressive, sluggish, helpless, and conﬁdent). Lazarus (2000b) has, for instance,
identiﬁed 15 core emotions. The PNA is quite liberal, and by releasing some constraints,
Hanin has developed a more comprehensive list of performance-related factors. To
account for this broad, liberal criterion, the term “feeling states” will be used throughout
this dissertation to describe emotions, moods, somatic sensations, and other feeling-
related states.

Affect. Affect is the experience of pleasant and unpleasant responses (valence) and
does not include any type of cognitive component (F rijda, 1993). Affect is a general term
that applies to both moods and emotions (Campos, Keltner, & Tapias, 2004; Ekkikakis &
Petruzzello, 2000). As Ortony, Clore, and Foss (1987) explained, “the word ‘affect’
entails that all emotions are affective conditions, but that not all affective conditions are

emotions” (p. 343). Unlike emotion, affect cannot be reduced to describe more speciﬁc

phenomenon (Ekkikakis & Petruzzello, 2000). For instance, emotions may be categorized
into different discrete forms, such as anger, joy, anxiety, etc. Affect does not comprise
speciﬁc categories in the ways that emotion and mood do because affect is a global
term—as F rijda described, it is a component of emotion and mood. However, affect can
be described in terms of the intensity of feeling and in terms of hedonic tone (also
referred to as valence or pleasantness).

Emotion. Emotion is probably the most relatable, yet the most difﬁcult term to deﬁne
or explain. It is a functional process in that it promotes adaptational responses to
environmental demands, and in turn, facilitates survival and well being (Campos, et al.,
2004; Ekkikakis & Petruzzello, 2000). Emotions are subjective in nature, involve
physiological responses, and result in behavioral effects (Botterill & Brown, 2002). Deci
and colleagues (Deci, 1980; Deci & Ryan, 1985) deﬁned emotion as “a reaction to a
stimulus event (either actual or imagined). . . [That] involves a change in the viscera and
musculature of the person; is experienced subjectively in characteristic ways [known as
affect]; is expressed through such means as facial changes and action tendencies; and
may mediate and energize subsequent behaviors” (p. 85) (Deci, 1980).

While these descriptions help to explain what emotion is, we can also explain
emotion by what it is not. What separates emotions from non-emotional phenomenon?
Lazarus (1991) noted that emotions can also be deﬁned by three components of reactions
proposed by Drever (1952). First, emotions elicit an impulse to act, such as ﬁght or ﬂight
responses—or typically in sport, reverting to habit. Second, there are patterns of somatic
changes that facilitate the aforementioned tendency in order to reach adaptational goals.

In sport, these changes include increased heart rate, butterﬂies in the stomach, etc. The

third component of an emotional reaction is the subjective experiential component—also
known as affect—which includes appraisal and evaluative judgment. An example of this
component would be a nervous athlete thinking, “this [anxiety] is good for me—I feel
energized and ready to perform.”

Emotions are unique from non-emotional phenomena such as reﬂexes, motivation,
cognition, or feeling states in the following four ways: (1) emotions involve the presence
of personally meaningful relational content; (2) emotions involve an appraisal of
potential harm or beneﬁt; (3) emotions involve the potential for action-readiness; and
most of all (4), emotions involve physiological (nervous system and hormonal) changes
(Lazarus, 2000a). Speciﬁc examples of “emotions” can be derived from Lazarus’ (2000b)
15 emotional categories, which include: anger, anxiety, fright, guilt, shame, sadness,
envy, jealousy, happiness, pride, relief, hope, love, gratitude, and compassion. Emotion
will be described in greater detail in the following sections. Though the focus of this
dissertation includes all feeling states, emotions have a particularly strong impact on
performance because emotions are more intense and dynamic than moods, and more
speciﬁc than affect or feeling states (Ekkikakis & Petruzzello, 2000; Mellalieu, 2003).
Speciﬁc theories of emotion will be provided in the following chapter.

Mood. Compared to emotions, moods are characterized as being less intense, longer
in duration (Albert & Rosen, 1990), and as having no speciﬁc target or source (Frijda,
1993). Moods may be considered to be sustained emotions. Common examples of mood
include being relaxed or depressed. The term “mood states” is somewhat oxymoronic
because mood is more trait-oriented and dispositional, whereas emotions are state-

oriented experiences. Like affect, mood is a too-often misrepresented term, which Levin

describes as “a clumsy phenomenological term” (Levin, October 3, 2003, p. 48). While
emotion and mood are generally believed to be cognitive in nature, and mediated by
appraisal processes, the critical difference between these states is the source of appraisal.
Moods are not a reaction to a speciﬁc event, but rather to a more holistic existential
evaluation of well-being (Ekkikakis & Petruzzello, 2000; Lazarus, 1991). As such,
moods are more stable than emotions, and they relate to more dispositional person-related
factors. Although mood is often believed to be a product of perceived physical energy,
the opposite is true: “psychic energy” is a result of mood (Morris, 1992). Another
difference between mood and emotion is that action tendencies are nearly irrelevant in
the case of moods (Lazarus, 2000b).

Anxiety, arousal and activation. Anxiety is a negative (i.e., unpleasant) emotion
resulting from perceived threat (Campos, et al., 2004; Raglin & Hanin, 2000). Anxiety is
characterized by vigilance and is closely associated with, but unique to, arousal.
According to Lazarus (1991), emotions (including anxiety) must elicit physiological
changes (e.g., autonomic nervous system stimulation, CNS stimulation, and hormonal
changes), but not necessarily arousal. Arousal is a condition of preparedness that
includes various levels of emotional display, physical energy, mental activity, and
associated physiological activity (Campos et al., 2004). Arousal can be described as an
activation level, ranging from low (e.g., sleepy) to high (e.g., manic). As well as being
synonymous with arousal, activation may refer to the intensity of an affective state. To be
more precise, activation describes the readiness or preparedness of a performer (Hardy,
Jones, & Gould, 1996). It is common to use the terms intensity, arousal, and activation

interchangeably (Lang, 2000; Zaichkowsky & Naylor, 2004).

The Emotion-Performance Relationship

One of the oldest and most researched topics in sport psychology is the anxiety?
performance relationship. Within this line of inquiry, the earliest, yet least researched
description of the arousal-performance relationship, was the Drive Theory (Spence &
Spence, 1966), which predicted a simple linear relationship between arousal and
performance. This relationship was explained by social facilitation theory (Zajonc, 1965),
which explains that higher activation is helpful for simple or well-learned tasks. This
model predicts that high arousal will facilitate activities like weight-lifting; and that an
experienced tennis player will serve better under pressure than a player who has just
learned how to serve. However, due to the over-simplicity of the model (i.e., its inability
to account for enough relevant factors), Drive theory has limited and equivocal support.

Researchers turned next to the Inverted-U hypothesis, or the Yerkes-Dodson law,
which describes the arousal-performance relationship as being curve-linear (See,
Landers, 1980; Yerkes & Dodson, 1908). In short, optimal performance occurs at a
moderate level of arousal (hence the inverted-U shape of the curve). Easterbrook’s (1959)
Cue Utilization theory lent a scientiﬁc explanation for the Inverted-U hypothesis. This
theory explains that high arousal is associated with narrow attentional focus, and that
very low arousal fosters a wider ‘spotlight’ of attention. Very wide attention (low
arousal) may cause athletes to process too much information, whereas very narrow
attention may cause an athlete to miss information.

Further exploration led sport scientists to realize that simple anxiety / arousal models
could not account for performance effects. As Landers (1980) explained, “anxiety is a

multidimensional phenomenon and we should use multi-method procedures to examine

it” (p. 217). This means that, in order to effectively proﬁle/predict performance, multiple
feelings must be accounted for. To this end, Martens, Vealey, and Burton (1990)
produced the Multidimensional Anxiety Theory (MDAT). The MDAT is the ﬁrst affect-
performance model to acknowledge that physical activation and mental activation should
be differentiated; and that inﬂuence of conﬁdence should also be considered. Although
there is general support for each dimension of the MDAT, examinations of the collective
postulates of the model have yielded inconsistent results. MDAT failed to account for the
simultaneous interactive effects of each dimension. More importantly, and like all models
up to this point, MDAT failed to acknowledge individual differences, favoring instead to
provide the same predictions for every athlete.

The Cusp-Catastrophe model (Hardy, 1990, 1996; Hardy & Frazey, 1987, June)
contends that performance is a function of interaction of physiological arousal and
cognitive anxiety. Based on complex mathematics, the Catastrophe theory succeeded in
modeling the anxiety-arousal-conﬁdence-perfonnance relationship within four-
dimensional space, but failed to yield consistent statistical support.

All of the theories presented above have two fundamental ﬂaws. First, as nomothetic
models they cannot account for individual differences. Second, they fail to recognize that
an even broader sample of emotions was necessary in order to explain how and why
affect inﬂuences performance. Hanin’s (Hanin, 1997, 2000a, 2003, 2007) Individual
Zones of Optimal Functioning (IZOF) model accounts for idiosyncrasies in affect-
performance relationships. Also, rather than being conﬁned to multidimensional anxiety,

Individual Zone proﬁles include any emotion that is personally relevant to an athlete.

However, before describing the IZOF model that is central to this dissertation, we should
preface the discussion by deﬁning emotions themselves.
The Individual Zones of Optimal Functioning Model

The theoretical foundation of this dissertation is the IZOF model (Hanin, 2000a,
2000d). Hanin’s original model emerged from his applied experiences consulting elite
Soviet athletes. These experiences contradicted the established (nomothetic) anxiety-
perforrnance models of the time. Based on his training and experiences, Hanin developed
the Zones of Optimal Functioning model (Hanin, 1978). With each series of studies
conducted by Hanin and others, the model was continually reﬁned, until it evolved into
what could be termed the modern Individual Zones of Optimal Functioning model
(Hanin, 1997, 2000a, 2003). The IZOF model establishes idiographic (individually
unique) proﬁles that describe the relationship between an individual athlete’s feelings and
best/worst performances. Each proﬁle incorporates a constellation of emotions (and other
psychobiosocial states), as well as the speciﬁc intensity zones associated with either
optimal functioning or with dysfunction. Within each proﬁle, emotions are organized into
four categories based on the ﬁmctional/dysﬁmctional effects and positive/negative
valence of each feeling state.

There are numerous ways to generate idiographic proﬁles (See, Hanin, 2003). In-
depth interviews, metaphoric description, and narrative procedures each have particular
advantages and weaknesses (Hanin & Starnbulova, 2002; Robazza, 2006; Ruiz & Hanin,
2004). However, the most commonly used method is called the Individual Emotion
Proﬁling (IEP) system. These stepwise procedures begin by having athletes recall their

personal best— and worst-ever performances. Then, based on these competitions, athletes

10

identify the positive and negative emotions associated with each performance and
estimate the intensity of the feeling state. Emotions associated with best performances are
deemed to be ‘ﬁmctional’ and worst performances identify ‘dysfunctional’ emotions. All
together, this information generates two graphic representations of individualized
emotion-performance relationships: a zone of optimal functioning (where intensity level
for functional categories are higher and dysfunctional affect is lower); and a zone of
dysﬁmction (with opposite characteristics).

Hanin has made three landmark contributions to the ﬁeld of sport psychology. Within
the topic of emotion-perforrnance relationships and in the realm of applied sport
psychology, Hanin has helped to shift the paradigm from exclusively nomothetic
perspectives, to consider individual differences. Secondly, more than any other
framework in Kinesiology, IZOF has promoted the study of emotions (Lazarus, 2000b).
And ﬁnally, for those who study emotion or regulate emotion in sport, Hanin has been a
major inﬂuence in how we see emotions. IZOF has promoted the view that emotions are
not inherently bad, but functional and adaptive—in this sense, Hanin’s proof that
negative emotions can be helpful to functioning is theoretically and practically important
to our ﬁeld. For instance, anxiety is a negative emotion that can be either helpful or
harmful to ﬁrnctioning—depending on the intensity level of the feeling, and on whether
or not an athlete interprets anxiety as being helpful or harmful to performance.

Despite these principal contributions, Hanin’s IZOF model cannot evade criticism.
IZOF-based analyses typically produce two proﬁles describing only poor and optimal
performance levels. The failure of this system to account for moderate performance

levels would prevent a graphic problem that undermines the predictive ability of the

11

model. Because typical proﬁles include gaps and/or overlaps between the optimal and
dysfunctional zones (See Figure 1), the IZOF model is sometimes ambiguous and cannot
predict performance. Figure l is a typical ﬁnal IZOF proﬁle. There are two zones
representing optimal (the dotted pattern) and poor (the striped pattern) performance.
Discrete items are located on the x-axis and organized into valence-function categories.
Intensity deﬁnes the parameters of each optimal and poor zone, and is described on the y-
axis. Note that there are gaps in the optimal feeling categories (P+ and N+) and overlaps
in the dysfunctional categories (N - and P-).

For example, Jack White’s optimal zone for conﬁdence is from 6-8 (on a 10-point
scale) and his dysfunctional zone ranges from 2-4. Would his performance at a “ﬁve” be
poor, optimal, or average? If the gap was larger, uncertainty would be even greater.
Likewise, what if John Lennon’s zones were 2-6 and 5-8? Intensity ratings of less than
two, ﬁve to six, and more than eight are undeﬁned or ambiguous. Hanin and colleagues
do not explain or even acknowledge these very fundamental problems. The existence of
gaps and of overlaps in idiographic proﬁles creates theoretical and practical problems for
researchers and practitioners in terms of the predictive ability of the IZOF model. These
issues are addressed by the Individual Affect-Related Performance Zone (IAPZ) model
described in the following section. One last critique of the IZOF model is that it may be
too complex. Speciﬁcally, a complete proﬁle for an individual athlete normally includes
16-20 discrete feeling items. The challenge for researchers and practitioners is always to
develop models that are complete, yet parsimonious. The IZOF model is more

comprehensive than it is parsimonious.

12

Figure 1
Example of an Individual Proﬁle Graph for a College Distance Runner

Reﬂnlng the Combined IZOF Proﬁle

 

Name: Sandy Context: COMPETITION
Date: Revised Location: 20034-2005 Time: BEFORE&DURING
Dysfunctional Optimal Emotions Dysfunction

 

 

 

Intensity (ca-10)

 

 

 

 

Note. The ﬁgure includes two zone proﬁles associated with optimal (zone with dots) and
poor performance (zone with diagonal stripes). Discrete feelings are organized into four
categories (negative-dysfunctional, negative-functional, positive-functional, and positive-
dysfunctional) and located on the x-axis; while the intensity of each feeling is described
on the y-axis. The intensity ratings act as parameters for optimal- and poor-zones.

The Individual Affect Performance Zones Model

As an extension of the IZOF model, IAPZ uses logistic regression to establish
probabilistic estimates for performance. Continuing from the previous example, IAPZ
could specify that at a “six,” Lennon has a 25% chance of functioning poorly, a 45%
probability of moderate performance, and a 30% chance of optimal functioning. This
would therefore be deﬁned as a moderate performance zone. In this way, it is clear to see

that IAPZ avoids the all-or-none descriptions of optimal and poor results predicted by

IZOF’s in-out of zone hypothesis. IAPZ has also made contributions to the ﬁeld by
conducting more direct assessments of emotion during competition (Cohen, Tenebaum,
& English, 2006; Edmonds, Mann, Tenenbaum, & Janelle, 2006; Johnson, Edmonds,
Moraes, Medeiros F ilho, & Tenenbaum, 2007; Medeiros Filho, Moraes, & Tenenbaum,
2008). Tenenbaum and colleagues’ Individual Affect Performance Zone (IAPZ) model
resolves the gap and overlap faults by establishing probabilistic estimates for optimal,
moderate, and poor performance levels across the entire range of emotional intensity
(Kamata, et al., 2002; Tenebaum, Kamata, & Hanin, 2002; Tenenbaum, etal., 2008).

However, relative to the body of IZOF literature, IAPZ studies have tended to be less
ecologically valid (Edmonds, et al., 2006; Tenebaum, et al., 2002) and less generalizable
to mainstream sports. For instance, because they use such ﬁne motor skills, both archers
(Johnson, Edmonds, Moraes, etal., 2007; Medeiros Filho, et al., 2008) and golfers
(Cohen, et al., 2006) aspire to reduce arousal as much as possible—an ironic
characteristic of two sports included in emotion research. This minor critique is most
applicable to negative feelings such as anxiety, anger, aggressions, and worry, etc.
Certainly, athletes in aiming sports generally desire high levels of positive feelings such
as conﬁdence or calmness.

An additional concern is that every study conducted thus far has assessed general
dimensional affect, rather than discrete emotions. In other words, rather than assessing
the intensity of relaxation, anger, or focus, IAPZ measures the valence (pleasant /
unpleasantness), arousal (similar to intensity), and sometimes also, the functionality of
their global affective state. This view of affect has empirical support (Russell, 1980,

1997; Russell, Weiss, & Mendelsohn, 1989; D. Watson, 2002), but there are also

14

signiﬁcant disadvantages with the approach (Lazarus, 1991, 2001). These concerns are
described in detail in the second chapter.
Issues with the IZOF and IAPZ Models Examined in this Study

This dissertation addresses the utility of using a combined IZOF and IAPZ model to
predict tennis player performance. In doing so, the following IZOF- and IAPZ-related
issues are addressed. (a) Probabilistic estimates were conducted on each of Hanin’s four
(function-valence) categories of discrete emotions to determine if IAPZ-based logistic
regressions can be used with discrete, rather than continuous, feelings. (b) Feeling states
and performance will be assessed during competition in a more vigorous and emotionally
intense sport (tennis) than has been used in the previous research. As such, the
transferability of applying the IAPZ framework to more vigorous, emotionally intense
sports than golf and archery (wherein the object of those sports is to suppress negative
feelings—be they functional or dysfunctional) are tested. (c) Building upon the strength
of previous IAPZ research to collect large numbers of observations (relative to smaller
sample sizes in IZOF studies), the proposed study attempted to acquire between 800-
1200 observations across nine collegiate tennis players. (d) Complete individual proﬁles
consisting of 10-14 emotional items were compared to more parsimonious models
featuring each athlete’s self-selected favorite four items. (e) Finally, though not addressed
previously in this introduction, the proposed study will integrate some qualitative
procedures—in the form of brief, occasional collaborative discussions with each player—
to reﬁne the IZOF proﬁles, to develop IZOF-based routines and psychological skills

goals, and to obtain feedback about the procedures.

15

Purposes and Hypotheses

Purposes

This dissertation had three objectives. The primary purpose of this study was to
integrate elements of two idiographic feeling-performance models and assess whether
this integrated model can predict athletic performance. Speciﬁcally, can ordinal logistic
regression (a foundation of the IAPZ method) be used to identify zones of functioning for
discrete feeling states (a foundation of the IZOF theory). The beneﬁt of the integrated
system would not necessarily be a better performing (i.e., more accurate) model, but
would have the theoretical and practical advantages of using discrete feeling items rather
than continuous descriptions of global affect. The second objective was to explore the
context dimension of the psychobiosocial model by comparing the zone proﬁles of during
service games to return of serve games. The third objective was to detemrine if a small
number of critical items determined by an athlete can predict performance as well as a
traditional, complete IZOF proﬁle. Based on these three purposes, there were three
corresponding hypotheses.
Hypotheses

Hypothesis 1. It was hypothesized that idiosyncratic feeling-performance proﬁles can
be developed for each participant using elements from both the IZOF and IAPZ methods.
Speciﬁcally, discrete feelings will be arranged into valence-function categories, and these
categories will predict the probabilities of poor, average, and good performances based
on ordinal logistic regression modeling. Separate tests will be performed for each feeling
category and performance level. Each test will compare the shape of each zone——

represented by the means and variances—between individuals. This hypothesis is

16

foundational for the entire study because it will determine if IZOF and IAPZ
methodologies can be integrated to produce meaningful idiosyncratic proﬁles.

Hypothesis 2. It was hypothesized that idiographic proﬁles will vary between serve
and return of serve contexts for each player. Feeling intensity and variance of proﬁles
will be compared within players across the two contexts for each feeling category and
performance level. This is an important test of the understudied dynamic context-
dimension of the psychobiosocial model put forth by Hanin (2000).

Hypothesis 3. It was hypothesized that critical the four discrete items chosen by
players as being their most important performance-feelings would not predict
performance as well as the full proﬁles (with 12 items) would. However, the four simple
“trigger items” would be nearly as effective in proﬁling and predicting performance. To
determine which were more effective, the accuracy of performance predictions for zones
in the four valence-function categories and the four trigger items were compared. Players
selected the “trigger” items from among the 12 items chosen in their proﬁle.

Definition of Terminologies

The following section provides summary deﬁnitions of important terminology that is
frequently used in subsequent chapters of this dissertation. These terms are organized into
two sections for affect-related phenomena and feeling-performance models.
Affect-Related Terms

Feeling States. The broadest category of affect-related terminologies, feeling states
are an umbrella term that includes affective conditions and physiological sensations such
as pain. Lazarus (1991, p. 57) clariﬁed the distinction between feeling and emotion

because the terms are too oﬁen “used interchangeably, but I think it wOuld be more

17

precise to speak of feelings as sensory perception, as in feeling pain, pleasure, and
distaste, rather than as emotion. . .to restrict the word feeling to the awareness of bodily
sensations.” Feelings such as reﬂexes are considered “pre-emotions” because they can
promote certain emotional states and occur without thought (Lazarus, et al., 1970). For
the purposes of this dissertation, feelings include performance-related descriptors like
quick, conﬁdent, and tired among others.

Affect. Affect is the experience of pleasant and unpleasant responses (also referred to
as valence, pleasantness, or hedonic tone). Affect does not involve a cognitive component
(Frijda, 1993), and is a global term that applies to both moods and emotions (Campos, et
al., 2004; Ekkikakis & Petruzzello, 2000). As Ortony, Clore, and Foss (1987) explained,
“the word ‘affect’ entails that all emotions are affective conditions, but that not all
affective conditions are emotions” (p. 343).

Emotion. Emotion is very difﬁcult to deﬁne. This term will be differentiated from
related concepts such as affect, mood, arousal, and feeling states in the beginning of the
second chapter. Emotions are adaptive situational responses to appraised threat that are
strongly mediated by motives, attributions, and coping skills (Lazarus, 1991, 2000b;
Scherer, 2001; Schorr, 2001 ). These adaptations can be ﬁrnctional or dysfunctional. A
mood is a less intense but more sustained feeling state than an emotion.

Valence. Also referred to as hedonic tone, valence describes the pleasantness of an
affective state (ranging from unpleasant to very pleasant). Examples of negative valence
include anxiety, anger, sluggishness, and fear. Examples of positive valence include

conﬁdence, happiness, and boldness (Lang, 2000; Smith & Lazarus, 1990, 1993).

18

Arousal. Arousal describes the degree of activation in the body, ranging ﬁ'om low
(sleepy) to high (manic) (Zaichkowsky & Naylor, 2004). Neurophysiological indicators
for arousal include changes in heart rate, skin conductance, and hormone levels. These
physiological responses to stressors are mediated by cognitive processes (Raglin, 2004).
Arousal is closely related to the intensity dimension of affect (Martens, et al., 1990).

Anxiety. Anxiety is an emotion characterized by threat or challenge that is perceived
to be greater than the person’s capacity (Campos, et al., 2004; Raglin, 2004). Anxiety is
considered a negative emotion in terms of hedonic tone (Zaichkowsky & Naylor, 2004).
Feeling-Performance Models

IZOF. Hanin’s (1997; 2000; 2003) Individual Zones of Optimal Functioning model.
This model describes the relationship between emotions and related psychobiosocial
factors and optimal functioning.

IAPZ. The Individual Affect-Related Performance Zones model (IAPZ; Kamata, et
al., 2002; Tenebaum, et al., 2002) is an extension of the IZOF model. IAPZ develops
probabilistic estimates for poor, moderate, and optimal performance levels.

Zones. Zones, in both the IZOF and IAPZ models, refer to the parameters (or
bandwidth) of feeling intensity that deﬁne the lower- and upper-bound at which a certain
performance level is most likely. For example, on an 11-point intensity rating for a
certain feeling, poor performance may be more likely than any other outcome from
intensities between 0-5.5; average performance is most likely from 5.6-7.9, and good
performance is the most likely outcome if feelings are in a zone between 8.0-11.0. These

are typically referred to as poor-zones, average-zones, and good-zones. Originally, Hanin

l9

deﬁned a zone of optimal functioning as one-half of a standard deviation above and
below the anxiety level associated with a best-ever performance.

Discrete Feelings. Feelings (emotions in particular) can be described by various
continuous dimensions or labeled as discrete items. Discrete feelings include categories
like happy, sad, and angry. A person can feel more than one discrete item at a given
moment. For example, before a tennis match, a player may simultaneously feel conﬁdent,
mentally calm, physically anxious, and excited.

Continuous Feelings. Feelings can also be described in terms of several continuous
dimensions. These dimensions describe characteristics of emotions or feeling states, such
as valence, arousal, and motivation-consistency, etc. Dimensional descriptions of feelings
provide a more descriptive account of an individual’s general state. For example, at a
given moment, a player who feels somewhat anxious may describe slightly negative
valence, and moderate arousal. While simpler than rating the intensity of several discrete
feelings, continuous descriptions cannot account for multiple co-existing feelings (such
as feeling simultaneously conﬁdent, calm, anxious, and excited).

Valence-Function Categories. Discrete feelings can be organized into categories
based on a two-by-two pleasant/unpleasant-functional/dysfunctional matrix. This matrix
results in four categories: positive-functional, negative-functional, negative-
dysfrmctional, and positive-dysfunctional groups. The IZOF model uses these groupings
to organize individual constellations of discrete feelings. All players in all sports can
organize their idiosyncratic proﬁles into these four categories.

Idiosyncratic Models. IZOF and IAPZ are idiosyncratic models of feeling-

performance relationships because they are founded on the premise that all athletes have

20

unique, individualized relationships between feeling and performance. Conﬁdence may
be important to John, but irrelevant to Paul. Anxiety may be important for both Ringo
and George, but Ringo plays great at moderate levels of anxiety, where George must have
very low levels of anxiety. These are examples of idiosyncratic differences.

Idiographic Differences. Idiographic differences are essentially a synonym for
idiosyncratic differences. In the IZOF and IAPZ literature, idio-graphic differences
generally refer to the shape of the individual proﬁles, or parameters of each zone. If, for
instance, if Woody’s average-zone is from 4.0-5.5, and Ledbelly’s average-performance
zone is from 5.5-9.0, their proﬁles are idiographically unique. In this case, they are
different in terms of their size (width) and location within the intensity continuum
(Ledbelly’s average-zone s twice as big and much higher).

Nomothetic Models. Early models of feeling-performance relationships were
nomothetic in nature because they described all individuals as sharing the same feeling-
perforrnance relationships. For instance, the Multidimensional-Anxiety Model (Martens, -
et al., 1990) predicts that all athletes will perform at their best when somatic-anxiety is
moderately intense. These nomothetic models contradict idiosyncratic models because

they treat all athletes the same and cannot account for interindividual differences.

21

CHAPTER TWO
REVIEW OF LITERATURE

This dissertation is designed to integrate elements of the IZOF and IAPZ models for
the purposes of identifying feeling-performance relationships; exploring these
relationships across different tennis contexts; and determining if a simple set of feeling-
items can predict performance as well as a more complete proﬁle. Given these purposes,
the review of literature is separated into three sections. The ﬁrst section provides a
theoretical foundation by describing the role that feelings and emotions have in human
functioning. The second section describes the ﬁrst of two feeling-performance models
that are foundational to this dissertation: Hanin’s (2000) IZOF. The third section
describes a newer model that is based on the idiosyncratic IZOF model: the IAPZ.

Why and How Feelings Affect Functioning

As explained in the previous chapter, the performance proﬁling that is fundamental to
this dissertation will not be exclusive to emotions. To determine what factors most
determine players’ performance levels, factors like conﬁdence, ﬁerceness, and quickness
must be considered along with true emotions such as relaxed, happy, nervous, and angry.
For this reason, general feeling states are of interest. However, due to a lack of research
and theory on feeling states, it is necessary that the focus of this review of literature now
be on emotions. In order to understand precisely how feelings inﬂuence performance, the
only option is to review theories of emotion and then to generalize this science to broader
feeling states. To that end, not only may some non-emotion feeling states affect
performance in a manner similar to emotions, but it is important to recognize that the

majority of items identiﬁed by athletes as being critical to performance are, indeed,

22

proper emotions. Consequently, the following section offers valuable theoretical insight
that is relevant to most, if not all, of the performance proﬁles developed by athletes.
Emotion and Adaptation

Evolution. Emotions are very advanced and complex extensions of stirnulus-response
reﬂexes that help all animals to survive in their environments (Darwin, 1998; James,
1884; Lazarus, 1991, 2000a). As life evolved, automatic responses to the environment
(i.e., reﬂexes) became less inﬂuential. Instead, humans appraise the threat potential of the
environment (as well as the opportunity potential). This appraisal process enables people
to process more information, consider past experiences, account for motivational factors,
and contemplate coping/response options. The evolution of cognitive operations has the
additional advantage of allowing us to predict environmental conditions and to prepare
response strategies (Csikszentrnihalyi, 1993). For instance, not only do emotions help
people to adapt to the biological demands of a situation, but also to the social demands.
Emotional processes resolve many of life’s challenges, such as cooperation, social
hierarchies, reproduction, and social interactions (Campos, et al., 2004; Ekkikakis &
Petruzzello, 2000; Goleman, 1995; Mayer & Salovey, 1997; Salovey & Mayer, 1990)

However, such that emotions are a product of appraisal, it is reasonable to say that
perception is more important than is reality (i.e., the physical environment or actual
threat). While there is still debate as to exactly what constitutes an emotion, the
“appraisal models” proposed by cognitive psychologists provide the strongest theories
available in my view (For reviews see, Campos, et al., 2004; F rijda, 1993; Frijda &
Zeelenberg, 2001; Lazarus, 1991, 2001; Lazarus & Folkman, 1984; Mayer & Salovey,

1997; Ortony, Clore, & Collins, 1988; Scherer, 2001; Schorr, 2001).

23

Relational meaning, valence, and appraisal. Emotions are driven by motivational
and perceptual dispositions that ﬁmction to establish a desired relationship between the
person and an environment (i.e., Relational Meaning) (F rijda & Zeelenberg, 2001). For
example, the more someone cares about a competition, the more emotional they will be.
To regulate relational meaning, emotions elicit action tendencies. By way of these action
tendencies, emotions facilitate behavioral responses to threat (e. g., ﬁght, ﬂight, or startle
responses). Action tendencies are an important component of emotional adaptation in
terms of coping (Lazarus, 1991, 2000b).

Valence (or hedonic tone) is another important attribute of emotions. In the distant
past, pleasure and positive emotional valence served evolutionary purposes (Lang, 2000;
Lazarus, 1991; Schorr, 2001). Positive emotions reinforce actions and objects that
promote the survival of our species. Likewise, negative emotions can be either aversive
(helping us to avoid negative things) or attacking (helping us to challenge negative
environment relationships). Positive valence is associated with superior health (Botterill
& Brown, 2002; Lazarus, 1991).

Emotions evolved to promote positive adaptations and they help people to reach their
goals—whether those goals are to avoid being prey, to win a ﬁght for food, to make a
body check in hockey, or to dribble a ball up court. But how, exactly, are emotions
evoked? Lazarus’ (1991) Cognitive Motivational Relational Theory describes the
appraisal process by which emotions are elicited. People appraise the relational meaning
of a situation or challenge in two stages. Primary appraisal is based on individual stake
(or motives). In some form, a person assesses three questions: do I care about the

outcome; is the outcome likely to be very good or bad; and how does the outcome affect

24

my ego? Secondary appraisal focuses on attributional (Weiner, 1985) considerations and
coping options (Lazarus & Folkman, 1984). A person considers the following three
questions: is my relational status attributed to personal or external causes; is my status
controllable; and is my status likely to remain stable? For example, if Dylan believes that
he can do something to either control his feelings or to change the situation that is
affecting his emotions, then his emotions are likely to be more helpful.

Adaptation. This biological perspective has an important practical implication in that
unpleasant emotions (i.e., “negative” states including fear, anger, anxiety, etc.) can be
helpﬁrl because they mobilize people to modify threatening or even harmful relationships
with the environment. From this perspective, we can see why emotions are considered to
be adaptational and fundamental to meeting the demands of a challenge. However, for
emotions to have this functional/adaptational signiﬁcance they must be more than just
“feelings”——they must be capable of promoting action. Emotions promote relational
changes with the environment by moderating attention and behavioral systems. In terms
of species survival, attention is primarily controlled by emotional-motivational systems
(Graham, 1979; Lang, 2000). Control over attention is important because it enables a
person to focus on threats, thus facilitating solution-development. Appetitive and
defensive emotions also have action tendencies associated with them (Davis, 1992;
Masterson & Crawford, 1982). Action tendencies are uniquely associated with each
emotion, which ensures that an emotional response will be adaptationally specialized for
the perceived threat.

Before a tennis match, for example, a player may feel jittery, butterﬂies in his/her

stomach, or sweaty palms, etc. According to scientiﬁc theory, this response is more likely

25

when the player is very close to reaching a goal (e.g., playing someone of relatively equal
ability, or a close match), and when the player cares more about the outcome. In this case,
human experience supports theory: athletes are typically more nervous in tight, highly
competitive matches where the outcome is important to the player. These physiological
responses are typical action tendencies for emotions like nervousness. They are part of
the adaptive quality of that emotion in that they help the players to generate energy and
redistribute blood ﬂow away from vital organs in favor of working tissue (muscles).
These adaptations are part of the ﬁght/ﬂight response, which helps athletes to perform
better in physical challenges. As such, a lack of nervousness can result in a lack of energy
and an inefficient distribution of blood circulation.

Maladaptive responses. Emotions are a product of highly-evolved cognitive
processes, but the underlying mechanisms that control these processes are outdated in
some ways (Csikszentrnihalyi, 1993; Lazarus, 1991, 2000b). Consequently, emotions are
imperfect in that they can cause maladaptive responses. Consider a scenario where a deer
is crossing a highway, but freezes in the headlights of an oncoming vehicle. This startle
response is normally adaptive and beneﬁcial because it enables an animal to go unnoticed
(camouﬂage) or appear dead to a stalking predator. It also provides animals with time to
make effective decisions. However, in this instance, evolution has failed to account for
highway trafﬁc and automobiles (Scherer, 2001). As a result, the deer’s startle response is
maladaptive because it assures the animal’s demise.

In sport, a common example of maladaptive feelings comes from how some athletes
perceive anxiety (or any other form of pre-competition activation). Athletes often become

worried and suffer performance decrements because of anxiety. This is because

26

symptoms of anxiety such as butterﬂies in the stomach and increased heart rate (which
are a helpful adaptation that enables athletes to mobilize more energy in order to
compete) are interpreted as being uncomfortable and bad. As such, athletes will lose
focus due to extreme worry. Appraisal can be distorted and reinforced by pleasure-
seeking (addictive) activities that may possess negative consequences—such as drug
abuse, unplanned pregnancy, eating disorders, and gambling debt, etc. (Csikszentrnihalyi,
1993). In sport, positive feelings (e. g., calm or overjoyed) can be maladaptive in that they
may reduce intensity or foster complacency. Negative feelings can be maladaptive if they
are overwhelming (e. g., being too nervous or too aggressive).
Characteristics of Emotion

Motivational and emotional intensity. Early theories in emotion—such as Instinct
Models and Drive Theories—were behavioral in orientation and generally neglected
cognitive factors. Ironically, physiology-based research has recently begun to support
cognitive approaches. Remember that emotions are associated with two biologically
driven motives: appetitive-approach (ﬁght), and defensive-avoidance (ﬂight) (James,
1884; Lang, 2000; Miller, 1959). Motivation is one of the most important and
fundamental mediators of emotion. From most points of view, motivation and emotional
intensity are closely associated and highly related to distance from a perceived threat,
beneﬁt, or goal (Lang, 2000). An athlete, for example, will generally be more anxious as
competition draws closer, or will be less motivated when they are losing by an
overwhelming score. Therefore, even from a biological/neurophysiological perspective,
emotion, motivation, and the relationship that a person has with his/her environment are

critical interrelated factors. Motivational and relational factors are highly cognitive in

27

nature. As such, biological perspectives support cognitive theories of emotion in some
ways (i.e., motivation). Cognitive Appraisal theories will be described in subsequent
sections. The motivational signiﬁcance of emotion is particularly relevant for, but not
exclusive to, appraisal models—motivation is a universal feature of affect.

Describing valence. Whether you prefer a cognitive or a physiological perspective,
valence is a universal component of emotion. People are hardwired, it seems, to organize
emotions hierarchically into pleasant (a.k.a., positive) and unpleasant (a.k.a., negative)
categories (Ortony, et al., 1988; Ortony, et al., 1987). Valence is associated with
motivation in that people are attracted to positive valence, and aversive to negative
valence. In other words, we are motivated to do things that feel good, and that we avoid
things that do not feel good. In this sense, emotions help to explain gambling, drug, and
sexual addictions, the obesity epidemic, and every-day concerns, such as why a person
may not attend to tedious chores. People are not only attracted to an object or experience
that inspires an emotional state (e.g., a rollercoaster, or scoring a goal), but to the emotion
associated with that experience. It is emotions, more so than experiences, people, or
objects, which inﬂuence human motives and behaviors.

Whereas emotional intensity is described within a continuum from low to high,
valence can be either continuous or categorical: one can describe a feeling in one of two
categories (pleasure and displeasure), or on a dimensional continuum from very pleasant
to very unpleasant (Lang, 2000; Lazarus, 1991). While the dimensional descriptions are
parsimonious, their simpliﬁed nature can obscure some of the important distinctions that
deﬁne emotions and make them unique. Most importantly from Lazarus’ (1991; 2000a;

2001; also Ortony et al., 1988) perspective, dimensional frameworks obscure the

28

relational meaning of emotions as separate categorical states. Examples of discrete
emotional “categories” include anger, fear, happiness, and pride. Discrete categories
retain intensity as a continuous factor, but view it as subordinate to the qualitative content
of each emotion. From a categorical perspective, it may be said that intensity does not
deﬁne an emotion—as it does from a dimensional approach—but describes an emotion
quantitatively (Ekkikakis & Petruzzello, 2000).

It is not a question of which is right or wrong, but of which is most appropriate for the
research question of interest (Lazarus, 2000a). A dimensional approach is better suited
for more general macro-descriptions and for exploratory studies of affective states;
whereas categorical descriptions are better for micro-descriptions of speciﬁc emotions
(Deiner, Smith & Frijda, 1995; Ekkikakis & Petruzzello, 2000; Russell, 1997).

Action-tendencies. Just as there are causes of emotions, there are also effects of
emotional expression. It is these effects that are, perhaps, the most important
characteristic of emotion in relation to sport performance. Emotional effects include
action tendencies, physiological changes, subjective affect, and subsequent (cognitive and
motivational) appraisals (Lazarus, 2000a). Action tendencies are a product of emotion
and can be part of the coping process. That said, coping is generally psychological,
deliberate, complex, and it involves planning. Action tendencies are “biologically driven
and therefore fairly rigid and [virtually] automatic (though capable of suppression and
transformation)” (Lazarus, 2000a, p.58). While action tendencies are fairly rigid, they are
probably the least deﬁnite and predictable component of the emotional process (Frijda &
Zeelenberg, 2001). Action tendencies are not consistent because they can be mediated by

coping processes (Lazarus, 2001). On the other hand, appraisal patterns—the “causes” of

29

emotions—are viewed as being more stable than action tendencies. The problem is that
action tendencies do not involve overt locomotor behaviors or the development of
speciﬁc motor programs, per se (Lang, 2000). Instead, “tendencies” include processes
such as attention focusing, increased arousal, felt readiness, and muscle priming. For this
reason, F rijda and Zeelenberg (2001) prefer to use the term “action readiness,” rather
than action tendency.
Appraisal Models

Having explained the purpose, and in part, the functional characteristics of emotions,
let us now consider the causes, or antecedents, of emotions. Early theories in emotion
proposed that emotions were elicited by the environment, and did not account for
evaluative potential of the person. For instance, Stimulus-Response Theories (J. B.
Watson, 1919), Physiological CNS Processes (Cannon, 1927), PNS Autonomic Activity
(James, 1894), F ight-Flight Theory (James, 1884; 1894), and Motivational Theories
focused on primitive hunger (Tomkins, 1962) and intimidation (Parkinson, 1997) drives.
These Biology-based models do not consider cognitive factors and are unable to explain
the idiosyncratic nature of emotion. Roseman and Smith (Roseman, 1996, 2001;
Roseman & Smith, 2001) compiled a list of questions that justify why cognitive models
are necessary to explain emotion. Why are there intra- and inter-individual differences in
emotional responses to the same stimulus? Why can a variety of different situations elicit
the same emotional response? What elicits an emotion when the environment does not
appear to be stimulating? Why and how is it that emotional responses can be both

functional and dysﬁmctional? How is it that emotions can be controlled, or modiﬁed?

30

As such, appraisal models developed ﬁ'om the ﬁeld of cognitive psychology were
developed to address the ﬂaws in purely biological and behavioral theories. Among these
appraisal models, the Cognitive Motivational Relational Theory (CMRT; Lazarus, 1991,
2000b) is arguably the strongest, but certainly the most accepted to sport researchers
interested in emotion-performance relationships.

The basic assumptions (Lazarus, 1991, 2000b; Roseman & Smith, 2001; Schorr,
2001) of Appraisal Theories are as follows: (a) emotions are mediated by cognitive
appraisals and each emotion (e.g., anger, joy, etc.) is a product of speciﬁc appraisal
patterns, or processes; (b) Appraisals initiate the emotional process, which includes
physical, neurological, and behavioral changes (Roseman, 1996, 2001; Weiner, 1985); (c)
Differences in appraisal patterns account for intraindividual (temporal) and
interindividual differences; ((1) A situation that elicits the same appraisal pattern will
evoke a consistent emotional response; (e) The appraisal process will generally ensure
that a response is appropriate—however, a misappraisal (caused by conﬂicting,
inappropriate, or incomplete processing) will result in an ineffective and maladaptive
emotional state (Lazarus, 2000a). It is important to recognize that appraisal can occur
through deliberate (volitional) control, or implicitly through automatic parallel processing
(Lazarus, 1991; 2001; Roseman, 2001).

The Cognitive Motivational Relational Theory (CMRT). Lazarus’ (1991)
Cognitive Motivational Relational Theory (CMRT) is probably the most accepted
appraisal model of emotion in the domain of sport psychology (Botterill & Brown, 2002;
Hanin, 2000a). The model is cognitive in nature as it is a form of appraisal theory.

Appraisal is the foundation of emotion and adaptation. “Feelings,” without appraisals, are

31

not emotions—they are mere reﬂexes. Though we cannot ignore the role of the
environment, emotions are strongly regulated by appraisal processes (Scherer, 2001). As
we saw from the review of biological perspectives in emotion, motivation is a critical
factor in determining emotional form, valence, and intensity (Lang, 2000). Accordingly,
Lazarus has incorporated motivational considerations into the CMRT. Relational
dynamics refer to the relationship between a person and his or her environment (or
perceived challenges, etc.), in terms of threat/well-being evaluations. Each discrete
emotion is derived from a unique relational meaning (Lazarus, 2000b; 2001). Relational
meaning is the core element of the CMRT. Lazarus (2000b) afﬁrmed this notion by
explaining: “appraisals and the meanings generated from them are always relational
because they must simultaneously take into account personal factors and environmental
demands, constraints, and opportunities” (p. 233). The term relational theme is used to
summarize the six components of appraisal (described in the following section). As such,
a “core relational theme” is the foundation of an “appraisal pattern.”

Lazarus (2000a) has identiﬁed three working assumptions for the CMRT. Firstly, an
emotion is deﬁned by its core relational theme, its appraisal pattern, and its action
tendency. Secondly, each distinct emotion is associated with unique physiological
changes. Thirdly, speciﬁcity of physiological responses are related to the action
tendencies associated with that emotion.

Structure of the Cognitive Motivational Relational Theory (CMRT). The
following section describes how the CMRT explains and predicts the inﬂuence that
emotions have on adaptation and performance. There are three important structural

components of the CMRT: primary appraisal, secondary appraisal, and coping.

32

Primary appraisal. There are two levels of appraisal in CMRT: primary and
secondary. Primary appraisal is based on individual stake in the object of focus (Lazarus,
1991; 2001a). There are three types of goal-related (stake) appraisals in this ﬁrst stage.
The ﬁrst is evaluation of “goal relevance.” If a goal is not important to a person,
appraisal will either elicit very weak emotional activation, or fail to stimulate emotion
altogether. This is why we see more tears in the Stanley Cup ﬁnals or Wimbledon than
we see in exhibitions. The second is “goal congruence,” which refers to the potential
beneﬁt or harm associated with the interpretation. As described earlier, goal congruence
will, therefore, determine the (positive or negative) nature of the emotion (i.e., valence or
hedonic tone). For example, an appraisal that anticipates victory will lead to pleasant
emotions. These emotions may result in facilitative pleasant feelings (motivation), or
inhibitory pleasant feelings (complacency). The third type of primary appraisal is the
“type of ego-involvement.” This is sometimes referred to as goal content (Lazarus, 1991).
Type of ego-involvement determines the more precise nature of the emotional state
(Lazarus, 2000a). Anger for instance, relates to ego-identity; shame is related to ego-
ideal; and guilt is a function of moral value.

Secondary appraisal. Just as personal relevance is an important basis for determining
emotional affect, so too is consideration of possible options (i.e., coping) related to the
challenge. This evaluation includes three factors that are comparable to Weiner’s
(Weiner, 1985) attributional model: Is my relational status attributed to personal or
external causes; is my status controllable; and is my status likely to remain stable?
Secondary appraisal, then, is deﬁned as assessments of coping options and prospects. The

ﬁrst of three secondary appraisals is evaluation of “blame and credit.” This is essentially

33

an attributional consideration. While Lazarus focuses on control (controllable /
uncontrollable) and causality (internal / external), interpretation of environmental
stability may also be relevant (Weiner, 1985). Attributing to blame or credit will, for
example, distinguish between feelings of anger and shame, respectively.

The next secondary appraisal involves “coping potential” as a prediction of how the
situation may change, for better or worse, because of the person’s ability to manage the
demands. There is a Tibetan saying that, ‘if you can ﬁx a problem, there is no need to
worry about it.’ A person will not feel extreme negative states as often if there is a great
deal of self-efﬁcacy and good coping skills (Lazarus, 2000b).

Finally, in addition to the ability to cope, “expectations of the future” will inﬂuence
secondary emotional appraisal. Expectations of the future include consideration of what
we think will change and whether or not we expect the change to be beneﬁcial or
harmful. This is similar to, but slightly different from, coping potential. Although both
involve an anticipated change in the current relational state, expectations of the future
involve a variety of sources for change, whereas coping potential involves more internal
and controllable factors. Ultimately, primary and secondary appraisals are driven by
motivational objectives, and core relational themes between (perceived) current and ideal
states (Lazarus, 1991; 2000b).

Coping. Effective adaptational coping requires the integration of motor, attentional,
motivational, and energy resources. There are two types of coping: emotion-focused and
problem-focused (M. V. Jones, 2003; Lazarus, 1991; Lazarus & Folkman, 1984). Coping
in sport can be emotion-focused, such as relaxed breathing before a free-throw (probably

an adaptive strategy), or yelling at a disgruntled fan (likely a maladaptive strategy) (M.

34

V. Jones, 2003; Lazarus, 2000a). Coping can also be problem-focused, and it is in this
instance that the adaptational potential of emotions in sport, and the importance of this
topic in sport psychology, are revealed (M. V. Jones, 2003; Lazarus, 2000a). There is an
old saying, ‘don’t get mad, get even.’ When done in a manner that is productive and
sportsman-like, this is sage advice. Rather than managing an emotion directly, we can
control valence and intensity by changing relational appraisal—this is the essence of
problem-focused coping. An athlete may say, “I am losing, but rather than quitting or
focusing on my frustration and sadness, let’s start winning!” But how? If the athlete
makes an appropriate decision (secondary appraisal) to work harder, smarter, focus on a
new strategy, relax, or focus on better cues, there is a great opportunity to positively
change both the level of performance (relational component) and resultant emotional
state. In this way, emotions and coping can help athletes reach goals and perform better
(Botterill & Brown, 2002; Lazarus, 2000a).
Summary of the Relationship between Feelings and Performance

One may ask, what is the point of considering all of these technical and semantic
items (e.g., levels of appraisal, dimensions, categories, subconscious processes,
motivational drives, affect, mood, etc)? In addition to providing a more complete
understanding of emotion, these descriptions illustrate the complexity and idiosyncratic
nature of feelings. To understand the relationship between affect and performance, we
must integrate the complexities of emotion and affect with the variability of performance.

In addition to being idiosyncratic and complex, emotions are adaptive, and to be
adaptive, they have to integrate a variety of actionable modalities (i.e., cognitive

appraisal, motor systems, attentional systems, motivation, energy production/activation,

35

and coping strategies). Motivation and activation (energy production) are critical
components of emotional processes, and important factors in performance. Because of
the role of motivation and activation, emotions help us to reach goals and to create
positive relational interpretations of beneﬁt/threat.

Emotions are not an unnecessary form of “baggage” that interferes with performance.
Emotions are vital and adaptive processes that integrate perception, motivation, action
tendencies, and coping behaviors. The only way to avoid emotions is: to not care (have
no stake); to not evaluate the meaning of things around us (have no thoughts); and to
believe every challenge is completely and effortlessly manageable. Obviously this is not
possible, so as long as there is competition in sport (be it task- or ego-oriented), there will
be emotion. Emotions, and the cognitive and behavioral components that they affect,
greatly inﬂuence functioning and performance. It is important to know why and how
emotions facilitate sport performance. However, before we can determine why and how
emotions (and other dimensions of psychobiosocial form) effect human functioning, we
must determine what emotions facilitate and inhibit functioning. To do this, we must
understand the relationship between affect and performance across time, context, age,
sport type, task type, and group dynamic.

Hanin’s Individual Zones of Optimal Functioning

Space limitations prevent an exhaustive review of all anxiety/arousal/emotion models

that explain the relationship between these states and performance. As such, the current

section focuses on the IZOF model, precluded by a brief description of its development.

36

Evolution of The “Zones” Model

Idiosyncrasies. As its name infers, the Individual Zones of Optimal Functioning
(IZOF) theory is an idiosyncratic model that describes the range and combination of
emotional (and other psychobiosocial) items that facilitate optimal functioning (and
dysfunction) (Hanin, 1995, 1997, 2000c, 2003, 2004; Robazza, 2006). The modern IZOF
model initially began as the Zone of Optimal Functioning (ZOF) model (Hanin, 1978).
The ZOF model was divergent from mainstream perspectives such as the Inverted-U
Theory (See Landers, 1980; Yerkes & Dodson, 1908), Multidimensional Anxiety model
(Martens, et al., 1990), Cusp Catastrophe theory (Hardy, 1990; Hardy & Frazey, 1987,
June), and other nomothetic anxiety-perforrnance theories because it postulated that
optimal anxiety was individualized and highly variable—not standardized at a “mid-
point” level (Hanin, 1978, 1986).

Zones. Where popular models described the optimal anxiety level as a single point,
Hanin postulated that there is actually a zone (a bandwidth of one standard deviation). It
is because of this feature that Hanin called his theory a “zones” model of functioning. As
the theory evolved, Hanin (1995) recognized that there is individual variability in both
the width of the zone and the intensity level of the zone’s midpoint. In other words, some
people may have a large optimal zone, others a very small bandwidth; some zones may
be located at a moderate intensity level, while others may be very high or low.

Multidimensional interactions. Multidimensional proﬁling uses several types of
feelings to predict performance (e. g., conﬁdence, anxiety, relaxation, anger, etc.) rather
than a single feeling, such as arousal. Multidimensional proﬁling is more complex, so it

provides a more complete understanding of feeling-performance relationships.

37

Unidimensional proﬁling is akin to assess how good a tennis player a woman is based on
her backhand—without assessing her forehand, volleys, or serve.

Contrary to popular criticism, Hanin’s early model did account for the
multidimensional nature of anxiety, as the State-Trait Anxiety Inventory (STAI;
Spielberger, 1988b) is a multidimensional measure of anxiety (Hanin, 2000c, 2007;
Raglin & Hanin, 2000). As Hanin (Hanin, 2000c, p. 74) explained, “that the STAI uses a
single combined score of state anxiety. ..1ed to the mistaken view that the scale was
unidimensional. However, one could argue that the combined score is an index of the
interaction effects of [emotion].” Also, after comparing the STAI and CSAI-2 across
three methods of IZOF proﬁles (i.e., recall, direct, and graphic methods), Pons, Balaguer,
and Garcia-Merita (2001, p. 9) concluded that the consistency between the two measures
“does not support the widely held view that sport speciﬁc psychological measures are
more useful than general measures.” However, because the STAI combines each
dimension and does not have independent subscales, Hanin’s early work did not examine
the separate effects and explicit interactions of cognitive and somatic anxiety.

Emotions in sport. In retrospect, the discussion over multidimensional interactions
between somatic anxiety, cognitive anxiety, and conﬁdence (Hardy & F razey, 1987,
June; G. Jones, Swain, & Hardy, 1993; Landers, 1980; Martens, et al., 1990; Woodman
& Hardy, 2001) was only scratching the surface of the real issue in affect-performance
research. The two most important developments in the study of anxiety-perforrnance
relationships are the (aforementioned) acknowledgement of idiosyncratic differences and
the realization that there is more to the equation than anxiety or arousal. Hanin (1993b,

2000a) recognized that researchers and practitioners must open the scope of their interest

38

to include a complete list of emotions (i.e., anxiety, anger, etc.). Not only must we
consider a variety of emotions (Botterill & Brown, 2002; Lazarus, 2000a, 2000b), but we
should explore the interactions of different emotions and feeling states (Hanin, 2000a;
Hardy, et al., 1996). These interactions are commonly referred to as emotional
constellations (Hanin, 2007) or recipes (Gould, Flett, & Bean, 2009; Hardy, et al., 1996).

The turning point for Hanin’s original ZOF theory came shortly after Gould and
Tuffey’s (1996) review, when Hanin published his seminal 1997 article. This 1997 article
transitions into the Individual Zones of Optimal Functioning model. The IZOF model is
more theory-based than its predecessor. It substantially expands the multidirnensionality
trend to include an inﬁnite number of discrete, idiosyncratic emotions.

Critique of Hanin’s ZOF. In 1996 Gould and Tuffey reviewed the literature on the
ZOF model and found general support for the ZOF model, but recommended more
precise methods and a stronger theoretical foundation.

Support. Although the explained variance in many ZOF studies was quite low, and
there are some exceptions to the “in/out of zone” principle, Gould and Tuffey concluded
that there was “fairly good” (p. 59) support for Hanin’s (1995) ZOF model. The authors
commended the practicality and intuitive appeal of the model. While the initial research
on ZOF was supported, Gould and Tuffey emphasized the need for more and better
research, and did so with a series of methodological recommendations described below.

Recommended methods. Gould and colleagues (Gould & Krane, 1992; Gould &
Tuffey, 1996) emphasized the need for intraindividual assessment of anxiety over time—
rather than interindividual comparison to group standards. While one carmot assume that

median intraindividual scores represent moderate anxiety, this method is, overall, more

39

appropriate than nomothetic comparisons. Another methodological criticism concerns the
use of recall methods that are potentially biased by outcomes, mood at the time of recall,
and forgetting. Though imperfect, the recall method is ecologically valid and pragmatic
(Hanin & Syrja, 1996; Jokela & Hanin, 1999; Tenebaurn & Elran, 2003; Tenenbaum,
Loyde, Pretty, & Hanin, 2002). Performance measures are also subject to the same
challenges. The key is to use the most appropriate method based on the research question,
and to be as vigilant and methodologically systematic as possible. For some of these
issues, there is no perfect solution. However, Gould and Tuffey proposed two simple
recommendations: larger sample sizes and sustained repeated measures assessment.

Recommended theory. Gould and Tuffey (1996) described Hanin’s ZOF model as
“an empirical ﬁnding in need of a theory” (p. 59). Ultimately, future research should
address two questions: why do zones of functioning exist; and how, ﬁ'om a mechanistic
perspective, do ZOFs inﬂuence performance? Multidimensional studies of affect items
and their interactions may help to answer these questions by revealing more about the
affect-performance relationship. Do ZOFs change over time? If so, then understanding
how and why they change may contribute to our understanding of why and how feelings
inﬂuence performance. Another way to better understand the theoretical underpinnings of
ZOFs may be to look at the generalizability of individual zones across contexts, sport
tasks, and non-sport domains.
The Modern Individual Zones of Optimal Functioning Model

The ZOF model was revised to the current Individual Zones of Optimal Functioning
Model over the course of three decisive publications (Hanin, 1997, 2000a, 2003). Over

these revisions, Hanin and his colleagues made several improvements and addressed each

40

of Gould and Tuffey’s (1996) critiques. The IZOF model emphasizes individual
differences, extends the model to include emotions and other psychobiosocial states,
reinforced the reﬂective method, and includes several different methods of determining
individual proﬁles (Hanin, 1997, 2000d, 2003). Hanin also identiﬁed two theoretical
bases: the mobilization-utilization theory and the application of a dynamical systems
model to organize psychobiosocial components of the model (Hanin, 2000a). Each of the
reﬁnements of the IZOF models are described in the proceeding section.

Defining characteristics of IZOF. The following section deﬁnes each component of
the acronym IZOF. In short, it is an idiographic feeling-performance model comprised of
(intensity) zones that represent various levels of functioning.

Individual. IZOF is in contrast with nomothetic models that ignore individual
differences in favor of general group trends. The individual aspect of the theory means
that each athlete’s optional state of functioning is associated with a unique combination
of emotions and intensities of emotions. The emotions that help John might not help, or
may even hurt, Paul’s performances. Likewise, even if Ringo and George had similar
critical emotions, Ringo might need to be moderately anxious, moderately conﬁdent, and
very focused, while George has to be very anxious, extremely conﬁdent, and moderately
focused. The idiosyncratic nature of the theory has overwhelming empirical support and
practical implications and is one of the most important contributions Hanin has made to
Sport Psychology (For reviews, see Hanin, 2000a, 2007; Robazza, 2006; Tenenbaum, et
al., 2008). Although the model favors an idiographic representation of emotion and

performance, “the model also attempts to generalize data across individuals, teams, and

41

[the] larger group” (Hanin, 2000c, p. 66). Such generalizations help to make the model
more practical.

lanes. “The zones principle implies a speciﬁc relationship between the perceived
intensity of optimal and dysfunctional emotional states and the quality of performance”
(Hanin, 20000, p. 66). The basic zone concept was described in the previous section and
did not vary during the evolution towards the IZOF model. The relevance of the zones
concept is elaborated upon in subsequent sections.

Optimal. Optimal feeling states are those that are most important for a particular
athlete in a particular context. Hanin uses the term optimal to differentiate his theory
from the ﬂow construct (Csikszentrnihalyi, 1994) and from ideal performance states
(Loehr, 1994; Orlick, 1986, 2008). These constructs focus on very exceptional
experiences and only on the effects of positive feelings. IZOF accounts for the helpful
and harmful effects of both positive and negative hedonic tone, as well as on high-
quality, but not necessarily extreme peak performances.

Function. One of the problems with early research in this area was the use of
outcomes as a dependant variable. Hanin has always focused on process measures of
performance because they are more controllable and more closely related to emotions as
predictors. As such, while outcome measures can be included, IZOF studies should use
performance criterion as the dependant variable. Performance measures vary depending
on sport, and there are many ways to assess performance. The simplest is to have athletes
rate their performance on various liken-type scales. Neurophysiological and

biomechanical measures of motor control are other examples of performance measures.

42

Theoretical underpinnings: Psychobiosocial dynamical model. In response to
Gould and Tuffey’s criticism, Hanin (2000) transformed the ZOF model into a more
complex and theoretically justiﬁed model. Whereas the ZOF model was a simple anxiety
theory, the IZOF model looks at multidimensional feelings and interactive .
psychobiosocial (PBS) states. Hanin’s current theoretical approach is inspired by
dynamical systems models, and in particular, the work of Ganzen (1984). The PBS model
incorporates three structural dimensions (intensity, form, and content) and two dynamic
dimensions (time and context) (Hanin, 2000; Robazza, 2006). Structural elements are
more person-oriented and dynamics are environment-oriented, making this an
interactionist-model. The ﬁve dimensions of the PBS model are described in the
following paragraphs.

Content. The qualitative, or informational, characteristics of the IZOF model are
described in the content dimension. The content dimension describes the functional
(facilitative/debilitative), hedonic (positive/negative), performance (optimal/suboptimal),
and task (task relevant/irrelevant) attributes of the model. The common example of how
this dimension is applied in the IZOF model is the distinction between positive-helpful,
positive-harmful, negative-helpful, and negative-harmful categories of emotions.

Form. Form is the foundation of the PBS-IZOF theory. The form dimension
organizes PBS feeling states into seven ‘psycho’ (cognitive, motivational, and
emotional/affect), ‘bio’ (bodily-somatic and movement/motor), and ‘social’
(performance, inter-personal, and communication) elements. Within each category of the
form dimension are discrete, descriptive items such as happy, anxious, focused, hard

working, teamwork, tense, etc. IZOF research has consistently examined

43

multidimensional emotions, rather than simple anxiety or arousal. IZOF is the only
performance model in sport that evaluates emotion, rather than basic anxiety or arousal.
Researchers have recently studied other “forms” of the PBS model, such as motivation
and bodily-somatic (Bortoli & Robazza, 2002; Bortoli, Robazza, & Nougier, 1997; Hanin
& Starnbulova, 2002; Robazza & Bortoli, 2003; Robazza, Bortoli, & Hanin, 2004; Ruiz
& Hanin, 2004). However, the other four forms have not been researched at this point.

Intensity. Intensity quantitatively describes the level, range, or zone of the form. The
intensity dimension describes a person’s effort, energy mobilizing (demobilizing), and
energy organizing (disorganizing) characteristics. Intensity is often determined using
Borg’s (1998) CR-10 Scale, which is described in detail in the third chapter. Most
importantly perhaps, the intensity dimension is essential to the in-out of zone concept—
which is also described later in this review.

Time. The time qualities of the PBS state have most often described experiences
before, during, or after competition. However, the time dimension may also include the
duration of the experience, time of description (i.e., predicted, actual, or recalled
experiences), the frequency of an experience, or the date of a performance within the
calendar year or periodized stage. While longitudinal studies have been more frequent in
recent IAPZ research, the time dimension remains relatively unstudied.

Context. Context refers to situational, interpersonal, inter-/intra-group, and cultural
considerations. The most common contextual distinction is between practice and
competition. Context is a drastically understudied yet promising area of research. For
instance, do athletes have different proﬁles for ‘regular’ games compared to ‘big’ games

and playoffs? Is a basketball player’s optimal recipe the same while defending as while

44

shooting free throws? Hanin (2002) theorized about some of the temporal and contextual
differences that might emerge in individual proﬁles, but little-to-no research has been
conducted in these areas. Comparing the structural and the dynamic components of the
PBS model, the dynamic elements—time and context—are by far the least researched.
Among other objectives, this dissertation examines the dynamic context dimension.
Theoretical Predictions of the IZOF Model

In-out of zone prediction. The in-out of zone hypothesis states that an athlete is
more likely to perform at a high level if she or he is within the idiographic zone of
optimal functioning. While the precise description of a zone is individually unique, the
in-out prediction is nomothetic, because it can be applied to all athletes. In line with
Gould and Tuffey (1996), a meta-analysis of ZOF research performed by Jokela and
Hanin (1999) revealed moderate effect-sizes supporting the “in-out of zone” hypothesis.
Subsequent IZOF research has generally had higher effect-sizes because of theoretical
changes to the model and better methodology. In particular, the modern IZOF model
includes two zones: a zone of optimal functioning (ZOF), and of dysfunction (ZDy). As
such, the in-out principle typically compares best and worst performance states.

According to this hypothesis, being ‘in-zone’ does not guarantee optimal
performance: it simply increases the probability of optimal performance. However,
Hanin’s model does not provide speciﬁc probabilistic estimates of performance success.
Another concern with the in-out principle is that it fails to account for a complete
working emotional range and a complete range of performances. The in-out principle

only accounts for intensity levels associated with best and worst performances: intensities

45

that lie outside of the optimal and dysfunctional zones are not accounted for and
moderate performances are not recognized.

Energy mobilization-utilization hypothesis. In addition to the PBS model and the
in-out of zone hypothesis, Hanin developed the energy mobilization-utilization
hypothesis to explain and predict how feelings determine performance and functioning.
Compared to some emotion-performance theories (Beilock & Carr, 2001; Easterbrook,
1959; Eysenck & Calvo, 1992; Hardy, 1990; Hardy & F razey, 1987; G. Jones, et al.,
1993; Wegner, 1994) the mobilization-utilization hypothesis is not a particularly
sophisticated model (for reviews see, Hanin, 2007; Hardy, et al., 1996; Janelle, 2002;
Williams, Davids, & Williams, 1999; Woodman & Hardy, 2001). While the hypothesis
has roots in earlier publications, Hanin described this hypothesis in the seminal 1997
article and did not fully articulate the hypothesis until 2000. The mobilization-utilization
hypothesis has not been researched beyond the scope of the IZOF model, has never been
tested directly, and its description is limited to three pages of Hanin’s 395-page book
“Emotions in Sport” (Human Kinetics, 2000). However, the hypothesis is effective,
intuitive, and based on a thorough review of relevant research.

The basic premise is that emotions regulate the mobilization and the organization of
energy—which is implicitly supported throughout mainstream psychology. One of the
foundations of Hanin’s hypothesis is Martens’ (1987) description of “psychic energy”—
which is synonymous with vigor, force, intensity of effort, and persistence (See Hanin,
20000, p. 84-86). In this sense, emotion is both energy and the direction of energy.
According to Hanin’s hypothesis, emotions have two functions: energizing and

organizing (Hanin, 2000c). Energizing mutually involves the mobilization and

46

demobilization of resources. Organizing involves efficient utilization (or misuse)—a
concept that shares an intriguing connection with Csikszentmihalyi’s (1994, 1996)
description of ‘psychic entropy’ in relation to optimal experience and mental functioning.
So how does the connection between emotion and energy determine performance? In
sport, optimal performance requires (a) enough energy for substantial physical exertion
and mental stamina, and (b) efﬁcient use of psychomotor resources. In short, emotions,
and more generally, feelings, help us to meet the energy and resource demands of sport.

How do speciﬁc emotions inﬂuence energy mobilization and utilization? The key to
Hanin’s hypothesis is the speciﬁc description of the effects that emotional valence and
function have on energy mobilization and utilization. Speciﬁcally, IZOF categories are
associated with the following functional energy effects. Positive-functional emotions
serve to mobilize and effectively organize energy. Negative-functional emotions assist
more in mobilizing than in utilizing energy. Positive-dysfunctional emotions reduce
effort and energy generation, as well reducing efficient utilization via inefﬁcient
processing and low alertness. Negative-dysfunctional emotions distract cognitive and
motor resources away from relevant and towards irrelevant cues. Clearly, the magnitude
of the effects of each category depends upon the intensity of each content item (emotion)
in the category. For instance, the motivational effects of negative hedonic tone may be
either facilitative or debilitative. This implication is supported by the cue-utilization and
processing efﬁciency theories (Easterbrook, 1959; Eysenck & Calvo, 1992).

According to the hypothesis, functional emotions are associated with the anticipation
of challenge, threat, or opportunity (a process focus). Dysfunctional emotions anticipate

positive or negative outcomes. Consequently, dysfunction emotions have a negative

47

effect because the athlete becomes a less active agent in determining personal
performance. Whether or not an athlete anticipates winning or losing, the effect is the
same: less effort and a more outcome- or ego-oriented focus. According to Hanin the
functionality of emotion and energy can describe two personality types, which are in line
with Eysenck and Calvo’s (1992) processing efﬁciency theory. “Smart users of energy”
do not require high activation levels (Hanin, 2000c, p. 86). Contrary to these low-energy
performers, high anxious athletes mobilize more energy resources and tend to be less
efﬁcient due to the strain placed on processing and attentional resources.

An important implication of this theory—an implication that is highly relevant to this
dissertation—is that the organization of feelings into functional and dysfunctional, and
positive and negative categories is essential in order to understand exactly how an
athlete’s feelings affect his or her performance. As noted earlier in this section, the
Energy Mobilization-Utilization Hypothesis is conceptually and intuitively strong, but
lacks formal empirical support. Though, it is beyond the purpose of this dissertation to
test the hypothesis, the organization of discrete feelings into valence-ﬁmction categories
is foundational as it allows comparisons between otherwise unique individual proﬁles. In
other words, let us say that Scott’s proﬁle has four items (intense, overjoyed, anxious,
and angry); and Shawn’s has conﬁdent, calm, unsure, and scared. Clearly, these unique
items cannot be compared. However, if each represented a different valence-function
category (e.g., positive-helpful, positive-harmful, negative-helpful, and negative-
harmful), comparisons could be made as both players have feelings that may be
organized within those valence-function categories. This is especially reliable if multiple

discrete items represent each category.

48

Methods and Measures of Performance Profiling

The following section examines two major topics in the area of assessing individual
proﬁles. The ﬁrst part will describe four established methods for determining individual
emotional recipes for performance excellence. The second part describes the positive-
negative affect scale, which is central to the assessment of individual proﬁles. The third
part reviews literature that has examined the reliability of emotional recall methods.

Methods of developing idiographic profiles. Several methods have been developed
to determine the exact recipes of feelings that most help and hurt a player perform. While
the method section will describe the techniques most relevant to this study, the
subsequent sections are intended to provide an overview of the options available to
researchers and practitioners alike. Hanin (2003) illustrated four methods: in-depth
interviews, individualized emotion proﬁling (Hanin’s stepwise procedures), self-
generated metaphors, and narrative descriptions. In addition to these relatively qualitative
approaches, many researchers have simply used psychometric questionnaires such as the
STAI, SAS, POMS, and CSAI-2 in either a recall or a real-time pregame protocol.
Though most researchers have focused primarily on precompetitive anxiety, these
methods may be used to assess pre-, mid-, or post-competition states. The proposed study
will primarily utilize the stepwise procedure (IEP) method—which will be augmented
with elements of interview and narrative procedures.

In-Depth interviews. The ﬁrst method of collecting individual data is the use of in-
depth interviews that include open-questions and questionnaires. One of the best-known
semi-structured questionnaires is Terry Orlick’s (1986) “competitive reﬂections” survey,

but psychometric tools may also be used to supplement interview data. Athletes can recall

49

their performance state at anytime of the competition, so this method is not restricted to,
for instance, pre-competition assessments only. As the most open-ended technique,
interviewing is very ﬂexible and sensitive to individual differences. Interviewees have the
opportunity to describe personally meaningful events and relevant feelings. However, this
method relies on the athlete’s ability to remember and to verbalize his/her thoughts.

For this reason, Hanin (2003; 2007) has consistently recommended that interviews be
conducted with experienced and more verbally skilled athletes, and that interviews focus
on more unique and memorable performances (e.g., best and worst ever). In addition to
being able to recall and verbalize their experiences, it is also important to ensure that
interviewees have had sufﬁcient experiences to draw upon—which is a limitation for any
IZOF-related assessment. If an athlete has not experienced a sufﬁcient number of
exceptional performances, how can she describe her ZOF? Other disadvantages of
interviewing include that it can be more time consuming, and that interviews may vary
signiﬁcantly from one athlete to another.

Individualized emotion profiling (IEP). The IEP, which is commonly referred to as
the stepwise procedures, evolved with the development of the modern IZOF model
(Hanin, 1997, 2000c). The IEP is the gold standard and most frequently used method of
IZOF assessment. The ﬁrst step in the IEP is to identify and describe previous best and
worst performances (See Figure 2 in the Appendix). While Hanin does not specify how
many best or worst performances should be included, most studies have included only
one example of each. Though timesaving, the use of only a single best, and a single worst
performance is concerning. Clearly, a single competition cannot describe two-

dimensional zones (i.e., range or bandwidth). The reﬂective stage offers a more concrete

50

description based on speciﬁc performances rather than conceptual or metacognitive
representations.

The second step uses the Positive and Negative Affect scale/list (PNA; Hanin, 1993;
Syrja & Hanin, 1997) to identify emotions and related feeling states that are personally
relevant for each athlete (Figure 3 in the Appendix). The PNA-77 is a list of positive and
negative emotion-oriented words that are associated with best and worst functioning
(Hagtvet & Hanin, 2007; Hanin, 1993b; Hanin & Syrja, 1995; Lukkarila & Hanin, 2001).
Details about the PNA are included in the next section. From the positive list, athletes are
generally instructed to choose 4 or 5 items that are relevant for best performances, and 4
or 5 items associated with worst performances. The same procedure is followed for
negative emotions. The result is an individual proﬁle of 16-20 idiosyncratically important
descriptors. This step allows participants to describe personally meaningful factors, rather
than being restricted by predetermined proﬁle items. Syrya and Hanin (1997) found that
as much as 85% of individually relevant emotion-type items were not included in
common emotion-performance scales.

The third step uses a modiﬁed Borg’s CR-10 scale (Borg, 1998) of perceived exertion
to estimate the intensity dimension of each emotion or related feeling state (Hanin,
2000c; Turmnavuori & Hanin, 2000). Borg’s scale is well-known to all disciplines of
kinesiology. It is best known to exercise scientists as the rate of perceived exertion scale.
The likert scale ranges from 0 (“nothing at all”) to 11 (“maximal possible”) and can be
used to describe physical effort or the intensity of any feeling state. Participants are
required to recall the intensity of each item from the PNA for each competition in their

best/worst list. An example of the CR-10 is provided in Figure 4 in the Appendix.

51

The ﬁnal phase of the IEP requires “visualizing” or graphing the intensity estimates.
The result of this procedure is the formal IZOF proﬁle. An example proﬁle is included in
Figure l. The graph includes four columned sections. Positive items (denoted P) are
included on the right side, and negative items (denoted N) are on the left. Items
associated with optimal functioning (denoted +) are located in the middle-two columns,
with dysfunctional items (denoted -) on the outside colurmrs. The columns therefore run
(from left to right) as N-, N+, P+, and P-. Because functional (+) items are generally
higher in intensity during optimal performances, and dysfunctional items are very mild,
the resulting optirnal-proﬁle ﬁts the traditional inverted-U or iceberg shape. Conversely,
the dysﬁmctional-proﬁle is typically U-shaped or concave.

The stepwise procedures balance the openness necessary to generate idiographic
patterns with more efﬁcient, structured procedures that permit interindividual
comparisons that can generate nomothetic models. However, it is the most time-
consuming method. Also, it is debatable how rigidly one should adhere to the protocols at
each step of the process.

Self-generated metaphors. As metaphors are not included in the assessment
procedures of this study, the description here will be brief. The use of metaphors to
establish performance proﬁle and to enhance performance is based on the Social
Cognitive Theory (Bandura, 1986). Hanin has endorsed the potential value of metaphoric
description (Hanin, 1997, 2000a) and has conducted three metaphor-studies with his
colleagues (Hanin, 1997, 2000a; Hanin & Stambulova, 2002; Hanin, Stambulova,
Lukkarila, & Tummavuori, 1997; Ruiz & Hanin, 2004). Athletes are asked to describe

what it felt like (using a metaphor) prior to, during, and after their best and worst

52

competitions. They are then asked to elaborate on each metaphor with the stem “in other
words, you felt like...” When at their best, athletes might say they feel like: a lion, a calm
breeze, an oak tree, a crouching red panther, etc. Descriptions of dysﬁmctional states may
include Garbage, sloth, a robot, and a black bug. Metaphoric descriptions may not be
effective for all personalities, but provide personally meaningful descriptions for many.

Narrative descriptions. Narrative data are simply stories, or accounts of experiences.
A common example of narrative data is the self-story, which is a “ﬁrst person narrative
which deﬁnes identity based on memories and personal perceptions of history, present
life, roles in various social and personal settings, and relationships” (Payne, 2000, p. 19).
As with previous methods, narrative description is ultimately another means of describing
factors that affect performance. In addition to very guided protocols, narrative
descriptions can also be obtained through joumaling (either during or after competitions).
Measurement Issues

The Positive-Negative Affect (PNA) scale. The PNA is an important component of
the recently described IEP, but can also be combined with many of the other methods.
The PNA was created for studies with elite Finish athletes that used the stepwise
Individualized Emotion Proﬁling system (Hanin, 1993b; Syrja & Hanin, 1997). The
Finish list was developed from Watson and Tellegen’s (1985) 10 global affect scales. The
English version includes 40 positive items and 37 negative items, each arranged in 14
synonymous rows. Examples of positive valence items include active, relaxed, happy,
conﬁdent, exhilarated, thrilled, daring, motivated, pleasant, and fast. Examples of
negative rows include afraid, aggressive, annoyed, worried, depressed, doubtful, unsafe,

sluggish, ﬁerce, sorry, tight, and tired. In addition to the 77 items in the list, participants

53

are urged to include their own words. This allowance enhances the idiosyncratic
capability of the PNA, and it promotes buy-in from athletes. Feelings such as attacking,
pissed off, and tough do not appear on the list, but may be appealing to some athletes.
The PNA-77 is shown in Figure 3 (See Appendix)—please refer to this ﬁgure for
examples of items, and to see how synonyms are grouped in each row. The PNA has been
translated into Spanish, Italian, German, and Russian (Hanin, 2003).

Syrja and Hanin (1997) compared the PNA-77 to popular anxiety inventories in sport
and found 85% item overlap with the STAI (Spielberger, 1988a), 60% overlap with the
PANAS, and 37% overlap with the POMS (Morgan, 1980; Morgan, O'Connor,
Ellickson, & Bradley, 1988). While the PNA-77 accounts for a reasonable number of the
items on these scales, the same cannot be said in reverse. Overlap ratings (similarity of
items ranging from 0 to 1.0) revealed that the aforementioned scales fail to account for
between 83% and 89% of the items on the PNA (Syrja & Hanin, 1997). A subsequent
study with hockey players revealed that the CSAI-2 failed to account for 85% of the
items identiﬁed by the PNA (Hanin, 2000d).

Extending Hanin’s list. There is an interesting—albeit unpublished—study that
contributes to our discussion of Hanin’s PNA list (See Gould, et al., Date unknown).
Two-hundred and eleven college athletes from a variety of male and female sports
completed surveys that identiﬁed feeling states that were critical for individual
performance. Affect was identiﬁed across eight categories based on combinations of
three factors: context (practice and competition), time (before and during), and
performance (good and poor). Because the authors used a very broad deﬁnition of

emotion—“those speciﬁc positive and negative feelings, thoughts, physical reactions, and

54

motivational tendencies” (Gould, et al., Date unknown, p. 8)——their results are consistent
with the more holistic psychobiosocial feeling state underlying the PNA-77.

Consistent with Hanin’s claim (Hanin, 2000d; Syrja & Hanin, 1997) that traditional
measures of competitive affect underrepresented athlete’s unique needs, Gould et al.
(Date unknown) identiﬁed 5448 different raw items (almost 26 per participant across the
eight contexts). Formal content analysis of the raw items yielded 65 general categories: of
which only 18 were included among the 77 PNA items. Interestingly, nine of the top ten
categories from Gould et al. are represented in the PNA. Gould et al. identiﬁed several
new items that represent the broader psychobiosocial state and are not all true emotions.

Critique of the PNA/IEP. Hanin (2003) acknowledged two criticisms of the IEP
system and of the PNA-77 list. First, the PNA-77 does not include pure emotions. There
are items that might be better categorized as somatic (e. g., jittery, tense, tired),
motivational (e.g., determined, willing, motivated), and cognitive (e. g., focused, alert).
The PNA-77 could, accordingly, be considered a list of broader, performance-oriented
psychobiosocial factors. As Hanin argued, “there is a clear need to extend the focus of
individual proﬁling ﬁ'om emotions to other modalities of an athlete’s state” (Hanin, 2003,
p. 12). Anyone with experience working with athletes in real-world competitive settings
cannot deny the validity of this broader inclusion of psychobiosocial feeling states.
However, the PNA could be organized into emotions, somatic items, and so forth. Future
research could combine Hanin’s PNA-77 with Robazza et al.’s (2004) somatic item list,
Gould et al.’s (Date unknown) grounded lists of items, and new research to provide a

more comprehensive list that is organized into various types of feeling state.

55

Reversals. One other criticism of the PNA-77 is that valence is predetermined for
each item. For example, anxiety is listed as a negative emotion and daring is presumed to
be positive. Numerous studies have established that negative emotions can be helpful or
harmful to performance (Edwards & Hardy, 1996; Hagtvet & Hanin, 2007; Hanin, 1993b,
1995; Hanin & Syrja, 1995, 1996; G. Jones, 1995; G. Jones, Hanton, & Swain, 1994; G.
Jones, et al., 1993; Robazza, Bortoli, & Nougier, 1998). This ﬁnding is referred to as
either a ﬁmctional reversal or directionality of anxiety. Additionally, one could also argue
that the hedonic tone of an emotion is determined by individual perceptions. If a person
feels that anxiety is pleasant, or that daringness is unpleasant, then the conventional
labeling of anxiety as having negative valence is arguable.

For instance, Robazza and Bortoli (2003) examined hedonic and functional
(directionality) reversals in the CSAI-2, and in somatic and affect PNA scales. Based on
their sample of 374 elite and non-elite Italian athletes, Robazza and Bortoli found
evidence for both functional and hedonic reversals. Speciﬁcally, 72% of the 90 affect and
somatic items included in this study were classiﬁed as both positive and negative.
However, upon closer inspection, emotions conventionally deemed to be positive were
perceived by participant to be positive 94% of the time, and unpleasant less than 6% of
the time. Traditional negative emotions were identiﬁed as negative in 87% of cases and
positive in only 13% of cases. Averaged together, hedonic reversals occurred in slightly
fewer than 10% of the 5608 cases in this study. These results are supported by Bortoli
and Robazza’s (2002) study of 50 Italian volleyball ofﬁcials which incorporated two

PNA-style lists: one list with 64 affect items, and a second list with 24 physiological /

56

somatic items. Along with Robazza et al. (1998) these studies, supported Kerr’s (1997)
description of hedonic reversals.

Hedonic reversals are noteworthy, but in the view of this dissertation, the traditional
assignment of valence to discrete feelings should not change because of a ﬁnding that
occurs only rarely and lacks theoretical explanation. Functional reversals, on the other
hand, have received nearly unequivocal empirical support. As such, these reversals
should always be considered when studying valance and performance.

Validity of recall. One ﬁnal measurement issue that is relevant for this study
involves the validity of performance recall. There have been numerous studies that have
examined the accuracy of both recall and prediction methods of assessing emotion. The
issue of reﬂective assessment procedures is integral to the validity of the IZOF model,
and any review of literature of this kind would be thoughtless to ignore it. However,
because it is not an essential focus of this project, this section will be as brief as possible.
Overall, recalling feeling states before and after performances is a method that is well
supported (Hanin & Syrja, 1996; Jokela & Hanin, 1999; Tenebaum & Elran, 2003;
Tenenbaum, et al., 2002).

Recall methods. Harger and Raglin (1994) assessed state anxiety in 34 college track
and ﬁeld athletes an hour before and two days after competition. The two values
correlated at .97 for females, and .96 for males. Recall was consistent for high (r = .96)
and for low (r = .97) performance ratings. Though not as high, other studies have
supported Harger and Raglin. Hanin and Syrja (1996) had 17 Olympic men’s soccer
players complete individualized precompetition PNA scales 24 hours before competition

(predicted), 40 minutes before competition (actual precompetition), and 30 minutes after

57

competition. Results showed signiﬁcant correlations for 14 of the 17 atlrletes’ predicted-
actual ratings, and for 13 of the 17 athletes’ recalled-actual ratings. While prediction and
recall were not reliable for all participants, the method, overall, was consistent for both
predicting as well as recalling actual precompetition emotions.

More recently, Tenenbaum and Elran (2003) compared precompetitive anxiety
reports 3, 7, and 14 days after competition using the CSAI—2 and the PNA-77. After a
small change in reports from the actual to the three-day recall point, the authors noted
that “correlations among the retrospective reports of emotions at Day 3, 7, and 14 were
extremely high, higher than the ones between actual and the three day delay report
correlations, indicating a high carryover effect from the ﬁrst report procedure”
(Tenenbaum & Elran, 2003, p. 283). On average, reported scores decreased only 2.37%
in the 14-day recall. These results were supported by a similar study of 52 university
athletes (Tenenbaum, et al., 2002). In studies of performance recall taken 3-22 months
after competition, correlations ranged between .75-.82 (Imlay, Carda, Stanbrough, &
O'Connor, 1995; Raglin & Turner, 1993; Turner & Raglin, 1996).

An alternative to recall assessments. Real time assessments can be too invasive
(Hanin, 1995; Hanin & Syrja, 1995; Raglin & Hanin, 2000). Such methods can interfere
with competitive routines and may be biased by the immediate moment. That is to say,
that rather than describing the general state prior to competition, real time assessments
are hyper-focused on a particular moment. Accordingly, researchers would be prudent to
collect a large number of such direct, real time assessments. Also, these direct
assessments are not as feasible or efficient as recall methods. Recall techniques avoid

these potential problems, but can be undermined by outcome bias. For example, a

58

successful outcome might bias recall of both emotions and of performance in a positive
direction (Gould, Tuffey, Udry, & Lochbaum, 1993). Recall is also subject to fading
effects in memory. In spite of these concerns, recall and even predictive methods of
assessing emotion are strongly supported empirically.

Because the proposed dissertation will utilize online, or real time, procedures rather
than recall methods, this section might appear to be irrelevant or even contradictory.
However, it is not my position that real—time assessment is superior or that reﬂection is
inferior. Predicted, real time, and reﬂective techniques are each valid and reliable—so
long as they are executed precisely. The problem with IZOF research up to this point is
not the use of recall designs, but the over-use of this method. The proposed method is
especially signiﬁcant to the body of literature not because it is better, but because it is
most appropriate given the purpose of the study and because it ﬁlls a relative void in the
literature with respect to online assessment.

Critique of IZOF Theory and Literature

While there are several possible criticisms of the IZOF model, our discussion will
focus on the faults that will be examined in the proposed study. These criticisms include
the failure to acknowledge moderate levels of performance, the failure to account for a
complete range of emotional intensity, and related to these two problems, the IZOF
theory does not account for overlap or gaps between optimal and dysfunctional proﬁles.
Other criticisms pertinent to this dissertation are described brieﬂy.

Exclusion of moderate performance. As described in the previous IZOF section,
Hanin’s stepwise procedure produces two zones. When an athlete is within his zone of

optimal functioning (ZOF), the probability of optimal performance is at its highest for

59

10;

mg

that individual. The further “out of zone” the athlete is, the worse their performance is
likely to be. The same principle applies for the zone of dysfunction (ZDy) and poor
performance. Generating proﬁles for only optimal performance and poor performance
leads practitioners and academics to an obvious question: why does the IZOF model
exclude moderate performance? The failure of the IZOF model to consider a fuller range
of sub-optimal performance levels is not trivial. The omission of moderate performance-
levels is fundamental to the gaps and overlaps problems described later in this section.

Incomplete functional range of predictors. In conjunction with the previous
concern that the IZOF model does not account for enough possible outcomes, the model
does not account for the complete range of intensity for each feeling state (aka, each of
the predictors). For example, reﬂecting on previous best and worst performances predicts
an IZOF for ‘aggression’ between 7-9, and an IZDy between 2-4. While intuition allows
us to reasonably assume that any level of aggression below 4 is likely to yield poor
performance, the IZOF system cannot determine what the performance effects are when
aggression is between 5-6 or above 9. If an athlete were to be a 5 or 6 for aggressiveness,
would this result in optimal, average, or poor performance? Likewise, when the athlete
exceeds a 9, is she able to sustain high performance, drop to moderate performance, or
does she experience a catastrophic poor level of functioning? The limited ability to
account for a complete functional intensity range on every predictor item and to describe
all levels of performance is clearly interrelated to one another. These concerns are related
to another, perhaps greater, problem with the IZOF model—gaps and overlaps.

Gaps and Overlaps. As a result of the above limitations, Hanin’s approach can result

in gaps and/or overlaps between ZOF and ZDy proﬁles. Gaps refer to the spaces in the

60

IEC

SIUC

Har

complete range of intensity that is not associated with either best or worst performance
(e.g., the ZOF is 3-5, and the ZDy is 8-10, establishing a gap from 6-7). When there is
overlap, some points on the intensity continuum are associated with both optimal and
dysfunctional performance (e. g., the ZOF is from 5-9 and the ZDy ranges from 0-6,
establishing an overlap from 5-6). These gaps and overlaps between ZOF and ZDy
proﬁles are depicted in Figure 1. Gaps are ambiguous—there is no way to determine their
meaning—but while gaps tell us nothing, overlaps tell us too much. Karnata et al. (2002)
point out that “the larger the overlap, the. less chance of predicting the quality of
performance from emotional states” (p. 193). For example, if the athlete is within the
overlapping range (e. g. 5-6 in the previous example), will they perform at an optimal or
poor level? If the model cannot answer this question, it provides little value for
researchers and practitioners. The IZOF model should resolve overlaps and must be able
to explain the meaning of (if not eliminate) gaps within proﬁles.

Kamata et al. (2002) addressed these shortcomings with their IAPZ model. By
establishing probabilistic estimates, IAPZ’s account for performance effects throughout
the entire range of valence, function, and emotional intensity—thus eliminating
ambiguous gaps between optimal and dysfunctional proﬁles. Additionally, Kamata et al.
recognized average performance outcomes—not just optimal and poor ones.

Brief aside: An anecdotal view on the overlapping phenomenon. Though empirical
studies suggested that overlap between ZOF and ZDy proﬁles was rare (For reviews, see
Hanin, 2000; 2007), the consulting experiences of the author in administering the IZOF
system to hundreds of athletes provided contrary anecdotal evidence. In fact, overlap was

so common that several practical solutions evolved. The solutions to this problem began

61

PFC

lute

with reﬂection and explanation. Why is there overlap? Kamata et al. (2002) made an
excellent point, which is that each level of an emotion is not going to elicit an exclusive
lOO-percent likelihood of poor, moderate, or optimal performance. At any level of
intensity, the probabilities of poor, moderate, and optimal performances likely coexist at
varying ratios. However, this is the only explanation offered by the IAPZ model.

Another reason for the overlap of intensity was that different items interact. The best
way to explain this is with an example. Say there was overlap in anxiety: when Adolf is
between 5-7 for anxiety, he experienced both best and worst performances. While anxiety
does not discriminate performance cleanly, it was an important predictor of performance.
The key in this example was the interaction between anxiety and conﬁdence. Anxiety
levels between 5-7 combined with low conﬁdence predict poor performance; while the
same anxiety levels combined with high conﬁdence create optimal performances. In this
sense, we can see how separate items could become feeling-conjunctions (i.e., conﬁdent-
anxiety or relaxed-nerves). Interacting feelings and the notion of optimal emotional
“recipes” are described in applied literature (Gould, et al., 2008; Hardy, et al., 1996).

An additional explanation for overlapping content is simply that the item does not
effectively predict performance! Upon initial reﬂection of either a best or worst
performance, an athlete might feel that an item (often a positive-harmful feeling such as
calmness or energetic) is important. However, after monitoring several competitions, it
became clear that calmness and feeling energetic are associated with both optimal and
poor performance; these items did not differentiate functioning. The stepwise IZOF
procedure can be considered analogous to backward linear regression in that you may

intentionally start with too many items in an effort to be comprehensive while moving

62

[0W1
pract
ortho

Othe

By o
3ij
the f

yer ;

C0111.

intra

mm“
is no
his h
idem
ThESt

M01

towards a model with only substantive non-redundant predictors. Researchers and
practitioners should not be afraid to eliminate less helpful items. We should never allow
orthodoxy and dogma to obscure our real experiences and special hmnan insight.

Other Issues Addressed in this Study

The following section provides a compilation of smaller criticisms and shortcomings
of the IZOF model. Each ﬂaw is addressed in the design of the current study. Firstly,
IZOF research has relied too heavily on recall of precompetition emotions. To balance
this preference, direct assessments of emotion during competition will be acquired.

Secondly, most IZOF studies include relatively small numbers of total observations.
By obtaining a larger number of repeated observations, more sophisticated statistical
analyses can be conducted that account for more predictors and interrelated factors. Of
the ﬁve dimensions of the PBS state, the context dimension is the most under-researched,
yet perhaps the most intriguing. The proposed study involves tennis, which can be broken
into two unique performing situations: returning serve and serving (where a player
controls the point and has an advantage). By obtaining a large number of observations,
intraindividual differences in proﬁles can be compared in each context.

Lastly, from an applied point of view, IZOF proﬁles are too big. It is important to be
comprehensive, but great scientiﬁc models must have a parsimonious balance. It simply
is not practical for an athlete to go into a competition with 16 or 20 emotional zones in
his head. In an attempt to bring simplicity to the IZOF model, athletes will be asked to
identify the two to four emotions that they feel are most important for their performance.
These critical items will be compared to the complete model to determine which predicts

Performance more reliably.

63

Ext

n8\l€
ﬁnpn

Anll

Klan
1003
IAPZ
toco
indu
new
Stud}
d.(7

h

1201

mode
ﬁonr
TEner

Pﬁsen

Extending the IZOF Model: Individual Affect-Related Performance Zones (IAPZ)

As explained in the introduction, this dissertation integrates the IZOF model with a
newer IAPZ model. The following section describes the IAPZ model; ways in which it
improves the IZOF model; and addresses shortcomings of the IAPZ method itself. .

An Introduction to Individual Affect Performance Zones (IAPZ)

The IAPZ model was established in 2002 after collaboration between Akihito
Kamata, IZOF creator Yuri Hanin, and Gershon Tenenbaum who has authored every
published IAPZ study (Cohen, et al., 2006; Edmonds, etal., 2006; Golden, Tenenbaum,
& Kamata, 2004; Johnson, Edmonds, Kamata, & Tenebaum, 2009; Johnson, Edmonds,
Moraes, et al., 2007; Johnson, Edmonds, Tenenbaum, & Kamata, 2007; Kamata, et al.,
2002; Medeiros Filho, et al., 2008; Tenebaum, et al., 2002; Tenenbaum, et al., 2008).
IAPZ is not a completely original theory. It is a reﬁnement of the IZOF model intended
to correct speciﬁc limitations of its predecessor. Research on this idiographic model
includes seven manuscripts and one conference proceeding. Of the nine manuscripts there
are: ﬁve empirical ﬁeld studies (two tennis, one golf, and two archery); one laboratory
study (racing video game); two review article; and the preliminary article by Kamata et
al. (2002), which introduces the probabilistic statistical procedure and compares IAPZ to
IZOF using two hypothetical cases. Each of these papers is described in Table 1.

IAPZ research is at a pivotal juncture in that there is enough support to make the
model credible; yet there is not enough research to identify IAPZ as a distinct theory
from the IZOF model. IAPZ and IZOF are intimately related systems. As such,
Tenenbaum’s model will be contrasted with Hanin’s IZOF theory, more so than it will be

presented as a stand-alone model in the forthcoming section.

64

.303 now... 5 2333030 mo 39:50 :38 05 38292 05200 .295? .20 38285 E 33.; .202

 

 

 

 

 

 

 

 

 

£9an 000$. 889$“. xoﬁsoo<
0% 590-0% ”common 5.5 30%.» 0833.0. wﬁuoonm H maggotom 8ND 38% 8:35:09
owﬁﬁo scuba “x8000 xaoﬁmmommn 00:00 am 65 8&2 rogue m 0% .3802 .055 8:032
95500895 ouggotom 300820 38:03.. Qoomv 8:35:02.
9:50 mowﬁﬁo .0509; 088mg 3:00am n ooﬁgotom :32 0% .0£E 0.00682 .3802
$808020 08300 SSE 30083330. 0315 mm: 6:0 8&2. boned... _ £00083 .0853.
EBSEoEV oocwquoﬁom Eo>0ow§ao wage a; n ooﬁgotom W50: Q83 SEEM 0% .833
9:50 mowﬁﬁo “coammommm 25:0 625 80% $55.2. v 0000;. .mccoﬁum £00.53.
ﬂaoﬁmmowwa 208388 20008538 08:. u maggotom 250 688
Ecoﬁcogwtge :0 manages; ”OH 0065882 82> 8m: 0:003 0% Essa—:05...
m0 033/ ”one: :03: 00 Z 35833.0.“ 25:0 83m HE EEO 80% wamoum m .582 .munoﬁwm
AME .mb ooggoﬁom “coﬁmmomﬂx 25:0 030 200m
:0 Hmm mocha Houstowﬁamoﬁ 02.2602 n ooggotom
$8: ”005.5808 .3 0.0 maowaoaam m«2.2 m$00 GE 38% imam
00 30:08.38 m0 Bootm 603003302 Hmm 60:0:an Bro 6&2. .200 N 0% .Eswnocoh .5200
32% 080m a055=§w $6883 ooggotoa a; u ooggotom
\ moon—2020 2.05 ”macaw—m 880008 00m $03808 m<an mmmm ENE G088 Suﬁsm
520-2% 80:00.35 Home :88 2508C. ”00:05:“ ”EH0 Hoots... mach m 0% .Esanaouoh .5200
008. ES 2088 ES m are usage 9% so: 388 55:
05 82609: 205.8 "EC. 3033 30306.09sz wﬁms 8mm Eouoﬁogm 0002 N 0% .Esgcocoh .Smﬁam
£332 £3910. \ Hyman 85932 Mad. 295m @036 £053..

 

 

 

 

 

 

 

oaoN ooSﬁEptom cows—oMéooE E22305 m0 «can 23 5 0.205 Rowan—m m0 began

_ 23

65

SOFT.

(301
incl
cor

1’01.

hi

hi

ICE

Pit

C0!

Dre

An example research study. However, before we explore these issues, and because
some may not yet have read an IAPZ study, a typical study is described. Johnson et al.
(2007) conducted a single-case study of an international female archer. The study
included numerous observations over the course of one indoor and one outdoor
competition. The competitions included ﬁve different shooting distances, with several
rounds at each distance, and several shots in each round. There were 348 observations in
this study. After each series of shots, the participant completed the affect grid (Russell, et
al., 1989) which assessed general affect and arousal (or activation), and heart rate values
were aggregated across the series. Functionality of affect was not rated.

Performance scores were based on accuracy of shooting in each series. Shooting
scores one-half standard deviation above the mean were labeled as optimal, scores one-
half standard deviation below the mean were labeled as poor, and all other scores
reﬂected moderate performance levels. In accordance with the inverted-U theory (Yerkes
& Dodson, 1908), poor and moderate performances were both split into two groups based
on whether or not each affect score was above or below the average value associated with
optimal performances. This resulted in ﬁve types of performance: poor-high affect, poor-
low affect, moderate-high affect, moderate-low affect, and optimal.

In keeping with Kamata et al.’s (2002) probabilistic method, ordinal logistic
regression was conducted separately at each competitive distance for each of the three
predictor variables (i.e., heart rate, arousal, and pleasure). None of the predictors were
combined or enabled to interact during the analysis. After the 15 graphs were developed,
predictor scores were essentially recoded into one of the ﬁve possible performance labels,

and placed in chronological order. This secondary analysis allowed the researchers to

66

Show
how r
acros.

“on

(1.6..
desc
char

stud

Bqua
use lc
Obser
comp
ofthe
and pc

PIOflle

show what zone the archer was in from one series of shots to the next. This ﬁgure showed
how often and when she was in-zone. The authors were thereby able to compare proﬁles
across distances (context analysis), and to provide a unique perspective on momentum.
How IAPZ Improves the IZOF Model

Because the IAPZ model is a relatively young model, and a comprehensive outline
(i.e., book chapter or similar forum) has not been presented, it is difﬁcult to deﬁne and
describe characteristics of the IAPZ model. It is especially difficult to separate
characteristics of IAPZ from the IZOF model. However, based on the ﬁve empirical
studies conducted thus far, the following characteristics of the theory are available.

Probabilistic estimates. The deﬁning feature of the IAPZ model is the use of
probabilistic estimates of performance outcomes using logistic regression. Where the
IZOF model conﬁnes itself to the intensity levels associated with two possible
performance states (optimal and dysfunctional zones), IAPZ accounted for complete
ranges of emotional intensity and three performance levels (poor, moderate, and optimal).
A standard IAPZ graph is presented in Figure 5 (note that at any point along the intensity
range (x-axis), the sum of the performance curves’ probabilities (y-axis) each will always
equal 1.0). Kamata et al. (2002) presented three different models (approaches) that each
use logistic regression to identify optimal zones. Two sets of 50 hypothetical
observations are presented so that probabilistic and traditional IZOF analyses may be
compared. The results of these comparisons supported the probability-based method. One
Of the most interesting results compared how often observations associated with optimal
and poor performances were correctly in- or out-of-zone (respectively). IZOF generated

Proﬁles were too narrow, and thus excluded too many optimal performances from being

67

includu

1181115 1

mislat

Flglll:

Exam

N0
(re
pl)

al'

included in-zone. In other words, the IAPZ method more accurately identiﬁed helpful
items as being associated with optimal performance. Likewise, the IAPZ method

mislabels out-of-zone observations less frequently.

Figure 5

Example of a Typical IAPZ Graph of Probabilistic Curves

 

Probabilistic Estimates of Performance by Intensity of Feeling

PIX=1) P(X=Zl

 

P(x=3) — — P(X=4J — ' -P(X=5)

 

 

 

H

Probability
0.099.053
0010\qu

99.0.
b—INOO

O

 

0

0.5
1
15
2
25
3

Q'LQLnLQOLQhLQGJI-QONLQO
si- L0 \o l\ 00 axi"

Ln H
«5 r-t

10.5

Intensity of Feeling

 

 

 

Note. This ﬁgure depicts the probability (y-axis) of various performance levels
(represented by different curves) across the range of feeling intensity. Curves p(x=l) to
p(x=5) represent the probabilities for poor-low intensity, average-low intensity, good,
average-high intensity, and poor-high intensity outcomes (consecutively).

Splitting suboptimal performances. Tenenbaum and colleagues break moderate (and

Some times poor) performances into two categories based on a high-low split of

emotional intensity. Although most IAPZ analyses have used this splitting procedure, the

68

[6350.

are C1

the!

p00

(ll

me
jus1
the

ICSI

obs

ex;

m
the
31‘!)
P5}-

31m.

reasons for this procedure were not explained in their introductory publication. Once data
are collected, the mean of emotional intensity is calculated for optimal performances.
Sub-optimal observations can then be designated as having intensities above and below
the optimal mean threshold. A typical IAPZ analysis yields four curves that represent:
poor performance (PP), moderate performance with above average-optimal intensity
(M/A), moderate performance with below average-optimal intensity (WB), and Optimal
performance (OP). In some studies, this procedure has also been applied to poor
performances (Johnson, Edmonds, Moraes, et al., 2007; Medeiros F ilho, et al., 2008).
The mean-split procedure is partly based on anxiety-performance theories, but due to the
limited amount of published material in this area, it is difficult to understand why this
method is included in most IAPZ analyses. Theoretical rationale aside, the true
justiﬁcation for the splitting of moderate results may be statistical in nature. Partitioning
the data may enhance the ﬁt of the model and thereby improve the overall quality of the
results, while taking advantage of the disproportionate numbers of moderate
observations.

Multimodal and online measurement. The following three characteristics are not
explicit attributes of the model, but they are common features of nearly all IAPZ studies
and are unique from nearly all IZOF research. From its inception, IAPZ research has
emphasized the use of multiple types of measurements, including physiological measures
(heart rate, skin conductance, etc.), immediate recalls (affect functionality, valence,
arousal, and performance), and post-competition reﬂection (PNA-77, FSS, etc.).
Psychophysiological measures have rarely been used in IZOF research, but are advocated

strongly in IAPZ research. Not only do these assessments measure affect from a different

69

perspective (explicit verbal versus implicit neurological), but they offer a measure that is
unbiased by memory and perception. Also, physiological measures can often be
monitored “online” while an athlete is performing. IAPZ research often combines online,
immediate recall, and reﬂection 30-60 minutes post-competition (Johnson, Edmonds,
Moraes, et al., 2007).

Of the ﬁve empirical research studies published by Tenenbaum and his colleagues,
three have assessed psychophysiological markers such as heart rate and skin conductance
during performance (Edmonds, et al., 2006; Johnson, Edmonds, Moraes, et al., 2007;
Medeiros F ilho, et al., 2008). Along with the many advantages of these objective
biological markers of emotion, there are disadvantages to these measures (Lang, 2000).
For instance, many of the necessary measurement devices cannot be utilized in
competitive sport settings because they are too sensitive, cumbersome, or fragile.
Additionally, physiological measures can be confounded in sport because physical
activity can inﬂuence readings independently ﬁom affect. As examples of these
limitations, IAPZ research has only measured neurophysiological factors during archery
and a simulated racing game.

It should be noted that liberties have been taken in describing the IAPZ emphasis as
being on multimodal measurement. This is not perfectly accurate. Authors have deﬁned
the issue as being one of multidimensionality. However, their terminology may be
mistaken. Where multidimensionality conventionally refers to the evaluation of multiple
unique forms of psychobiosocial content, IAPZ research focuses almost exclusively on
one form: arousal. For instance, Edmonds et al. (2006) utilized several online and

immediate recall measures, such as heart rate, skin conductance, and verbal report of both

70

pleasantness and arousal. These are not four unique dimensions or forms of affect. More
so, they are different perspectives on a single form of affect (arousal). As such, IAPZ
research prefers to evaluate global affect using different measurement modalities, rather
than truly assessing multiple unique dimensions of affect.

Intraindividual longitudinal and multi-context research. As we move further
away from explicit, deﬁning attributes of the IAPZ model and closer to trends in this line
of research, we arrive at the ﬁnal two features. More so than in IZOF literature, IAPZ
research has emphasized more long-terrn (season-long) repeated measures designs.
Researchers have also excelled in comparing APZ’s for individual athletes across
different contexts/tasks (e.g., shooting distance in archery). These studies have offered
unique information about the dynamics of emotions during a competition, across different
tasks within a competition, and between competitions over the course of a season.

For instance, in their case study of an intemational-level female archer, Johnson et al.
(2007) collected 348 observations over ﬁve shooting distances in two competitions. Each
shooting distance was associated with a unique APZ proﬁle. Generally, optimal zones
and related probabilities decreased as shooting distance increased. Medeiros Filho et al.
(2008) conducted a similar study of three archers over a 10-month season. At least 40
observations were collected at each distance for all three participants. Results supported
JOhnson et al.’s conclusion that APZs ﬂuctuate within individuals across various
distances and during the duration of the season as well. Speciﬁcally, optimal zones
narrowed as distance increased (i.e., task difﬁculty increased). Their study did not reveal
momentum for optimal performances (i.e., optimal performance did not increase the

Probability of subsequent optimal performance). It is worth noting that this study and all

71

other IAPZ research have conﬁrmed the basic premise of idiosyncratic zone theories: that
individuals have unique Optimal zones of functioning and vary from athlete to athlete.
Critique and Shortcomings of IAPZ
Compared to the up to 20 items assessed through Hanin’s procedure, IAPZ
measures are simpler and less invasive. By reducing the assessment of affect to only two
or three questions (rating valence, intensity, and functionality), measurement is less
invasive. However, important information is lost (or never acquired). The criticisms of
IAPZ can be considered a product of IAPZ’s methodological catch-tvventy-two. In
summary, the IAPZ model may be less ecologically valid; it has abandoned discrete
emotional items in favor of a broader focus on global affect; and does not examine the
interactions between discrete items (i.e., how conﬁdence, anxiety, and focus interact).
Deemphasized ecological validity. Because it is early in the development of the
IAPZ model, much of the preliminary research has involved more controlled
experimental designs (e.g., video racing games and archery). This methodology is not a
problem at this point in the IAPZ’s development. However, it is very important to
remember that these idiosyncratic models originated from the applied realm, and should
continue to remain ecologically valid. IAPZ must use experimental tasks that are
generalizable to popular competitive sports that are emotionally intense—like hockey,
football, and basketball.
Global dimensions versus discrete categories. Since its inception, and in every

StUdy since, researchers have assessed IAPZ within the framework of Russell’s (1980)
circumplex model of affect. In keeping with this perspective, researchers have

consistently used the affect grid to measure feeling states. The affect grid (Russell, et al.,

72

1989) is a 9 X 9, 81-square graph. The two axes of the graph represent pleasure (from
“displeasure” to “highly pleasurable”) and arousal (from “sleepiness” to “highly
aroused”). The grid can be deployed in a number of ways, such as by: having participants
respond to two separate nine-point likert scales (Cohen, et al., 2006); having participants
respond verbally while performing (Edmonds, et al., 2006); or in the traditional manner
of checking the one square of the 81 that describes their feeling state. Some studies have
included a third axis to assess the functionality of the person’s state (from very unhelpful
to very helpful).

Weaknesses of continuous dimensions. As mentioned previously, IAPZ research has
favored Russell’s circumﬂex model, which focuses on global measures of affect. IAPZ
authors have presented this as an effort to achieve “parsimony and idiosyncratic
generalizability” (Johnson et al., 2007, p.318) by shifting the focus from emotions to
affect. However, this move towards parsimony neglects important information about the
emotions. The distinction between discrete categorical and continuous descriptions of
affect is a complex issue in psychology. This debate is ongoing and a review of the
literature on this topic suggests that both approaches are relevant (Lazarus, 1991; 2000).
It is not a question of determining which is right, but of understanding the strengths and
Weaknesses of each. Though its parsimony makes the affect grid convenient for ﬁeld
research, it does provide a comprehensive representation of emotion. Lazarus (1991)
cXplained the difference succinctly:

Conceiving of emotions as discrete categories, each of which can be placed on a

dimension from weak to strong, is very different from thinking of them as

73

overlapping dimensions, in which many categories are reduced to a few and their

distinct qualities lost or blurred. (p. 59-60)

Consider the following example of sadness and anger, two very different emotions. A
person could perceive sadness and anger as being equally unpleasant, and thus give them
the same valence score. Furthermore, each might be considered equally helpful (or
harmful). As such, measuring only valence and functionality will not distinguish between
unique emotions if those emotions are similar in terms of valence and functionality, but
differ in every other parameter of affect.

Watson (2002) recommends that global dimensions of affect be assessed in
combination with lower-level assessments of speciﬁc affect categories. Not only does this
approach evade the conﬂict between the discrete and continuous positions in science, but
it allows researchers to capture more information—which is the ultimate goal of scientiﬁc
inquiry. Speciﬁcally, this dual-method approach accounts for both a simpliﬁed depiction
of higher-order characteristics (e.g., valence and functionality), as well as the unique
personal signiﬁcance of discrete emotions (e. g., sadness, joy, conﬁdence, anxiety, etc).
While IAPZ research has focused almost exclusively on global measures of affect, the
traditional (PNA-based) stepwise method is more ﬂexible. Hanin’s method assessing
IZOF meets Watson’s recommendations because discrete emotions are categorized by
their valence and functionality (e.g., N-, N+, P+, and P- form categories). Despite citing
Watson’s work, IAPZ researchers neglect Watson’s call for multilevel representation of
affect in favor of only global descriptions.

Deemphasizing interactions between feeling states. Along with the disadvantages

mentioned in the previous paragraph, there is a second disadvantage of depicting affect

74

along continuous axes. When you chose a point on an axis, you are saying that in that
given moment, the person had a single feeling state—that they can only experience one
type of affect at a time. This constraint is substantial: IAPZ does not consider the
interaction of multiple feeling states at a given moment (i.e., being conﬁdent yet anxious,
or angry but focused).

Furthermore, and more frustratingly, IAPZ researchers have chosen not to have their
predictors (e. g., pleasantness, arousal, and functionality) interact. Pleasantness (valence)
scores could be combined with functionality ratings to identify affective content similar
to that in the IZOF model. Currently, by separating these indices, researchers often ﬁnd
that more pleasant and more functional emotions are associated with optimal
performance. The problem with his approach is that it does not differentiate between
relevant combinations such as negative-functional and negative-dysfunctional feeling
states. This is just one differentiation that has been neglected in IAPZ studies up to this
point. In particular, Cohen et al. (2006), Edmonds et al. (2006), Golden et al. (2004),
Johnson, Edmonds, Tenenbaum, et al. (2007) and Johnson, Edmonds, Moraes, et al.,
(2007) could have coded pleasantness ratings below the midpoint as “negative” and
above the midpoint as “positive,” repeated the procedure for functionality ratings, and
then combined these to establish four groupings similar to Hanin’s N-, N+, P+, and P-.

Misrepresenting IZOF research. Unfortunately, there have been instances where
IAPZ authors have misrepresented the IZOF model in such as way a to favor IAPZ
during comparison. These critiques of IAPZ do not represent problems with the IAPZ

model itself, but are important considerations nonetheless.

75

The ﬁrst example comes from Kamata et al. (2002), who only presented one method
of estimating IZOF proﬁles. Not only are there numerous IZOF methods, but the method
presented was the most outdated (e. g., recall of optimal states using questionnaires, +/-
either ‘A or '/2 of a standard deviation). Also, while IZOF research has over-utilized recall
procedures, Kamata et al. present this as the only method used in IZOF studies.

Golden et al. (2004) seem to overstate the reliance of IZOF research on best-/worst-
ever reﬂections. Again, while many studies have asked participants to reﬂect on previous
best and worst performances, some IZOF studies established optimal zones by
monitoring repeated competitions. Golden et al. also exaggerated the focus on pre-
competitive states and failed to acknowledge IZOF studies that have assessed emotions
during performance.

In addition to these ﬂawed statements, Golden and colleagues made a rhetorical
comment that cannot be ignored. The authors explained that “the purpose of [their] study
was to replace Hanin’s (2000) concept of affect-performance linkage by deploying a
probabilistic method” (Golden, et al., 2004, p. 37). The notion that IAPZ is a sufficient
model to replace IZOF is highly overstated. IAPZ is arguably insufficient to justify even
being considered unique from IZOF. This concern is explained in the following section.
Conclusion: Is IAPZ a method or a model?

Considering the above review of IAPZ literature, and in spite of high regards for
Tenenbaum’s research on the topic, one must ask if the IAPZ is truly a unique affect-
performance theory? Does the IAPZ separate itself sufﬁciently from the IZOF model? It
is the position of this dissertation that the IAPZ is a method of assessing individual zones

of optimal functioning. The IAPZ method lacks unique procedures for determining

76

individual proﬁles, does not have a unique theoretical or conceptual foundation, and
simply does not sufficiently distinguish itself from the work of Hanin.

Though there is greater emphasis on the use of multiple methods of assessing optimal
zones, the IAPZ does not present a unique set of procedures for collecting performance
data. IAPZ provides an excellent alternative method for analyzing this data once
obtained, but little more. Also, with the exception of three brief paragraphs describing the
Individual Psychological Crisis Theory (IPCT, See Tenenbaum et al., 2008), IAPZ does
not have a unique theoretical foundation. Tenenbaum et al. (2008) introduced the IPCT
only recently—describing it as “conceptually linked to the main assumptions of the
IZOF” (p. S18). The IPCT also recognizes the probabilistic nature of phenomenon.
However, very little else is offered about how this theory explains IAPZ’s, and there is no
comparison between IZOF and IAPZ models with respect to the predictive and
explanatory ability of the IPC Theory.

IAPZ’s use of logisitic regression to determine probabilistic estimates is an important
contribution to IZOF research. In fact, Tenenbaum and colleagues’ IAPZ is one of the
most important contributions to the IZOF model. However IAPZ is only minimally
unique from IZOF. The differences in methods supplement, but do not supplant the work
of Hanin, Robazza, and others. Furthermore, the use of OLR to determine individual
proﬁles is not something that all practitioners are capable of computing due to the
statistical complexity and the amount of data necessary for this analysis. Consequently,

even this most-beneﬁcial aspect of IAPZ carmot replace the IZOF model completely.

77

Summary: Integrating the IZOF and IAPZ Frameworks

The literature review has thus far provided a foundation for the primary purpose of
this study: to examine the effects of feeling states on tennis performance by integrating
and contrasting the IZOF and IAPZ frameworks. Among the various types of feeling
states, emotions are the most researched, most relevant to individual performance
proﬁling, and have the strongest theoretical foundations. Accordingly, special emphasis
was placed on the role of emotions and related theories. By their very nature, emotions
are important performance variables because they inﬂuence goal-achievement, cognitive
processing, and motor functioning. However, the focus of this dissertation is not limited
from other important performance factors. Feelings of conﬁdence, vitality, speed, and
focus (etc.) are important to athletic performance, and generally function in a similar
manner as emotions (recall Hanin’s mobilization-utilization hypothesis).

The ﬁrst and most fundamental hypothesis of this study is that the IAPZ’s OLR
method will identify idiosyncratic probabilistic estimates for each function-valence
category of each athlete’s individual proﬁle. Testing this hypothesis requires a large
number of observations that are relatively balanced across a range of performance-levels,
along with a variety of emotional feelings and intensities.

Because the complete IEP-based IZOF proﬁle can include 16-20 items—which is not
sufficiently simple for applied practice—a small number of “critical” trigger emotions
will be determined by each athlete, and compared to the predictive accuracy of the four
ﬁrnction-hedonic categories. Thus, a second hypothesis will compare the accuracy of
each trigger item and the four valence-function categories to determine how accurately

each item/category predicts poor, average, or best performance based on feeling intensity.

78

For the ﬁnal hypothesis, the dynamic context component of the ﬁve-dimensional
psychobiosocial model will be examined by comparing service games to service return
games. More speciﬁcally, means and variances of feeling intensity for three perfonnance-
levels and four feeling categories will be compared across the two contexts within each
participant. While intraindividual differences are expected due to the different demands
of serving and returning serve, the outcome of this hypothesis is not as important as the
mere fact that it is being assessed. In order to understand the dynamics of feeling-

performance relationships, all dimensions of the psychobiosocial model must be tested.

79

CHAPTER THREE
METHOD

This investigation was designed to assess the effects of feeling states on performance
during competitive tennis matches for the purpose of contrasting and integrating the
IZOF and IAPZ models of emotions and performance. Also, intraindividual differences
in performance-zones were compared between serve and return of serve context. Finally,
a complete proﬁle with four feeling categories (based on functionality and valence) were
compared to four simple idiosyncratic discrete feeling states identiﬁed by each player as
being critical for his performance.

Ten male NCAA tennis players were recruited and provided informed consent to
participate in the study without compensation. Participants created individualized proﬁles
that describe the relationship between performance and the intensity of various feeling
items they experience during poor and good performances. Based on these individual
proﬁles, monitoring forms were created for each player. The form also assessed the
outcome, context (service/retum), set score, performance rating, perceived challenge and
skill, and intensities of individual emotions. The challenge and skill ratings of the form
were used to assess ﬂow (Csikszentrnihalyi & Larson, 1987; Csikszentrnihalyi, Rathunde,
& Whalen, 1993; Jackson & Csikszentrnihalyi, 1999), and were not pertinent to the
purposes of this dissertation. Participants completed the monitoring form during the
changeovers of each intra-squad challenge match. Each participant played between 5-8
(M = 6.90) matches and completed 52-138 forms (M = 91.8). An average of 13.3

observations were recorded for each match. The IAPZ method was used to determine

80

bandwidths of feeling intensities for individual zones of poor, average, and good
performance. Upon establishing individual zone proﬁles, hypotheses were tested.
Participants

Ten male players from an NCAA Division I tennis team were recruited to participate
in this study. The study and its purpose were described to the head coach of the team, and
his approval was obtained before requesting consent from each player.

The players in this study represented an ideal sample because they were experienced,
highly trained and relatively elite athletes. On average, participants were 20.1 years of
age (SD = 1.66), had 13.7 years of tennis experience (SD = 2.58), and 2.5 years of
college tennis experience (SD = 1.27). They also had previous experience working with a
sport psychology consultant, although none of the participants had experience with the
IZOF model prior to this study. Because of these characteristics of the sample, it was
expected that participants would reﬂect a high level of emotional self-awareness and
tennis experience. Both of these attributes are essential when assessing individual zones
of ﬁmctioning because the method and the theory are based on elite athletes (Hanin,
2000b) and individuals who are aware of their emotions. Hanin explains that experienced,
competitive athletes offer more reliable recall and emotional awareness than do novice or
inexperienced athletes (Raglin & Hanin, 2000).

Also, tennis represented an ideal sport for the purposes of this study. Players compete
alone on court, so team dynamics and social factors do not have a strong inﬂuence within
the psychobiosocial model—those dimensions are essentially controlled for by choice of
Sport. Tennis is a difﬁcult sport where momentum shifts regularly, so there is generally a

great deal of variability in terms of emotion and performance. There are two unique tasks,

81

or contexts in tennis—serving and retuming serve—that require different skills,
strategies, and perhaps different optimal feeling states. These differences permitted
comparisons of the context dimensions of the psychobiosocial model. Finally, the rules of
tennis require intermissions in each set in order to change sides after every odd-numbered
game. Interrnissions during sets are approximately 90 seconds and breaks between sets
are approximately three minutes in duration. These changeover breaks were used to self-
assess performance without interfering with the normal conduct of play, as each
observational form required only 25-40 seconds to complete.
Measures of Feeling

The foundation of the current methodology involved the development of individual
proﬁles of optimal (IZOF) and dysfunctional (IZDy) performance zones. IZOFs and
IZDys are identiﬁed through stepwise procedures known as Individualized Emotion
Proﬁling (IEP; Hanin, 1997, 2000c, 2003). The current study slightly modiﬁed the
stepwise procedures based on the applied experiences of the investigator, and integrated
them with Hanin’s narrative-interview method (Hanin, 2003). The stepwise procedures
for the identifying and measuring individual proﬁles will be described in further detail in
the procedure section.

Positive-negative affect list. The Positive and Negative Affect List (PNA-77; Hanin
& Syrja, 1995, 1996) was established by Yuri Hanin as part of his stepwise procedure of
assessing individual proﬁles. The PNA is a list of feeling items, which athletes chose
from to describe their experiences. The PNA-77 is depicted in Figure 3 of the Appendix.
With respect to hedonic tone, there are 40 positive and 37 negative items. Items on each

lrne / row are synonyms, so participants were permitted to choose no more than one item

82

from any row. In addition to the 77 items in the PNA, participants were urged to include
their own adjectives or key word items.

Borg CR-10. Borg’s Category Ratio scale (CR-10; Borg, 1998) has been used to
describe the intensity of feeling states in the IZOF model since the model’s revision in
1997. The CR-10 allows athletes to identify the intensity of feeling states, and thus, to
establish zones of optimal functioning and of dysfunction. In spite of its title, the CR-lO
is a 12-point scale that includes 0 (“nothing at all”) and 11 (“maximal possible”—
sometimes coded as ‘#’ or ‘10""). The scale begins at 0, 0.5, and 1, increasing by integers
of 1, up to 10, and ends with the value 10* which is identiﬁed as 11 in the current study.
The descriptive anchors for each level on the continuum are described in Figure 4 (See
Appendix) as part of the monitoring form for participants during matches.

The traditional CR-lO scale was modiﬁed slightly by Hanin and his colleagues
(Hanin & Syrja, 1995; Turnmavuori & Hanin, 2000). The reliability of the CR—lO scale in
IZOF research was established by Hanin and Syrja (1996). In a later study of 200 athletes
providing 46,934 intensity ratings, the distribution of verbal anchors along the intensity
continuum were conﬁrmed to be appropriate (Tummavuori & Hanin, 2000).

Monitoring form. The primary source of data for this dissertation was the
monitoring forms completed during competition (See Figure 4). The monitoring forms
included three sections. The ﬁrst section identiﬁed the context (service or return),
outcome of the game (won / lost), and score in the set. Next, perceptions of performance,
Challenge, and skill were assessed across three separate 11-point likert scales (from 0-10).
The descriptive words “poor,” “average,” and “good” were used to anchor the likert scale

at 0, 5, and 10 respectively. The likert scale was partitioned into three groupings to

83

represent each anchor (e.g., 0-3, 4-6, and 7-10). The third and most important set of
information on the form was the intensity ratings for each content item. Discrete items
were self-selected, and intensity was determined based on the Borg CR—lO scale—these
content and intensity dimensions were described in the previous sections. Participants
were urged to limit their monitoring form to only 10-14 items, with at least two items in
each of the four content categories (e.g., helpful-positive, harmful-negative, etc.).
Participants chose 12-13 items.

Procedure

Michigan State University’s Internal Review Board (IRB) for ethical and legal
research conduct approved the study protocol. The primary investigator met with the
head coach of an NCAA Division I Men’s Tennis program prior to recruiting
participants. The NCAA tennis season begins with a semester-long preseason in the
autumn. In the ﬁrst week of the preseason, players attended an orientation meeting at
which time informed consent was provided and IEP’s were developed.

Adaptations to Hanin’s procedures. The traditional stepwise procedures for
identifying IZOF proﬁles were modiﬁed slightly based on the applied experiences of the
Primary investigator, who had used Hanin’s method for hundreds of athletes prior to
conducting this study. Three changes were made, which increased the number of
Performances recalled, reduced the number of items in each proﬁle, and reﬁned each
proﬁle based on face-to-face interviews. Rather than reﬂecting on a single best- and
Worst-ever performance, participants were asked to identify two or three such ‘poor’ and
‘good’ experiences. Also, rather than identifying as many as 20 unique content items,

Participants were only permitted to include 10-14 relevant items. After completing the

84

stepwise procedures, individual proﬁles were further reduced to 12 items—including at
least two, but no more than four items in each of the four categories described in Chapter
2 (e.g., positive-helpful, negative-helpful, etc.). One participant preferred to have a 13-
item proﬁle. Although the full stepwise process utilizes 16-20 items in each athlete’s
proﬁle, Hanin recommends that ﬁeld research reduce proﬁles to only 12 items (Hanin,
2000d). Also, more recent studies have used the PNA-10, a list of the 10 most prominent
items based on previous research (Hagtvet & Hanin, 2007; Hanin & Syrja, 1995;
Lukkarila & Hanin, 2001). As such, it is acceptable to simplify athletes’ proﬁles.
Stepwise procedure. Given the aforementioned modiﬁcations to the traditional
stepwise procedures, the following process was employed in order to develop individual
proﬁles. Participants began by identifying and describing their best and worst
performances in the previous 2 years (more recent performances were preferable). Next,
based on these performances, participants used the PNA to identify helpful positive and
negative emotions (based on previous best performances), and harmful positive and
negative emotions (based on previous worst performances). Thirdly, the Borg CR-lO
scale was used to estimate the intensities of each discrete content item (the positive and
negative feeling states from the PNA) during each best and worst performance. It is
important to note that, given the purpose of the current study, participants were asked to
recall intensities during competition—rather than pre-competition and post-competition
states. Fourthly, the intensities of each content item were combined across best-ever
performances to create a bandwidth, or zone, of optimal functioning (IZOF). Intensities
were also combined across worst-ever performances to create a zone of dysfunction

(IZDy). IZOF and IZDy proﬁles were depicted on the same graph and compared (See

85

Figure l). The graphic representation, or “visualization,” of the proﬁles was the ﬁnal
procedure before individual meetings were conducted to reﬁne the proﬁle.

Interviews. A ﬁnal revision of the stepwise procedures involved individual
interviews. During these interviews, players were allowed to reﬁne intensity estimates
and eliminate erroneous items. This is important because by the end of the procedure,
many athletes felt that some of the initial items were not predictive of, or important for
performance, or that the original zone estimates (based on previous performances) should
be raised or lowered by a point or two. Participants also identiﬁed four critical “trigger”
items from among the 12-13 in their ﬁnal proﬁle during their interviews.

Data collection: monitoring forms. Having established and reﬁned each IZOF
proﬁle, identiﬁed trigger items, and explained the in-vivo data collection procedures; the
next stage of the procedure involved monitoring feelings and performance during
competitive intra-squad matches. Every player on the team played a round-robin of
matches against his teammates. These challenge matches were important because they
partly determined rankings within the team (only the top six players play singles matches
in the NCAA). Because of the importance of these matches, and the relative parody
within the team, these challenge matches were highly competitive.

In a post-study survey of players, the average rating of how important the challenge
matches were was 7.05 (on a scale from 1-10; SD = 1.32), and the average rating of how
competitive the matches were was 6.40 (SD = 2.31). Six of the 10 ratings of
competitiveness were between 7-8. There were two ‘outliers’ of sorts who rated
cOmpetition as a 2-3: neither player performed as well as he had hoped. One participant

rated the importance of the matches lower than the rest of the team (i.e., a 4 out of 10).

86

The challenge matches were played as typical NCAA matches, which included
changeovers with a 90 second break after every odd service-game of each set (note that
there is no break after the ﬁrst game, only a changing of ends). Though a 90-second break
is not permitted during the ﬁrst changeover (after the ﬁrst game of each set), participants
completed forms during the ﬁrst changeover in order to collect more data. Also, without
this procedure, the ﬁrst game of each set would likely have been forgotten by the end of
the third game and therefore either lost to the dataset records or recalled less accurately.

In tennis, players alternate service games. Serving to start each point of a game is an
advantage in tennis. When a player is not serving, he/she is said to be “returning serve.”
As such, with the exception of the ﬁrst changeover of each set, there is one service game
and one return game played between each changeover. Participants were asked to
monitor one of the two previous games during each changeover. Speciﬁcally, participants
were asked to complete the monitoring form for the game that they felt had the extreme
(very positive or very negative) performance. If both games had comparable
performances, then participants were urged to choose the context (service or return) that
they had least often monitored. If both games had extremely good or poor performances,
players were urged to complete a form for both, so long as two forms could be completed
without excessively distracting the player or delaying the conduct match. Occasionally,
players felt that the previous service and return games were identical in their feelings and
performance. This was more common for tiebreakers (where the two players alternate
serves every two points). In these cases, players circled both serve and return on their

forms—data were entered accordingly.

87

In both pilot testing and in the actual data collection, forms were completed in 25-40
seconds, and were considered to be minimally distracting. In a post-study survey of
players, the average rating of how distracting the forms were from their performance was
4.05 (on a scale from 1-10; SD = 2.813). Two players found completing the form to be
quite distracting, rating it as a 7.5 and a 9. The participant who provided the latter rating
withdrew from the study after ﬁve matches. Though moderately-to-very distracting for
some players, all participants felt that they completed the forms very accurately (M = 8.2;
SD = 1.42). In fact, the attention to detail and desire to be accurate was the primary cause
of the negative, distracting effects identiﬁed by the aforementioned two players. Both of
these participants described very high accuracy, and showed throughout the data
collection process that they were reporting their feelings and performance ratings as
accurately as possible. These participants may have essentially been trying too hard. All
players were told to complete the forms quickly, so as not to over-think or distract from
performance. The primary investigator attended all challenge matches to ensure that
forms were available and that participants’ concerns could be addressed.

Ethical Considerations

The principal investigator had a dual-relationship with the team as both researcher
and mental training consultant. Right to refuse was especially important because the
investigator was in his third year as the team’s mental trainer and while positive rapport
and trust had been developed over the years, it was important that players not feel forced
t0 participate. It is clearly unethical to force or coerce people to participate in a research
Study. Furthermore, coercing participation is likely to result in untrustworthy data. The

investigator was considerate of the effects that the monitoring procedure had on the

88

participants’ performance—particularly because of the importance that the matches could
have on each player’s ranking. As such, all players on the tennis team were given the
opportunity to exclude themselves from the study at any time. Also, in order to ensure
that players were not coerced into participating, an independent party met with the
participants without the researcher or coaches being present. The independent party
reminded participants that players could withdraw at any time. Due to ﬁ'ustration over his
performances and several injuries, one player withdrew after his ﬁfth match (one match
before the ﬁnal round of play).

Measures were also taken to make the collection of data less intrusive. Each player
was given his own binder and pen with copies of the monitoring form. Binders were set
on each player’s bench / chair to make it easier to write. The investigator reviewed forms
during each match to make sure that players had not made mistakes (e.g., incomplete
data, etc.). Most importantly, when participants chose not to complete a form after very
frustrating points, the investigator did not force players to comply. This occurred very

rarely, and when it did so, it was most often at the end of a match after a loss.

89

CHAPTER FOUR
RESULTS

Treatment of Raw Data: The Basics

Performance levels. Participants rated their performance on an 11-point scale from
0-10. By convention and for the sake of simplicity, these ratings were converted to
ordinal data by deﬁning ratings between 0-3 as ‘poor’ performances, 4-6 as ‘average’
performances, and 7-10 as ‘good’ performances. For two participants, the ordinal
parameters were modiﬁed due to extreme distributions of ‘good’ performances.
Speciﬁcally, Player I originally had only nine good performances (out of his 97
observations). Contrarily, Player 3 rated 70 out of 101 observations as good
performances. Player 3 also had an extremely low number of poor performances (e.g.,
only ﬁve). After consulting with each player, both agreed that their standard were
excessively high or low. Player 1’s standard for good performances was adjusted to 6-10;
and Player 3’s parameters were set at 0-4, 5-7, and 8-10. The revised distributions of
poor, average, and good performance for each participant are described in Table 2.

Aggregated feeling categories. Each of the 12-13 discrete feelings identiﬁed by
participants was grouped into one of four possible categories based on the functionality
and valence of the item. These categories were determined through the Individual
Emotion Proﬁling process that included the PNA-77 list of feelings (described in the
previous chapter). Rather than conducting analyses for each discrete item for each player,
analyses were conducted for the aggregated value of items in each category. Aggregating
items within categories helped to create parsimony in the models: as developing

individual curves for all 121 discrete items selected by the 10 players would not be

90

productive. For example, Player 2’s ratings for energetic, conﬁdent, and determined, and

Player 5’s selection of relaxed, quiet, alert, and overjoyed were aggregated into the

positive-functional category. This not only simpliﬁed each player’s proﬁle, but it also

made it possible to compare individuals because while few players selected the same

discrete items, every player had the same aggregated function—valence categories.

Analyses were also performed for the four discrete items identiﬁed by each player as

being most important for his performance (i.e., trigger items). As such, the raw values

entered by players were not directly considered in the results. Ultimately, feeling

intensities were represented by eight categories: positive-functional, negative-functional,

negative-dysfunctional, positive-dysfunctional, and Triggers 1-4.

Table 2

Distribution [and Percentage Proportion] of Performances for Each Participant

 

Performance Levels

 

 

Participant Poor Average Good Missing Observations
l 32 [33.0] 31 [32.0] 34 [35.1] 0 97
2 37 [36.6] 44 [43.6] 20 [19.8] 0 101
3 8 [7.9] 38 [37.6] 55 [54.5] 0 101
4 11 [14.7] 42 [56.0] 22 [29.3] 0 75
5 13 [13.4] 67 [69.1] 17 [17.5] 1 98
6 22 [42.3] 14 [26.9] 16 [30.8] 0 52
7 60 [58.8] 26 [25.5] 16 [15.7] 0 102
8 16 [20.0] 53 [66.3] 11 [13.8] 3 83
9 20 [29.0] 44 [63.8] 5 [7.2] 2 71
10 21 [15.2] 83 [60.1] 34 [24.6] 0 138
Totals 240 [26.3] 442 [48.5] 230 [25.2] 6 918

 

 

91

Descriptive Data

The total number of valid observations collected for this study was 918. Fewer than
10 observation forms were deemed unreliable. Such forms were excluded due to
excessive incomplete data, or because the intensity ratings for every item were identiﬁed
with a single circle (out of frustration). If a player circled both ‘serve’ and ‘return’ on the
same form, the observation was entered into the dataset twice. This occurred very rarely:
most often for tiebreakers. There were very few tiebreakers, and even fewer were double-
coded. The total number of observations for each participant ranged from 52 to 138.

Player 6 had only 52 observations because he missed three matches due to an injury
and then asked to be removed from the study before the ﬁnal round of matches due to
frustration from under-performing. Completing the forms did not cause his frustration,
but did exasperate it. All players completed nearly every possible form. The range in
observations per player was simply due to the number of games and sets played in each
match. Participants sometimes failed to complete the last form of the match after a loss.
Given the extraordinary compliance and quality of data provided by participants—as well
as the disappointment associated with many losses—the investigator did not follow up on
these missing forms. Descriptive data for performance levels and intensity ratings of
feelings are described in the subsections below.

Distribution of performance. As described in Table 2, there was a relatively normal
distribution of performance ratings across participants. In terms of the statistics, the ratio
0f skewness to standard error was 0.198 (0.016 / 0.081), and kurtosis ratio was -6.531 (-
1.058 / 0.162) well above the conventional threshold of 2.0 (Raykov & Marcoulides,

2008). While signiﬁcant negative kurtosis was observed for the ordinal performance

92

variable, this was reﬂective of the parameters established before the study. That is to say
that, in order to conduct reliable logistic regressions, a larger nmnber of extreme
performances (i.e., non-average ones) were necessary. As such, more poor and good
performances were desirable, and thus a ﬂat kurtosis distribution. This violation of
normality does not implicate the subsequent analysis as nonparametric statistics do not
require normality (Mueller & Hancock, 2008; Pallant, 2009), and AN OVAs are robust
enough to overcome negative kurtosis (Vincent, 1999). Also, many of the ANOVAs that
were conducted accounted for each performance level separately.

In total, 240 (26.3%) of the 918 performances were poor, 442 (48.5%) average, and
230 (25.2%) were good (See Table 2). This distribution varied slightly among players,
including the following noteworthy exceptions: Player 3 had very few poor performances
(7.9%); Player 6 had more poor (42.3%) and fewer average (26.9%) performances; and
Player 9 had relatively few good performances (7.2%).

The frequency of raw performance scores across all participants is of less
consequence to the study’s purpose and hypotheses, but worthy of a brief description.
Each performance rating from 0-10 had the following percentage distribution: 0 (9.5%), l
(3.8%), 2 (4.8%), 3 (7.8%), 4 (10.0%), 5 (21.3%), 6 (18.9%), 7 (11.5%), 8 (7.7%), 9
(3.7%), and 10 (0.9%). The very low occurrence of extremely high performance levels
may have some consequence on the results and will be discussed in the next chapter.

Intensity of feelings. Feeling intensities are described separately for each of the four
feeling categories in Tables 3-6, which are organized into the three performance levels
for each participant. Figures 6-9 (See Appendix) depict the distribution of intensity scores

aggregated across participants for each feeling category. These raw descriptive tables and

93

ﬁgures are not critical to the objectives of the study, but do provide interesting
background information for the interindividual comparisons.

Of particular interest are the variability of means and standard deviations within each
column (i.e., within performance levels between participants). For the most part, means
appear to be different across performance levels for each player, but formal AN OVAs
were not conducted to examine this as it is neither relevant to the study’s purposes, nor is

it required by convention.

Table 3

Individual Means and Standard Deviations for Positive-Functional Feeling Intensities
across Performance Levels

 

 

 

 

 

Poor Average Good

Players Mean SD Mean SD Mean SD
1 4.83 1.34 5.62 1.08 6.86 0.67
2 0.82 1.09 4.39 1.75 6.13 2.22
3 4.84 1.97 6.50 11.711 8.70 1.27
4 5.36 1.06 6.110 1.55 6.71 1.76
5 3.10 1.28 5.80 1.64 6.67 1.30
6 1.66 1.51 4.43 0.79 5.22 0.27
7 1.39 1.78 5.55 1.32 8.59 0.78
8 3.61 1.89 5.74 1.53 7.65 1.49
9 6.30 0.65 7.14 0.80 8.27 0.28
10

3.63 2.29 6.31 1.07 8.03 0.87

l

94

Table 4

Individual Means and Standard Deviations for Negative-Functional Feeling Intensities
across Performance Levels

 

 

 

 

 

 

Poor Average Good
Players Mean SD Mean SD Mean SD
1 2.27 2.66 2.32 2.50 1.19 0.95
2 0.56 0.94 3.83 2.30 5.20 2.95
3 6.56 0.94 6.71 1.79 7.17 1.27
4 5.97 0.86 6.23 0.96 6.23 0.85
5 5.28 2.18 3.51 1.33 3.51 0.58
6 6.16 1.25 4.55 1.38 3.89 0.79
7 1.48 1.88 5.67 1.36 8.44 1.03
8 6.52 3.09 2.65 2.50 2.39 3.94
9 5.18 0.82 5.75 0.91 8.27 0.41
10 3.16 1.82 5.92 1.09 7.39 0.62
Table 5

Individual Means and Standard Deviations for Negative-Dysfunctional Feeling Intensities
across Performance Levels

 

 

 

 

 

Poor Average Good
ﬂayers Mean SD Mean SD Mean SD
1 2.14 1.80 3.10 1.73 2.26 1.31
2 4.94 2.34 2.94 1.87 l.89 1.85
3 5.11 2.60 2.65 2.10 0.91 0.96
4 3.30 0.35 2.61 0.58 2.24 0.52
5 6.38 1.98 4.05 1.18 3.49 1.04
6 7.20 1.18 4.81 1.18 5.20 1.35
7 1.09 1.39 0.71 0.69 0.51 0.19
8 2.86 1.98 0.96 1.03 0.38 0.64
9 4.49 2.24 2.81 1.38 3.20 1.02
10 5.98 2.51 2.63 1.23 2.95 0.59

l

 

 

95

Table 6

Individual Means and Standard Deviations for Positive-Dysfunctional Feeling Intensities
across Performance Levels

 

 

 

 

 

Poor Average Good
Players Mean SD Mean SD Mean SD
1 4.34 1.11 4.90 1.13 3.97 1.73
2 6.09 1.49 4.27 1.58 3.48 1.44
3 4.88 2.60 3.71 1.96 3.70 1.72
4 2.36 1.01 2.35 1.04 2.70 0.91
5 1.54 1.18 4.63 1.69 6.47 1.30
6 5.55 1.32 5.77 0.89 5.60 0.63
7 0.01 0.07 0.10 0.32 0.00 0.00
8 1.02 1.92 2.75 2.50 3.61 2.89
9 0.58 0.63 0.69 0.72 1.00 1.00
10 5.33 2.55 2.49 1.41 1.85 1.07

 

Treatment of Raw Data: Ordinal Logistic Regressions

Ordinal logistic regressions. As the foundation for the IAPZ method and this
dissertation, ordinal logistic regressions (OLR) were conducted on the four feeling
categories and four trigger items. Parameter estimates from the regressions were used to
establish performance curves (probabilistic estimates of performance across feeling
intensity), which were then used to identify poor, average, and good performance zones.
These performance zones were used to predict performance based on feeling intensities
for each category or trigger.

Mean splitting performance levels. Initially, OLR was conducted using three
dependent measures of performance described earlier. However, because none of the
IAPZ graphs identiﬁed a typical catastrophe-type result (e. g., performance decrements

occur at the highest levels of intensity) the OLR and curves were conducted again with

96

ﬁve outcome variables. [As described more thoroughly in the method, poor and average
performances were each split into two subgroups: based on feeling intensities above and
below the mean intensity for good performances. The resulting ﬁve performance levels
acted as dependent measures in the OLR and included the following performances
variables: poor-low, poor-high, average-low, average-high, and good.

Conducting ordinal logistic regression. For each participant, eight separate OLRs
were conducted for the four feeling categories and four trigger items. Within each
regression, as many as ﬁve dependant variables were included to establish an
independent set of parameter estimates. Five parameter estimates were identiﬁed from
each regression, including intercepts for performance outcomes one through four (the
ﬁfth dependent measure acted as a control), and the slope for the overall output. The
probability of each performance (0;) is depicted mathematically below with or
representing the intercepts, Bx the slope, and IV represented feeling intensity from 0.00 to
11.00 increasing by 0.01 for precision:

01=1/[1+exp(a1+Bx*IV)]

92:1/[1+exp(a2+Bx*IV)]-0.

03:1/[1+exp(or3+l3x*IV)]-91-92

04:1/[1+exp(or4+Bx*IV)]-91-92-93

05=1-01-02-03-04
Note that the probability of the ﬁfth outcome (good performance) is established as the
remaining probability (e. g., 1.0 minus the sum of average- and poor-performance thetas).

Establishing individual proﬁle curves. A screenshot of one of the 80 spreadsheets is

provided in Figure 10. In the top row, 01 to 05 are represented correspondingly by p(x=1)

97

to p(x=5). Parameter estimates on to on were inserted in the second row with the ﬁnal cell
in the row representing the slope (Bx). The ﬁrst column of the ﬁgure included feeling
intensities (the independent variable). To interpret the spreadsheet: at the intensity-level
of 0.0, the probability of poor-low performance was 0.4%, poor-high was 5.7%, average-
low was 11.0%, average high was 39.0%, and good performance was most likely at
43.9%. As such, 0.0 is clearly part of the good performance zone. However, the zone of
good-functioning is short-lived for this individual, as average-high performance is more
likely than good performance above the intensity level of 0.43 (circled in Figure 10). The
values inside of the box inserted at the bottom of the ﬁgure show that good performance
moves from being more-to-less probable than average-high performance from 0.43 to
0.44. Each spreadsheet was represented with visual graphs (See Figures 11-18 in text for
examples of probabilistic curves).

Performance prediction parameters. In total, 80 curves were generated with three
to ﬁve performance outcomes on each graph. Figures B6-B13 display examples of unique
IAPZ curves. Each curve was visually inspected to identify a rough approximation of
each performance zone, or bandwidth. From these approximations, the original
spreadsheets were consulted to identify the precise parameters of each zone (See Figure
10). This process was repeated for each performance zone of the 80 probabilistic curves,
then manually programmed into SPSS to establish performance predictions for each
category and trigger item. The performance prediction variable is described in the

forthcoming accuracy section of the results.

98

Figure 10

Example of an Excel Spreadsheet used to Generate OLR Performance Curves

 

 

 

- “ml"?

6.4629 .

0.00422:
i 0.00424;
' 0.00425
0.00426
“ 0.00428, ‘
"0.004292

0.00431 ‘

0.0043 2?

0.00434 ‘ 4

“0.001435”
0.00436 3

0.057134616

0.057310751 ;
0.05748739
0057664536 '
0.057842187 .

‘ 0058020347-
0058199016
”0058378195
‘ 0.058557886 ’ "
‘ ” 0.05873809 *
0.058918807; '

, 0 mﬁiiwf’
2727712942 -

 

71351 I

4.57666

0.109911
0.110189
0.110467
0.110746 ”
' 0.111026
0.111306
0.111586 ' ‘
” 01118677 ‘

0.112148

. 041243 ,
0.112712;

elm-4)

‘ 0246188383
0389969724}
0390312713 '
0390654356
039099465 "
0391333588
0391671166 “
0.39200738
0392342224
’ 0392675694?
0393007784 ‘ ‘ ‘
' 039333849

““5! +_
0.32945

0.438762
0.437951 *'
’ 043714
' 0.43633
‘ 0.43552 ”'

. 014339. 1.
0.433091

0.432283 ’

0.431474

6.430555" ‘

 

0.00474 ;

" 0.004751
3 0.004774

0.00478

7 0.0048
  90°43?
900433

0053507353
0.063801874 .

0.063996932

0064192538 '
0064388694
0.064585401 '
006478266
"0064980473"

0.119915
0.120209. ’
0.120504-
0.120798-
0.1210947
0.121389
"0121685-
‘0121982 ‘

0401142507 3
0401435585 : '
0401727155 1
0402017212 5
0402305753 ‘
0402592772 ‘ ”
0402878266
0403162231

0.410598

0.409801 *

0.409004
0.408208

’ 0.407413

0.406617 N
0.405823

0.405029 ’_

0.43471 ’ ‘

0.404235
0.403442
0.402649 ’

-_.®‘o-oo486
.044-.. 030483
. _ 0.45 000489

' 006517884
0065377763 1

1 0.122874,

0.122576 0403725554:
0404004905 )

0.122279 [0403444662

 

 

Note. This image is a screen shot of an actual data set (Player 10’s Positive-Dysfunctional
curve). The image was modiﬁed to include the circle and box located at the bottom of the
ﬁgure, indicating a change in zone. Also, the plane of the spread sheet was frozen after
the independent value 0.1 to save space. The ‘IV’ column represents feeling intensity
from 0.0 to 11.0, increasing by one-hundredth of a point. Columns ‘p(x=1)’ to ‘p(x=5)’
represent the parameter estimates (in row 2) and probabilities (in all subsequent rows) for
poor-low, poor-high, average-low, average-high, and good performances respectively.

99

Interindividual Differences

F our methods were deployed to assess interindividual differences in feeling-
performance proﬁles. These methods included: qualitative comparisons of the feeling
items chosen by players; visual inspection of individual graphs; and statistical
comparisons of the means and variances of each zone.

One of the core hypotheses of this dissertation is that there are individual differences
in performance-feeling proﬁles. By nature of these proﬁles, there are two ways in which
proﬁles can vary—in either location or in width—and thus two possible tests of
individual differences. First, zones can vary in terms of their location on the continuum.
For instance, Kurt’s optimal zone for positive-functional feelings may be higher than
Krist’s. The second method compares the width, or size, of each zone. For example,
Kurt’s optimal zone might range from 8-10, while Krist’s ranges from 5-9—thus
differing in both size and location.

Differences in the location of each zone can be represented by assessing differences
in central tendency (e. g., ANOVA’s and post-hoc t-tests). There is no conventional
method to assess whether or not the bandwidth of one player’s zone is different from
another’s. Without an established statistical procedure to follow, the variances of
intensity ratings were assessed simultaneously for all participants. Twelve such separate
tests were conducted for each feeling category and performance level. Variances for each
player were compared by dividing the one player’s variance score by the other’s, and
referring to the standardized F-Tables using the number of observations for each player

(i.e., n-l) to determine the critical value ("Comparing Variances," 2001).

100

Before presenting these statistical methods, there are two qualitative means of
assessment of examining individual differences. Firstly, the variety of discrete items
selected by each participant to represent each feeling category can be examined.
Secondly, visual comparisons of probabilistic performance proﬁles can be made.

Qualitative differences in individual proﬁle items. The most idiosyncratic element
of feeling-performance proﬁles may be the choice of discrete feelings. Recall that while
participants shared four common feeling categories (e. g., positive-functional, etc), the
exact items within each category were self-selected and unique across participants. The
question is: how often did participants choose the same discrete feeling states. It would
be unreasonable to consider one player’s choice of ‘angry’ as being different from
another’s choice of ‘furious’ to be substantively different. As such, before conducting this
comparison, synonymous items were grouped and compared to unique groupings. The
grouping of items adhered to the presentation of the PNA-77 (Hanin, 1993a; Syrja &
Hanin, 1997) in the Individual Emotion Proﬁling system (EIP; Hanin, 1997, 2000a). Both
positive and negative items are presented in Table 7.

For each set of positive and negative feeling items, there are 14 rows from which to
choose. A majority of players (ﬁve or more) supported six rows from the positive list and
eight rows from the negative list. While some rows received very strong support and
could be generalized as being important for most players, the distribution of player
support shown in Table 7 supports the notion of idiosyncratic differences in players’
choices of critical performance-related feelings. The rows that received strong support
(i.e., seven or eight players) included being satisﬁed, conﬁdent, aggressive, and irritated.

Surprisingly, rows with items like relaxed, fast, energetic, determined, nervous, intense,

101

888 <20 0.50: so @0003 00w00b0 0003 080: 00050005? no 0390 .3 003 020> 0.80000 00000—0 82502 .082

 

 

102

0 000—00000 .000 £50002 £00000: Shem _ 0000: 0908 0.000000
_ 00030030 0303000 608000 60000008 N 0003803 0000388 0000000
m :5 0.83 00000058 $0003 .005 N 32000 03000 .23 .0205
m 0N0— .nmmwwim 6.5005 N 03000.00 0000020 .003
m 03000.5 .0003: .0200: m 00:90 60:80
0 00:83 03000500000 £5080 m 0000000300 0002.026 00:00:00
m Ewt .Ew: 605000 .0003 v 0030 00:00:05 40.000000 .800
m $030.00 .0000: .2520: .0085. v 30.8w? .0U0w0000 .3800? .0360
0 0000c 6000:: m 00000000 6020053:
0 002800.: 63000005 05000:: 433000 c :20 .0000 .2000 £030
0 00000000 60030000 600500000 0 0020000 .0030..." .000 005800000
0 33:00 600000 .3000 .2000 0 0000 030008800 0000—00
0 000000000 68005 60.8000 0. 030 .5008 00005000
0 000—03 .0009: 03000030 .300 w 000000000 0000.000 60000—0 00%
0.8005 080: 032002 to00sm 080: 03:80

 

0.8090809 80: E00000 030002 000 03:80 Agogcoamv .«o 30m 0000 80 to00=m 0000—0

N. 030,—.

and discouraged were not the most strongly supported themes. Finally, all 28 rows of
positive and negative items were chosen by at least one player.

Visual comparisons across performance and category. In keeping with how
Tenenbaum and colleagues have typically displayed IAPZ results, each player’s unique
probabilistic curves (Figures 11-18 later in text) were combined to produce simpliﬁed
performance zone proﬁles (Figures 19-22 later in text). The disadvantage of the
simpliﬁed ﬁgures is that they do not describe the probability of each performance level
across intensity. Rather, the most probable performance level at each level of intensity is
given absolute recognition. For example, rather than showing a 10% chance of poor
performance, a 40% chance of average performance, and a 50% likelihood of good
performance at a certain level of intensity, that point would be described as being a part
of the player’s good-performance zone. The advantage of these ﬁgures is that they allow
researchers to see a parsimonious comparison of all 10 players on a single ﬁgure.

Probabilistic IAPZ curves. The importance of these ﬁgures is to ﬁrst show that the
study was successful in combining IZOF and IAPZ methodologies to create probabilistic
feeling-performance proﬁles. The second objective of these ﬁgures is to represent the
idiographic nature of these differences.

Figures 1 1-13 are included to show how individual proﬁles can vary in terms of the
location and/or the width of each zone. Figure 11 shows a typical probabilistic proﬁle
with a relatively high (in terms of location of the x-axis) and small (in terms of
bandwidth) average zone. As was often the case in this dissertation, two of the ﬁve
performance curves had very low probabilities (e. g., the curves associated with poor-high

and average-high performance-intensities). Note that the legend describes each curve in

103

Figure 11

Positive-Functional IAPZ Curves for Player 1

«=p(x=1) apnea ——-p(x=3) —-- p(x=4) --p(x=5)
0.9 3
0.8 ~
0.7 .
0.6 -
0.5
0.4 3
0.3 .
0.2 ;
0.1 L.

Probability

 

 

10.5
11

Intensity of Feeling

Note. Curves labeled p(x=l) and p(x=2) represent the probabilities of poor performances
with low and high intensity ratings (respectively). The curves p(x=3) and p(x=4)
represent average performances in the same way; and p(x=5) is the good-perfonnance
curve. The purpose of this curve is to illustrate a typical IAPZ proﬁle. Note that only
three of the ﬁve performance curves are prominent.

statistical notation: with p(x=l) representing poor-low performance, through p(x=5)
representing the probability of good performance. Figure 12 has a much wider and lower
average performance zone. From an applied perspective, this is generally a preferable
proﬁle because the range for poor performance is smaller, the good-zone is larger, and
there is a nice buffer between the two where performances are average—which mean that
the athlete does not drastically slip from good to poor performance. Figure 13 is actually

quite similar to Figure 12. The means for average performance in both ﬁgures are

104

approximately 4.25. If the only method of assessing individual differences were to
compare mean values, the zones would likely be deemed equal. This would clearly be a
wrong assessment. As such, assessing the mean location and width of each zone are

necessary.

Figure 12

Positive-Functional IAPZ Curves for Player 2

==plx=1) - - - - p(x=3) - ° ' p(x=4) —p(x=5)

 

 

9.09
001$

Probability

 

 

 

 

 

Intensity of Feeling

Note. The curve labeled p(x=l) represent the probabilities of poor performance. The
curves p(x=3) and p(x=4) represent average performances with low and high intensity
ratings (respectively); and p(x=5) is the good-performance curve. The purpose of this
curve is to illustrate how ‘normal’ proﬁles can vary from one another. Note that the
location and breadth of each zone varies from the previous proﬁle.

105

Figure 13

Positive-Functional IAPZ Curves for Player 6

=ptx=ll ---- p(x=3) - ~ p(x=4) --p(x=5)
1
0.9
0.8
0.7 l 7
0.6
0.5 .
0.4 ;
0.3
0.2
0.1 Q ~
0 .

OLrgHLqNugr-nvgﬁ-
o H N m

 

Probability

 

 

Ln

m. “’2 ”'2““‘?°°‘-Q°‘U'!° "‘
<- m \D I\ oo 0““ "‘

10.5

Intensity of Feeling

Note. The curve labeled p(x=1) represent the probabilities of poor performance. The
curves p(x=3) and p(x=4) represent average performances with low and high intensity
ratings (respectively); and p(x=5) is the good-performance curve. The purpose of this
curve is to illustrate that idiographic differences can occur with the location (central
tendency) or bandwidth (variance) of zones. As with the previous proﬁle (Figure 13), the
mean of the moderate zone at approximately 4.25, however, the width of this zone is
clearly much narrower.

Figure 14 is interesting because it shows four prominent performance curves rather
than three. In this case, there are two inﬂuential average-performance curves. It is
uncommon to the established literature in this ﬁeld that the probability of good
performance increases through the intensity continuum—yet this result was common for
this study. Figure 14 presents the issue clearly. Rather than the average-high performance

curve being to the right of the good performance curve, it is only barely to the right of the

106

average-low performance curve. There were not signiﬁcant distributions of sub-optimal

performances at high intensity levels of functional proﬁles.

Figure 14

Positive-Functional IAPZ Curves for Player 4

=Plx=1l - ' - ' P(X=3) — ° ' P(X=4) —P(X=5)

 

Probability

 

 

 

intensity of Feeling

Note. The curve labeled p(x=1) represent the probabilities of poor performance. The

curves p(x=3) and p(x=4) represent average performances with low and high intensity
ratings (respectively); and p(x=5) is the good-performance curve. The purpose of this
curve is to illustrate a large bandwidth for average performance, with both high- and low-
intensity average performance curves featuring prominently. It is also an example of how
the high- and low-intensity average-curves are generally in close proximity.

Figures 15 and 16 provide examples of typical dysfunctional proﬁles. The probability
of poor performance increases as feeling intensity increases. Though each of these
Figures has only three prominent performance curves, the inﬂuence of the other two

performance curves is greater than in Figure 11. It is interesting to note, for instance, that

107

the probability of poor performance appears to decrease beyond the intensity of 10.0 in
Figure 15. However, the true probability of poor performance is actually the combination
of poor-low (p(x=1)) and poor-high (p(x=2)) curves. Consequently, the probability of
poor performance at the intensity-level of 11.0 is approximately 81% (0.52 + 0.29). For
applied interests, it should be noted that the bandwidth for poor performance in Figure 16
is very large (approximately 3-11). Keeping negative-dysﬁanctional emotions very low

would be critical for this type of player.

Figure 15

Negative-Dysfunctional IAPZ Curves for Player 3

=ptx=11 =5p(x=2) - - - p(x=3) - ' - p(x=4) -—-p(x=5)

Probability

 

 

 

 

Intensity of Feeling

Note. Curves labeled p(x=1) and p(x=2) represent the probabilities of poor performances
With low and high intensity ratings (respectively). The curves p(x=3) and p(x=4)
represent average performances in the same way; and p(x=5) is the good-performance
Curve. The purpose of this curve is to illustrate good-performance being most probably at
lower intensity levels. There is also an interesting interaction and inﬂuence from all ﬁve
Performance curves.

108

Figure 16

Negative-Dysﬁmctional IAPZ Curves for Player 2

=Plx=ll ==='P(X=2) — - - P(X=3) — ° ° P(X=4) —P(X=5)

Probability
O
4:.
l
l
i
l
l
|
I

 

 

 

 

 

 

 

 

Intensity of Feeling

Note. Curves labeled p(x=1) and p(x=2) represent the probabilities of poor performances
with low and high intensity ratings (respectively). The curves p(x=3) and p(x=4)
represent average performances in the same way; and p(x=5) is the good-performance
curve. The purpose of this curve is to illustrate an example of good performance being
most probably at lower intensity levels. Also note that this proﬁle has a very large
bandwidth for poor performance (i.e., p(x=1)).

Figures 17 and 18 display linear curve proﬁles. It is rarely, but not unheard of for the
IAPZ method to result in linear proﬁles (See Edmonds, et al., 2006). Particularly for
Figure 18, this type of an output had a negative impact on the subsequent accuracy
analysis because Player 1’s zone proﬁle predicted good performance at all levels of
intensity. As such, every observation for Player 1 that was not rated as a good
performance would result in an inaccurate prediction score. Fortunately, there were very

few proﬁles like Figure 18.

109

Figure 17

Negative-Functional IAPZ Curves for Player 3

=p(x=1) ==ptx=21 ---p(x=3) --' p(x=4) ——p(x=5)

Probability

 

 

 

 

 

 

Intensity of Feeling

Note. Curves labeled p(x=1) and p(x=2) represent the probabilities of poor performances
with low and high intensity ratings (respectively). The curves p(x=3) and p(x=4)
represent average performances in the same way; and p(x=5) is the good-performance
curve. The purpose of this curve is to illustrate a rare linear relationship between intensity
and probability for p (x=5). Also note the absence of a poor performance zone.

110

Pig

Po.

pu

1h;

Figure 18

Positive-Dysfunctional IAPZ Curves for Player 1

===p(x=1) ====plx=21 ---p(x=3) - °- p(x=4) —p(x=5)

 

Probability
o
4:.

 

.ﬂ‘LQLnLQOLthqleO‘l-QOLQH
<1- Ln \0 l\ co oF‘SH

1
15
2
2.5,
3.
35

Intensity of Feeling

Note. Curves labeled p(x=1) and p(x=2) represent the probabilities of poor performances
with low and high intensity ratings (respectively). The curves p(x=3) and p(x=4)
represent average performances in the same way; and p(x=5) is the good-performance
curve. The purpose of this curve is to display a unique proﬁle where only one
performance level is predicted across the entire range of intensity. Outcomes are also
linear, rather than curvilinear.

For personal interest, readers can connect the probabilistic proﬁles in Figures “-18
to the zone proﬁles in Figures 19-22. For example, the summary zone proﬁle for Figure
18 can be found at the bottom of Figure 22. These ﬁgures support the ideographic nature
of feeling-performance relationships and the related hypothesis (Hypothesis #1).

Performance zone ﬁgures. Figures 19-22 are more meaningful to the study’s
purpose. The performance zone proﬁle for positive-functional feelings were identical in

that all players went from poor to average to good performance levels as intensity

111

Figure 19

Positive-Functional Affect Performance Zones for Each Participant

 

I Poor Average I Good

#10
#9
#8
#7
#6
#5
#4
#3
#2
#1

 

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

 

 

 

Note. Participants are displayed as rows assigned to the y-axis. The x-axis represents the
intensity of positive-functional feelings. Performance levels are assigned in grey-scale,
with: good-zones in black, poor-zones in dark grey, and average-zones in light grey.

112

Figure 20

Negative-Functional Affect Performance Zones for Each Participant

 

I Poor Average I Good

#10

 

#9

 

#8

 

#7

 

#6

 

 

 

#5

 

.1 .1
fl '1
n h
. 4-
5.1! 14:
a “.—
._.._ .—...,‘<' 31....— .._..__. _..__. _.._ _
. .
-. .
.
‘ .
.. .

 

#4 stirs-472:? " .: ~-~' .' « ~L"."‘zzi ' ---':=::
b . wv . .. A 1‘. VA
.-

 

#3

 

#2 xbmmﬁm

#1

 

Irrrjl‘lYTVTl1YITTTI‘1'YrTFIYUITIT1TYTYW‘TII'YTIYYTYTTTjT]

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

 

 

Note. Participants are displayed as rows assigned to the y-axis. The x-axis represents the
intensity of negative-functional feelings. Performance levels are assigned in grey-scale,
with: good-zones in black, poor-zones in dark grey, and average-zones in light grey.

113

 

Figure 21

Negative-Dysfunctional Affect Performance Zones for Each Participant

 

I Poor Average I Good

. ,...';.~...'. '—. -, ,....', l”__.,,_w__ ,
N‘ ‘) ‘ 1 5'...‘{.::'".3",..‘ ‘;,; ’.:§t_.T‘:-.1.',".s¥ '-...f'.'§.élgg,‘_'.ﬁi'
- .3 ’ , ’ ‘ . .i. ’ . i ‘ l '
l
.
l

#10
#9

.—
_—.—l
l l.
3' _——4 ._
. ...'.
. :,.1
- u
,
i...— __ .—

7 _ J..i._,_.__._.. ..

7 *2}. a- .3 .1 .' .-—- '..

# 1'... . .1 11.7“ *4 '1 ‘- wt. "1
.i I":~‘....‘*r ....:. W' r '7’}? . ‘ NEE-r171; 3.: : .‘.......--l .;<__. . ..----- ...__ -'--=.'- =__,...\ .
# 'l — -“11'§“’“i?“ii~ “rt-5'29: 'a‘iqs.i33"’.t:. "“t 3.9.55"- , ---' "-----. ' ”Vt-1'15"}
-ﬂ.ot_1‘ r. '1. .1... imam “324-: .:.:..+.b ' .. . - -. 1“.

«I Ea'lf-l' " _ ‘ . .
ﬂ 1 Bi? -‘ ._'. ."l..'~‘~~‘.~~‘:-Z';:I:'-_: ~, 1:159-" zit-3;: ...".-_;*.'u.:'a.,"s.'

 

#3 mare-n»... .. _.
#2 -‘.‘ -na 'r’. i. _l ..,. myam were“ ---"-*‘.'i--".-‘.'.':'.?-"-.'r~':-.'.-:¥---'--;-:-.:~z~';h..-.-‘ ‘..'--'~'-.ﬂ" 2:13 §:’-;-L.m‘..v-*-.'...='
o_- . . . ' . , .A : 7 > . 1.. _. - A .. L . .

 

#1 1:33..

I‘1'11‘TI‘IrjtrrYTVII'rVYTjr'r"1'1TrY'Vl‘VTVITWYTITﬁ‘ITj

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

 

 

 

 

Note. Participants are displayed as rows assigned to the y-axis. The x-axis represents the
intensity of negative-dysﬁmctional feelings. Performance levels are assigned in grey-
scale, with: good-zones in black, poor-zones in dark grey, and average-zones in light

grey.

114

Figure 22

Positive-Dysfunctional Affect Performance Zones for Each Participant

 

 

 

 

I Poor Average I Good
#10 h I I I I I l...»- J 4.; i. .J____.-.:J» ..
#9 d I |
#8 4 l _ l ---—

 

 

~ . .. 1.:rL-tﬁeuw~;mw.ﬁsém:n = . my“maria-rz—vauhwanmaa.

 

 

 

"HAHN?! 51.; -
# 5 ifﬂ‘ﬂmas'vt-‘L. . .1”
iii-yh'J-Sv.“ .-~~.: Inﬁ-‘é-ﬂi

 

 

 

 

ITTTTTrTrY[ITY'IITTYIfI'TTIrWTII’ITWTIITTTIrTTIIVYTVIIUIII

0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 11.00

 

 

 

Note. Participants are displayed as rows assigned to the y-axis. The x-axis represents the
intensity of positive-dysfunctional feelings. Performance levels are assigned in grey-
scale, with: good-zones in black, poor-zones in dark grey, and average-zones in light

grey.

increased (Figure 19). In no case was there a “catastrophe effect” wherein the highest
intensity levels were associated with performance decrements. This result is unique from
proﬁles presented in both IZOF and IAPZ literature. Beyond these two similar patterns,
individual proﬁles appear to differ in terms of the location and/or bandwidth of each
zone. However, it is also worth suggesting that some players’ proﬁles are quite similar.
For instance, Player 3 and 4 are alike, as are Players 7, 10, (and to some extent) 5, and 8.
Performance zone proﬁles for negative-functional feelings displayed more

idiosyncratic patterns (Figure 20). Most players’ performances improved as intensity

115

increased. However, Players 1 and 8 only had one performance zone; and Players 3, 5,
and 6 only had two performance zones in their proﬁles. Performance decreased as
intensity increased for Players 5 and 6. This result is atypical because functional feelings
of greater intensity normally improve performance—whether they are positive or
negative in terms of valence. Once again, while most participants’ proﬁles were unique,
Players 7, 9, and 10 were relatively similar.

. Performance zone proﬁles for negative-dysfunctional feelings displayed even more
idiosyncratic patterns (Figure 21). Player 9 had only one performance zone. Players 1, 5,
6, 7, and 8 had only two zones (Players 5’s good-zone is so small that there were
effectively only two likely outcomes). Players 2, 3, 4, and 10 had three zones, and each of
their proﬁles appears to be different. Performance diminished as intensity increased for
most participants. Atypically however for dysfunctional emotions, Players 1 and 7
displayed improved performance over intensity gains.

Performance zone proﬁles for positive-dysfunctional feelings displayed erratically
idiosyncratic patterns (Figure 22). Player 1 and 3 had only one performance zone. Players
4, 6, 7, 8, and 9 had only two zones. Players 2, 5, and 10 had three zones, and each of
their proﬁles appears very different. What makes these results especially peculiar is that
all but two players displayed improved performance with increased intensity of
dysfunctional feelings—including Players 1 and 3 whose lone zones predicted good
performance. Only Players 2 and 10 performed better with lower intensity. Of the four
categories, the positive-dysfunctional set had the most unique—if not strange—results.

Although visual-qualitative comparison of individual proﬁles does not provide

objective statistical comparisons, there are clearly many advantages to this method. In

116

fact, in much of IZOF and IAPZ research, visual comparisons are the lone method of
comparison. Because of the large number of observations collected in this study, it was
possible to complement these qualitative comparisons with the forthcoming statistical
analyses.

Central tendencies across performance. Understanding What the individual
proﬁles look like, and how probabilistic curves proﬁles translate into zone proﬁles, a
quantitative assessment of the location and breadth of zones is now presented. As
described in the introductory paragraph of this section describing interindividual
differences, two statistical tests were conducted in order to assess differences in the
location (central tendency) and width (variance) of zones. The results from these analyses
are presented in Tables 8-19 (See Appendix) and summarized in a later table.

Twelve separate one-way AN OVAs were conducted for each feeling category and
performance level, with feeling intensity as the dependent measure and participants as the
nominal predictor. Each of these ANOVAs revealed signiﬁcant differences between
participants’ mean intensity values. For all 12 analyses, Levene’s Test for homogeneity
of variances was violated (p < .001). The second value listed in the degrees of freedom
for each AN OVA vary considerably between performance-levels for each of the valence-
feeling categories because the number of poor, average, and good performances varied.

Based on these violations of homogeneity of variances, post-hoe comparisons were
made between each participant using Dunnett’s C. This test was selected because it
accounts for unequal variances (violations of homogeneity), unequal samples, and is
more conservative than comparable tests (e. g., Games-Howell) and thus controls the

alpha level more effectively and reduces the chance of Type I error (Pallant, 2009).

117

Furthermore, the critical p-value for variance analyses were adjusted to a more
conservative .01. Forty-ﬁve comparisons were made between the 10 players for each of
the 12 feeling-category and performance level combinations (See Tables 8-19 in the
Appendix for each comparison). In order to assess every possible interindividual
comparison, these large numbers of t-tests were necessary.

Clearly, this many t-tests increases the likelihood of Type I error. It is important to
note that—unlike most post-hoe tests—this dissertation was not concerned with any one
particular t-test. To study the extent of idiosyncratic variability, a general representation
of the proportion (percentage) of signiﬁcant interindividual differences was required. A
Bonfferoni correction for the 45 tests would require an alpha level of < .001 , which was
not practical. After consulting with a statistical advisor from Michigan State University, it
was recommended that a .01 alpha-level and the choice of the Dunnett’s C provided
sufﬁciently reliable post-hoe results. As such, results should be interpreted as a general
index of idiosyncrasy resulting from a balance of pragmatism and conservative rigor.

Performances across positive-functional feelings. A one-way between subjects
analysis of variance was conducted to explore inter-player differences in feeling
intensities during poor performances. Signiﬁcant individual differences in zone-
intensities were identiﬁed at a p < 0.05 level: F (9, 230) = 36.57, p < .001. An identical
AN OVA was conducted for average performances that revealed signiﬁcant differences: F
(9, 432) = 13.41, p < .001. Analysis of good performances also revealed signiﬁcant
individual differences: F (9, 219) = 18.43, p < .001. Mean differences from Dunnett’s C
post-hoe analyses are presented in Tables 8-10 with signiﬁcant values (p < 0.01) in bold.

Of the 45 post-hoc tests for each figure (or AN OVA), there were 23, 19, and 23

118

signiﬁcant comparisons across poor, average, and good performance levels
respectively—meaning that zones were slightly more similar across individuals during
average performance-levels.

Performances across negative-functional feelings. A one-way between subjects
analysis of variance was conducted to explore inter-player differences in feeling
intensities during poor performances. Signiﬁcant individual differences were identiﬁed at
a p < 0.05 level: F (9, 230) = 37.19, p < .001. An identical ANOVA was conducted for
average performances that revealed signiﬁcant differences: F (9, 432) = 38.01, p < .001.
An analysis of good performances also revealed signiﬁcant individual differences: F (9,
219) = 60.96, p < .001. Mean differences from Dunnett’s C post-hoe analyses are
presented in Tables 11-13 with signiﬁcant values (p < 0.01) in bold. Of the 45 post-hoe
tests for each ﬁgure (or AN OVA), there were 26, 24, and 28 signiﬁcant comparisons
across poor, average, and good performance levels respectively.

Performances across negative-dysfunctional feelings. A one-way between subjects
analysis of variance was conducted to explore inter-player differences in feeling
intensities during poor performances. Signiﬁcant individual differences were identiﬁed at
a p < 0.05 level: F (9, 230) = 33.16, p < .001. An identical ANOVA was conducted for
average performances that revealed signiﬁcant differences: F (9, 432) = 27.40, p < .001.
An analysis of good performances also revealed signiﬁcant individual differences: F (9,
218) = 33.55, p < .001. Mean differences from Dunnett’s C post-hoe analyses are
presented in Tables 14-16 with signiﬁcant values (p < 0.01) in bold. Of the 45 post-hoe
tests for each ﬁgure (or ANOVA), there were 21, 27, and 25 signiﬁcant comparisons

across poor, average, and good performance levels respectively.

119

Performances across positive-dysfunctional feelings. A one-way between subjects
analysis of variance was conducted to explore inter-player differences in feeling
intensities during poor performances. Signiﬁcant individual differences were identiﬁed at
a p < 0.05 level: F (9, 230) = 87.59, p < .001. An identical ANOVA was conducted for
average performances that revealed signiﬁcant differences: F (9, 432) = 45.22, p < .001.
An analysis of good performances also revealed signiﬁcant individual differences: F (9,
217) = 28.57, p < .001. Mean differences from Dunnett’s C post-hoc analyses are
presented in Tables 17-19 with signiﬁcant values (p < 0.01) in bold. Of the 45 post-hoe
tests for each ﬁgure (or AN OVA), there were 28, 34, and 26 signiﬁcant comparisons
across poor, average, and good performance levels respectively.

Unlike most research analyses, no particular post-hoe comparison was critical for the
hypotheses. Rather, the sum total of signiﬁcant paired differences was used as an
indicator of how much interindividual variability there was in the location of zones
within the sample. For each of the four feeling categories and three performance levels,
12 separate analyses were conducted between the 10 participants (i.e., 45 interindividual
comparisons per analysis). Of these 540 total comparisons, 304 signiﬁcant differences
were identiﬁed—meaning that 56.3% of the possible comparisons were statistically
unique. Before examining this degree of idiosyncratic difference, differences in the width
of zones must be considered, and are in the following section.

Variance across performance. Before conducting interindividual comparisons to
determine how many individuals differ from one another in terms of the variability (or
bandwidth) of their zones, Levene’s test of homogeneity of variances was assessed. Just

as the F -test of an AN OVA provides an overall assessment of mean differences while

120

accounting for family-wise error, Levene’s test offers a big-picture perspective on the
homogeneity of variance (i.e., that the squared stande deviations of each participant are
statistically equal. The Levene statistics for poor performances across feeling categories
were 4.59 (positive-functional), 8.56 (negative-functional), 4.66 (negative-dysfunctional),
and 15.11 (positive-dysfunctional). The Levene statistics for average performances across
feeling categories were 6.95 (positive-functional), 11.84 (negative-functional), 10.068
(negative-dysﬁmctional), and 15.83 (positive-dysfunctional). The Levene statistics for
good performances across feeling categories were 10.43 (positive-functional), 18.33
(negative-fimctional), 6.41 (negative-dysfunctional), and 9.173 (positive-dysﬁmctional).
All 12 of these tests were signiﬁcant (p < .001 for all), meaning that there were
meaningful differences in variance scores across participants.

Once more, the Stun of Dunnett’s C tests were used as an indicator of general
variability in the size of zones across players. Like the comparisons of central tendencies,
540 comparisons of variance were conducted. Of these comparisons, 289 signiﬁcant
differences were identiﬁed—meaning that 53.5% of the possible comparisons were
statistically unique. In other words, approximately half of the individual zones were
signiﬁcantly bigger or smaller than one another.

Summary of individual differences. Of the possible interindividual comparisons
across feeling categories and performance levels, 76.3% showed signiﬁcant differences in
central tendency, variance, or both (See Table 20). Good performance zones displayed
the most idiosyncratic tendencies (averaging 82.2%), followed by average (75.6%) and
poor performances (71.1%). Zone proﬁles for positive-dysfunctional feelings had the

most signiﬁcant individual differences (83.0%), followed by negative-ﬁmctional (80.7%),

121

negative-dysﬁmctional (73.3%), and positive-functional (68.1%) feelings. In summary,
players displayed greater individual differences across good-performance zones and
reversal categories (positive-dysfunctional and negative-functional combinations).
Likewise, players were more similar to one another for poor-performance zones across

non-reversal feelings.

Table 20

Summary of Individual Differences across Performance and Feeling-Category.

 

 

 

Performance

Feeling Category Poor Average Good Subtotal
Positive- 28 29 35 92
Functional [62.2] [64.4] [77.8] [22.3]
Negative - 36 32 41 109
Functional [80.0] [71.1] [91.1] [26.4]
Negative- 27 36 36 99
Dysfunctional [66.0] [80.0] [80.0] [24.0]
Positive- 37 39 5 36 112
Dysfunctional [82.2] [86.7] [80.0] [27 .2]

128 136 148 412
Subtotal [31.1] [33.0] [35.9] [76.3]

 

Note. This table indicates the degree of interindividual idiosyncrasy in feeling-
performance zones. Values represent the number of inter-individual differences in central
tendency or variance. The total possible number of comparisons was 45 (for non-subtotal
cells), 135 for the feeling-category subtotal column, and 180 for the performance subtotal
row. Values in brackets represent the percentage of signiﬁcant individual differences
based on the potential number for that cell. The percentage value in the bottom-right cell
is based on a total number of 540 comparisons.

122

The performance-feeling grouping with the highest variance was good-zone for
negative-functional feeling (91.1% different). Meaning that, more so that for any other
area of the proﬁles, the tennis players in this sample were most unique from one another
when playing well in terms of negative-functional. The two groupings with the least
number of individual differences were the poor- and average-zones for positive-
functional feelings (with 62.2% and 64.4% differences respectively). Meaning that the
players in this study were most similar to each other in their suboptimal performances in
the positive-functional category.

Given that the number of participants and the limited range of feeling intensities for
zones to distribute across, the fact that 76.3% of interindividual comparisons showed
differences in either the location or the size of zones was surprisingly high. Though
idiographic differences are limited (i.e., there are not 100% differences between each
player), these analyses provide strong support for the notion of idiosyncratic feeling-
performance relationships across different players.

Intraindividual Context Differences

The second major objective of this study was to compare individuals’ proﬁles across
two separate tennis contexts: serving and returning serve. It was hypothesized that there
would be signiﬁcant differences in feeling-performance zones across service and return
of service contexts. To test this statistically, means and variances for feeling intensity
were compared within individuals across category and performance level. Like the
previous comparison between players, zones were compared in terms of the location
(central tendency) and breadth (variability). However, where as the predictor in the

previous analysis had 10 levels (i.e., 10 players), the independent factor in the current

123

analysis of context had only two levels. Accordingly, post-hoc comparisons were not
necessary to determine where differences in signiﬁcant F-tests occurred within the
predictor.

Central tendencies across performance. The ecological design of the study did not
permit for service data to be paired with return of serve data across observations. As
such, while each AN OVA was conducted within subjects, the analysis had to be unpaired
(not a repeated measures analysis). One analysis of variance was conducted for each
participant, performance level, and feeling category. Tables 21-23 present the F -values
for each of the 120 AN OVAs. Signiﬁcant differences (p < .01) between service and
retum zones are expressed in bold.

There was only one signiﬁcant F -test within the positive-functional category; one
signiﬁcant F -test within the negative-functional category; two signiﬁcant F -test within
the negative-dysfunctional category; and two signiﬁcant F -tests within the positive-
dysﬁmctional category. These six signiﬁcant tests represent a mere 5.0% of the total
context comparisons—an especially low proportion given the decision to not make a ﬁll]
Bonfferoni adjustment to the alpha-level (to p < .001). Though not essential for the
objectives of this dissertation, it is worth noting that the six differences were evenly split
among poor and good performances.

Variance across performance. Because the predictor variable had only two levels,
the Levene’s Test to assess family-wise inequality of variances across context was not
necessary. Rather, the same method of comparing variances used in the interindividual
analysis was repeated here. These values are reported in Tables B20-23 as the “s F

Ratio” with signiﬁcant differences (p < 0.01) expressed in bold.

124

Table 21

Intraindividual Comparisons of Central Tendency and Variance F-Ratios across Return
and Serve Contexts for the Positive-Functional Feeling Category

 

 

 

 

 

Poor Average Good
Player Mean F 32 F ratio Mean F 82 F ratio Mean F s2 F ratio

1 0.53 1.88 0.95 1.62 0.53 1.26
2 0.02 1.35 0.00 1.70 0.14 2.86
3 0.07 3.05 0.45 1.26 0.89 1.59
4 0.98 5.51 0.01 1.01 4.83 2.10
5 0.02 1.01 1.26 1.16 0.60 11.99
6 0.04 1.09 0.05 6.47 0.79 1.29
7 0.05 1.64 0.46 2.46 0.04 1.34
8 1.13 1.13 1.20 1.19 0.18 2.20
9 2.98 1.63 0.98 1.45 0.20 4.00
10 0.20 1.14 0.20 1.02 0.04 1.01

 

Note. Mean values are represented by F-Ratios across context. Signiﬁcant values at a p <
.05 level are represented in bold. Variance values are represented by the F-Ratio of
variances in each context. Signiﬁcant values at a p < .01 level are represented in bold.

125

Table 22

Intraindividual Comparisons of Central Tendency and Variance F-Ratios across Return
and Serve Contexts for the Negative-Functional Feeling Category

 

 

 

Poor Average Good
Player Mean F 32 F ratio Mean F s2 F ratio Mean F 52 F ratio

1 0.1 1 2.86 0.70 3.55 0.29 1.62
2 0.31 1.70 0.03 1.07 0.77 2.54
3 0.25 1.14 0.15 1.53 0.56 1.10
4 0.35 3.60 2.29 1.02 2.29 1.45
5 5.20 1.96 1.93 1.46 0.11 4.44
6 0.79 1.04 0.69 1.12 0.99 3.12
7 0.03 1.46 0.07 1.39 0.06 1.65
8 5.22 1.22 3 .67 1.13 3.01 69.02
9 2.95 2.00 0.42 1.62 0.11 0.00
10 0.07 1.07 0.01 1.00 0.05 1.01

 

Note. Mean values are represented by F -Ratios across context. Signiﬁcant values at a p <
.05 level are represented in bold. Variance values are represented by the F-Ratio of
variances in each context. Signiﬁcant values at a p < .01 level are represented in bold.

126

Table 23

Intraindividual Comparisons of Central Tendency and Variance F-Ratios across Return
and Serve Contexts for the Negative-Dysﬁmctional Feeling Category

 

 

 

 

Poor Average Good
Player Mean F s2 F ratio Mean F 32 F ratio Mean F 32 F ratio
1 0.02 2.54 0.01 1.74 0.18 2.57
2 0.00 1.02 0.57 1.69 0.29 1.09
3 6.80 19.28 3.92 1.85 0.00 1.05
4 1.38 2.50 0.47 1.75 1.46 2.63
5 1.05 1.79 2.90 1.40 0.79 1.55
6 0.14 3.60 0.04 2.59 0.36 2.14
7 0.09 1.01 0.17 1.22 1.44 1.13
8 0.03 1.35 1.93 3.57 0.01 1.87
9 1.74 2.96 0.02 1.23 0.05 1.86
10 0.06 1.05 0.12 1.04 0.00 1.00

 

Note. Mean values are represented by F-Ratios across context. Signiﬁcant values at a p <
.05 level are represented in bold. Variance values are represented by the F-Ratio of
variances in each context. Signiﬁcant values at a p < .01 level are represented in bold.

127

Table 24

Intraindividual Comparisons of Central Tendency and Variance F-Ratios across Return
and Serve Contexts for the Positive-Dysfunctional Feeling Category

 

 

 

Poor Average Good
Player Mean F 32 F ratio Mean F s2 F ratio Mean F s2 F ratio
1 0.86 1.24 0.60 1.59 0.00 2.29
2 0.41 1.48 0.30 1.46 0.51 3.36
3 0.01 8.17 0.55 1.29 5.23 1.72
4 1.07 2.25 0.01 1.84 1.41 1.77
5 1.82 8.08 0.51 1.19 0.08 1.04
6 4.50 1.39 0.66 3.41 3.55 1.56
7 0.58 11.00 0.56 2.20 0.00 0.00
8 1.92 24.68 1.41 1.97 0.35 4.43
9 0.34 1.25 2.95 1.48 0.20 4.00
10 0.32 1.13 0.00 1.01 0.54 1.16

 

Note. Mean values are represented by F-Ratios across context. Signiﬁcant values at a p <
.05 level are represented in bold. Variance values are represented by the F-Ratio of
variances in each context. Signiﬁcant values at a p < .01 level are represented in bold.

There were only two signiﬁcant F -tests within the positive-ﬁmctional category; three
signiﬁcant F -tests within the negative-functional category; one signiﬁcant F -test within
the negative-dysfunctional category; and four signiﬁcant F -tests within the positive-
dysfunctional category. These 10 signiﬁcant tests represent a mere 8.33% of the total
context comparisons. Though not essential for the objectives of this dissertation, it is
worth noting that four differences existed for poor performances, three for average

performances, and three for good performance levels.

128

Summary of context differences. Especially compared to the interindividual
analysis that explored the degree to which idiosyncratic differences existed in
performance-feeling proﬁles, the current analysis of context eﬁects showed very little
variability. While 76.3% of individuals’ zones were different, only 15 out of the 120
context comparisons (12.5%) were different. This result does not support the notion that
individuals have unique feeling-performance proﬁles for serving compared to returning
service. Players’ zones were generally stable across each tennis context.

Though differences were few, poor (17.5%) and good (15.0%) performance zones
were more variable than average (7.5%) performance zones. With only 16 total
differences, each player could be said to have 1.6 differences across the 12 comparisons.
Based on that average, Player 3 (three context differences) and Player 8 (four context
differences) had relatively more differences. As such, it may be possible that the context
effect was at least somewhat applicable for two of the ten tennis players in this study.
Accuracy of Discrete Categories versus Trigger Items

The third major objective of this dissertation was to determine if a small number of
critical trigger items (feelings) could be as effective as a full proﬁle of 12-13 items
representing four fundamental categories of feeling. To answer this question, the
accuracy of eight variables (four feeling categories and four trigger items) were
determined and compared. After conducting OLRs on all eight variables, parameters
were identiﬁed for each performance level. These performance zones were used to
predict performance ratings. Accuracy was determined by comparing the predicted level
of performance to the actual performance rating for each observation. The process of

detennining accuracy and comparisons of accuracy results are presented below.

129

Creating performance prediction variables. Performance predictions were
generated from the OLR curves and zone bandwidths depicted graphically in Figures 19-
22. The process of establishing this data was described earlier in the section titled
“Treatment of Raw Data: Ordinal logistic regressions.” Continuing from where this
section ﬁnished, eight new variables (columns) were created in the SPSS ﬁle. These
variables combined the bandwidth parameters from each player’s four feeling categories
and four trigger items with the actual aggregated intensity value in each category and
trigger. From this process, performance predictions were established for each
observation. For example, Player 5’s zones for positive-functional feelings were poor
(0.0-3.70), average (3.71-4.87), and good (4.88-11.0). If the average positive-functional
(PF) intensity rating was 3.0 for a certain observation, that observation’s PF performance
prediction would be ‘poor,’ if the aggregated feeling intensity was 4.5, the prediction
would be for ‘average’ performance, and ‘good’ performance would be predicted if the
rating was a 5.0 or a 10.0, etc.

Determining accuracy of OLR-based zones. Determining accuracy was a simple
process of comparing the predicted performance values in each of the eight columns to
the actual performance level for that observation. Continuing from the previous example,
If the actual performance rating was ‘average,’ than only positive-functional scores
between 3.71 and 4.87 would be coded as accurate predictions. When the predicted and
actual performance levels were identical, the accuracy column for that category/trigger
was coded as a 1 (accurate), if not, it was a 0 (inaccurate).

Comparing accuracy results. The accuracy of the traditional feeling categories and

the idiographic trigger items are described in Table 25. The analyses of accuracy ratings

130

Table 25

Accuracy of Performance Predictions for Feeling Categories and Trigger Items

 

 

Category / Item Rank Accuracy Z value Signiﬁcance
Positive-Functional 1 73.2% 1 .84 0.066
Trigger 2 1 71.4% 0.08 0.938
Trigger 1 1 71.1% 2.19 0.029
Trigger 3 4 67.6% 2.85 0.004
Negative-Functional 5 63.4% 2.09 0.037
Negative-Dysfunctional 6 60.2% 2.18 0.030
Positive-Dysfunctional 7 56.9% 1 .03 0.303
Trigger 4 7 55.7% - -

 

Note. ‘2 value’ and ‘Signiﬁcance’ columns describe the comparisons between that row
and the subsequence row in the rank-ordered list based on Wilcoxon Signed Rank Test.
Signiﬁcant differences at the p < .05 level are in bold.

were one of the few statistical tests that were conducted across all participants
simultaneously. While each player had unique feeling items and emotion-performance
proﬁles, the accuracy of the zone-based predictions could be collapsed across individuals.
Because the outcome variable in this analysis was nominal with only two possible
outcomes (accurate or not), a non-parametric equivalent of a one-way repeated measures
AN OVA was used (Pallant, 2009). This non-parametric option is called a Friedman Test.
Because each participant’s observations were combined, and non-parametric analyses are
conditional on the assumption of independent observations, it was very important to use
the Friedman Test rather than the Kruskal-Wallis between-groups test. Table 25 depicts

the eight feeling categories and items, their ranking, percentage accuracy, Z score for the

131

difference between each row and the row below it (less accurate), and the p-value for the
paired Z score comparison at the critical alpha level of .05.

The results of the Friedman Test indicate signiﬁcant differences in the accuracy
scores of the eight variables, x2 (7, n = 903) = 247.060, p < .001. Just as a signiﬁcant F -
Test in an AN OVA justiﬁes post-hoe t-tests, the non-parametric equivalent tests (the
Wilcoxon Signed Rank Test) were conducted. Seven paired comparisons were made
along the ranked order of the eight variables, such that the most accurate variable was
compared to the second most, and the second most was compared to the third most
accurate, etc. As Table 25 shows, the three most accurate variables were statistically
equal ranging from 71.1% to 73.2% accurate. To explain each row, the positive-
ﬁmctional category was accurate in predicting performance 73.2% of the time, but not
statistically more accurate than the Trigger 2 variable (71.4%), 2 = 1.84, p = .066, at a
critical alpha level of 0.05. As Table 25 shows, Positive-Functional feelings, Trigger 2,
and Trigger 1 were the most accurate; Trigger 3 (67.6%) was next; Negative-Functional
feelings (63.4%) were less accurate; Negative-Dysfunctional feelings (60.2%) were less
accurate; and Positive-Dysfunctional feelings (56.9%) and Trigger 4 (55.7%) were
equally least accurate.

Simply stated, the trigger items were signiﬁcantly more accurate than the aggregated
traditional feeling categories. Three of the top four most accurate variables were trigger
items. It should be noted that functional categories were more accurate than dysfunctional
ones; and that most individuals would probably beneﬁt from three, rather than four,

critical trigger items.

132

CHAPTER FIVE
DISCUSSION

This investigation was designed to assess the effects of feeling states on performance
during competitive tennis matches for the purpose of contrasting and integrating the
IZOF and IAPZ models of emotions and performance. Secondly, intraindividual
contextual differences in performance—zones were compared within individuals. Thirdly,
four discrete ‘trigger feelings’ were compared to a complete proﬁle with four valence-
function categories of feeling to determine which predicted performance most accurately.
Each hypothesis is brieﬂy reviewed in the following paragraph, and expanded upon
through the discussion chapter.

The ﬁrst hypothesis predicted that discrete feelings organized into four valence-
function categories would form idiosyncratic feeling-performance proﬁles that accurately
predict performance. This hypothesis was supported by the results. The second
hypothesis predicted that idiographic proﬁles would vary between serve and return of
serve contexts for each player. The results failed to support this hypothesis. The ﬁnal
hypothesis proposed that the simpliﬁed “trigger” items would be comparable to a
complete proﬁle in terms of predictive ability. This assumption was not only met but was
clearly exceeded, based on the results of this study.

Integrating Elements of IZOF with IAPZ

The essential objective of this dissertation was to test whether or not valuable
elements of the IZOF model could be integrated with the IAPZ method. Based on the
successful production of probabilistic idiographic curves and the reasonable accuracy

performance predictions, it appeared ﬁom this study that Tenenbaum and colleagues’

133

IAPZ model (Johnson, et al., 2009; Kamata, et al., 2002; Tenebaum, et al., 2002;
Tenenbaum, et al., 2008) can accommodate discrete feeling states. Results supporting this
conclusion are presented.

Establishing accurate probabilistic curves. Based on the results of this study, OLR
analysis was clearly capable of identifying performance zones from aggregated groups of
discrete feelings (See example Figures 6-13). Overall, the zones depicted in Figures 19-
22 were 65.0% accurate (valence-function categories were 63.5%; and zones for trigger
items were 66.4%). Accuracy results were more than twice the likelihood of chance (i.e.,
33.3% based on three possible outcomes), and therefore considered to be strong. Ideally,
these accuracy scores would be compared to accuracy reported in previous IAPZ
research. However—with the exception of the introductory paper, by Tenebaum et al.
(2002), which used a theoretical data set to compare the accuracy of the proposed IAPZ
model to the established IZOF model—no subsequent research has reported the accuracy
of predictive proﬁles. It would beneﬁt this ﬁeld of study to consistently report the
accuracy of the proﬁles to predict performance.

Individual differences. The identiﬁcation of individual differences offered further
statistical support that discrete feeling items can be used with IAPZ methods to develop
probabilistic models. Both qualitative and quantitative support for the idiosyncratic
nature of feeling-performance relationships were presented. Statistical comparisons of
zone locations and bandwidths supported traditional visual comparisons (See Figures 19-
22). This is an important result because IAPZ and IZOF studies have often relied on
visual comparisons without producing statistical evidence in support. While idiographic

differences were identiﬁed, these differences should not be overstated. As the description

134

of Figures 19-22 showed, some players had effectively similar proﬁles. There certainly
was not complete, 100% uniqueness across participants. In the same way that the
Transtheoretical model of physical activity groups people into ﬁve motivational
categories (e. g., pre-contemplative, maintenance), perhaps athletes’ feeling-performance
relationships can be grouped into a parsimonious number. As robust as the visual and
statistical evidence was, differences in individual selection of items appeared to be the
greatest source of idiosyncratic differences.

Discrete items. Hanin (2000c, 2007), Lazarus (1991, 2000b), Pekrun (1992, 2000;
Pekrun, Elliot, & Maier, 2006) and others have argued that discrete emotions (e.g.,
anxiety, relaxation, anger, etc.) that are grouped by similar valence and ﬁmctionality
(e.g., as being either helpful or harmful to performance) act in similar ways. They further
contend that different valence-function groupings have unique mechanistic effects on
functioning (i.e., see Hanin’s Energy Mobilization-Utilization hypothesis described in
Chapter 2). As such, there are valuable theoretical reasons to study discrete feelings that
can be organized into valence and function categories.

There are also practical beneﬁts to using discrete emotions, rather than measuring
affect using the continuous, two-dimensional construct of the affect grid (Russell, et al.,
1989). People do not identify their feelings in terms of continuous valence and arousal /
intensity. They can, but they do not do it naturally. Athletes likely relate better to an
individual proﬁle with personalized, familiar, and meaningful items like conﬁdence,
relaxed, aggressive, fast, than they do to describing the hedonic tone of their state.

In addition to being more tangible and meaningful to the athlete, discrete items

provide another advantage to the traditional IAPZ method. No IAPZ study published up

135

to this point has developed proﬁles based on integrated valance-functionality feeling
states. In fact, only two studies have even assessed functionality of feelings (See, Cohen,
et al., 2006; Golden, et al., 2004). In other words, IAPZ studies have developed separate
individual proﬁles for valance ratings, activation ratings, and functionality ratings——but
have not integrated these measures. By failing to combine measures, IAPZ research has
not compared reversals (positive-dysfunction and negative functional categories).
Reversals have provided important sources of idiosyncratic feeling-performance
information. Most noticeably for example, practitioners now appreciate the beneﬁts of at
least moderate levels of negative feelings like nervousness and aggression; while being
mindful of the potential detriments of positive states (e.g., being too relaxed or
overjoyed). Hence, negative feelings in particular should always be organized into
helpful and harmful subclasses.
Intraindividual Context Differences

With only 12.5% of the context comparisons reaching statistical signiﬁcance, the data
from this study fails to support the notion that individuals have. unique zones of
functioning (or IAPZs) across the context dimension of the psychobiosocial model in
terms of serving and returning serve. This ﬁnding supports the idea that tennis players
want to remain centered during matches—with a consistent, stable focus. Aside from
each player’s ﬁrst shot of the point (the actual serve or return), both players are required
to execute the same skills: groundstrokes and volleys. In other words, beyond the serve or
return that starts the point, rallies require the same skills from each player. The

differences between serving and returning serve are not as physically or mentally unique

136

as the differences in playing offense and defense in basketball or hockey, for instance.
The differences in task demands between serve and return of serve may be quite subtle.

It is possible that there are differences—but that they are quite subtle, and that they
may exist for only some tennis players. Because each feeling item was aggregated into
one of four feeling categories, the precise impact of each item (with the exception of the
trigger items) was not identiﬁed. Here is an example of how the limited sensitivity of the
aggregated analysis might hide a subtle effect. Player 8 identiﬁed 12 discrete items in his
proﬁle, including four negative-dysfunctional items. Of these four, three may have
identical effects when serving as when returning serve. However, the fourth feeling
(indecisive) may be critically important for serving well. The differences in optimal
zones for this one item across context could be hidden by the statistical ‘noise’ created by
the other three items. Likewise, an item like ‘determination’ might be most relevant when
returning serve—which is generally a more defensive posture—while other positive-
functional feelings may not vary across context. This hypothesis could be tested by using
the same method to analyze the four feeling-categories; but doing so separately for each
discrete item. This hypothesis and potential analysis are, however, beyond the objectives
of this study.

Based on their psychological make up or playing style, some players might ﬁnd that a
shiﬁ in their zones or their mental focus will beneﬁt their performance as they transition
from dictating the point with their serve to counter-punching and defending on return.
However, based on the results of this study, coaches and sport science practitioners may
consider each task to require the same mental/emotional approach. However, other types

of contextual comparisons are possible within the psychobiosocial framework (Hanin,

137

2000), such as being down/up a break (the score), their previous record against the
current opponent, or the importance of the match. These comparisons may yield
substantive ﬁndings.

Simpliﬁed Feeling-Performance Profiles

In developing scientiﬁc models, there is a constant struggle to balance
comprehensiveness with parsimony. In structural equation modeling, factor analysis, and
other statistical methods, researchers have to make somewhat arbitrary decisions about
how many factors should be included in a model. A researcher may ﬁnd that six things
are statistically signiﬁcant, but that only three of them are substantively meaningful. Does
that researcher present a more complete model, or a more succinct, powerful model?

The results ﬁom this dissertation clearly argue that most tennis players can self-
identify three critical trigger items that have a strong inﬂuence on their performance.
Moreover, these three items are more accurate at predicting performance than the
complete proﬁle with four valence-function categories (e. g., 70.0 % compared to 63.5%).
With the exception of the positive-functional category, the top three trigger items were
more accurate than all of the other feeling categories. This result was surprising. It was
expected that accuracy would have to be sacriﬁced in order to simplify the model. This
was not the case: a simple proﬁle with three or four discrete, self-selected items is more
accurate than completing a 12-item proﬁle. This result is particularly relevant for
consultants, coaches, and athletes—as well as for scientists interested in less intrusive

methods of collecting data.

138

It is worth noting that ‘complete’ proﬁles in this study included 12-13 items rather
than the typical 16-20 originally proposed by Hanin (2000). An even bigger proﬁle would
likely include some less impactful items, and therefore be even less accurate overall.

The value of categories. The results from this study should not be considered as an
exclusive endorsement of simple trigger-item performance modeling. There is
considerable value in more comprehensive proﬁles. Firstly, although the trigger items
were statistically more accurate, the four categories from the full proﬁle were not
drastically less accurate. Based on the values in Table 25, the average accuracy of trigger
items was 66.45% and the mean accuracy of the feeling categories was 63.43%.

More importantly, a complete proﬁle with valence-ﬁrnction categories presents
theoretical value for both practitioners and to researchers. Both researchers and
practitioners can beneﬁt from being able to organize emotions and other feeling states
into valence-functionality categories. Feelings have different performance effects based
on their pleasant/unpleasant (or attraction/avoidance) and functional/dysfunctional nature.
As such, negative-dysfunctional feelings may be more important when studying athletes
in slumps; and positive-dysfunctional feelings may be more important when examining
periodized training—two issues not examined in this study (See, Hanin, 2002).

Considering both functional and dysfunctional feelings is beneﬁcial because coaches
and consultants are able to identify what players should strive towards, as well as what
they should avoid—cg, barriers to optimal performance. While it is well known that
athletes should strive to achieve speciﬁc goals, or ideal performance states (IPS; Orlick,
2008), the ability to be aware of and to avoid barriers to self-actualized performance is

foundational to humanistic consulting paradigms (Hill, 2001; Rogers, 1980). For

139

instance, as much as Jean may want to be conﬁdent, calm, and aggressive, she may not
want to be irritated and worried. Developing strategies and routines that prevent irritation
and worry may be more helpful to Jean than only striving to be conﬁdent, etc. If the
theoretical and practical beneﬁts of organizing feelings into valence- and functionality-
based categories are important to a scientiﬁc study or to an applied intervention, then
trigger items, alone, may not sufﬁce.

With all of those points in mind, the results of this study suggest that all categories
are not equal. Some categories appear to inﬂuence performance more strongly or more
clearly than others. For instance, positive-functional feelings were less idiosyncratically
unique across participants and were more accurate at modeling (predicting) performance
ratings. On the contrary, positive-dysfunctional feelings were relatively poor predictors of
performance. It could be concluded that, for some participants in this study, positive-
dysfunctional feelings were not relevant or helpful to their idiographic proﬁle.

An interesting question might be to investigate if the trigger items selected by each
player were, in fact, their most accurate discrete predictors. Although the trigger items
were more accurate than the complete proﬁle, they were not necessarily the most accurate
(i.e., the most effective predictor) of the 12 self-selected items. This could be determined
by conducting separate OLRs on each of the 12 items in each player’s proﬁle,
establishing the bandwidth zones, and examining the accuracy of each zone. Again,
however, such an analysis was not necessary for the objectives of this dissertation.
Strengths and Limitations

Strengths of the study. Before discussing the critical limitations of this dissertation,

it is worth acknowledging its strengths. Most importantly, the data collected was reliable.

140

There were times when a player’s frustration made completing the forms impossible, and
in those rare cases, the participant’s frustration was always respected. The rapport
between the researcher and player-participants helped to ensure data of the highest
quality. It is also important to the study that an exceptionally. large number of
observations were collected.

Finally, the theoretical and practical foundation for the study seems to have been
supported. For instance, understanding relevant theories (i.e., IAPZ, IZOF, CMRT, etc)
led to the integration of probabilistic analysis (OLR) with discrete items that could be
combined into valence-function categories. The researcher’s practical experience in
applying the IZOF model was foundational to the interest in more parsimonious trigger-
item methods. The real value of this dissertation is in the contributions in bridging the
IZOF and IAPZ frameworks. From an IZOF perspective, the study offers a very large
dataset, assessments taken during performances rather than upon reﬂection, better
statistical analyses (including the use of OLR), and the alternative of simpliﬁed trigger-
item proﬁling. From an IAPZ perspective, it is clear that discrete items from the PNA-77
could be used as an alternative to the Affect Grid. Perhaps most importantly of all,
interactions of valence and function were studied, which recognized the effects of
reversals (e.g., the different effects of negative-helpful and negative-harmful feelings).
The study also provides another tennis context to the IAPZ literature (which is more
transferable than the laboratory-based and aiming sport contexts used in nearly all IAPZ
research.

Limitations. All forms were completed during intra-squad challenge matches. These

matches were very competitive—more competitive than practices, and many exhibition

141

matches. However, challenge matches were likely not as competitive as conference
matches or major tournaments. This likely reduced the frequency of very extreme feeling
intensities. Though there inevitably would have been more intensity, it is impossible to
know if the relationship between performance and intensity of feelings would have
changed. The intensity of feelings was also less extreme because of the culture of
tennis—which participants had been socialized in for a long time. In tennis, extreme
displays of emotion are frowned upon. Compared to hockey, basketball, football, and
many other sports, tennis players generally strive to be more centered. That said, tennis
players generally express more emotional intensity than athletes in aiming sports such as
curling or small bore, as well as some track and ﬁeld events.

The design of this study generated an unavoidable degree of dependency within the
dataset. Speciﬁcally, since both players in each match completed forms, the performance
and feelings of one participant were interrelated to the other participant’s data in that
match. If, for instance, if Bob was to play a near-perfect game with intense emotion,
Dylan (his opponent) is more likely to report poor performance, and his feelings may
have been inﬂuenced by Bob’s intensity. With this limitation in mind, there is no reason
to think that interindividual dependency would inﬂuence the relationship between
feelings and performance within individuals. Additionally, while there is total
dependence for variables like outcome (i.e., if one player wins, the other must lose),
outcomes were not considered in the analysis of this dissertation—performance was
analyzed. Performance ratings were thoroughly described for participants. This measure
enabled participants to focus on process, rather than results, which is inﬂuenced less by

what one’s opponent is doing. Also, since participants often had two games to chose from

142

when completing the monitoring form, each player in a match did not necessarily chose
the same game. This would have also reduced the dependency within the dataset.

Another lirnitation—or more so, a difference from previous literature—is that ‘good’
performances were analyzed rather than ‘best-ever,’ ‘optimal,’ or ‘excellent’ levels of
functioning. Comparing good performances with poor and average performances is less
drastic than if the study had examined more extreme performance levels. If a greater
frequency of very high performance ratings had been collected, a separate analysis could
have grouped performances differently. Perhaps, for instance, four performance levels
could have been studied (e.g., poor, average, good, and great). Such a nuanced
breakdown of performance has not been conducted in previous studies.

Finally, while the participants in this study qualify as expert tennis players with more
than 10 years of experience in the sport, they are not world-class athletes. Hanin (2000a,
2007) has argued that emotion-performance modeling is more effective with more
experienced, self-aware, ultra-elite athletes. Relative to many of the IZOF and IAPZ
studies in the ﬁeld, the current sample could be considered less elite. A more elite sample
might have improved the accuracy of the models, and provided more very high
performance ratings.

Dual-role: investigator and consultant. The primary investigator and author of this
dissertation worked with most of the participants in this study as the tennis team’s mental
training consultant for two years prior to data collection (the study began in the author’s
third year with the team). Over the years, a variety of sport psychology topics were
discussed and skills were developed. It should be noted that none of the participants had

previous experience in developing feeling-performance proﬁles. It is, nevertheless,

143

I

7

important to acknowledge the duel-role that the investigator held in this study. The
rapport between the players and the consultant/investi gator certainly improved the buy-in
and dedication of the participants. However—in spite of instructions by the investigator
to provide the most accurate and honest responses possible——the duel-role of the
investigator may have also increased the chances of socially desirable responses.

In keeping with the integrated, applied-research, nature of this study, it is interesting
to note that the process of collecting data also acted as a sort of a sport psychology
intervention for many players. Upon completion of the study, most of the participants
explained that completing the forms (i.e., monitoring their performance and their
feelings) forced them to be more emotionally self-aware. For all but one player, this was
a beneﬁcial thing, as it helped them to correct harmful feelings and to focus more on
helpful feelings. From a research perspective, this confound is neither rare, nor is it
overwhelmingly problematic. Measuring variables—whether by interview, psychometric
assessment, force plate, or respirometer—oﬁen has some affect on the participant. In the
case of this dissertation, this inﬂuence was unavoidable. However, it should be
acknowledged that monitoring feelings and performance during matches may have
inﬂuenced the actual feelings and performance of players during those matches. Future
studies should assess participants’ emotional regulation skills and emotional intelligence
before and after the monitoring process.

Future Research Directions

Future analyses of this data set. The data collected for this dissertation has the

potential for a great deal of further analysis. Extensive ‘data mining’ provides a dubious

foundation for the development of generalizable theories due to the possibility of

144

capitalizing on chance. This study offers a unique sample size that is sufﬁcient for more
elaborate statistical analyses. In addition to the future directions identiﬁed throughout the
discussion chapter, other analyses with the current data set are described below.

For instance, it may be possible to nest feelings states within players or within
valence-function categories using multilevel models or latent variable models.
Additionally, researchers interested in momentum could conduct time series analyses
within matches (combining participants). Does being in-zone in one game inﬂuence the
probability of being in-zone or of performing well in the subsequent game? How long
does being in-ﬂow, in-zone, or performing exceptionally well, have a positive inﬂuence
on subsequent performances?

Most of all, because each observation assessed challenge and skill ratings (standard
indices of challenge-skill balance (See, Csikszentrnihalyi & Larson, 1987;
Csikszentrnihalyi, et al., 1993; Jackson & Csikszentrnihalyi, 1999)) a number of ﬂow-
related studies could be conducted. This could include a direct comparison of how feeling
states and ﬂow states inﬂuence performance. For example, which predicts performance
better, being in-ﬂow, or being in-zone. Comparing effect sizes of ﬂow and feeling data
would be particularly interesting from a statistical perspective. Another important line of
inquiry includes the relationship between ﬂow and affect. Based in Csikszentrnihalyi and
colleague’s (Csikszentrnihalyi, 1994, 1997; Csikszentrnihalyi & LeFevre, 1989; Jackson
& Csikszentrnihalyi, 1999) descriptions of ﬂow, it is clear that ﬂow is a positive state that
may be associated with performance, but is not required to. F low is a state of optimal
experience, not optimal functioning—though Jackson seems to favor the performance

beneﬁts of ﬂow more than Csikszentrnihalyi (Gould, Eklund, & Jackson, 1992; Jackson,

145

2000; Jackson & Kimiecik, 2008; Jackson & Roberts, 1992). Consequently, ﬂow should
correlate: positively with positive valence; negatively with negative valence; and the
relationship with valence should not vary based on functionality of feelings. In short, if
ﬂow is a positive, optimal experience and not about functioning, ﬂow should correlate
more strongly with valence than function.

Future research beyond this study. It would be interesting to study context effects
in a sport where the contrast between contexts are greater than in tennis. Offense-defense
comparisons in basketball, or comparing high school football players who play multiple
positions would make for excellent samples. However it is executed, more should be
known about the context dimension of the psychobiosocial state.

Nearly all studies published in the area of emotion-performance relationships have
focused on establishing zones and comparing the size and location of zones. While
valuable, these questions may be ignoring much more important questions about how the
zones are controlled. Researchers ask, ‘how big is the zone?’ but they should also be
asking: ‘how often are you in-zone?’ and ‘how much control do you have over each
feeling?’ Does it matter how small the zone is if the player is able to get into that zone
almost all of the time? How much control an athlete has over each feeling in his or her
proﬁle may be more important in designing an intervention or psychological skills
training program. From an applied perspective, athletes would beneﬁt most from research
that can: (a) determine which items are most inﬂuential; (b) how much control athletes
have over those items; (c) and then developing skills and routines that facilitate self-

regulation of these critical performance factors.

146

Future publications in this area of research should consider the beneﬁts of
establishing standard descriptive statistics for IZOF/IAPZ models. Such data might
include the distribution of performance ratings in either scale or ordinal form. Another
valuable statistic would be the percentage of signiﬁcant interindividual differences in
zone shapes (i.e., intensity and size). This statistic would offer a general index of the
degree to which idiographic differences exist. Finally, it is critically important for this
ﬁeld of research to develop individual models that accurately predict performance. There
is little value in establishing individual zones if the proﬁles cannot accurately predict
performance. To this end, it is surprising and frustrating that the accuracy statistics from
this dissertation cannot be compared to other published studies. Such a statistic would
likely over-estimate the accuracy of each proﬁle because the zones were estimated from
the same sample of performances that the accuracy of the model was tested against.
Regardless, this is an important statistic, which helps to validate the models in each
study. In future studies, such a capitalization on chance could be assessed by developing
individual proﬁles from a series of initial performances before testing the accuracy of the
zones against a series of subsequent performances for each participant.

Applied Implications

The tennis laboratory. Tennis is an excellent sport to conduct research in. It requires
both athleticism and skill. There are long intervals of play (tennis points are longer than
the average play in football and baseball, or a shift in hockey), followed by short breaks
between points and longer breaks between games and sets. It is an individual sport, so
participants’ performance levels and feelings are more “controlled for” than in team

sports. Although tennis is not as “emotionally intensely” as rugby, hockey, football, and

147

many other sports, it is more emotional than typical aiming sports, track and ﬁeld, etc.
Also, tennis is a long rollercoaster. A player can be performing at his absolute best in one
set, and then drop the next set at love (0-6). A player may go from ﬂow to frustration in a
matter of just one or two games. Accordingly, tennis is a sport that provides an excellent
laboratory for feeling-performance research.

However, if a researcher has sufﬁcient buy-in from a team, many other sports could
provide unique insight into the relationship between feeling states and athletic
functioning. For instance, hockey players could monitor performance and feelings
between shifts; football players could do so between possessions; and volleyball and
basketball players could complete forms while sitting on the bench after substitutions.
More researched should be conducted in a variety of sports. Such information might
challenge existing theories of emotion and feeling in athletic competition.

What these results mean for practitioners. What, do the results from this
dissertation mean for coaches, athletes, and sport science consultants? Feeling-
perforrnance proﬁling is a fascinating area of research because it is highly theoretical,
complex, and it requires advanced statistical analyses—yet it also has enormous practical
implications and must be disseminated in basic, straightforward terms for practitioners.
Four very important results could be disseminated to coaches, athletes, parents, and
administrators such as coaching educators and league officials.

Idiosyncrasy. First, there is further support for the essentially unequivocal ﬁnding
that individuals differ in their feeling-performance proﬁles. Not only do the size and
intensities of idiographic zones vary, but so to do the choice of discrete feeling items. In

fact, individuals may vary more in terms of the factors that they feel are most inﬂuential

148

to their performance, than they do in the shape of their idiographic proﬁles. Everyone in
sport should understand that emotions are not barriers to optimal performance: they are
essential for optimal performance. Athletes are better served by understanding and
working with their feelings than from trying to ignore them.

Discrete feelings. In addition to being an important result for researchers, the fact that
discrete feelings can be used to develop probabilistic proﬁles is important in a practical
sense. First, discrete descriptions of feeling states are more personally meaningful to
people than valence and arousal ratings. Second, the development of probabilistic proﬁles
has a very simple and profoundly important meaning. Though Hanin has always
emphasized such probabilistic relationships, his proﬁles and in-/out-of zone hypothesis
reinforce a more black-and-white relationship between zones and performance. The
beauty of probabilistic proﬁling is that it promotes an important truth for practitioners:
preparation gives athletes a better chance to perform well—but nothing can be certain. As
such, the more we do to prepare and be ‘in zone,’ the more likely we are to succeed.

Can an athlete have multiple zone proﬁles? The results from this study clearly
support the notion that athletes have similar proﬁles for when they serve compared to
when they return serve. Tennis players want to remain in a fairly stable zone during a
match. However, athletes from all sports should identify when they may beneﬁt from
adjusting their feelings, emotional constellations, or emotional thermostats. Are their
times when a player should be a little more aggressive, try to be extra-conﬁdent, or very
relaxed? Basketball provides a clear and indisputable example. Even the most intense
players must shift their attitude and feeling while shooting free-throws. A football player

might beneﬁt from breaking the huddle and surveying the defense (before the start of the

149

play) in a more relaxed state so as to make the right reads and decisions—then raise his
intensity during the play. Athletes should not only know what feelings comprise their
individual proﬁles, but also know how to adjust their feelings in certain situations. To
some degree, athletes may need different attitudes for different contexts (e. g., playoffs
versus regular season, rivalry games, starts of matches versus the second half).

Keep it simple. The ﬁnal implication for practitioners in sport is to keep it simple.
Based on the results of this study, a combination of three personally meaningful, athlete-
selected, discrete feelings will more accurately predict performance than a compete
Individual Emotion Proﬁle. Coaches are responsible for an incredible number of things.
They are stretched in every possible direction by parents, administrators, sport scientists,
and most of all, by their players. As such, sport science consultants must be careful not to
overwhelm coaches with every possible sport psych, biomechanics, or nutritional topic
on earth. Emotional proﬁling offers two essential things for coaches: it is important and it
can be simple. Simply stated, coaches and athletes get a great bang for their buck.
Proﬁling does not demand too much time, yet the information gained is truly essential to
optimizing performance. All coaches, players, teammates, and consultants should know
the 2-4 factors that most strongly inﬂuence each athlete’s performance within a
team/training group. If I know that Judy needs to be calm and conﬁdent, and that Harold
needs to be aggressive but not be angry, I am better informed to help Judy and Harold
reach their performance goals. I recommend that athletes share this information with their
coaches and teammates. This discussion of applied information for coaches continues in

the ﬁnal section of this dissertation.

150

What coaches can do. Though highly theoretical, this research can be useful to
coaches. In fact, while the full procedure used in this study may not be pragmatic, the
process of developing proﬁles is a useful tool for coaches and parents. Developing
proﬁles with athletes involves reﬂecting on previous best and worst performances,
considering the importance of various feeling states, and sharing beliefs about what it
takes to perform at one’s best. This process improves emotional and metacognitive
awareness in athletes. But more importantly, it can help coaches to understand their
players better while also building rapport. In my view, rapport between coaches and their
players is the single most important factor in effective coaching. Talking about how
athletes feel shows athletes that their coaches care about them as people. Along with the
necessary discussion about emotions, these proﬁling activities can segue into teaching
emotional-regulation and life skills to athletes. Finally, coaches and athletes can use the
information gained from this process to develop more effective and individualized
precompetition routines and coping strategies (See Gould, et al., 2009).

Not only does developing individualized feeling-performance proﬁles provide
coaches, mental trainers, and athletes with important information about how athletes
perform at their best, but the process of acquiring this knowledge also helps practitioners
to build closer relationships and better routines within teams. There are so many
unanswered questions in sport psychology! Sport psychologists must continue to work
more closely with coaches and athletes in order to share their knowledge, as well as to
learn from their more applied counterparts. In my experiences, working with athletes will

also help researchers to identify new and important questions. Our future is very bright.

151

 

APPENDIX

Figure 2

Example of the First Form from the Stepwise Individual Emotion Proﬁling (IEP) System

 

IZOF-Based Emotion-Proﬁling: Stop-Wise Procedures

You are asked to complete a series of forms that will help to identify your individualized zone of
optimal functioning in sport. The ﬁrst form has you reﬂect on your best- and worst-ever
performances over the past year or two. If possible, try to focus on more recent performances.
When reﬂecting these competitions, focus mostly on what you thought, did, and felt during the
performance. Please do not reﬂect on post-match events; we want to focus our assessment on
feelings during and before competition. Subsequent forms will help to describe your zone of
optimal functioning in terms of the descriptors that matter for you, and the intensity range that is
related to best and worst performances! Finally, we will create a visual ﬁgure that describes the
zones that are associated with your most functional and dysfunction performance states.

Step 1: Identify BEST EVER and WORST EVER Performances

 

BEST EVER Competition:

WORST EVER Competition:

 

 

 

Date: Location: Date: Location:
Competition: Competition:
Result: Result:

 

Details: (environment, opponents, feelings, etc)

Details: (environment, opponents, feelings, etc)

 

 

 

 

 

 

p ------------------------- - ------------------------- d
Date: Location: i Date: Location:
Competition: Competition:
Result: Result:

 

Details: (environment, opponents, feeergs, etc)

Details: (environment, opponents, feelings, etc)

 

 

 

 

 

L_._ —————————————————————— I

 

 

Date Location: Date: Location:
Competition: Competition:
Result: Result:

 

Details: (environment, opponents, feelings, etc)

Details: (environment, opponents, feeliigs, etc)

 

 

 

 

 

 

 

 

 

152

 

Figure 3

The Positive—Negative Affect (PNA-77) Scale: The second IEP form

Step 2: Identify HELPFUL-Positive and HELPFUL-Negative Emotions

Go over the list of HELPFUL-positive (pleasant) emotions below and select ﬁ'om the list 2-4
words that describe the emotions you felt before your best ever match in the recent past. You may
select only one word on the same line. Circle the words you select. If you don't ﬁnd a word
describing an emotion that is important to you, you may add you own word to the list. Follow the
same procedure for HELPFUL-Negative emotions: these are normally unpleasant emotions that you
feel help you. Please select at least two (2) items for each positive and negative list.

W):
active, dynamic, energetic, vigorous
relaxed, comfortable, easy
calm. peaceful, unhurried, quiet
cheerful, merry, happy

conﬁdent, certain, sure

delighted, overjoyed, exhilarated
determined, set, settled, resolute
excited, thrilled

brave, bold, daring, dashing

glad, pleased, satisﬁed, contented
inspired, motivated, stimulated
lighthearted, carefree

nice, pleasant, agreeable

quick, rapid, fast, alert

Your own emotion:

 

HELPFUL-NEGATIVE FEELINGS (N+):
afraid, fearful, scared, panicky
angry, aggressive, furious, violent
annoyed, irritated, distressed
anxious, apprehensive, worried
concerned, alarmed, disturbed,

dissatisﬁed

discouraged, dispirited, depressed
doubtful, uncertain, indecisive, irresolute
helpless, unsafe, insecure

inactive, sluggish, lazy

intense, ﬁerce

jittery, nervous, uneasy, restless

sorry, unhappy, regretful, sad, cheerless
tense, strained, tight, rigid

tired, weary, exhausted, worn out

Your own emotion:

Step 3: Identify HARMFUL-Negative and HARMFUL-Positive Emotions

Follow the same procedure as for step 2, select between two to ﬁve (2-4) words to describe the
HARMFUL-Negative and HARMFUL-Positive emotions you felt before your WORST EVER
competition. When complete, you should have between 12-16 items.

HARMFUL-NEGATIVE FEELINGS (N-):
afraid, fearful, scared, panicky
angry, aggressive, furious, violent
annoyed, irritated, distressed
anxious, apprehensive, worried
concerned, alarmed, disturbed,
dissatisﬁed
discouraged, dispirited, depressed
doubtful, uncertain, indecisive, irresolute
helpless, unsafe, insecure
inactive, sluggish, lazy
intense, ﬁerce
jittery, nervous, uneasy, restless
sorry, unhappy, regretful, sad, cheerless
tense, strained, tight, rigid
tired, weary, exhausted, worn out
Your own emotion:

 

153

HARMFUL—POSITIVE FEELINGS (P-j:
active, dynamic, energetic, vigorous
relaxed, comfortable, easy
calm, peaceful, unhurried, quiet
cheerful, merry, happy

conﬁdent, certain, sure

delighted, overjoyed, exhilarated
determined, set, settled, resolute
excited, thrilled

brave, bold, daring, dashing

glad, pleased, satisﬁed, contented
inspired, motivated, stimulated
lighthearted, carefree

nice, pleasant, agreeable

quick, rapid, fast, alert

Your own emotion:

Figure 4

Example of the Monitoring Form used by Participants

 

Type of Game: Serve / Return

Result: Won / Lost

Set Score:

 

 

 

 

 

 

 

 

 

you him
Rate your feelings about...
My performance was: Poor Average Good
Performance: 0 l 2 3 4 5 6 7 8 9 10
ﬁzchallenge of the game (task) Poor Average Good
Challenge: 0 1 2 3 4 5 6 7 8 9 10
My skills to meet my goals for the Poor Average Good
game were.
Skill: 0 1 2 3 4 5 6 7 8 9 10
. Emotional Intensig
! Intensity Scale (P+) 0 .5 l 2 3 4 5 6 7 8 9 10 11*
E 0 nothing at all (P+) 0 .5 l 2 3 4 5 6 7 8 9 10 11*
30.5 veryverylittle (P+) 0 .5 1 2 3 4 5 6 7 8 9 10 11*
i 1 very little (P+) 0 .5 1 2 3 4 5 6 7 8 9 10 11*
i
izmﬂ" (N+) 0.51234567891011“
E3 moderate (N+) 0.51234567891011*
i 4 (N+) 0 .5 1 2 3 4 5 6 7 8 9 10 11*
i 5 much (N+) 0 .5 l 2 3 4 5 6 7 8 9 10 11*
l
E6 (N-) 0.51234567891011“
i 7 verymuch (N-) 0 .5 1 2 3 4 5 6 7 8 9 10 11*
is (N-) 0.51234567891011“
. 9 .(N-) 0 .5 l 2 3 4 5 6 7 8 9 10 11"
I 10 very very
much (P-) 0 .5 1 2 3 4 5 6 7 8 9 10 11*
11* maximal (P-) 0 .5 l 2 3 4 5 6 7 8 9 10 11*
g ________ 19 35531313 ........ _(P-) 0.51234567891011“
(P-) 0 .5 l 2 3 4 5 6 7 8 9 10 11“

 

Comments (Not necessary, but please include if there is something you would like to note):

 

 

 

154

Figure 6

Frequency of Combined Positive—Functional Intensity Ratings for All Participants

 

200‘“

150‘

100‘

Frequency

50-

  
  
 

3“ g. ’ '
6 7 8 9 10 11
Intensity Rating

   

 

 

 

Note. Values within each column represent the number of observations across feeling
intensity. The distribution of intensity ratings includes poor, average, and good
performances.

 

155

Figure 7

Frequency of Combined Negative-Functional Intensity Ratings for All Participants

 

150‘

100-

Frequency

50‘

 

 

4 5 6
Intensity Rating

Note. Values within each column represent the number of observations across feeling
intensity. The dislribution of intensity ratings includes poor, average, and good
performances.

156

 

Figure 8

Frequency of Combined Negative-Dysfunctional Intensity Ratings for All Participants

 

 

 

200—

150-

310%

III.

50-
0__ . - ~.:- E in
01234567891011

Intensity Rating

Note. Values within each column represent the number of observations across feeling
intensity. The distribution of intensity ratings includes poor, average, and good
performances.

157

 

Figure 9

Frequency of Combined Positive-Dysﬁmctional Intensity Ratings for All Participants

 

150- , Z * ‘

100-

Frequency

50-

 

 

0 1 2 3 4 5 6 7 8 9 10
Intensity Rating

Note. Values within each column represent the niunber of observations across feeling
intensity. The distribution of intensity ratings includes poor, average, and good
performances.

158

 

Table 8

Interindividual Comparisons of Mean Difference [and Variance] for Positive-Functional
Feeling Intensities during Poor Performances

 

 

1 2 3 4 5 6 7 8 9
1 -
4.00
2 [1.50] '
3 -0.02 4.02 _
[2.17] [3.27]
4 -054 4.54 -0.52 _
[1.60] [1.06] [3.47]
5 1.73 -2.28 1.74 2.26 _
[1.09] [1.39] [2.36] [1.47]
6 3.17 -0.83 3.18 3.70 1.44 _
[1.28] [1.92] [1.70] [2.04] [1.39]
7 3.44 -0.56 3.46 3.98 1.72 0.27
[1.76] [2.65] [1.23] [2.82] [1.91] [1.38]
8 1.21 -2.79 1.23 1.75 -051 -l.96 -2.27 _
[1.99] [3.00] [1.09] [3.18] [2.16] [1.56] [1.13]
9 -1.47 -5.48 -1.75 -0.94 -320 -4.64 -4.91 -2.69
[4.27] [2.84] [9.27] [2.67] [3.93] [5.46] [7.52] [8.50]
,0 1.19 -2.81 1.21 1.73 -053 -1.98 -225 -0.02 2.67
[2.92] [4.39] [1.34] [4.67] [3.17] [2.28] [1.66] [1.47] [12.46]

 

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoe comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F -Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

159

Table 9

Interindividual Comparisons of Mean Difference [and Variance] for Positive-Functional

Feeling Intensities during Average Performances

 

 

1 2 3 4 5 6 7 8
1 -
1.23
2 [2.62] '
3 -0.88 -2.11 _
[2.51] [1.05]
4 -O.48 -1.72 0.40 _
[2.06] [1.27] [1.22]
5 -O.18 -1.42 0.70 0.30 _
[2.30] [1.14] [1.09] [1.11]
6 1.19 -0.04 2.07 1.67 1.37 _
[1.87] [4.90] [4.69] [3.86] [4.30]
7 0.07 -l.l6 0.95 0.56 0.26 -1.12 _
[1.49] [1.76] [1.68] [1.39] [1.54] [2.78]
8 -0.12 -1.35 0.76 0.36 0.06 -131 -019
[2.00] [1.31] [1.25] [1.03] [1.15] [3.75] [1.35]
9 -1.52 -2.76 -0.64 -1.04 -1.34 -2.72 -1.60 -1.40
[1.84] [4.82] [4.62] [3.80] [4.23] [1.02] [2.74] [3.69]
10 -0.68 -1.92 0.19 -020 -0.50 -l.88 -0.76 -0.57
[1.01] [2.65] [2.54] [2.09] [2.32] [1.85] [1.51] [2.03]

 

vﬁszJII . 7

Note. Mean differences (player on leﬁ column subtracted from player on top row) are
represented by Dunnett’s C post-hoc comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

160

Table 10

Interindividual Comparisons of Mean Difference [and Variance] for Positive-Functional
Feeling Intensities during Good Performances

 

N

DJ

10

 

1 2 3 4 5 6 7 8 9
0.73

[11.06]

-1.84 -2.56 _

[3.65] [3.03]

0.15 -0.58 1.98

[6.99] [1.58] [1.91]

0.19 -053 2.03 0.05

[4.36] [2.54] [1.19] [1.60]

1.64 0.91 3.47 1.49 1.44

[6.00] [66.36] [21.92] [41.95] [26.18]

-1.73 -2.45 0.11 -1.87 -1.92 -3.36

[1.36] [8.10] [2.68] [5.12] [3.20] [8.19]

-079 -152 1.04 -094 -0.98 -2.43 0.93

[5.01] [2.21] [1.37] [1.40] [1.15] [30.07] [3.67]

-1.41 -2.13 0.43 -155 -l.60 -3.04 0.32 -0.62
[5.69] [62.96] [20.79] [39.79] [24.83] [1.05] [7.77] [28.53]
-1.17 -1.90 0.67 -1.32 -1.36 -2.81 0.56 -0.38 0.24
[1.69] [6.55] [2.16] [4.14] [2.58] [10.14] [1.24] [2.97] [9.62]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoe comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

161

Table 11

Interindividual Comparisons of Mean Difference [and Variance] for Negative-Functional

Feeling Intensities during Poor Performances

 

\O

10

 

1 2 3 4 5 6 7 8 9
1.70 _

[8.00]

.430 -6.00 _

[7.99] [1.00]

-3.70 -540 0.59

[9.55] [1.19] [1.20]

-3.02 -4.72 1.28 0.69

[1.49] [5.36] [5.36] [6.40]

-3.89 -5.60 0.40 -019 -0.88 _

[4.57] [1.75] [1.75] [2.09] [3.06]

0.79 -0.91 5.09 4.49 3.81 4.68 _

[2.01] [3.98] [3.97] [4.75] [1.35] [2.27]

-4.26 -5.96 0.04 -055 -1.24 -0.36 -505

[1.34] [10.74] [10.72] [12.82] [2.00] [6.14] [2.70]

.2.92 -4.62 1.38 0.79 010 0.98 .3.71 1.34
[10.56] [1.32] [1.32] [1.11] [7.08] [2.31] [5.25] [14.17]

-0.89 -2.60 3.40 2.81 2.12 3.00 -1.68 3.36 2.02
[2.15] [3.73] [3.72] [4.45] [1.44] [2.13] [1.07] [2.88] [4.92]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoc comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in hold.

162

“I
.

Table 12

Interindividual Comparisons of Mean Difference [and Variance] for Negative-Functional
Feeling Intensities during Average Performances

 

 

1 2 3 4 5 6 7 8 9
1 -
-151
2 [1.18] '
3 -4.39 -2.88 _
[1.94] [1.65]
4 -3.83 -232 0.56 _
[6.71] [5.70] [3.46]
5 -1.18 0.33 3.20 2.64 _
[3.50] [2.98] [1.80] [1.92]
6 .223 -0.71 2.16 1.60 -1.04 _
[3.26] [2.77] [1.68] [2.06] [1.07]
7 -3.35 -l.84 1.04 0.48 -2.17 -1.12 _
[3.39] [2.88] [1.75] [1.98] [1.03] [1.04]
8 -033 1.18 4.06 3.50 0.86 1.90 3.02
[1.01] [1.19] [1.95] [6.76] [3.53] [3.29] [3.41]
9 -3.43 -1.91 0.96 0.40 .224 —1.20 -0.08 -310
[7.47] [6.35] [3.85] [1.11] [2.14] [2.29] [2.21] [7.53]
10 -3.60 -209 0.79 0.23 -2.42 -1.38 -0.25 -327 -0.17
[5.22] [4.44] [2.69] [1.29] [1.49] [1.60] [1.54] [5.26] [1.43]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoe comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

163

 

Table 13

Interindividual Comparisons of Mean Difference [and Variance] for Negative-Functional

Feeling Intensities during Good Performances

 

 

1 2 3 4 5 6 7 8 9
1 -
2 -4.01 _
[9.57]
3 -5.99 -1.98
[2.61] [3.66]
4 -5.04 -103 0.95 _
[1.26] [12.01] [3.28]
5 -l.65 2.36 4.34 3.38
[2.71] [25.96] [7.09] [2.16]
6 -2.70 1.31 3.29 2.34 -104
[1.44] [13.80] [3.77] [1.15] [1.88]
7 -7.25 -3.24 -1.26 -221 -559 .455
[1.17] [8.18] [2.23] [1.47] [3.17] [1.69]
8 -120 2.81 4.79 3.83 0.45 1.49 6.04
[17.08] [1.79] [6.54] [21.45] [46.36] [24.65] [14.61]
9 -6.48 -2.47 -O.48 -1.44 -4.82 -3.78 0.77 -5.27
[5.44] [52.07] [14.22] [4.34] [2.01] [3.77] [6.37] [92.99]
,0 -6.20 -219 -021 -1.16 -4.55 -350 1.04 -500 0.27

[239] [22.88] [6.25] [1.91] [1.13] [1.66] [2.80] [40.87] [2.28]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoc comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F -Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

164

m

Table 14

Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Dysfunctional Feeling Intensities during Poor Performances

 

 

1 2 3 4 5 6 7 8 9
1 -
-2.80
2 [1.70]
3 -297 -0.18
[2.09] [1.23]
4 -l.l6 1.63 1.81
[26.74] [45.37] [55.81]
5 -4.24 -1.45 -1.27 -3.08 _
[1.21] [1.40] [1.72] [32.42]
6 -5.06 -2.27 -2.09 -3.90 -0.82 _
[2.33] [3.95] [4.85] [11.50] [2.82]
7 1.05 3.85 4.02 2.21 5.30 6.12 _
[1.68] [2.85] [3.51] [15.89] [2.04] [1.38]
8 -072 2.08 2.25 0.44 3.53 4.35 -1.77 _
[1.21] [1.40] [1.72] [32.36] [1.00] [2.82] [2.04]
9 -2.35 40.44 0.62 -1.19 1.89 2.71 -3.41 -1.63
[1.55] [1.10] [1.35] [41.36] [1.28] [3.60] [2.60] [1.28]
19 -3.84 -1.04 -0.87 -2.68 0.40 1.22 -4.90 -3.12 -1.49
[1.94] [1.14] [1.07] [51.93] [1.60] [4.52] [3.27] [1.60] [1.26]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoe comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

165

Table 15

Interindividual Comparisons of Mean Difference [and Variance] for Negative-

Dysfunctional Feeling Intensities during Average Performances

 

10

 

1 2 3 4 5 6 7 8 9
0.16 _

[1.16]

0.45 0.29

[1.47] [1.27]

0.49 0.33 0.04 _

[8.82] [10.22] [12.96]

-095 -1.11 -1.40 -1.44 _

[2.18] [2.52] [3.20] [4.05]

-1.72 -l.88 -2.16 .220 -0.76 _

[2.17] [2.51] [3.19] [4.06] [1.00]

2.39 2.23 1.94 1.90 3.33 4.10

[6.27] [7.26] [9.21] [1.41] [2.88] [2.89]

2.14 1.98 1.69 1.65 3.09 3.85 -025

[2.83] [3.27] [4.15] [3.12] [1.30] [1.30] [2.22]

0.28 0.13 -0.16 -020 1.24 2.00 -2.10 -l.85

[1.59] [1.84] [2.33] [5.56] [1.37] [1.37] [3.95] [1.78]

0.47 0.31 0.02 -002 1.42 2.18 -1.92 -l.67 0.18
[1.98] [2.29] [2.90] [4.46] [1.10] [1.10] [3.17] [1.43] [1.25]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoc comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F -Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

166

 

 

Table 16

Interindividual Comparisons of Mean Difference [and Variance] for Negative-
Dysfunctional Feeling Intensities during Good Performances

 

\l

00

9

10

 

1 2 3 4 5 6 7 8 9
0.38 _

[1.98]

1.36 0.98

[1.87] [3.70]

0.02 -035 -1.33

[6.47] [12.82] [3.47]

-1.23 -1.60 -2.58 -1.25 _

[1.59] [3.14] [1.18] [4.08]

-294 —3.31 .429 -2.96 -1.71

[1.05] [1.88] [1.97] [6.82] [1.67]

1.76 1.38 0.40 1.73 2.98 4.69

[50.65] [100.32] [27.12] [7.82] [31.91] [53.35]

1.89 1.51 0.53 1.87 3.12 4.83 0.13

[4.21] [8.34] [2.25] [1.54] [2.65] [4.44] [10.29]

-094 -131 -2.29 -0.96 0.29 2.00 -2.69 -2.83
[1.66] [3.29] [1.12] [3.89] [1.05] [1.75] [30.47] [2.53]

0.42 0.04 -093 0.40 1.64 3.36 1.34 -1.47 1.36
[4.92] [9.75] [2.63] [1.32] [3.10] [5.18] [10.29] [1.17] [2.96]

 

Note. Mean differences (player on leﬁ column subtracted from player on top row) are
represented by Dunnett’s C post-hoc comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s

variance. Signiﬁcant values at a .01 level are represented in bold.

167

 

Table 17

Interindividual Comparisons of Mean Difference [and Variance] for Positive-Functional
Feeling Intensities during Poor Performances

 

 

1 2 3 4 5 6 7 8 9
1 -
4.75
2 [1.79] '
3 -053 1.22 _
[5.51] [3.07]
4 1.98 3.73 2.51
1.22] [2.18] [6.70]
5 2.81 4.55 3.34 0.83
[1.13] [1.58] [4.86] [1.38]
6 -1.20 0.54 -0.67 -3.18 4.01
[1.41] [1.27] [3.90] [1.72] [1.25]
7 4.33 6.08 4.86 2.35 1.52 5.53
[245.8] [441.2] [1353.6] [202.0] [278.8] [347.2]
8 3.33 5.07 3.86 1.35 0.52 4.53 -100
[2.99] [1.66] [1.84] [3.63] [2.63] [2.11] [734.2]
9 3.77 5.52 4.30 1.79 0.96 4.97 -0.56 0.44
[3.11] [5.58] [17.13] [2.56] [3.53] [4.39] [79.00] [9.29]
19 -0.99 0.76 -0.46 -2.97 -3.79 0.211 -532 432 -4.76

[5.31] [2.96] [1.04] [6.46] [4.68] [3.76] [1304] [1.78] [16.51]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoe comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

168

Fain-'39 Haw—J
' a

Table 18

Interindividual Comparisons of Mean Difference [and Variance] for Positive-Functional
Feeling Intensities during Average Performances

 

\l

00

9

10

 

1 2 3 4 5 6 7 8 9
0.63 _

[1.96]

1.19 0.56

[3.02] [1.54]

2.55 1.92 1.37

[1.19] [2.33] [3.59]

0.27 -0.36 -092 -229

[2.24] [1.14] [1.35] [2.67]

-0.86 -150 -205 -3.42 -1.13

[1.61] [3.16] [4.86] [1.35] [3.61]

4.81 4.18 3.62 2.25 4.54 5.67

[12.76] [25.04] [38.52] [10.73] [28.60] [7.93]

2.15 1.52 0.97 -040 1.88 3.02 -2.65 _

[4.88] [2.49] [1.62] [5.80] [2.18] [7.85] [62.26]

4.21 3.58 3.02 1.66 3.94 5.07 -0.60 2.06

[2.45] [4.82] [7.41] [2.06] [5.50] [1.53] [5.20] [11.97]

2.42 1.79 1.23 -014 2.15 3.28 -239 0.26 -1.79
[1.57] [1.25] [1.93] [1.86] [1.43] [2.52] [20.00] [3.11] [3.85]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoe comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

169

 

 

Table 19

Interindividual Comparisons of Mean Difference [and Variance] for Positive-Functional
Feeling Intensities during Good Performances

 

 

1 2 3 4 5 6 7 8 9
1 -
2 0.49 _
[1.44]
3 0.27 -0.21 _
[1.01] [1.43]
4 1.27 0.79 1.00 _
[3.58] [2.49] [3.55]
5 -2.50 -2.99 -2.77 -3.77 _
[1.75] [1.22] [1.74] [2.04]
6 -1.63 -2.12 -1.90 -290 0.87 _
[7.47] [5.19] [7.42] [2.09] [4.26]
7 3.97 3.48 3.70 2.70 6.47 5.60
[0.00] [0.00] [0.00] [0.00] [0.00] [0.00]
8 0.36 -0.13 0.08 -0.92 2.86 1.99 -3.61
[2.80] [4.03] [2.82] [10.02] [4.92] [20.92] [0.00]
9 2.97 2.48 2.70 1.70 5.47 4.60 -1.00 2.61
[2.99] [2.08] [2.97] [1.20] [1.70] [2.50] [0.00] [8.37]
19 2.12 1.63 1.85 0.84 4.62 3.75 -1.85 1.76 -0.85
[2.59] [1.80] [2.57] [1.38] [1.48] [2.88] [0.00] [7.26] [1.15]

 

Note. Mean differences (player on left column subtracted from player on top row) are
represented by Dunnett’s C post-hoc comparisons. Signiﬁcant values at a .05 level are
represented in bold. Variance values are represented by the F-Ratio of each player’s
variance. Signiﬁcant values at a .01 level are represented in bold.

170

 

 

 

REFERENCES

Albert, M., & Rosen, A. (1990). A systematic analysis of the various ways that the terms
‘affect,’ ‘emotion,’ and ‘mood’ are used. Journal of Communication Disorders,
23, 237-246.

Bandura, A. (1986). Social foundations of thought and action." A social cognitive theory.
Englewood Cliffs, NJ: Prentice Hall.

Beilock, S. L., & Carr, T. H. (2001). On fragility of skilled performance: What governs
chocking under pressure? Journal of Experimental Psychology: General, 130,
701-725.

Borg, G. (1998). Borg's perceived exertion and pain scales. Champaigrr, IL: Human
Kinetics.

Bortoli, L., & Robazza, C. (2002). Idiosyncratic performance affect in volleyball
referees: An extension of the IZOF-emotion model proﬁling. Journal of Sport
Behavior, 25(2), 115-133.

Bortoli, L., Robazza, C., & Nougier, V. (1997). Emotion in hockey and rugby. Paper
presented at the Proceedings of the 9th World Congress of Sport Psychology.
Innovations in sport psychology: Linking theory and practice, Part I, Wingate,
Israel.

Botterill, C., & Brown, M. (2002). Emotion and perspective in sport. International
Journal of Sport Psychology, 33, 38-60.

 

Campos, B., Keltner, D., & Tapias, M. P. (2004). Emotion. In C. D. Spielberger (Ed.),
Encyclopedia of applied psychology (V 01. 1, pp. 713-722). Oxford, UK: Elsevier
Academic Press.

Cohen, A. B., Tenebaum, G., & English, R. W. (2006). Emotions and golf performance:
An IZOF-based applied sport psychology case study. Behavior Modification,
30(3), 259-280.

Comparing Variances. (2001). Retrieved February 16, 2010, from
ht_tp://www.ltcconline.net/g;eenl/courses/201/reggession/comparingVarianceshtm

Csikszentrnihalyi, M. (1993). The evolving self: A psychology for the new millennium.
New York: Harper Collins Publishers Inc.

171

Csikszentrnihalyi, M. (1994). Flow: The psychology of optimal experience. New York:
Simon & Schuster, Inc.

Csikszentrnihalyi, M. (1996). Creativity: Flow and the psychology of discovery and
intention. New York: HarperCollins Publishers Inc.

Csikszentrnihalyi, M. (1997). Finding ﬂow: The psychology of engagement with everyday
life. New York: Basic Books.

Csikszentrnihalyi, M., & Larson, R. (1987). Validity and reliability of the Experience
Sampling Method. Journal of Nervous and Mental Disease, 1 75, 526—536.

Csikszentrnihalyi, M., & LeFevre, J. (1989). Optimal experience in work and leisure.
Journal of Personality and Social Psychology, 56(5), 815-822.

Csikszentrnihalyi, M., Rathunde, K., & Whalen, S. (1993). Talented teenagers: The roots
of success and failure. New York: Cambridge University Press.

Darwin, C. (1998). The Expression of the Emotions in Man and Animals. In P. Ekman
(Ed.) (originally published 1872 ed.). Oxford, UK: Oxford University Press.

Davis, M. (1992). The role of the amygdala in conditioned fear. In J. P. Aggleton (Ed.),
The Amygdala: Neurobiological Aspects of Emotion, Memory, and Mental
Dysﬁmction (pp. 255-306). New York: Wiley.

Deci, E. L. (1980). The psychology of self-determination. Lexington, MA: Health.

Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and self-determination in human
behavior. New York: Plenum.

Drever, J. (1952). A dictionary of psychology. London: Penguin.

Easterbrook, J. A. (1959). The effect of emotion on cue utilization and the organization of
behavior. Psychological Review, 66, 183-201.

Edmonds, W. A., Mann, D. T. Y., Tenenbaum, G., & Janelle, C. M. (2006). Analysis of
affect-related performance zones: An idiographic method using physiological and
introspective data. The Sport Psychologist, 20, 40—57.

Edwards, T., & Hardy, L. (1996). The interactive effects of intensity and direction of
cognitive and somatic anxiety and self-conﬁdence on processing bias. Journal of
Sport & Exercise Psychology, 18, 296-306.

Ekkikakis, P., & Petruzzello, S. J. (2000). Analysis of the affect measurement conundrum

in exercise psychology I: Fundamental issues. Psychology of Sport and Exercise,
1 , 7 1 -8 8.

172

Eysenck, M. W., & Calvo, M. G. (1992). Anxiety and performance: The processing
efficiency theory. Cognition and Emotion, 6(6), 409 - 434.

F rijda, N. H. (1993). The place of appraisal in emotion. Cognition and Emotion, 7, 357-
387.

Frijda, N. H., & Zeelenberg, M. (2001). Appraisal: What is the dependent? In K. R.
Scherer, A. Schorr & T. Johnstone (Eds.), Appraisal Processes in Emotion:
Theory, methods, research (pp. 141-156). New York: Oxford University Press,
Inc.

Ganzen, V. A. (1984). Systemnje opisanija v psikhologii [Systems descriptions in
psychology]. Leningrad: Leningrad University Press.

Golden, A. S., Tenenbaum, G., & Kamata, A. (2004). Affect-related performance zones:
An indiographic method for linking affect to performance. International Journal
of Sport and Exercise Psychology, 2, 24-42.

Goleman, D. (1995). Emotional intelligence: Why it can matter more than IQ. New York,
NY: Bantam Books.

Goleman, D. (Ed.). (1997). Healing emotions: Conversations with the Dalai Lama on
mindﬁtlness, emotions, and health. Boston, MA: Shambhala Publications, Inc.

Gould, D., Eklund, R. C., & Jackson, S. A. (1992). 1988 USA Olympic wrestling
excellence II: Competitive cognition and affect. The Sport Psychologist, 6(4),
383-402.

Gould, D., F lett, M. R., & Bean, E. (2009). Mental preparation for training and
competition. In B. W. Brewer (Ed.), The handbook of sports medicine and
science: Sport psychology (1 ed., pp. 53-63). Hoboken, NJ: Wiley-Blackwell
Publishing Ltd.

Gould, D., Hardy, L., Medbery, R., Dieffenbach, K., Lauer, L., & Jones, G. (Date
unknown). Identifying the range of affective elements invovled in sport
acheivement striving. Unpublished manuscript, University of North Carolina
Greensboro.

Gould, D., Tuffey, S., Udry, E., & Lochbaum, M. (1993). Multidimensional state anxiety
and middle distance running performance: An exploratory examination of Hanin's
(1980) zones of optimal functioning hypothesis. Journal of Applied Sport
Psychology, 5, 85-95.

Graham, F. K. (1979). Distinguishing among orienting, defense, and startle reﬂexes. In

H. D. Kimmel, E. H. van Olst & J. F. Orlebeke (Eds), The Orienting Reﬂexes in
Humans: An International Conference Sponsored by the Scientiﬁc Aﬂairs

173

Division of the North Atlantic Treaty Organization (pp. 137-167). Hillsdale, NJ:
Lawrence Erlbaurn.

Hagtvet, K. A., & Hanin, Y. L. (2007). Consistency of performance-related emotions in
elite athletes: Generalizability theory applied to the IZOF model. Psychology of
Sport and Exercise, 8, 47-72.

Hanin, Y. L. (1978). A study of anxiety in sports. In W. F. Straub (Ed.), Sport
Psychology: An analysis of athlete behavior (pp. 236 - 249). Ithaca, NY:
Movement Publications.

Hanin, Y. L. (1986). The state-trait anxiety research on sports in the USSR. In C. D.
Spielberger & R. Diaz-Guerrero (Eds.), Cross-Cultural Anxiety (V 01. 3).
Washington, DC: Hemisphere.

Hanin, Y. L. (1993a). Optimal performance emotions in top athletes. In S. Serpa, J.
Alves, V. Ferreira & A. Paula-Brito (Eds), Sport psychology: An integrated
approach. Proceedings from the V111 World Congress of Sport Psychology (pp.
229-232). Lisbon, Portugal: ISSP.

Hanin, Y. L. (1993b). Optimal performance emotions in top athletes. In S. Serpa, J.
Alves, V. Ferreira & A. Paula-Brito (Eds), Sport psychology: An integrated
approach. Procedings ﬁ‘om the V111 World Congress of Sport Psychology (pp.
229-232). Lisbon, Portugal: ISSP.

Hanin, Y. L. (1995). Individual Zones of Optimal Functioning (IZOF) model: An
ideographic approach to performance anxiety. In K. Henschen & W. Straub
(Eds), Sport psychology: An analysis of athlete behaviour (pp. 103-119).
Longrneadow, MA: Movement Publications. ‘

Hanin, Y. L. (1997). Emotion and athletic performance: Individual zones of optimal
functioning model. European Yearbook of Sport Psychology, 1 , 29-72.

Hanin, Y. L. (2000a). Emotions in sport. Champaign, IL: Human Kinetics, Inc.

Hanin, Y. L. (2000b). Emotions in Sport. Champaign, IL: Human Kinetics.

Hanin, Y. L. (2000c). Individual Zones of Optimal Functioning (IZOF) Model: Emotion-
performance relationships in sport. In Y. L. Hanin (Ed.), Emotions in sport (pp.
65-90). Champaign, IL: Human Kinetics.

Hanin, Y. L. (2000d). Successful and poor performance and emotions. In Y. L. Hanin
(Ed.), Emotions in sport (pp. 157-188). Champaign, IL: Human Kinetics.

174

Hanin, Y. L. (2002). Individually optimal recovery in sports: An application of the IZOF
model. In M. Kellrnan (Ed.), Enhancing recovery: Preventing underperformance
in athletes (pp. 199-218). Champaign, IL: Human Kinetics.

Hanin, Y. L. (2003). Performance Related Emotional States in Sport: A Qualitative
Analysis. Qualitative Social Research, 4(1),

Hanin, Y. L. (2004). Emotion in sport: An individualized approach. In C. D. Spielberger
(Ed.), Encyclopedia of applied psychology (Vol. 1, pp. 739-750). Oxford, UK:
Elsevier Academic Press.

Hanin, Y. L. (2007). Emotions in sport: Current issues and perspectives. In G.
Tenenbaum & R. C. Eklund (Eds.), Handbook of sport psychology (3 ed., pp. 31-
58). Hoboken, NJ: John Wiley & Sons, Inc.

Hanin, Y. L., & Stambulova, N. B. (2002). Athlete-generated metaphors as descriptors of
performance states. Svensk Idrottspsykologi (Swedish Sport Psychology), 2(1), 7-
9.

Hanin, Y. L., Stambulova, N. B., Lukkarila, J., & Tummavuori, M. (1997). Perforrnance-
related states: A metaphoric description. Journal of Sport Science, 19(8), 650-651.

Hanin, Y. L., & Syrja, P. (1995). Performance affects in junior ice hockey players: An
application of the individual zones of optimal functioning model. The Sport
Psychologist, 9, 169-187.

Hanin, Y. L., & Syrja, P. (1996). Predicted, actual, and recalled affect in Olympic-level
soccer players: Ideographic Assessments on Individualized Scales. Journal of
Sport and Exercise Psychology, 18(3), 325-335.

Hardy, L. (1990). A catastrophe model of performance in sport. In G. Jones & L. Hardy
(Eds.), Stress and performance in sport (pp. 81-106). Chichester, England: Wiley.

Hardy, L. (1996). A test of catastrophe model of anxiety and sports performance against
multidimensional anxiety theory models using the method of dynamic differences.
Anxiety, Stress, and Coping, 9, 69-86.

Hardy, L., & F razey, J. (1987). The inverted-U hypothesis: A catastrophe model for sport
psychology? Paper presented at the Annual Conference of the North American
Society for the Psychology of Sport and Physical Activity, Vancouver, Canada.

Hardy, L., & Frazey, J. (1987, June). The inverted-U hypothesis: A catastrophe model for
sport psychology? Paper presented at the Annual Conference of the North
American Society for the Psychology of Sport and Physical Activity, Vancouver,
Canada.

175

Hardy, L., Jones, G., & Gould, D. (1996). Understanding psychological preparation for
sport: Theory and practice for elite performers. Chichester, England: Wiley.

Harger, G. J ., & Raglin, J. S. (1994). Correspondence between actual and recalled
precompetition anxiety in collegiate track and ﬁeld athletes. Journal of Sport and
Exercise Psychology, 16, 206-211.

Hill, K. L. (2001). Frameworks for sport psychologists: Enhancing sport performance.
Champaign, IL: Human Kinetics.

Imlay, G. J ., Carda, R. D., Stanbrough, M. E., & O'Connor, P. J. ( 1995). Anxiety and
performance: A test of optimal function theory. International Journal of Sport
Psychology, 26, 295-306.

Jackson, S. A. (2000). Joy, fun, and ﬂow state in sport. In Y. L. Hanin (Ed.), Emotions in
sport (pp. 135-156). Champaign, IL: Human Kinetics, Inc.

Jackson, S. A., & Csikszentrnihalyi, M. (1999). Flow in sports: The keys to optimal
experiences and performances. Champaign, IL: Human Kinetics, Inc.

Jackson, S. A., & Kimiecik, J. C. (2008). The ﬂow perspective of optimal experience in
sport and physical activity. In T. S. Horn (Ed.), Advances in sport psychology (3rd
ed., pp. 377-399). Champaign, IL: Human Kinetics, Inc.

Jackson, S. A., & Roberts, G. C. (1992). Positive performance state of athletes: Toward a
conceptual understanding of peak performance. The Sport Psychologist, 6, 156-
171.

James, W. (1884). What is emotion?. Mind, 4(188-204).

Janelle, C. M. (2002). Anxiety, arousal and visual attention: A mechanistic account of
performance variability. Journal of Sports Sciences, 20, 237-251.

Johnson, M. B., Edmonds, W. A., Kamata, A., & Tenebaum, G. (2009). Determining
individual affect-related performance zones (IAPZs): A tutorial. Journal of
Clinical Sport Psychology, 3(1), 34-57.

Johnson, M. B., Edmonds, W. A., Moraes, L. C., Medeiros F ilho, E. S., & Tenenbaum,
G. (2007). Linking affect and performance of an international level archer
incorporating an idiosyncratic probabilistic method. Psychology of Sport and
Exercise, 8, 317-335.

Johnson, M. B., Edmonds, W. A., Tenenbaum, G., & Kamata, A. (2007). The
relationship between affect and performance in competitive intercollegiate tennis:
A dynamic conceptualization and application. Journal of Clinical Sport
Psychology, 1, 130-146.

176

‘5 l n'.‘ m . u “92er —J
_.

Jokela, M., & Hanin, Y. L. (1999). Does individual zones of optimal functioning model
discriminate between successful and less successful athletes? A meta-analysis.
Journal of Sports Sciences, 1 7, 873-887.

Jones, G. (1995). More than just a game: Research developments and issues in
competitive anxiety in sport. British Journal of Psychology, 86(449-478).

Jones, G., Hanton, S., & Swain, A. (1994). Intensity and interpretation of anxiety
symptoms in elite and non-elite sports performers. Personality and Individual
Diﬁerences, 17(5), 657-663.

Jones, G., Swain, A., & Hardy, L. (1993). Intensity and direction dimensions of
competitive anxiety and relationships with performance. Journal of Sports
Sciences, 1 I (525-532).

Jones, M. V. (2003). Controlling emotions in sport. The Sport Psychologist, 1 7, 471-486.

Kamata, A., Tenenbaum, G., & Hanin, Y. L. (2002). Individual Zone of Optimal
Functioning (IZOF): A Probabilistic conceptualization. Journal of Sport and
Exercise Psychology, 24(2), 189-208.

Kerr, J. H. (1997). Motivation and Emotion in Sport: Reversal theory. East Sussex, UK:
Psychology Press.

Landers, D. M. (1980). The arousal-performance relationship revisited. Research
Quarterly for Exercise and Sport, 51, 77-90.

Lang, P. (2000). Emotion and motivation: Attention, perception, and action. Journal of
Sport and Exercise Psychology, 20, $122-$140.

Lazarus, R. S. (1991). Emotion and Adaptation. New York: Oxford University Press, Inc.

Lazarus, R. S. (2000a). Cognitive-motivational-relational theory of emotion. In Y. L.
Hanin (Ed.), Emotions in sport (pp. 39-64). Champaign, IL: Human Kinetics, Inc.

Lazarus, R. S. (2000b). How emotions inﬂuence performance in competitive sports. The
Sport Psychologist, 14, 229-252.

Lazarus, R. S. (2001). Relational meaning and discrete emotions. In K. R. Scherer, A.
Schorr & T. Johnstone (Eds.), Appraisal processes in emotion (pp. 37-67). New
York: Oxford University Press.

Lazarus, R. S., Averill, J. R., & Opton, E. M., Jr. (1970). Towards a cognitive theory of

emotion. In M. B. Arnold (Ed.), Feelings and emotions: Loyola symposium (pp.
207-232). New York: Academic Press.

177

 

Lazarus, R. S., & F olkman, S. (1984). Stress Appraisal and Coping. New York: Springer.
Levin, T. (October 3, 2003). Mood vs. affect. Psychiatric News, 38(19), 48.
Loehr, J. E. (1994). The new toughness training for sports. New York: Penguin Books.

Lukkarila, J ., & Hanin, Y. L. (2001). Emotion dynamics in the winning team during the
1999 world junior ice hockey championship. Journal of Sport Science, 19(8), 652-
653.

Martens, R. (1987). Coaches guide to sport psychology. Champaign, IL: Human Kinetics.

Martens, R., Vealey, R. S., & Burton, D. (1990). Competitive Anxiety in Sport.
Champaign, IL: Human Kinetics.

Masterson, F. A., & Crawford, M. (1982). The defense motivation system: A theory of
avoidance behavior. The Behavior and Brain Sciences, 5, 661-696.

Mayer, J. D., & Salovey, P. (1997). What is emotional intelligence? In P. Salovey 8c D.
Sluyter (Eds), Emotional development and emotional intelligence: Educational
implications (pp. 3-31). New York: Basic Books.

McNair, D. M., Lorr, M., & Doppleman, L. F. (1981). Proﬁle of Mood State manual. San
Diego: Educational and Industrial Testing Service.

Medeiros Filho, E. S., Moraes, L. C., & Tenenbaum, G. (2008). Affective and
Physiological States during Archery Competitions: Adopting and Enhancing the
Probabilistic Methodology of Individual Affect-Related Performance Zones
(IAPZs). Journal of Appl ied Sport Psychology, 20(4), 441-456.

Mellalieu, S. D. (2003). Mood matters: But how Much? A comment on Lane and Terry
(2000). Journal of Applied Sport Psychology, 15, 99-1 14.

Morgan, W. P. (1980). Test of champions: The iceberg proﬁle. Psychology Today, 14,
92-99.

Morgan, W. P., O’Connor, P. J ., Ellickson, K. A., & Bradley, P. W. (1988). Personality
structure, mood states, and performance in elite male distance runners.
International Journal of Sport Psychology, 19, 247-263.

Morris, W. M. (1992). A functional analysis of the role of mood in affective systems. In

M. S. Clark (Ed.), Review of personality and social psychology (pp. 256-293).
Newbury Park, CA: Sage.

178

 

Mueller, R. 0., & Hancock, G. R. (2008). Best practices in structural equation modeling.
In J. W. Osborne (Ed.), Best practices in quantitative methods (pp. 488-508);
Thousand Oaks, CA: Sage.

Orlick, T. (1986). Psyching for sport. Champaign, IL: Human Kinetics.
Orlick, T. (2008). In pursuit of excellence (4th ed.). Champaign, IL: Human Kinetics.

Ortony, A., Clore, G. L., & Collins, A. (1988). The cognitive structure of emotions.
Cambridge, UK: Cambridge University Press.

Ortony, A., Clore, G. L., & Foss, M. A. (1987). The referential structure of the affective
lexicon. Cognitive Science, 11, 341-364.

Pallant, J. (2009). SPSS survival manual (3rd ed.). NY: Open University Press.

Parkinson, B. (1997). Untangling the appraisal-emotion connection. Personality and
Social Psychology Review, I, 62-79.

Payne, M. (2000). Narative therapy: An introduction for counsellors. London: Sage.

Pekrun, R. (1992). The impact of emotions on learning and achievement: Towards a
theory of cognitive/motivational mediators. Applied Psychology: An International
Review, 41(4), 359-376.

Pekrun, R. (2000). A social-cognitive, control-value theory of achievement emotions. In
J. Heckhausen (Ed.), Motivational psychology of human development (pp. 143-
163). New York: Elsevier Science.

Pekrun, R., Elliot, A. J., & Maier, M. A. (2006). Achievement goals and discrete
achievement emotions: A theoretical model and prospective test. Journal of
Educational Psychology, 98(3), 583-597.

Pons, D., Balaguer, I., & Garcia-Merita, M. L. (2001). Is the Breadth of Individualized
Ranges of Optimal Anxiety (IZOF) Equal for all Athletes? A Graphical Method
for Establishing IZOF. The Spanish Journal of Psychology, 4(1), 3-10.

Raglin, J. S. (2004). Anxiety and optimal athletic performance. In C. D. Spielberger
(Ed.), Encyclopedia of applied psychology (Vol. 1, pp. 139-144). Oxford, UK:
Elsevier Academic Press.

Raglin, J. S., & Hanin, Y. L. (2000). Competitive anxiety. In Y. L. Hanin (Ed.), Emotions
in sport (pp. 93-111). Champaign, IL: Human Kinetics, Inc.

179

Raglin, J. S., & Turner, P. E. (1993). Predicted, actual, and optimal precompetition
anxiety in adolescent track and ﬁeld athletes. Scandinavian Journal of Exercise
and Science in Sports, 2, 148-152.

Raykov, T., & Marcoulides, G. A. (2008). An introduction to applied multivariate
analysis. New York: Routledge Taylor and Francis Group.

Robazza, C. (2006). Emotions in Sport: An IZOF perspective. In S. Hanton & S. D.
Mellalieu (Eds), Literature Reviews in Sport Psychology (pp. 127-158).
Hauoppage, NY: Nova Science.

Robazza, C., & Bortoli, L. (2003). Intensity, idiosyncratic content and functional impact
of performance-related emotions in athletes. Journal of Sports Sciences, 21, 171-
189.

Robazza, C., Bortoli, L., & Hanin, Y. L. (2004). Pre-Competition Emotions, Bodily
Symptoms, and Task-Speciﬁc Qualities as Predictors of Performance in High-
level Karate Athletes. Journal of Applied Sport Psychology, 16(2), 151-165.

Robazza, C., Bortoli, L., & Nougier, V. (1998). Performance-related emotions in skilled
athletes: hedonic tone and functional impact. Perceptual and Motor Skills, 87,
547-564.

Rogers, C. R. (1980). A way of being. New York: Houghton Mifflin Company.

Roseman, I. J. (1996). Why these appraisals? Anchoring appraisal models to research on
emotional behavior and related response systems. In N. H. F rij da (Ed.),
Proceedings of the Ninth International Conference of the international Society for
Research on Emotions (pp. 106-110). Toronto: International Society for Research
on Emotions.

Roseman, I. J. (2001). A model of appraisal in the emotion system: Integrating theory,
research, and applications. In K. R. Scherer, A. Schorr & T. Johnstone (Eds),
Appraisal Processes in Emotion (pp. 68-91). New York: Oxford University Press.

Roseman, I. J ., & Smith, C. A. (2001). Appraisal theory: Overview, assumptions,
varieties, controversies. In K. R. Scherer, A. Schorr & T. Johnstone (Eds),
Appraisal Processes in Emotion (pp. 3-19). New York: Oxford University Press.

Ruiz, M. C., & Hanin, Y. L. (2004). Metaphoric description and individualized emotion
proﬁling of performance related states in high-level karate athletes. Journal of
Applied Sport Psychology, 16(3), 1-16.

Russell, J. A. (1980). A circumplex model of anxiety. Journal of Personality and Social
Psychology, 39, 1161-1178.

180

Russell, J. A. (1997). How shall an emotion be called? In R. Plutchik & H. R. Conte
(Eds), Circumflex models of personality and emotions (pp. 205-222).
Washington, DC: American Psychology Association.

Russell, J. A., Weiss, A., & Mendelsohn, G. A. (1989). Affect gn'd: A single-item scale
of pleasure and arousal. Journal of Personality and Social Psychology, 5 7, 493-
502.

Salovey, P., & Mayer, J. D. (1990). Emotional Intelligence. Imagination, Cognition and
Personality, 9, 185-211.

Scherer, K. R. (2001). The Nature and Study of Appraisal: A review of the issues. In K.
R. Scherer, A. Schorr & T. Johnstone (Eds), Appraisal Processes in Emotion:
Theory, methods, research (pp. 369-392). New York: Oxford University Press.

Schorr, A. (2001). Appraisal: The evolution of an idea. In K. R. Scherer, A. Schorr & T.
Johnstone (Eds), Appraisal Processes in Emotion (pp. 3-19). New York: Oxford
University Press.

Smith, C. A., & Lazarus, R. S. (1990). Emotion and adaptation. In L. A. Pervin (Ed.),
Handbook of Personality: Theory and research (pp. 609-637). New York:
Guilford.

Smith, C. A., & Lazarus, R. S. (1993). Appraisal components, core relational themes, and
the emotions. Cognition and Emotion, 7, 233-269.

Spence, J. T., & Spence, K. W. (1966). The motivational components of manifest
anxiety: Drive and drive stimuli. In C. D. Spielberger (Ed.), Anxiety and behavior.
New York: Academic Press.

Spielberger, C. D. (1988a). Manual for the State-Trait Anger Expression Inventory
(ST AIAO. Odessa, F l: Psychological Assessment Resources.

Spielberger, C. D. (1988b). Manual for the state-trait anxiety expression inventory
(STADO. Odessa, F l: Psychological Assessment Resources.

Syrja, P., & Hanin, Y. L. (1997). Individualized and group-oriented measures of emotion
in sport: A comparative study Annual Congress of the European College of Sports
Science. Book of abstracts (pp. 682-684). Copenhagen: University of
Copenhagen.

Tenebaum, G., & Elran, E. (2003). Congruence between actual and retrospective reports

of emotions for pre- and post-competition states. Journal of Sport and Exercise
Psychology, 25, 323-340.

181

Tenebaum, G., Kamata, A., & Hanin, Y. L. (2002). The individual zones of optimal
functioning. Journal of Sport and Exercise Psychology, 24, S l 23.

Tenenbaum, G., Edmonds, W. A., & Eccles, D. W. (2008). Emotions, coping strategies,
and performance: A conceptual framework for deﬁning affect-related
performance zones. Military Psychology, 20, SI 1-S37.

Tenenbaum, G., & Elran, E. (2003). Congruence of actual and recalled retrospective
reports of emotions for pre- and postcompetition states. Journal of Sport and
Exercise Psychology, 25, 323-340.

Tenenbaum, G., Loyde, M., Pretty, G., & Hanin, Y. L. (2002). Congruence of actual and
recalled precompetition emotions in equestrians. Journal of Sport and Exercise
Psychology, 24(3), 271-288.

Tomkins, S. S. (1962). Affect, imagery, consciousness: The positive aﬂects (Vol. 1). New
York: Springer.

Tummavuori, M., & Hanin, Y. L. (2000). Quantifying emotion intensity on the modified
Borg ’s CR-10 scale. Paper presented at the The 5th Annual Congress of the
European College of Sport Science, Jyvaskyla, Finland.

Turner, P. E., & Raglin, J. S. (1996). Variability in precompetition anxiety and
performance incollege track and ﬁeld athletes. Medicine and Science in Sport and
Exercise, 28(3), 378-385.

Vincent, W. J. (1999). Statistics in kinesiology (2nd ed). Champaign, IL: Human
Kinetics.

Watson, D. (2002). Positive affectivity. In C. R. Snyder & S. J. Lopez (Eds), Handbook
of Positive Psychology (pp. 106-119). New York: Oxford University Press.

Watson, D., & Tellegen, A. (1985). Toward a consensual structure of mood.
Psychological Bulletin, 98, 219-235.

Watson, J. B. (1919). Psychology ﬁom the Standpoint of a Behaviorist. Philadelphia, PA:
Lippincott.

Wegner, D. M. (1994). Ironic processing of mental control. Psychological Review, 101,
34-52.

Weiner, B. (1985). An attributional theory of achievement motivation and emotion.
Psychological Review, 92, 548-573.

Williams, A. M., Davids, K., & Williams, J. G. (1999). Visual perception and action in
sport. New York: E & FN Spon Routledge.

182

Woodman, T., & Hardy, L. (2001). Arousal and performance. In R. N. Singer, H. A.
Hausenblas & C. M. Janelle (Eds), Handbook of Sport Psychology (2nd ed., pp.
319-339). Hoboken, NJ: John Wiley & Sons, Inc.

Yerkes, R. M., & Dodson, J. D. (1908). The relation of strength of stimulus to rapidity of
habit-formation. Journal of Comparative Neurology and Psychology, 18, 459-
482.

Zaichkowsky, L. D., & Naylor, A. H. (2004). Arousal in sport. In C. D. Spielberger (Ed.),
Encyclopedia of Applied Psychology (V 01. 1, pp. 155-161). Oxford, UK: Elsevier

Academic Press.

Zajonc, R. B. (1965). Social facilitation. Science, 149, 269-274.

183

R".
"I
I."
"
H
NIH
“
Al”
H

31293 03063 5191