(«.5 .1959. 3
Jr: a:
43:3“.
‘.
.. in:

.2 I..
.iri
.15...

.xuMuﬁﬁneh ‘4...th

iizgi:
A

:3» a
“thug“? in... n
.n 1-. I til-xii“!
:ﬁfﬂuﬂhﬂuﬁm
'Y o k “I
. an»...

5..“ a..|.—ii§.vr¢|
i‘l—I h.‘l’.u1v‘-Jy‘ I
is. it‘lsunummt! )2.“
ﬁ‘
a. a

 

.. {I4

 

.uﬂmn. .nxmm.u#§quﬂh.l§.§ “Mm“

 

“"5 LIBRARY
or coal Michigan State
University

 

 

 

This is to certify that the
dissertation entitled

DECISION-MAKING IN SPORT: AN EXAMINATION OF THE
TAKE THE FIRST HEURISTIC AND SELF-EFFICACY
THEORY

presented by
TERI J. HEPLER

has been accepted towards fulﬁllment
of the requirements for the

PhD. degree in KinesiologL

DollaAi Ml;

Major Professor’s Signature

 

R,C\'\\r‘r~—W\LA.,{L ‘\ 1“; 2 C‘0 ((5

Date

MSU is an Afﬁnnative 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:/Prolecc&Pres/CIRC/DateDue indd

DECISION-MAKING IN SPORT: AN EXAMINATION OF THE TAKE THE FIRST
HEURISTIC AND SELF-EFFICACY THEORY
BY

Teri J. Hepler

A DISSERTATION
Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Kinesiology

2008

ABSTRACT
DECISION-MAKING IN SPORT: AN EXAMINATION OF THE TAKE THE FIRST
HEURISTIC AND SELF-BF F ICACY THEORY
By
Teri J. Hepler

The purpose of this study was to explore the role of self-efﬁcacy and the Take the First
('ITF) heuristic in option-generation and decision-making performance in a simulated
sports task. Participants (N = 72) performed a basketball video-based decision-making
task. The option-generation task asked participants to generate options regarding what
move the player with the ball should make next and subsequently decide which option
represented the optimal move. In addition, participants' decision-making self-efﬁcacy
was assessed using a 10-item questionnaire comprising various aspects of decision-
making in basketball. Similarly, the decision-making performance task required
participants to watch video situations and make a decision as fast as possible. Participants
rated their degree of conﬁdence, from 0 (not at all conﬁdent) to 10 (extremely conﬁdent),
to perform various tasks related to decision-making. Speciﬁc hypotheses investigated
whether: participants used the 'I‘TF heuristic, option quality decreased with serial
position, participants who use "FTP make decisions that are faster and/or better, and self-
efﬁcacy inﬂuenced any of these relationships. Results of the study supported many of the
tenets of the 'ITF heuristic, such that people used the heuristic on a majority of the trials
(70%) and earlier generated options were better than later ones. Mixed results were found
on dynamic inconsistency and decision conﬁdence as each relates to the number of

Options generated. Results did not support the notion performance would have been

better, on average, by generating only one option. Exploratory questions indicated that
while using TTF did not produce higher quality decision, it was associated with faster
decisions and greater conﬁdence in one’s decisions. Self-efﬁcacy was also shown to be
signiﬁcantly related to the TTF heuristic. Participants with higher self-efﬁcacy beliefs
used TTF more frequently and they generated fewer options than those with low self-
efﬁcacy. Likewise, self-efﬁcacy was also related to several aspects of option-generation
and decision-making. After controlling for basketball knowledge and competitive
basketball playing experience, self-efﬁcacy was signiﬁcantly related to decision
conﬁdence on both tasks and decision quality on the option— generation task. However,
efﬁcacy beliefs were not signiﬁcantly related to decision quality on the decision-making

performance task or decision speed on either of the tasks.

DEDICATION

To my mom, whose love and strength inspire me everyday and who showed me what it
means to be an extraordinary woman.

iv

ACKNOWLEDGEMENTS

A special thank you to my advisor and dissertation committee chair, Dr. Deborah L.
Feltz, for her guidance, support, and patience. Thank you to my dissertation committee
members, Drs. Dan Gould, Tim Pleskac, and Jere Brophy, for all of their insightﬁil
guidance on this project. Also, thank you to Graig for all those late-night brainstorming
sessions and for helping to keep my sanity.

To Molly, I would have never accomplished this without your love and support!

TABLE OF CONTENTS

LIST OF APPENDICES .................................................................................................. viii

LIST OF TABLES ............................................................................................................. ix

LIST OF FIGURES ............................................................................................................ x

CHAPTER 1

INTRODUCTION
Nature of Problem ................................................................................................... 1
Statement of the Problem ........................................................................................ 4
Purpose of the Study ................................................................................................ 7
Option-Generation Hypotheses/Research Questions .............................................. 8
Decision-Making Performance Hypotheses/Research Questions ........................... 9
Delimitations ......................................................................................................... 10
Deﬁnitions ............................................................................................................. 10

CHAPTER 2

REVIEW OF LITERATURE
Decision-Making in Sport ..................................................................................... l3
Decision-Making and Self-Efﬁcacy ...................................................................... 18
Bounded Rationality and Heuristics ...................................................................... 19
Fast and Frugal Heuristics ..................................................................................... 22
Take the First Heuristic ......................................................................................... 28
Summary ................................................................................................................ 34

CHAPTER 3

METHOD
Participants ............................................................................................................ 36
Self-Report Questionnaires ................................................................................... 37
Video Situations .................................................................................................... 39
Option-Generation Task ........................................................................................ 41
Decision-Making Performance Task ..................................................................... 43
Scoring of Tasks .................................................................................................... 43
Procedures ............................................................................................................. 44
Treatment of the Data ............................................................................................ 47
Data Analysis ......................................................................................................... 48

CHAPTER 4

RESULTS
Demographic Information ..................................................................................... 50 .
Descriptive Statistics ............................................................................................. 51
Correlations .......................................................................................................... 52

vi

Preliminary Analyses ............................................................................................. 53

Option-Generation Hypothesis 1 ........................................................................... 55
Option-Generation Hypothesis 2 ........................................................................... 57
Option-Generation Hypothesis 3 ......................................................................... 58
Option-Generation Hypothesis 4 ........................................................................... 59
Option-Generation Hypothesis 5 ........................................................................... 59
Option-Generation Hypothesis 6 ........................................................................... 60
Option-Generation Hypothesis 7 ........................................................................... 61
Option-Generation Hypothesis 8 ........................................................................... 61
Option-Generation Research Question 1 ............................................................... 62
Option-Generation Research Question 2 ............................................................... 62
Decision-Making Performance Hypothesis 9 ........................................................ 63
Decision-Making Performance Hypothesis 10 ...................................................... 63
Decision-Making Performance Hypothesis 11 ...................................................... 64
Decision-Making Performance Research Question 3 ............................................ 65
Decision-Making Performance Research Question 4 ............................................ 65
Decision-Making Performance Research Question 5 ............................................ 66
CHAPTER 5
DISCUSSION
Take the First and Option-Generation ................................................................... 67
Self-Efﬁcacy and Option-Generation .................................................................... 73
Self-Efﬁcacy and Decision-Making Performance ................................................. 75
Take the First and Decision-Making Performance ................................................ 78
Gender Differences ................................................................................................ 81
Implications ........................................................................................................... 82
Future Research Direction ..................................................................................... 84
Conclusions ........................................................................................................... 86
APPENDICES ................................................................................................................... 88
TABLES ............................................................................................................................ 97
REFERENCES ................................................................................................................ 105

vii

LIST OF APPENDICES

Appendix A. Demographic and Importance Questionnaire ............................................... 88
Appendix B. Basketball Knowledge Test .......................................................................... 89
Appendix C. Decision-Making Self-Efﬁcacy Questionnaire ............................................ 91
Appendix D. Example Option-Generation Score Sheet ....................... 92
Appendix E. Example Decision-Making Score Sheet ....................................................... 93
Appendix F. Informed Consent Form ................................................................................ 94
Appendix G. Study Directions ........................................................................................... 95

viii

LIST OF TABLES

Table 1. Means and Standard Deviations of all Independent and dependent
Variables .............................................................................................................. 98

Table 2. Descriptive Statistics for all Independent and Dependent Variables and
Various Demographics by Gender ....................................................................... 99

Table 3. Correlations between all Independent and Dependent Variables ...................... 100

Table 4. Means and Standard Deviations of Option Quality according to Serial
Position .............................................................................................................. 101

Table 5. Post Hoc Test Comparing Option Quality according to Serial Position ............ 102

Table 6. Post Hoc Test Comparing TTF Frequency according to Number of Options
Generated ........................................................................................................... 103

Table 7. Post Hoc Test Comparing Decision Conﬁdence according to the Number of
Options Generated ............................................................................................. 104

ix

LIST OF FIGURES

Figure 1. Timeline of Events for Order A .......................................................................... 47

CHAPTER 1
INTRODUCTION

Nature of the Problem

The 1993 National Collegiate Athletic Association Men’s Basketball
Championship was the stage for one of the most unforgettable moments in the history of
college sports. The Michigan Wolverines and the North Carolina Tar Heels battled back
and forth all game. However, North Carolina surged to a 3 point lead with 46 seconds left
in the game, forcing Steve Fischer’s Wolverines to call their ﬁnal time-out. As the
players headed back onto the court to resume play, Coach Fischer reminded his players
that they were out of time-outs. With 20 seconds on the clock and his team down by 2
points, Michigan’s super sophomore Chris Webber came up with a crucial rebound.
Following some confusion in the backcourt, and a missed traveling call by the referees,
Webber frantically brought the ball upcourt. As he dribbled to the right sideline, two Tar
Heel defenders trapped him as he picked up his dribble. Sensing he was in trouble and the
clock‘winding down, Webber had only a split-second to make a decision about what to do
next. Immediately, Webber called a time-out. There was only one problem: Michigan
was out of time-outs. The play resulted in a technical foul against Michigan and
effectively sealed the national championship for the North Carolina Tar Heels. This
incident, dubbed as the “time-out that never was”, is so infamous that it ranks 11th on
ESPN’s list of the “biggest chokes in the last 25 years.”
(http://sports.espn.go.com/espn/espn25/storﬂpage=listranker/25biggestchokes).

As can be seen ﬁom this example, performance in sport is determined not only by

physical skill execution, but also by the quality of decisions made throughout the

competition. One decision could mean the difference between victory and defeat. The
purpose of this dissertation was to explore the process of decision-making in sport.
Speciﬁcally, it examined the different options athletes generated for an upcoming move,
the rules they used to choose among those options, and the role that self-efﬁcacy played
in these processes. This chapter brieﬂy summarizes the decision-making research in
sport, introduces heuristics, and concludes with discussion of the tenets and limitations of
the Take the First heuristic — a particular heuristic for making quick decisions.

Decision-making, which is deﬁned as a process by which an individual selects
one action ﬁ'om among a set of alternatives in a given situation, has been studied
extensively in the sport psychology literature (Tenenbaurn, 2004). In fact, journals such
as the International Journal of Sport and Exercise Psychology and the Psychology of
Sport & Exercise have devoted entire issues to the topic. However, most of the research
on decision-making in sport has been rooted in the expert-novice paradigm. Accordingly,
this line of research explores differences in the perceptual systems and information
processing mechanisms of expert and novice performers. Results of this research suggests
that, compared with novices, experts make faster, more accurate decisions, recall game
structured information better, detect game-related signals faster, and predict events from
advanced cues better (Starkes & Ericsson, 2003)

While this line of research has yielded valuable information regarding the
decision-making of experts and novices, it has not provided further insight into the
process of decision-making. However, one area that might help shed light on the
cognitive processes involved in decision-making is the study of heuristics. Practically

speaking, a heuristic is a rule of thumb that a person uses to make a decision. While there

are many different heuristics that people use in different environments, fast and frugal
heuristics (FF H) are particularly relevant to sport (Gigerenzer, Todd, & ABC Research
Group, 1999). FF H are especially useﬁrl in fast-paced, dynamic sports because they rely
on limited information to make quick decisions.

While there are many different heuristics in the fast and frugal family, this
dissertation explored the FFH known as the Take the First (TTF) heuristic. This rule of
thumb, which was proposed by Johnson and Raab in 2003, suggests that for familiar, yet
ill-deﬁned tasks, a person should simply choose one of ﬁrst options that comes to mind,
instead of generating and evaluating all possible options. One of the appeals of TTF is
that it describes the processes of option-generation and decision choice. These processes
are important in situations requiring people to generate potential options, rather than
choose from a set of predeﬁned choices, and then quickly determine which option is best.
There are many real-life situations when people are presented with speciﬁc, predeﬁned
options. For instance, a person who walks into a grocery store to buy a box of breakfast
cereal will ﬁnd many different kinds of cereal on the shelf from which to choose. .
However, many other real-world circumstances do not present peOple with these clear, .
explicit options. Sport is one such environment, as athletes must freely generate options
for each situation that arises. A major assumption of this heuristic is that options are
generated in a sequential, meaningful way based on option similarity, experience,
strategy, and environmental factors. Based on the sequential order of option-generation,
earlier options represent better decisions than later ones. Thus, a person should choose

the ﬁrst option generated because it likely represents the best decision.

As TTF is a relatively new heuristic, there have only been two published studies
on the topic (Johnson & Raab, 2003; Raab & Johnson, 2007). Both of these studies
required team handball players to watch offensive attack situations in handball, cite
potential moves the player with the ball could perform (e. g., shoot at the goal), and then
choose the best decision among the generated options. Results of these studies support
the main predictions of TTF. In both studies, the ﬁrst Option was chosen approximately
60% of the time across all participants. Moreover, the ﬁrst options generated were, on
average, of higher quality than later options (Johnson & Raab, 2003; Raab & Johnson,
2007)

Statement of the Problem

While these early studies have provided support for the use of the 'I'I‘F heuristic in
sport decision-making, there are many limitations and knowledge gaps. For instance, a
major assumption of these studies is that participants actually verbalized their ﬁrst,
intuitive options. However, there was no manipulation, such as time pressure, to ensure
that participants stated their ﬁrst inclination. Rather, participants may have contemplated
several options and only indicated the one they felt was the best. This potential confound
would explain why the ﬁrst option was of such high quality and why participants selected
that option as the best decision in a majority of the cases. I

Another major knowledge gap in this area of research is how various psychosocial
factors inﬂuence a person’s tendency to use TTF in sport. One psychosocial factor that
has been linked to heuristics and decision-making is self-efﬁcacy. Self-efﬁcacy
represents people’s beliefs in their capabilities to successfully perform a task (Bandura,

1997). Research has found that efﬁcacy beliefs are positively related to decision—making

performance in various settings, including business (Bandura & Wood, 1989; Wood &
Bandura, 1989) and sport (Hepler & Feltz, 2008; Tenenbaurn, Levy-Kolker, Slade,
Lieberman, & Lidor, 1996). According to Bandura (1997), self-efﬁcacy is related to
heuristics in that it inﬂuences the options a person considers, what information is
collected, how information is interpreted, and how information is used to make a
decision. This relationship between self-efﬁcacy and heuristics was discussed in a
concept paper by Wood, Atkins, and Tabemero (2000). The authors suggest that the
inﬂuence of self-efﬁcacy on the performance of complex tasks is mediated by the use of
judgmental heuristics. In particular, Wood et a1. (2000) posit that people with low
perceptions of self-efﬁcacy are more vulnerable to the biases of heuristics than are high
efﬁcacious people. For example, people with low self-efﬁcacy beliefs may be more
susceptible to the availability bias because they limit their external search and rely more
on memory. The proposed link between efﬁcacy beliefs and heuristics is supported by the
ﬁndings of a study on entrepreneurial decision-making (Bryant, 2007). Using a mixed-
method design, Bryant (2007) found that self-regulatory factors, including self—efﬁcacy,
were related to the use of various decision-making heuristics by entrepreneurs. For
instance, results indicated that self-efﬁcacy was positively related to participants’ use of
the “strategic ﬁt” heuristic. In this study, the “strategic ﬁt” heuristic was used to
determine whether or not an entrepreneurial opportunity was worth exploring and also
how the opportunity ﬁt in with the overall vision of the company.

While self-efﬁcacy may be related to heuristics in general, it is also likely that
efﬁcacy beliefs inﬂuence the use of the speciﬁc TTF heuristic. According to Bandura’s

theory, self-efﬁcacy inﬂuences cognitive ﬁmctioning. Speciﬁcally, self-efﬁcacy

inﬂuences people’s perceptions of a situation, as well as the quality of their analytic
thinking. People with high efﬁcacy beliefs tend to perceive situations as realistic
challenges, visualize success, and exhibit efﬁcient analytic thinking. Conversely, those
who doubt their capabilities typically dwell on the risk of failure, visualize failure, and
display inefﬁcient analytic thinking (Bandura, 1997). It is likely that people with low
efﬁcacy who constantly envision failure will doubt the ﬁrst option that comes to mind
and attempt to cognitively process several other options. On the other hand, a strong
belief in one’s abilities may make it quite easy to simply trust one’s instincts and accept
the ﬁrst option. In fact, there are a few research ﬁndings that provide indirect support for
the link between self-efﬁcacy and the TTF heuristic. For instance, Bouffard-Bouchard
and colleagues (1991) found that when matched on ability, students with high self-
efﬁcacy were less likely to prematurely reject correct solutions than were students with
low self-efﬁcacy. In terms of 'ITF, this would translate into a positive relationship
between self-efﬁcacy and choosing the ﬁrst option, assuming it was an acceptable option.
Similarly, another study on entrepreneurial decision risk-taking found that people with
high self-efﬁcacy beliefs perceived situations as presenting more opportunities and fewer
threats than did those with low ability beliefs. In turn, high self-efﬁcacy participants took
more risks when making decisions than did participants with low self-efﬁcacy. Many
would consider choosing the ﬁrst option, with little or no consideration of alternatives, to
be an example of very risky decision behavior.

In addition, sport success is often determined by an athlete’s ability to use early
predictive cues in order to guide their own actions. For instance, a baseball batter must

predict where the pitch will cross the plate before deciding whether or not to swing.

Likewise, a football quarterback must anticipate where the defense will go before
deciding which receiver to throw the ball to. Expert-novice research has indicated that
experienced players are more efﬁcient and accurate at utilizing advanced cues than are
inexperienced athletes (Starkes & Ericsson, 2003). Furthermore, experts are more
conﬁdent in their predictions than are novices (Tenenbaum et al., 1996). Thus, Bandura
concluded that “the cognitive side of athleticism is not just matter of gaining predictive
knowledge but of gaining the self-assurance to act on it unhesitatingly (Bandura, 1997,
pp. 375). The TTF heuristic epitomizes action without hesitation, as it suggests that when
facing a familiar situation, a person should simply go with their ﬁrst inclination. The
importance of having conﬁdence in one’s instinctive decision-making capabilities can be
seen in the aforementioned example involving Chris Webber. During the entire play,
Webber seemed to be very confused and indecisive about what to do with the ball. Aﬁer
grabbing the rebound, his instinctive decision was to pass the ball to a guard, but he
second-guessed that decision, causing him to travel. Webber’s self-doubts regarding his
ability to make good decisions may have also contributed to his calling the infamous
time-out. Perhaps he did not feel capable of making a decision that could help lead his
team to victory and called the time-out to transfer the decision-making responsibility to
his coach.
Purpose of the Study

The purpose of this study was to explore the role of self-efﬁcacy and the TTF
heuristic in option- generation and decision-making performance on a simulated sports
task. As TTF is a relatively new heuristic, it is important to provide further evidence that

this heuristic is in fact used by athletes and to explore how using the heuristic relates to

decision-making performance. Moreover, this dissertation expands on previous research
by examining how self-efﬁcacy inﬂuences option-generation and decision-making
performance using the framework of the TTF heuristic.

This study investigated several hypotheses and research questions. Many of the
predictions were based on the tenets of the TTF heuristic, as well as previous research
ﬁndings on TTF (Johnson & Raab, 2003; Raab & Johnson, 2007). Additionally, several
hypotheses were based on self-efﬁcacy theory and research (Bandura, 1997; Bandura &
Wood, 1989; Hepler & F eltz, 2008; Wood & Bandura, 1989). In addition, this
dissertation examined several research questions that are rooted in theory, but have not
yet been empirically tested.

Option-Generation Hypotheses

1. Participants will choose the ﬁrst option generated as the best decision in more
than half of the trials.

2. Option quality will be negatively related to serial position

3. As the number of generated Options increases, the likelihood that the ﬁrst option
will be chosen decreases.

4. Conﬁdence in ﬁnal decision will be negatively related to the number of options
generated.

5. On average, participants would make better choices if they only generated one
option.

6. Decision-making self-efﬁcacy will be positively related to ﬁnal decision quality,
after controlling for basketball knowledge and competitive basketball playing

experience.

7. Decision-making self-efficacy will be negatively related to option-generation
speed, in which higher efﬁcacy beliefs will predict faster generation speeds, after
controlling for basketball knowledge and competitive basketball playing
experience.

8. Decision-making self-efﬁcacy will predict the conﬁdence in ﬁnal decision, after
controlling for basketball knowledge and competitive basketball playing
experience.

Option-Generation Research Questions

1. Does self-efﬁcacy predict the use of TTF heuristic, after controlling for basketball
knowledge and competitive basketball playing experience?

2. Does self-efﬁcacy inﬂuence the number of options generated, after controlling for
basketball knowledge and competitive basketball playing experience?

Decision-Making Performance Hypotheses

9. Decision-making self-efﬁcacy will be positively related to decision-making
quality, after controlling for basketball knowledge and competitive basketball
playing experience.

10. Self-efﬁcacy will be positively related to decision-making speed, after controlling
for basketball knowledge and competitive basketball playing experience.

11. Self-efﬁcacy will be positively related to decision conﬁdence on the decision-
making performance task, after controlling for basketball knowledge and
competitive basketball playing experience.

Decision-Making Performance Research Questions

3. DO participants who have a high tendency to use TTF make better decisions than
those who have a low tendency to use TTF?
4. Do participants who have a high tendency to use TTF make faster decisions than
those who have a low tendency to use TTF?
5. Are participants who have a high tendency to use TTF more conﬁdent in their
decisions than are those who have a low tendency to use TTF?
Delimitations
The ﬁndings are limited to a population of undergraduate and graduate students
who have previous experience or knowledge of basketball. These results may not
generalize to youth sport participants or athletes in sports other than basketball. Likewise,
the ﬁndings may not be applicable to closed or independent sports and situations. Finally,
as the decision-making tasks utilize video situations, these results may not apply to
decision-making during actual game play.
Deﬁnitions
1. Basketball knowledge test score: participants’ score on the 10-item basketball
knowledge test.
2. Decision conﬁdence: average conﬁdence in ﬁnal decision on the 13 task trials.
3. Decision-making: a process by which an individual selects one course of action
from among a set of two or more alternatives in a speciﬁc situation.
4. Decision-making performance task: experimental task requiring participants to
watch 13 video clips and make a decision regarding the best move in each

situation.

10

5. Decision-making self-eﬁ‘icacy: participants’ conﬁdence in their ability to make
decisions on the basketball video test.

6. Decision quality: average decision quality (0-4) of all 13 ﬁnal decisions in each
task.

7. Decision response time (or speed): average response time, in seconds rounded to
the nearest hundredth) on the 13 trials in each task.

8. Dynamic inconsistency: switch in preference from ﬁrst Option to a different ﬁnal
decision.

9. Ecological rationality: the ability of a heuristic to exploit the structure of
information in the environment.

10. Fast and frugal heuristics (FFH): a subset of heuristics which allow people to
make adaptive decisions in the real world by minimizing the amount of time,
knowledge, and computation necessary to make decisions.

11. Heuristic: a rule of thumb; a mental device that can solve a class of problems in
situations with limited knowledge and time.

12. Option-generation: process of generating alternatives from scratch.

13. Option-generation task. experimental task requiring participants to watch 13
video clips, generate potential moves for each clip, and choose one option as the
best decision for each situation.

14. Self-efficacy: people’s judgments of their capabilities to organize and execute
courses of action needed to achieve designated types of performance.

15. Take the first (TIT) heuristic: a heuristic that suggests that in familiar, yet ill-

deﬁned tasks, a person should choose one of the initial Options generated, rather

-

11

than exhaustively generating all possible options and subsequently processing
them deliberately.
16. Take the ﬁrst score: number of times (out of 13) that participants chose the ﬁrst

option on the option- generation task.

12

CHAPTER 2
REVIEW OF LITERATURE
The purpose of this chapter is to provide a review of literature that is relevant to
the variables and procedures used in this study. Speciﬁcally, the chapter introduces
decision-making in sport, summarizes the relevant research on decision-making in sport,
and examines the inﬂuence of self-efﬁcacy on decision-making. Next, the chapter
focuses on heuristics by providing a general overview of heuristics, outlining the TTF
heuristic, and discussing some of the limitations of previous research on TTF.
Decision-Making in Sport
“Success in athletic competition requires more than physical skills. It is now
widely recognized that cognitive factors play an inﬂuential role in athletic development
and functioning" (Bandura, 1997, pp. 369). Accordingly, one of the key cognitive
components of athletic performance is decision-making (Chamberlain & Coelho, 1993;
Thomas, 1994). Decision-making is a process by which an individual selects one action
from among a set of two or more alternatives in a speciﬁc situation (Tenenbarnn, 2004).
Speciﬁcally, a decision represents the course of action an athlete chooses to pursue and
when to execute that action. The importance of decision-making in sport is highlighted by
the fact that every voluntary action is preceded by a decision to perform that action
(Grehaigne, Godbout, & Bouthier, 2001). For example, a soccer match could involve
over 2,500 touches, with a tactical decision preceding each touch (Morris, 1981).
Decision-making is especially important in complex in open, interdependent sports such
as football, basketball, hockey, volleyball, and soccer because the environment and game

. conditions are constantly changing. In these types of sports, an athlete must assess a

13

rapidly changing situation, make a decision about what action to take, and then execute
that action. All of these things must occur within a matter of seconds, or even less, as the
dynamic nature of these sports imposes a strict time constraint.

As decision-making is such a critical component Of sport performance, a great
deal of research has been conducted on the topic. The majority of this research has
focused on the differences between expert and novice performers. This research has used
a number of approaches, such as examining differences in perceptual systems,
knowledge, signal detection, information recall, and speed and accuracy of tactical
decisions. The earliest line of research investigated the hardware hypothesis (Clark &
Warren, 1935; Olson, 1956; Starkes & Deakin, 1984). The hardware hypothesis suggests
that experts have more highly developed perceptual systems than do novices, which
contribute to superior decision-making capabilities. These studies compared various
sensory and perceptual systems such as depth perception, visual acuity, and reaction time
of experts and novices. For instance, one study investigated the eye movements of expert
and novice baseball players. In this study, participants were shown videos of different
pitches and the task was to identify the type of pitch (i.e., fastball or curveball) as fast as
possible. Phototransistors were used to determine participants’ eye movement reaction
time, deﬁned as the amount of time it took for the eyes to move once the pitcher had
released the ball. Results indicated that there were no differences in eye reaction time
between experienced and non-experienced players. Another study by Starkes (1987)
investigated several “hardware” aspects of female ﬁeld hockey players ﬁ'om three
competitive levels. There were no differences in dynamic visual acuity or coincident

anticipation timing between the national, varsity, and novice players. Interestingly,

l4

national-level players (i.e., experts) had slower simple visual reaction time than did
varsity and novice players. The ﬁndings of these studies, and many others, have yielded
equivocal results that provide no solid link between hardware and expertise (Shank &
Haywood, 1987).

Another framework that has been used to investigate expert-novice differences in
decision-making is the software hypothesis (Starkes & Ericsson, 2003). The software
hypothesis posits that, when compared to novice performers, experts are better decision-
makers because they have a more extensive knowledge base and use their knowledge
more efﬁciently. Studies based on the software hypothesis have examined differences
between experts and novices in visual search patterns, advanced cue utilization, signal
detection, and information recall. For example, while the previously mentioned study of
baseball batters found no expert-novice differences in eye movement reaction time,
results did suggest that the two groups differed in their visual search patterns. Experts
had a tendency to ﬁxate on the point of release and to only attend to the oncoming ball.
Meanwhile, novices had a more active visual search pattern, as their gaze often shifted to
non-relevant cues, such as the pitcher's head, just before the ball was released.
Consequently, the experienced players were better able to identify the type of pitch that
was thrown. This suggests that experts perform better because they attend to more
relevant cues (Shank & Haywood, 1987). Support for differences in cue utilization can
also be evidenced by studies by Abernethy and Russell (1987). These studies utilized
temporal and spatial occlusion to investigate advanced cue utilization in experienced and
inexperienced racket sport players. The task was to predict the landing point of different

badminton shots presented in a video clip. According to the results, the racket arm side

15

provided the most pertinent cues and experts were able to extract relevant information
from earlier events than were novices.

Signal detection and information recall studies are also common methods to
investigate the issue of expert-novice software. One study involved expert and novice
volleyball players who were shown slides of structured and unstructured volleyball
situations. Structured activities were deﬁned as game situations while unstructured
situations represented non-game activities, such as warm-ups and time-outs. As a signal
detection study, the objective was to determine, as quickly as possible, whether or not a
volleyball appeared in the picture. The volleyball players responded more quickly than
non-players to both the structured and unstructured scenes. However, there were no
differences between the two groups on the accuracy of signal detection (Allard & Starkes,
1980). Another project assessed the information recall capabilities of expert and novice
basketball players. Participants were shown slides of both structured and unstructured
situations in a basketball game. After viewing the scenes, they were asked to diagram the
positioning of the players depicted in the slide. The researchers found that experts more
accurately recalled structured situations than did novices. However, there were no
differences in the recall of unstructured scenarios (Allard, Graham, & Paarsalu, 1980).

Another area of research on decision-making in sport has investigated the role of
expertise in determining the speed and accuracy of tactical decisions. These studies have
typically used slides, diagrams, or videos to present participants with tactical decision-
making situations. For instance, Tenenbaum and colleagues (1993) presented team
handball players with slides Of various handball situations. Participants in this study were

asked to indicate what move the highlighted player should make. Decision-making

.

16

performance was deﬁned as the accuracy of decision, which was rated by handball
experts. According to the results, experience was positively related to decision-making
performance (T enenbaum, Yuval, Elbaz, Bar-Eli & Weinberg, 1993). In another study,
Starkes (1987) examined the speed and accuracy of decisions made by female ﬁeld
hockey players. The elite-level players made mOre accurate decisions than did varsity or
novice athletes. There were no signiﬁcant differences between the groups on decision-
making speed. Helsen and Pauwels (1988) investigated decision-making in soccer. In this
study, the researchers used special lenses to project life-size video images of soccer
situations on a large screen. Participants stood 7 m in front of the screen with a soccer
ball directly in front of them. One moment in each clip depicted the ball being kicked
back towards the participant. At that moment, participants executed their tactical decision
by kicking the ball to the desired target (e. g., shoot at the goal, pass to a teammate).
Across all of the trials, experience did not predict speed or accuracy of decisions.
However, when examining only the correct decisions, experienced players made faster
decisions than did inexperienced players.

Overall, research on decision-making in sport has found support for some
consistent differences between experts and novices. Experts tend to have greater recall
for structured game information, detect game-related signals faster, are better able to
predict events from advanced cues, and make faster, more accurate decisions than do
novice performers. It is important to understand that expert-novice differences only
pertain to the speciﬁc context in which the person has expertise. For example, an expert
soccer player may have a decision-making advantage in soccer, but that advantage does

nOt generalize to other sports or non-soccer situations (Chamberlain & Coelho, 1993).

17

Decision-Making and Self-Efﬁcacy

Despite the fact that there has been extensive research focusing on the relationship
between experience and decision-making in sport, there has been little examination Of the
role of self-perceptions in decision-making. For instance, one self-perception that has
been linked to both cognitive and physical performance is self-efﬁcacy. Self-efﬁcacy
refers to people’s judgments of their capabilities to organize and execute courses of
action required to attain designated types of performances (Bandura, 1977). According to
Bandura and Wood (1989), self-efﬁcacy beliefs can be especially inﬂuential on decision-
making in complex, dynamic enviromnents. In a series of studies investigating
managerial decision-making in a simulated business organization, these researchers found
that self-efﬁcacy was positively related to decision-making quality. The results of these
studies led Bandura to conclude that: "... people who believe strongly in their problem-
solving capabilities remain highly efﬁcient in their analytic thinking in cOmplex decision-
making situations. Quality of analytic thinking, in turn, fosters performance
accomplishments" (Bandura, 1997, pp. 452).

While there has been evidence linking self-efﬁcacy and decision—making in
organizational settings, there have only been a handful of studies that have explored self-
efﬁcacy beliefs and decision-making in sport. Tenenbaum and colleagues conducted a
study involving expert, intermediate, and novice tennis players. Participants in this study
viewed video footage of various temris strokes. There were different experimental
conditions where the video was occluded at various points prior to, during, or after the
racket made contact with the ball. The task was to judge the ﬁnal landing spot of each

stroke. After deciding where the ball would land, participants rated how conﬁdent they

18

were in their decisions. The researchers found that experts were signiﬁcantly more
conﬁdent in their decisions than were intermediate players or novices on shots where the
video was occluded shortly after contact (Tenenbaum et al., 1996). While this study did
measure conﬁdence, it did not investigate how self-efﬁcacy inﬂuences decision-making
performance. Two studies by Hepler and colleagues have investigated this relationship.
In the ﬁrst study, Hepler and Chase (2008) examined the relationship between task self-
efﬁcacy, decision-making self-efﬁcacy, and decision-making performance on a video test
involving softball situations. While the results indicated that decision-making self-
efﬁcacy was related to physical performance, there was no signiﬁcant relationship
between self-efficacy and decision-making performance. However, the authors cited
several methodological limitations that may have inﬂuenced the results. Thus, Hepler and
Feltz (2008) conducted a follow-up study utilizing improved methodology. In this
research, participants had to decide which base to throw the ball to in 10 video
simulations Of defensive baseball situations. Decision-making performance was deﬁned
as the product of the decision speed and accuracy, as determined by baseball experts.
Results of this study indicated that self-efﬁcacy was a positive and signiﬁcant predictor
of decision-making performance on a simulated sport task.
Bounded Rationality and Heuristics

Despite the considerable amount of research on decision-making in sport, there
has been little emphasis on examining the mechanisms through which athletes make these
decisions. One area which could be useful is the study of heuristics. A heuristic,
commonly called a rule of thumb, can be deﬁned as a simple “mental device that can

solve a class Of problems in situations with limited knowledge and time” (Raab &

l9

Gigerenzer, 2005, p.188). These simple rules of thumb can be particularly useful is
decision research because they describe the actual process of problem-solving, not just
the outcome (Gigerenzer, 2004). Heuristics are rooted in Simon’s (1957) notion Of
bounded rationality. According to bounded rationality, adaptive behavior can only be
understood by jointly considering the limited capacities of the human mind and the
structure of the environment. Simon has likened these two factors (i.e. cognition,
environment) to blades on a scissors (Simon, 1990, p. 7). In this manner, it would not be
possible to understand how a scissors works by simply examining only one blade of the
too]. However, inspecting both blades, and Observing the interaction between them,
provides all of the infOrmation needed to understand the cutting mechanism of a scissors;
In this manner, intelligent behavior can best be understood by considering people’s
cognitive capabilities in the context of the task environment. Overall, bounded rationality
seeks to explain human behavior in real world conditions, which typically involve limited
time, knowledge, and computational capacity (Gigerenzer, 2004).

Simon’s view of bounded rationality is in direct opposition to the classical,
normative theory of unbounded rationality. The perspective of unbounded rationality
assumes that people have unlimited resources (i.e. knowledge, time, processing
capabilities) which can be used to make optimal, rational decisions. However, in order to
perform these inﬁnitely complex calculations, humans would have to possess some form
of superintelligence. An example of cognition based on unbounded rationality is the
expected utility theory. Use of the expected utility theory requires a person to identify
every possible consequence associated with an action, assign a numerical utility value to

each consequence, and multiply that utility value by the probability that the consequence

20

will occur. As a'result of this process, the option with the highest expected utility value
would represent the optimal, rational choice (Gigerenzer, Todd, et al., 1999). Conversely,
methods associated with bounded rationality suggest that either an optimal solution does
not exist or that it is not feasible to calculate the optimal choice. Instead, people Often
choose a “satisﬁcing” strategy (Simon, 1956). Satisﬁcing is a heuristic in which a person
chooses among a set of alternatives by selecting the ﬁrst option which exceeds a speciﬁed
aspirational level (Simon, 1990). For instance, a person may decide that he or she wants
to eat a hamburger for lunch. Using a satisﬁcing strategy, the person would drive past
Taco Bell and Kentucky Fried Chicken, but stop at the very ﬁrst restaurant, such as
McDonald’s, that served hamburgers.

According to Gigerenzer, Todd, and colleagues (1999), satisﬁcing and other
heuristics are tools that are contained within the adaptive toolbox of the mind
(Gigerenzer, Todd, etal., 1999). Speciﬁcally, the adaptive toOlbox refers to the
“collection of specialized cognitive mechanisms that evolution has built into the human
mind for speciﬁc domains of inference and reasoning” (Gigerenzer, Todd, et al., 1999,
pp. 30). This toolbox is adaptive because people can choose which tool to use in order to
meet the demands of the environment, they can identify situations in which a tool is
effective or ineffective, and they can revise and create new tools when needed. Tools in
the adaptive toolbox are domain-speciﬁc (i.e. specialized, not generalized), they work
within cognitive limitations, and they are only useﬁrl if they match the environment (i.e.
ecologically rational) (Martignon & Schmitt, 1999).

In order to operate within the constraints of the mind and the environment, the

toolbox utilizes resources from three different layers. The ﬁrst level contains evolved

21

capacities such as depth perception or face recognition. These evolved capacities are used
to form the basis of the second layer, known as building blocks. These building blocks,
which are discussed in detail below, comprise rules regarding search, stopping, and
decisions. Finally, these building blocks are used to construct heuristics, or tools, which
comprise the last layer in the adaptive toolbox. In keeping with the toolbox metaphor, the
evolved capacities serve as the metal which is molded into various nuts and bolts, or
building blocks. Subsequently, these nuts and bolts (i.e., building blocks) are assembled
together to create the tools of heuristics (Gigerenzer, 2007). Moreover, new tools can be
fashioned out of existing heuristics or nested within other tools. One family of tools in the
adaptive toolbox that might be particularly useful in the study of decision-making in sport
is fast and frugal heuristics (FFH).

Fast and Frugal Heuristics (FFH)

FFH are a subset of heuristics which allow people to make adaptive decisions in
the real world by minimizing the amount of time, knowledge, and computation necessary
to make decisions. Before providing a more detailed description of FFH, a few examples
of these tools are presented ﬁrst. One commonly cited example Of FFH is the recognition
heuristic (Goldstein & Gigerenzer, 2002). The recognition heuristic can be used to infer
which alternative has the higher value on a particular criterion. Speciﬁcally, this tool
states that when one alternative is recognized and the other is not, then it can be inferred
that the recognized alternative has the higher value on the speciﬁed criterion. For
example, research has demonstrated the power of this heuristic in an inference task
requiring participants to decide which city is more populous. In one study, students from

Germany and the United States were asked to determine which US. city had the larger

22

population: San Diego or San Antonio. While two-thirds Of the students from the US.
answered the question correctly, an astounding 100% of the German students correctly
selected San Diego as the larger city. The recognition heuristic can help to explain this
result. According to the researchers, the German students reported that they recognized
San Diego, but had never heard of San Antonio. Based on the fact that they recognized
one city, but not the other, they simply inferred that the city they recognized, San Diego,
was more populous. Conversely, students in the United States were at least somewhat
familiar with both cities, therefore they could not rely on the recognition heuristic
(Goldstein & Gigerenzer 2002). Thus, recognition heuristic is only ecologically rational
in situations where a person is ignorant Of one or more alternatives.

While the American students in this study had too much knowledge to use the
recognition heuristic, they could have relied on another F FH, known as the Take the Best
(TTB). The TTB heuristic posits that when choosing between two alternatives, a person
will search cues in order of their cue validity (i.e., correlation with criterion), from
highest to lowest, until they ﬁnd a cue that discriminates between the choices.
Subsequently, they will choose the alternative that has the highest value on the
discriminating cue (Gigerenzer & Goldstein, 1996). In the aforementioned study, two
cues that most likely would have been searched ﬁrst, as they are typically highly
correlated with population, are the city’s status as a state capital and the presence of a
professional football team. In this example, the state capital cue would not distinguish
between the two cities, as neither San Diego nor San Antonio is a state capital. Thus,
students would extend their search to the next cue, which is whether or not the city has a

professional football team. Using this cue would successfully discriminate between the

23

cities, as only San Diego has an NFL team. Thus, San Diego would be selected as the
largest city because it has the highest value on the cue that discriminates between the two
options.

The recognition heuristic and TTB heuristic demonstrate two key aspects of FF H:
speed and frugality. These heuristics are fast because they use simple rules which do not
attempt to optimize through the weighting and integration of information. Likewise, these
rules are frugal because they use very limited information to make decisions. Thus, FFH
are especially useful for making decisions in dynamic environments involving time
pressure (e.g., sports). A heuristic can be considered fast and frugal if it displays four
central characteristics (Bennis & Pachur, 2006; Gigerenzer, 2004). The ﬁrst aspect of
FFH is that they exploit evolved capacities. These capacities can be biologically evolved

(e.g., depth perception), culturally evolved (e. g., technological savvy), or the result of
learning (e. g., pattern recognition). F or' instance, the recognition heuristic utilizes the
evolved capacity of recognition memory. A second characteristic of F FH is that they
exploit the structure of the environment. This is the notion of ecological rationality, in
that the effectiveness (i.e., rationality) of a heuristic depends on how well it matches the
environment. In this manner, a rational heuristic ﬁts the environment, whereas an
irrational one does not match the structure of the environment. As a result, FFH seek to
deﬁne the environments in which a heuristic would be effective, as well as those where it
would be ineffective. For instance, the recognition heuristic is proposed to be effective in
situations where the decision-maker is ignorant of one or more of the alternatives, but is

not effective when a person recognizes (or fails to recognize) all of the options.

24

Search rules, stopping rules, and decision rules, collectively known as process
rules, are another vital aspect of F FH. These rules represent the building blocks, or ABCs
of FFH (Gigerenzer, Todd, et al., 1999). Search rules deﬁne what information or cues
will be searched and in what order that search will be conducted. Stopping rules dictate
when the search for new information will be terminated. Likewise, decision rules denote
how decisions will be made based on the available information. It is easy to illustrate
these rules in the TTB heuristic. In TTB, search rules dictate that the cues should be
searched in decreasing order of validity, stopping rules terminate the search once a single
discriminating cue has been found, and the decision rule states that the option with the
highest value on the last cue should be selected. One ﬁnal characteristic of FF H is that
they are simple. The simplicity of F F H is reﬂected by the fact that these heuristics do not
attempt to optimize through the weighting and integration of information. Moreover, the
clearly deﬁned process rules which govern F FH allow these simple rules to be
computationally modeled. As such, the decision processes of FFH are highly transparent
and the algorithms are easily testable (Gigerenzer, Todd, et al., 1999).

As FF H are simple rules based on limited information, many people might
question the accuracy of these heuristics. The accuracy of heuristics is typically evaluated
by comparing their performance with that of computationally complex models that take
into account all of the available information. Overall, research has found that FF H
perform as well as most normative methods, such as multiple regression. In their book
Simple Heuristics that Make Us Smart, Gigerenzer, Todd, and colleagues (1999) devote a
great deal of attention addressing the accuracy of FFH. In fact, an entire chapter is

dedicated to reporting tests comparing the performance of two FFH, TTB and minimalist

25

(same as TTB, but with a random search order) with that of two linear strategies, multiple
regression and Dawes’ rule. Each of these four methods was used to predict which of two
alternatives would score higher on the criterion. The data represented 20 different
environments in various disciplines such as psychology, economics, and transportation.
Some of the speciﬁc tasks required the various techniques to predict differences in
mortality, housing prices, attractiveness, and high school drop-out rate. The ﬁndings
showed that the FF H were indeed fast and frugal, as the minimalist and TTB heuristics
searched for 2.2 and 2.4 cues, respectively. On the other hand, the linear methods used all
of the available information, with a mean of 7.7 cues over the 20 tasks. Additionally, the
authors reported that the beneﬁt of speed and frugality did not cost much in terms of
accuracy. While the multiple regression technique was the most accurate, predicting the
correct answer 77% of the time, TTB performed almost as well with an accuracy rate Of
75%. Meanwhile, the predictions of Dawes’ rule and the minimalist heuristic were
correct 73% and 69% of time, respectively (Gigerenzer, Todd, et al., 1999). Additional
research also supports the notion that F FH can simultaneously be fast, ﬁ'ugal, and
accurate (Gigerenzer, Czerlinski & Martigrron, 1999; Gigerenzer & Goldstein, 1996).
Previous research has examined a few FF H in a sporting context. For instance, the
gaze heuristic has been cited as a tool to help judge and catch ﬂy balls in sports such as
baseball (Gigerenzer, 2004). Judging a ﬂy ball can be a difﬁcult task, as the ﬂight path is
inﬂuenced by several factors such as trajectory, wind resistance, and velocity of the ball.
In order to investigate how athletes perform this difﬁcult task, McLeod & Dienes (1996)
conducted two experiments involving skilled ball catchers. Results of the study indicated

that athletes did not perform complicated calculations (e.g., trajectory, wind resistance) to

26

properly position themselves to catch a ﬂy ball, but rather they used a very simple rule:
keep the angle of the gaze constant. While the study was not an examination of heuristics,
other researchers have subsequently referred to the ﬁndings of the study as the “gaze
heuristic” (Gigerenzer, 2004). Thus, the gaze heuristic states that the perceptual
component of catching a ﬂy ball simply requires an athlete to adjust his or her running
speed so that the angle of gaze does not change. Consequently, this strategy allows
players to intercept the ball at the apprOpriate time and place without conducting complex
mental computations. However, one must also consider the types of environments or
tasks on which this heuristic would be useful. Accordingly, the gaze heuristic is only
ecologically rational in environments in which the ball is high up in the air, on its
descent, and in front of the athlete. For example, a ﬁelder who used the gaze heuristic to
judge a ball that is still on an upward trajectory would end up over-running the ball.
Likewise, a shortstop would not want to use the gaze heuristic to try to ﬁeld a ground ball
(Gigerenzer, 2004).

Furthermore, the recognition heuristic discussed above has also been applied to
the sport setting. For example, Snook and Cullen (2006) examined the recognition
heuristic on a task where participants were presented with a pair Of professional hockey
players and asked to decide which player scored more points in his career. Additionally,
participants were asked to indicate whether or not they recognized the players. In the
event that participants recognized a speciﬁc athlete, they were instructed to write down
what they knew about the athlete. Based on these responses, the researchers created three
different categories of recognition: not recognized, merely recognized, and recognized

with additional knowledge. According to the results, when faced with a choice between a

27

merely recognized player and an unrecognized player, participants chose the recognized
athlete 95% of the time. This answer was correct on 81% of the trials. Similarly, in pairs
where one player was recognized with additional knowledge and the other athlete was not
recognized, the athlete recognized with additional knowledge was selected as the high
scorer in 98% of the occasions. This response was accurate 94% of the time (Snook &
Cullen, 2006). Likewise, the recognition heuristic has also been explored in tasks
requiring participants to predict which team or athlete would win a particular
competition. One such study examined participants’ predictions regarding professional
tennis matches. Participants who recognized only one of the players in a match predicted
that the recognized player would win the match in 90% of the cases (Serwe & Frings,
2006)
Take the First Heuristic

While these examples illustrate how heuristics can be used in the sport domain,
they fail to explain how athletes generate options and choose among them. This is the
speciﬁc aim of the Take the First (TTF) heuristic (Johnson & Raab, 2003; Raab &
Johnson, 2007). According to TTF, information and concepts can be represented by
nodes. Nodes are interconnected and the strength of these connections is determined by
the degree of similarity between the nodes. Accordingly, the more similar a node is to
another, the stronger the connection between the two will be. When a node is activated,
spreading activation stimulates other similar nodes. Spreading activation is sequential in
that only one node is activated at a time, and it operates in order of decreasing similarity
(i.e., highest to lowest). In the context of TTF, the rules governing the spreading

activation are determined by strategy. The strategy a person uses depends on his or her

28

goals, task constraints, and other environmental factors that are speciﬁc to the particular
domain of functioning. For example, Johnson and Raab (2003) proposed that handball
players use functional and spatial strategies. Functional strategies focus on the functional
result (e. g., pass, shoot); whereas, spatial strategies refer to spatial orientation (e. g., left,
right). These strateg'es inﬂuence spreading activation" by deﬁning node similarity. For
instance, a spatial strategy in basketball would closely link “pass to the left elbow” to
“pass to the left baseline” but not to “shoot”. Consequently, “pass to the left baseline”
would be strongly activated, whereas “shoot” would only be slightly activated or it may
not be stimulated at all. In this manner, strategy determines the order in which Options are
generated. Likewise, the repetitive activation of a node, such as through practice in
sports, strengthens its association in the associative network. When faced with a similar
situation, the links with the strongest connections, which should represent high quality
decisions, will be retrieved ﬁrst. Consequently, a person should simply choose the initial
option generated, as that should represent an appropriate decision.

Simply put, the TTF heuristic states that when making a decision in a familiar, ill-
deﬁned task, a person should choose one of the ﬁrst Options that springs to mind. The .
basic underlying principle of this heuristic suggests that options are not generated in
random order, but rather that they are meaningfully generated, with better options coming
ﬁrst. Speciﬁcally, high quality options are generated ﬁrst simply because they have the
strongest connections within the associative network, which is where a person searches
for Options.

Earlier in this chapter, the four central characteristics of F FH were discussed.

These factors suggested that FFH exploit evolved capacitiesas well as the environment,

29

they involve process rules, and they are simple. Thus, it is important to understand how
TTF satisﬁes these qualities. First, TTF exploits the evolved capacities of the associative
memory network, as well as the ability to recognize patterns. Moreover, the heuristic is
proposed to be effective in environments with which the performer is familiar. TTF also
uses process rules. Speciﬁcally, search through the associative network occurs in order of
similarity, from highest to lowest, to the initial option stemming from the stimuli. The
stOpping rule suggests that search should be terminated after only a few options have
been generated. Finally, the decision rule dictates that the ﬁrst Option should be taken.
Furthermore, the TTF heuristic also conforms to the simplicity aspect of FFH, as it makes
no effort to deliberately process, weight, or integrate any of the information.

To date, only two published studies have empirically investigated the TTF
heuristic in sport (Johnson & Raab, 2003; Raab & Johnson, 2007). The ﬁrst study
involved 85 male handball players representing six club teams from Germany and Brazil
(Johnson & Raab, 2003). All of the participants were between the ages of 13 and 18.
Participants in this study performed a simulated sports decision-making task in which the
objective was to generate Options regarding what course of action the player with the ball
should take and subsequently decide which Option was the best one in that situation. The
simulated sports task involved 31 videotaped offensive attack situations in handball.
Approximately 10 s into each clip, the video was frozen while a player in the middle of
the court had possession of the ball. At that time, participants were asked to pretend that
they were the player with the ball and to give three verbal responses. First, they were to
name, as quickly as possible, the ﬁrst Option that intuitively came to mind. Next, they

were asked to give other appropriate options that the player could perform. Finally,

30

participants had to choose which course of action they felt was the best decision in that
situation. The quality of each of the options that was generated by study participants was
evaluated by handball experts. The experts rated the choices on a 5-point Likert scale
ranging ﬁom 0 (inappropriate) to 4 (best possible solution). Option-generation strategy
was evaluated by a computer program that examined all consecutive pairs of Options to
detect patterns supporting either a functional or spatial strategy. The number of
consecutiye pairs reﬂecting a speciﬁc strategy was summed across all trials for each
participant. Subsequently, a participant’s strategy score was calculated by subtracting the
number of strategy pairs from functional pairs. Based on this score, a median split was
used to categorize participants as using a spatial or functional strategy of option-
generation. As was mentioned previously, a person using a functional strategy focused on
the action, such as passing or shooting, whereas participants who used a spatial strategy
were primarily concerned with the direction of the action (i.e., left, right, middle).
Results of the study provided support for TTF. First, the ﬁndings demonstrated
that participants did in fact use the TTF heuristic, as they selected the ﬁrst Option on 60%
of the trials. Moreover, Johnson and Raab (2003) found that the serial position of an
option was negatively related to the quality of the option. In other words, earlier options
were of higher quality than were later ones. It is interesting to note that participants
generated relatively few options (M = 2.3) for each trial. Moreover, option-generation
strategy was found to inﬂuence the number and quality of options that were generated.
Speciﬁcally, participants who exhibited a functional strategy generated more options and

resulted in poorer ﬁnal decisions than did the spatial strategy (Johnson & Raab, 2003).

31

A follow up study conducted by Raab and Johnson (2007) provided a ﬁrrther
exploration of TTF. In particular, this study focused on two aspects that were not directly
addressed in the ﬁrst study: search strategy and expertise. This longitudinal study
involved 69 adolescent male and female handball players who were classiﬁed as expert,
near expert, or non-expert. The study was conducted over a period of 2 years with
measures taken on 4 occasions spaced 6 months apart. Participants in this study
performed a video decision-making task that was adopted ﬁom the original study by
Johnson and Raab (2003). Unlike the original study, the authors sought to get direct,
formal measures Of search strategy and expertise. Eye-tracking data were used to classify
participants into spatial or functional search strategy groups. According to the
researchers, a spatial search strategy was characterized by eye ﬁxations in one area of the
screen; whereas, attending to the entire screen was indicative of a functional strategy.
This direct measurement is in contrast to the ﬁrst study, which inferred search strategy
based on the pattern of options that were generated. Moreover, the researchers also used a
tactical questionnaire to measure tactical expertise for attack situations. Many Of the
results in this study replicate the ﬁndings of Johnson and Raab (2003) thereby supporting
the TTF heuristic. Results. revealed that people had a tendency to choose the ﬁrst option,
that earlier generated options were of higher quality than were later generated ones, and
, people did come up with very many options (M = 3.3).

As a new heuristic, there has been limited research on 'I'I'F. Thus, there are many
limitations and knowledge gaps regarding the role of TTF in Option-generation and
decision-making. A major limitation of the research that has been conducted on TTF is

the failure to take measures to ensure that participants cite their ﬁrst inclination. Johnson

32

and Raab (2003) state that they did not utilize time pressure in their experimental task
because they wanted to foster creative option generation. However, the same study found
that slower-generated ﬁrst options were of higher quality than were initial options that
were generated quicker. This ﬁnding could simply be an illustration of the speed-
accuracy tradeoff, but it could also reﬂect that participants were evaluating some options
before giving their ﬁrst response. Thus, in order to get an accurate evaluation of the TTF
heuristic, it is important that researchers implement measures, such as time pressure, to
try to ensure that participants honestly name the ﬁrst intuitive option.

As mentioned previously, this line of research has not attempted to identify
psychosocial factors, such as self-efﬁcacy, that might inﬂuence a person’s use of TTF.
Previous research has suggested that individual differences, such as intelligence (BrOder,
2002), personality and happiness (Schwartz, Ward, Monterosso, Lyubomirsky, White, &
Lehman, 2002), and expertise and knowledge (Johnson & Raab, 2002; Raab & Johnson,
2007) inﬂuence the use of heuristics. In order to advance our understanding of decision-
making and TTF, it is important to identify the various psychosocial factors that inﬂuence
a person’s tendency to use this speciﬁc heuristic. Related to this concern, previous studies
have not examined TTF in varying game conditions. Research has found that quality Of
decision-making decreases during the ﬁnal, critical moments of competition (Bar-Eli &
Tractinsky, 2000). However, the studies conducted by Johnson and Raab did not vary the
game conditions to determine if participants used TTF during pressure situations.

Another limitation is that previous research has required participants to verbally
state all of their options. However, these options were not written down or recorded for

participants to review. Consequently, the ﬁnal decision of participants may reﬂect biased

33

recall of the generated options, such as occurs in the recency and primacy effects. The
scoring system is also a weakness of the two previous studies on TTF. In most situations
in sport, there are many decisions which would be considered at least minimally
acceptable. For instance, in basketball, almost any decision that doesn’t result in a
turnover could be considered appropriate. However, many of these options vary in the
quality of the decision, ranging from the best decision to one that is simply adequate.
Despite employing a rating system ranging ﬁom 0 to 4, the previous studies have
evaluated decisions by counting only the best decision or by tallying the frequency of all
appropriate decisions. Neither of these methods accurately reﬂects the varying quality of
different decisions.

This dissertation addressed several of the limitations of the previous research. In
particular, the current study used time pressure to encourage participants to give their ﬁrst
intuitive response, investigated the inﬂuence of self-efﬁcacy on TTF, required
participants to write down the potential options, varied game conditions to include
. pressure-ﬁlled, end-of-the-garne situations, and utilized a more comprehensive scoring
system to evaluate decision quality. These facets could help advance the current
knowledge and understanding of decision-making in sport.

Summary

Decision-making is an important determinant of performance in sport. Previous
research has mainly focused on understanding decision-making differences between
expert and novice performers. However, self-efﬁcacy has also been found to be an
inﬂuential factor in determining the speed and accuracy of decisions. Likewise, there has

been little research examining the rules, or heuristics, that people use to make decisions

34

in sport. The TTF heuristic is a promising tool to help explain the process of option-
generation and decision-making in familiar, dynamic environments. The aim of the
present study was to examine the ‘I'I'F' heuristic in sport, as well as how self-efﬁcacy

inﬂuences TTF, option-generation, and decision-making.

35

CHAPTER 3
METHOD

In order to investigate the TTF heuristic, self-efﬁcacy, and decision-making, a
sport-speciﬁc study was devised. This study utilized a video-based decision-making task
in the sport of basketball. This chapter begins by identifying the participants and
instruments involved in the study. Next, the speciﬁc study procedures are detailed.
Finally, the chapter concludes by discussing the treatment of the data and how the data
will be analyzed.
Participants

One hundred and ﬁve undergraduate and graduate students were recruited for the
study. However, only 72 met the criterion of 1 year of competitive basketball playing
experience in which competitive experience was deﬁned as membership on a team that
had a coach and where records and/or standings kept. This criterion was used because
familiarity with a task is a basic tenet of TTF. All participants between the ages of 18 and
30 (M = 21.23, SD = 2.32) participated in the study. Participants were recruited ﬁ'om
various classes Offered in the Kinesiology department. Based on the conceptual
framework self-efﬁcacy theory, participants were required to have requisite incentive to
perform well on the task. Thus, in order to be included in the ﬁnal data analysis,
participants were required to indicate that it was at least moderately important for them to
do well in the basketball decision-making test. Importance ratings were pertinent
because, according to self-efﬁcacy theory, self-efﬁcacy inﬂuences performance only
when the performer has requisite incentive to do well. Similar to the inclusion criteria of

other studies, participants who selected an importance rating below 5 (as described

36

below) were not included in the ﬁnal data analysis (Chase, 2001; George, 1994; Hepler &
Chase, 2008). Consequently, data from two participants was excluded from the ﬁnal
analysis. Therefore, the results reported in this study are based on the ﬁnal sample size of
70 (34 males, 36 females).
Self-Report Questionnaires

Demographic and importance questionnaire. A demographic questionnaire was
used to collect background information from the participants. Speciﬁc items included
age, gender, and year in school. In addition, information was collected regarding
participants’ basketball playing experience and viewing habits. Participants’ playing
experience, at both the competitive and recreational levels, was assessed. Total number
of years of competitive basketball experience, including the number of years at various
competitive levels (e. g., junior high, high school varsity), and the total nrnnber Of years of
recreational basketball experience was assessed. Recreational basketball was deﬁned as
any playing experience that did not meet the criteria for competitive experience (i.e.,
coach, standings). Moreover, participants were asked to indicate the average number Of
hours spent per week watching basketball, either on television or in person. Finally,
participants rated how important it was for them to be successﬁrl in the basketball
decision-making test. Importance ratings were based on an 11-point scale, with O
denoting "not important at all," 5 "somewhat important," and 10 "very important." As
mentioned above, participants who marked an importance rating below “5” were viewed
as lacking the requisite incentive; thus, those participants were not included in any data

analyses. The demographic questionnaire is presented in Appendix A.

37

Basketball knowledge test. In order to assess participants’ knowledge about the
Offensive rules, strategies, and tactics of basketball, a 10-item test was constructed for
this study (see Appendix B). An example question from the test asked: “When a
defensive player is fronting (standing in front of) an offensive low-post player, what pass
can be used to get the post player the ball?” The test was a multiple-choice format with
four possible answers for each question. Three collegiate coaches took the basketball
knowledge test and all received a perfect score. Participants’ basketball knowledge was
deﬁned as the number of correct responses on the 10-item basketball test. The measure
exhibited acceptable internal consistency (or = .70).

Decision-making self-eﬂicacy questionnaire. The decision-making self-efﬁcacy
Of participants was measured through a task-speciﬁc questionnaire designed speciﬁcally
for this study (see Appendix C). The self-efﬁcacy questionnaire comprised 10 items
reﬂecting various aspects of offensive decision-making in basketball. The instructions
asked participants to rate their conﬁdence in their ability to perform various decision-
making tasks in the upcoming basketball video test. Example items included: “know who
to pass the ball to” and “make decisions quickly.” Participants’ degree of certainty was
scored on an 1 1-point scale ranging from 0 (not at all conﬁdent) to 10 (extremely
conﬁdent). The questionnaire was constructed in accordance with Bandura’s guidelines
(2005). Self-efﬁcacy strength was deﬁned as the average efﬁcacy rating across all items
(i.e., sum of all scores divided by 10). The scale exhibited high reliability, as Cronbach’s
alpha was .96.

Rating of conﬁdence in ﬁnal decision. A single-item measure was used to assess

participants’ conﬁdence in each of their decisions. This question asked: “How certain are

38

you that this is the best decision for this situation?” Ratings were based on an ll-point
scale ranging from 0 (not at all conﬁdent) to 10 (extremely conﬁdent). Participants’
option-generation conﬁdence score represented their average conﬁdence on the 13 task
trials, while decision-making performance conﬁdence score reﬂected average conﬁdence
ratings on the 13 decision-making performance trials.
Video Situations

A total of 26 video clips depicting various offensive basketball situations were
used in the study. The video situations were compiled from edited video footage Of a high
school boy’s basketball game. All videos were edited using Windows MovieMaker. High
school boys basketball was used because it is a sport with which most college students
are familiar, as either a player or a spectator. Moreover, using ﬁlm from an actual game
ensured that the clips illustrated real, competitive situations in basketball. The videotape
was ﬁlmed ﬁ'om above, such that all of the relevant offensive and defensive players were
visible in the frame. For each trial, the video clip played and then froze suddenly when a
member of the offensive team was in possession of the ball. The task in both the option-
generation and decision-making performance phases of the study was for participants to
watch a video clip and decide what the player with the ball should do next (e. g., pass the
ball to a speciﬁc tearmnate, drive to the basket, etc.). All videos were viewed on a laptop
computer. Video clips ranged from 4.80 — 21.90 seconds (M = 12.20, SD = 5.13).
Participants were not given information regarding the length of each video.

The video clips used in the study were determined by three collegiate basketball
coaches. Prior to conducting the study, a set of 50 basketball clips were assembled and

sent to the coaches. In addition to the videos, the coaches were also sent a list of potential

39

moves for each speciﬁc situation. After viewing each video situation, the coaches rated
the appropriateness of each of the possible Options on a 5—point Likert scale: 0 =
inappropriate, 1 = somewhat appropriate, 2 = appropriate, 3 = very appropriate, 4 = best
possible. Only situations in which the coaches unanimously agreed on the best possible
Option were eligible for use in the study. Based on these ratings, 26 video situations were
retained for use in the study. Coaches’ ratings showed a high amount of agreement on the
other Options, as the overall intraclass correlation coefﬁcient was .90, with agreement on
individual situations ranging ﬁom .79 — 1.0. Likewise, the experts also rated, from 0 (not
difﬁcult at all) to 5 (very difﬁcult) how difﬁcult it is to make the best decision for each
video clip. Diffrculty ranged from 1.67 — 4.67 for the 26 videos used in the study, with an
average difﬁculty rating of 2.76 (SD = 1.06).

In an attempt to eliminate decision-making advantages based on participants’
primary position, dominant hand, or playing style preference, special care was taken to
try to balance the video clips that were used in the study. Balancing the videos meant
including relatively equal sample of clips that depicted the ball: on the left and right side
of the court, on the inside and outside of the low post (or paint area), and in possession of
guards or post players. While the clips were not perfectly balanced in all Of these aspects,
all Of these variables are adequately represented. Out of the 26 clips, the ball ended up on
the left side of the court 9 times, the middle of the court 4 times, and on the right-hand
side of the court 13 times. Additionally, the ball was located inside (i.e., low post area or
free throw lane) in 12 of the video clips, while 14 clips ended with the ball outside (i.e.,

outside of the free throw lane). Finally, guards (i.e., point guards, shooting guards, off

40

guards) had possession of the ball on 16 occasions with post players (i.e., forwards,
centers) controlling the ball in 10 video situations.

All video clips were presented with speciﬁc game conditions, such as score, half,
and time remaining. These game conditions were important because decisions often
depend on various aspects that are unique to the speciﬁc situation. For example, an
Opportunity for an Open, break-away lay—up is typically the best decision in most
situations. However, a chance at a quick, easy score may not be the best decision in some
situations, such as those at the end of a game where the team with the ball has a slim lead
and needs to run time Off the clock. Most of the video clips (20) used the same game
conditions, which stated that there were approximately 15 rrrinutes remaining in the game
and the score was tied 28-28. However, in order to explore decision-making and TTF in
critical situations, six video clips used different game conditions. These clips depicted
critical situations in which the game was on the line. All of these critical situations took
place with less than a minute to play in the game with no more than three points
separating the two teams. For instance, one situation stated that there were 34 seconds left
to play in the game with the opposing team leading by 3 points.

Option-Generation Task

This study involved two different phases: an Option-generation phase and a
decision-making performance phase. The option—generation part of the study required
participants to generate options regarding what the player with the ball should do next,
decide which option represented the best move, and rate their conﬁdence in that decision.
As soon as the video stopped, participants were instructed to verbally state, as quickly as

possible, the ﬁrst option that came to mind. These responses were required to be stated

41

aloud in order to measure response time. After verbally stating their initial response,
participants immediately wrote the answer down on the provided form. Next,
participants were given approximately 45 s to write down any other options that they felt
were acceptable in that situation. Once participants generated all of their options, they
selected the Option that represented the best possible move and rated how conﬁdent they
were in thatdecision. For an example score sheet from the Option-generation task, see
Appendix D.

An important element of the option-generation task was the time pressure. In
order to evaluate the TTF heuristic, it is imperative for researchers to get an accurate
assessment of participants’ ﬁrst intuitive options. This requires participants to honestly
indicate the ﬁrst idea that pops into their heads. In this study, time pressure, in the form
Of a loud buzzer, was used to encourage participants to state their true ﬁrst intuitive
response. Speciﬁcally, participants were informed that if they failed to give a response in
2 s, then a loud, annoying buzzer would continue to go off until an answer was given.
The buzzer was intended to be an unpleasant stimulus that participants would want to
avoid; thereby, giving them incentive to quickly and honestly state the ﬁrst option that
came to mind. The time frame of 2 s was determined from a small pilot study involving
experienced and novice basketball players. In this pilot study, participants viewed
situations involving various levels of time pressure ranging from 1-5 3. According to the
feedback from the pilot study participants, 2 s was the optimal amount of time, as it
allotted ample time to generate an option, but was short enough to prevent people from

evaluating that Option or generating subsequent choices.

42

Decision-Making Performance Task

The second facet of the study was the decision-making performance phase. In
this task, participants viewed a video clip and then stated their decision as fast as
possible. Participants were informed that this was to be their ﬁnal decision and that they
could not change it. As in the Option-generation phase, decisions were ﬁrst stated
verbally and then recorded on paper. Finally, they rated their conﬁdence in the ﬁnal
decision. Appendix E illustrates an example score sheet for the decision-making
performance task.
Scoring of Tasks

Various performance scores were used in the option-generation and decision-
making performance tasks. The ﬁrst score was decision quality. The decision quality
score for each possible option was determined by the ratings of the collegiate coaches.
Speciﬁcally, the average coaches’ rating was calculated for each potential move in the 26
situations. Subsequently, the average rating was used as the quality score for the
corresponding option. For example, if two coaches rated a potential move “3” (very
appropriate) while one coach gave the same option a quality rating Of “2” (appropriate),
then participants citing that option would receive a decision quality score of 2.67. In the
option-generation phase, each option that participants came up with was given a decision-
quality score, as was the ﬁnal decision. Likewise, the decisions made during the decision-
. making performance tasks were also assigned a decision quality score. The decision
quality score in the option-generation phase was the average decision quality Of all ﬁnal
decisions in the option-generation phase. These scores could range from O to 4.

Additionally, decision quality according to serial position was also calculated by the

43

average score for each position (e. g., 1”, 2‘“) across the 13 Option-generation trials. For
the decision-making performance task, the decision quality performance score
represented the average decision quality score of all the decision-making performance
trials.

The second performance score used in this study was response time.

The computer was used to measure response time, which was recorded in seconds,
rounded to the nearest hundredth. Speciﬁcally, the experimenter was responsible for
recording response time by pressing the computer’s spacebar when participants gave their
responses. This was an important aspect because pressing the spacebar stopped the clock
on the computer. Response time measured how long, from the end of the video scene, it
took participants to give their ﬁrst response (option-generation task) or ﬁnal decision
(decision-making performance task).The ﬁnal response time score for each task was the
average response times across all trials in the corresponding phase.

A few additional performance measures were obtained from the Option-generation
task. One of these scores was TTF ﬁequency. The TTF ﬁequency score simply
represented how often participants choose the ﬁrst option. As such, scores could range
from 0 to 13. Another performance measure from the option-generation phase was the
average number of options-generated for each trial. This score, which was the sum of
generated options divided by the number of trials (i.e., 13), could range from 1-5.
Procedures

Before conducting this study, permission was obtained from the Institutional
Review Board for Human Subjects Research. Participants were recruited from various

classes Offered in the Department of Kinesiology. Prior to initiating any data collection

procedures, informed consent was Obtained for all participants (see Appendix F). Upon
arriving at the testing site, participants completed the basketball knowledge test followed
by the demographic and decision-making self-efﬁcacy questionnaires. Next, participants
performed the decision-making video trials. The two video tasks, option-generation and
decision-making performance, were counterbalanced. In this manner, half of the
participants ﬁrst performed the Option-generation phase followed by the decision-making
performance task; whereas, the other half performed the tasks in the reverse order. While
the order to tasks was counterbalanced, the order of the videos remained the same.
Accordingly, all participants watched videos 1-13 in the ﬁrst phase Of the study, with half
of the participants performing the option-generation task for those videos and the other
half performing the decision-making performance task. Likewise, participants performed
the opposite task for videos 14-26.

There are two ways in which counterbalancing was important in this study. The
ﬁrst was to account for a possible priming effect. In other words, to control for the
possibility that performing the option-generation task might prime participants to perform
better on the decision-making performance task. Another way in which counterbalancing
the order of the tasks was important was to provide another test of the TTF heuristic in
sport. Thus far, research has only taken a within-person approach to examine use of TTF.
However, it may also be fruitﬁrl to take a between-persons approach. Applying the
between-persons perspective in this study, the results of the option-generation task Of one
group of participants can be used to predict the decisions made by the other group of
participants on the performance task. For instance, imagine that 70% of the participants

who performed the option-generation task cited “Pass to A” as their ﬁrst response to

45

Video 1. Based on these results, it is possible to infer that participants used TTF in the
decision-making performance task if they decided “Pass to A” was the best decision for
that same video situation. Assuming the two groups of participants are equal, it would be
expected that approximately 7 0% of the participants performing the decision-making task
would make the decision to “Pass to A”. It should be noted that it is not being suggested
that the between-persons approach be used alone in research on TTF, but rather that it
may provide additional information about TTF when combined with the within-person
approach.

As there were two different performance orders in this study, the remainder of this
section will outline the procedures for those participants who performed the Option-
generation task followed by the decision-making task (i.e., Order A). After completing
the decision-making self-efﬁcacy questionnaire, participants performed the option-
generation task. The option-generation task, as described earlier in this chapter, required
participants to name their ﬁrst, intuitive option, generate other acceptable options, select
the option representing the best decision, and rate their conﬁdence in that decision. Prior
to perforrrring the option-generation task, the experimenter read the task instructions (see
Appendix G) to all participants and answered any questions regarding the task. After
completing all 13 of the option-generation trials, participants then completed the
decision-making performance task. Similar to the previous phase, the experimenter read
aloud the instructions for the decision-making performance task (also see Appendix G)
and answered any questions. Finally, all participants were debriefed about the details of
the study. To better understand the procedures, an outline of the protocol is provided

below in Figure 1.

46

Figure 1. Timeline of Events for Order A.

1.

2.

Informed consent

Basketball knowledge test

Demographic questionnaire

Self-efﬁcacy questionnaire

Option- generation task

4a. Review game conditions

4b.

4c.

4d.

4e.

4f.

4g.

4h.

Watch Video 1

Verbally state ﬁrst option (within 2 8)

Write down ﬁrst option

Write down other acceptable options (within 45 3)
Circle Option representing the best decision

Rate conﬁdence in decision

Repeat for Videos 2-13

Decision-making performance task

5a.

5b.

5c.

5d.

5e.

Watch Video 14
Verbally state decision
Write down decision

Rate conﬁdence in decision

Repeat for Videos 12-26

6. Debrief participant

Treatment of the Data

47

Before conducting the primary analyses, the data were screened for outliers,
normality, linearity, and homogeneity. Data screening was conducted in accordance with
the recommendations of Tabachnick and Fiddell (2001). Boxplots and standardized
scores were used to identify univariate outliers. A case was deemed an outlier if it was
more than 3 standard deviations away from the mean. Normality was evaluated by
histograms and skewness and kurtosis statistics. A non-normal distribution was indicated
if the skewness statistic divided by the standard error exceeded :t 2. Bivariate scatterplots
were used to evaluate linear relationships between the variables. Linearity is indicated by
oval-like plot patterns. Finally, homogeneity between the two test orders was evaluated
by Levene’s test of homogeneity.

Data Analysis

Following data screening, descriptive statistics (i.e., mean, standard deviation)
and bivariate correlations were calculated for all independent and dependent variables.
Study hypotheses and research questions were evaluated using various statistical
methods. Speciﬁcally, one-way ANOVAs (Option-Generation Hypothesis 2, Option-
Generation Hypothesis 3, Option-Generation Hypothesis 4, Decision-Making Research
Question 3, and Decision-Making Research Question 4, Decision-Making Research
Question 5), hierarchical multiple regression analyses (Option-Generation Hypothesis 6,
Option-Generation Hypothesis 7, Option-Generation Hypothesis 8, Option-Generation
Research Question 1, Option-Generation Research Question 2, Decision-Making
Hypothesis 9, Decision-Making Hypothesis 10, and Decision-Making Hypothesis 11),
paired t-test (Option-Generation Hypothesis 5), and frequencies (Option-Generation

Hypothesis 1) were used to evaluate the various hypotheses and research questions. All

48

statistical analyses were conducted in SPSS 17. An alpha level of .05 was used for all

statistical tests.

49

CHAPTER 4
RESULTS

The purpose of this study was to examine the TTF heuristic and self-efﬁcacy in
option—generation and decision-making in sport. This chapter presents the results in ﬁve
main sections. The ﬁrst section presents detailed demographic information about the
study participants. In the second section, descriptive ﬁndings, such as the means and
standard deviations, of relevant study variables are cited. Next, the correlations among
various study variables are discussed. The fourth section on preliminary analyses outlines
results regarding data screening and major assumptions. Finally, the chapter concludes by
detailing the results of the various statistical analyses used to evaluate the study
hypotheses and research questions.
Demographic Information

Out of the entire sample of 105 participants, 33 were omitted from data analyses
due to lack Of competitive basketball experience, while 2 other participants were dropped
because of low importance ratings. Thus, the ﬁnal sample comprised 70 participants (34
males, 36 females) who were, on average, 21.19 years of age (SD = 2.25). Participants
reported a moderate amount of playing experience, with almost 5 years of competitive
experience (M = 4.47, SD = 3.33) and over 6 years of recreational basketball playing
experience (M = 6.09, SD = 5.65). In addition, participants represented a wide range of
playing positions, as 28 participants reported “forward” as their primary position,
followed by 19 “shooting guards”, 12 “point guards”, and 8 “centers”. Three participants
did not indicate a primary position. As a whole, participants were avid basketball fans,

watching an average of 3.39 hours of basketball each week (SD = 3.08). Overall, the

50

participants thought that it was quite important to do well on the basketball decision-
making test (M = 7.04, SD = 1.72).
Descriptive Statistics

Means and standard deviations for basketball knowledge, decision-making self-
efﬁcacy, TTF frequency, and quality, conﬁdence and response time scores for the option-
generation and decision-making tasks are presented in Table 1. Overall, participants
demonstrated an adequate knowledge of basketball (M = 6.84, SD = 2.20) and a
moderately high sense of decision-making self-efficacy (M = 6.82, SD = 1.61).
Moreover, participants performed similarly on the option-generation and decision-
making performance tasks in terms of decision quality. On both tasks, participants made
decisions that were classiﬁed as “very appropriate”. Likewise, participants reported
similar levels of conﬁdence in their decisions on the option-generation task as they did on
the decision-making performance task. However, participants made decisions faster
during the option-generation phase (M = 1.79, SD = .43) than in the decision-making
performance task (M = 2.71 , SD = 1.18). This likely reﬂects the time pressure imposed
by the buzzer during the Option-generation task.

Performance (i.e., decision quality, conﬁdence, and response time) on the option-
generation and decision-making tasks was examined according to basketball playing
position. Participants’ use of TTF was also compared by position. Speciﬁcally, a one-way
MANOVA was used to determine if guards (i.e., point guards and shooting guards)
differed from post players (i.e., forwards and centers) on decision-making performance or
use of TTF. Results indicated that the two groups were signiﬁcantly different on only one

variable, decision-making performance quality, F (1, 65) = 4.05, p = .048. As this result

51

was only marginally signiﬁcant and there were no other signiﬁcant differences between
the two groups, position was not included in any of the data analyses.

Additionally, a one-way MANOVA was used to compare male (n = 34) and
female (n = 36) participants on all independent and dependent variables, as well as
pertinent demographic information. Results indicated that males and females differed
signiﬁcantly on several variables. In particular, females had less recreational basketball
experience, watched fewer hours of basketball, thought it was less important to do well
on the task, scored lower on the basketball knowledge test, had lower self-efﬁcacy
beliefs, used TTF less frequently, made lower quality decisions on both tasks, expressed
less decision conﬁdence on both tasks, and generated lower quality ﬁrst Options than did
male participants. Table 2 presents descriptive statistics for all variables by gender.

While there were signiﬁcant gender differences on many study variables, neither
self-efﬁcacy theory nor the TTF heuristic suggest that gender should inﬂuence any of the
relationships among these variables. Consequently, all study hypotheses were evaluated
by two separate analyses. The ﬁrst analysis evaluated each hypothesis as stated, with no
consideration for the inﬂuence of gender. A second analysis included gender as a
covariate (ANCOVA) or control variable (regression). However, gender very rarely
changed the substantive interpretation of study hypotheses. Thus, results of the second
analysis controlling for gender will only be reported when they differ from the original
analysis.

Correlations
In order to gain a better understanding of the relationships among variables, the

correlations between all independent and dependent variables were examined. These

52

results indicated that many of the study variables were signiﬁcantly correlated. For
instance, self-efﬁcacy was signiﬁcantly related (p < .001) to gender (r = .38),
competitive basketball experience (r = .36), basketball knowledge (r = .47), TTF
ﬁ'equency (r = .47), decision quality on both tasks (r = .48 Option-generation, r = . 36
decision-making performance), decision conﬁdence on both tasks (r = .69 option-
generation, r = . 70 decision-making performance), and decision-making performance
response time (r = -.37). In addition to self-efﬁcacy, TTF frequency was signiﬁcantly
correlated with gender (r = .40), competitive basketball experience (r = .26), basketball
knowledge (r = .32), option-generation decision quality (r = .24), decision conﬁdence on
both tasks (r = .39 option-generation, r = .37 decision-making performance), and average
number of generated Options (r = -.42). Table 3 presents the correlations among all
relevant study variables.
Preliminary Analyses

As was mentioned in Chapter 3, the data were screened for outliers, normality,
linearity, and homogeneity prior to conducting any statistical analyses. Based on the
boxplots and z-scores, three instances of univariate outliers were identiﬁed. However,
univariate outliers are not typically problematic unless they contribute to multivariate
outliers (Tabachnick & Fiddell, 2001). Using Mahalanobis distance, it was determined
that the data did not contain any multivariate outliers; thus, all cases were retained.
Moreover, the skewness statistics revealed several violations of univariate normality.
Speciﬁcally, TTF frequency (-2.20), option-generation quality (-2.57), option-generation
conﬁdence (-4.50), decision-making performance conﬁdence (-2.84), and decision-

making performance response time (-4.66) were all signiﬁcantly and negatively skewed.

53

Option-generation number, or the average number Of options generated for each trial,
demonstrated signiﬁcant positive skewness (2.78). Inspection of histograms and the
Kolrnogorov-Smimov tests conﬁrmed these signiﬁcant deviations ﬁom a normal
distribution. Thus, the data were transformed using a square root transformation (TTF
frequency, option-generation quality, decision-making performance conﬁdence, and
Option-generation number) and a logarithmic transformation (Option- generation
conﬁdence and decision-making performance response time). Following transformation,
all variables were normally distributed. Transformed variables can be difﬁcult to interpret
because they lose the original metric. Thus, all data analyses were conducted twice: once
using the skewed raw variables and once using the normally distributed transformed
variables. As the main interpretation Of the analyses did not differ between the raw and
transformed variables, and the original variables are easier to interpret, all results are
based on analyses using the raw variables.

Another statistical assumption that was explored was linearity. All combinations
displayed an oval-like pattern, indicating that the data met the assumption of linearity.
Finally, it was necessary to examine homoscedasticity between the two task orders.
Levene’s test of homogeneity of variance was used to evaluate this assumption. Results
indicated that all variables met this criterion except for decision-making quality (p <
.001). Consequently, all data analyses involving this variable assumed unequal variances.
Intercorrelations among variables were examined to screen for multicollinearity. Based
upon the recommendation Of Pedhazur (1982), the minimum tolerance for
multicollinearity was set at .80. Inspection of the correlations revealed two potential

instances of multicollinearity. In particular, the correlation between Option-generation

54

conﬁdence and decision-making conﬁdence was .89, while quality of ﬁrst option and
Option-generation decision quality were also highly related (r = .81). Due to this high
correlation, these two sets of variables will not be simultaneously entered as predictors in
any of the analyses. All other correlations fell within acceptable ranges (.02 - .70).

As there were two separate orders in this study, it was necessary to determine if
there were any differences between the groups that would prevent the data from being
combined. A one-way MANOVA was used to determine if the groups differed on
basketball knowledge, self-efﬁcacy, or any of the performance measures (quality,
conﬁdence, or response time) on the option-generation and decision-making performance
tasks. The MANOVA indicated that the two orders were signiﬁcantly different on only
one factor: Option-generation response time, F (1 , 68) = 7.49, p = .008. Speciﬁcally,
participants in Order B (who performed the decision-making performance task ﬁrst)
responded 0.27 s faster than those in Order A. However, this difference was relatively
small and only occurred on one variable. Thus, it was deemed appropriate to combine
data from both goups for ﬁrrther data analyses.

Primary Analyses
Option-Generation Hypothesis 1 : Participants will choose the first option generated as
the best decision in more than half of the trials.

Participants’ use of TTF was evaluated by two separate analyses. The ﬁrst simply
calculated the frequency with which participants chose their ﬁrst option as their ﬁnal
decision during the option-generation phase. On 658 out of 910 total trials, participants
made their ﬁnal decision in concordance with their ﬁrst option. In other words,

participants used TTF on 72.3% of the option-generation trials. A closer inspection of .

55

frequency of use of TTF according to gender indicated males used the heuristic (80.0%)
more often than did females (66.0%). However, on average, both male and female
participants used TTF on a majority of the trials.

A second between-persons analysis was also used to evaluate the prevalence of
TTF. For this analysis, one group’s most frequently cited ﬁrst option in the option-
generation task was compared to the ﬁ'equency with which that same option was used by
the other group in the decision-making task. For example, on Situation 1 the most popular
ﬁrst option stated by participants in Order A was “shoot”. A total of 21 out of 38, or
55.3%, of the participants gave this answer. If participants in Order B used the TTF
heuristic when performing the decision-making performance task, it should be expected
that the percentage of participants stating “shoot” would be similar to the results of Order
A (55.3%). For this situation, the ratios were relatively close, as 68.8% (22 out of 32) of
Order B participants decided that “shoot” was the best option. Across all option-
generation trials, the most popular answer for each trial was given 594 out of 988 times
(60.1%). Likewise, Order A participants stated those same options 58.3% of the time
(485 out of 832).

It is also interesting to note that the most popular answer given during the option-
generation phase corresponded with the most popular answer cited in the decision-
making performance task on 22 out Of 26 trials (84.6%). A chi-square test of
independence, based on a 2 x 2 contingency table, was used to statistically compare the
proportion of responses in Order A with those in Order B. The results suggested that the
two groups did not differ in their responses, )8 (l, N = 70) = .63, p = .43. The overall

- percentage of trials on which participants chose the ﬁrst option in the Option-generation

56

phase (72.3%) combined with the congruence between ﬁrst options in the option-
generation task with that of those decisions being made in the decision-making
performance phase suggest that participants often relied on the TTF heuristic when
making decisions in dynamic, time pressured sporting environments.
Option-Generation Hypothesis 2: Option quality will be negatively related to serial
position.

Another major proposition of the TTF heuristic is that the quality Of an Option will
decrease as serial position increases. An inspection Of the raw means supports this
consistent decrease in option quality, as earlier Options (i.e., ﬁrst and second) were of
higher quality than later ones (i.e., fourth and ﬁfth). See Table 4 for the means and
standard deviations for option quality at each serial position. In order to statistically
compare option quality based on the order in which it was generated, a one-way ANOVA
was conducted. In this analysis, option quality was the dependent variable with serial
position serving as the group factor. Results of the AN OVA suggested signiﬁcant group
differences, F (4, 1802) = 47.24, p < .001, n2 = .10, so post hoc tests were used to identify
these differences. As the number of Options at each position varied a tremendous amount
(910 ﬁrst Options, 8 ﬁfth options) and there was heterogeneity Of variances (Levene’s
statistic [4, 1802] = 17.30, p <.001), the Garnes-Howell statistic was used for post hoc
analyses. As illustrated in Table 5, the post hoc tests revealed that the ﬁrst option was
signiﬁcantly better than all other Options, except the ﬁfth. Similarly, the second option
was signiﬁcantly better than the third and fourth options, while the third option was of
greater quality than the fourth option. It is important to point out that the ﬁfth Option was

not signiﬁcantly different from any other options, but that is most likely due to low

57

number of options that were generated in that position. Out of 910 Option-generation
trials, there were only 8 occasions on which participants generated ﬁve different options.
A simple inspection of the means, as well as the results of the one-way ANOVA,
supports the second hypothesis. However, the effect was rather small (r12 = .10).
Option-Generation Hypothesis 3: As the number of generated options increases, the
likelihood that the ﬁrst option will be chosen decreases.

According to TTF, dynamic inconsistency, or the incongruence between the ﬁrst
Option and ﬁnal decision, increases with the number of options a person has to choose
from. In other words, the more options a person generates, the less likely that he or she
will select the ﬁrst option as the ﬁnal decision. A one-way ANOVA was used to compare
participants’ use of TTF based on the number of Options generated. Results indicated
overall signiﬁcant differences, F (4, 905) = 65.33, p < .001. However, this effect was
small, n2 = .22. According to the Games-Howell post hoc test, the only signiﬁcant
differences in dynamic inconsistency were found on situations in which only one option
was generated (see Table 6). Therefore, when participants generated only one option,
they had signiﬁcantly lower rates of dynamic inconsistency than when they generated
two or more options. This ﬁnding makes intuitive sense, as a person simply must choose
the ﬁrst option if that is the only one that they have generated. Based on the post hoc
tests, the signiﬁcant results of the ANOVA should be interpreted cautiously. These
ﬁndings provide limited support for the hypothesized dynamic inconsistency of TTF.
Option-Generation Hypothesis 4: Conﬁdence in ﬁnal decision will be negatively related

to the number of options generated.

58

The fourth hypothesis is based on Johnson and Raab’s (2003) assertion that a
decrease in conﬁdence may be an underlying mechanism for dynamic inconsistency.
Based on this assertion, it would be expected that decision conﬁdence would decrease as
the number of options generated increased. The raw means demonstrate a relatively
consistent decrease in decision conﬁdence. For instance, participants were most conﬁdent
in their decisions when they generated only one option (M = 8.36, SD = 1.72) and
expressed the lowest conﬁdence when four different Options were generated (M = 7.28,
SD = 1.78). The one way ANOVA used to evaluate these means revealed signiﬁcant
differences in decision conﬁdence, F (4, 905) = 14.01, p < .001. However, this effect was
very small, n2 = .06. Similar to the results on the previous hypothesis regarding dynamic
inconsistency, post hoc analyses revealed that conﬁdence was signiﬁcantly different on
situations where one option was generated as compared to situations where two or more
Options were generated (see Table 7). Accordingly, participants had higher conﬁdence
when they came up 'with only one option than when they generated multiple Options.
However, conﬁdence did not continue to decrease as more options were generated. In
other wOrds, participants’ decision conﬁdence was statistically equivalent on situations
where they came up with two alternatives as when they generated ﬁve Options. Thus,
Hypothesis 4 was partially supported.

Option-Generation Hypothesis 5: On average, participants would make better choices if
they only generated one option.

This hypothesis tested one Of the central tenets of TTF: less is more. Based on the
heuristic, it is beneﬁcial to stop option generation in the early stages, as the earlier an

option is generated, the higher the quality of the option. Accordingly, this hypothesis

59

sought to compare the quality of the ﬁrst Option with the quality of the ﬁnal decision.
Simply in terms of means, the average ﬁnal decision (M = 2.94, SD = 0.48) was of
slightly higher quality than participants’ ﬁrst option (M = 2.71, SD = 0.55). The paired t-
test indicated that this difference was signiﬁcant, t(69) = 5.75, p < .001 and that the effect
was moderate-to-large, Cohen's d = .69. Moreover, separate t-tests by gender suggested
that difference between the ﬁrst and ﬁnal Options was signiﬁcant in both males, t(3 3) =
3.78, p = .02, and females, t(35) = 4.71 , p < .001. Further inspection of the data
suggested that the difference was based on participants making a positive switch.
Throughout the study, there were 252 instances where participants did not choose the ﬁrst
Option as their ﬁnal decision. In 161 of those cases, participants switched from a poor
ﬁrst option to a better ﬁnal decision. Conversely, participants switched from a good ﬁrst
option to a poor ﬁnal decision on only 84 occasions. Option-Generation Hypothesis 5
was not supported.

Option-Generation Hypothesis 6: Decision-making self-eﬁicacy will be positively related
to ﬁnal decision quality.

This hypothesis tests the notion that self-efﬁcacy will predict quality Of the ﬁnal
decision. The correlation between the variables (r = .48) suggests that self-efﬁcacy is
positively related to decision quality on the option-generation task. However, in order to
get a more detailed understanding between efﬁcacy and decision quality, it is necessary
to control for variables that might inﬂuence participants’ decision quality. In this study,
variables expected to inﬂuence quality, speed, and conﬁdence of decisions were
basketball knowledge and years of competitive basketball experience. Thus, a

hierarchical regression, with years of competitive basketball experience and basketball

60

knowledge entered at Step 1 and self-efﬁcacy at Step 2, was conducted. Results of this
analysis indicated that self-efﬁcacy was a signiﬁcant and positive predictor of ﬁnal
decision quality on the option-generation task, ,8 = .27, t(66) = 2.40, p = .019, after
controlling for the effects of competitive experience (,3 = .10, t(66) = 0.91 p = .366) and
basketball knowledge (,8 = .374, t(66) = 3.22, p = .002). The overall model predicted 37%
of the variance in option-generation decision quality. Accordingly, self-efﬁcacy
accounted for an additional 6% of the variance in ﬁnal decision quality, R2 change = .06,
F(1, 66) = 5.75, p =.019. Hypothesis 6 was supported.

Option—Generation Hypothesis 7: Decision-making self-efficacy will be negatively related
to option-generation speed, in which higher eﬂicacy beliefs will predict faster generation
speeds.

This option-generation hypothesis examined the predictive validity of self-
efficacy on option-generation speed. First, the correlation between self-efﬁcacy beliefs
and speed of option-generation were examined. Self-efﬁcacy and option-generation speed
were not signiﬁcantly related (r = -.21, p = .075). Likewise, the hierarchical regression,
controlling for competitive playing experience and basketball knowledge, also indicated
that self-efﬁcacy was not a signiﬁcant predictor of speed of option-generation, ,6 = -.11,
t(66) = -0.81, p = .42. These results did not support Option-Generation Hypothesis 7.
Option-Generation Hypothesis 8:Decision-making self-eﬂicacy will predict the
conﬁdence in ﬁnal decision.

The relationship between decision-making self-efﬁcacy and decision conﬁdence
on the option-generation task was examined by correlations and a hierarchical regression

analysis. Based on the correlation, self-efﬁcacy was signiﬁcantly and positively related to

61

decision conﬁdence (r = .69, p <.001). Next, a hierarchical regression was conducted to
examine if self-efﬁcacy beliefs predicted decision conﬁdence in the option-generation
phase after controlling for competitive playing experience and basketball knowledge. The
regression indicated that self-efﬁcacy was a positive and signiﬁcant predictor of decision
conﬁdence in the option generation phase, ,B = .67, t(66) = 6.49, p < .001. This model
predicted almost half of the variance in decision conﬁdence on the option generation task,
R2 = .48. Self-efﬁcacy beliefs accounted for a signiﬁcant amount of additional variance in
option-generation conﬁdence, R2 change= .33, F (l, 66) = 42.13, p < .001. In other words,
self-efﬁcacy predicted 33% of the variance in decision conﬁdence after controlling for
playing experience and knowledge. Thus, Hypothesis 8 was supported.
Option-Generation Research Question 1: Does self-efficacy predict the use of IT F
heuristic?

Point biserial correlations suggested that self-efﬁcacy was signiﬁcantly related to
use of TTF, r = .46, p < .001. Hierarchical regression analysis was used to further
investigate. Once again, competitive playing experience and basketball knowledge were
entered as control variables in Step 1 of the analysis and self-efﬁcacy beliefs served as
the only predictor variable in Step 2. After controlling for experience and knowledge,
self-efﬁcacy was shown to be a signiﬁcant predictor of TTF, ,6 = .39, t(66) = 3.1 1, p =
.003, accounting for 11.2% of the variance, F(1, 66) = 9.66, p = .003. Accordingly,
higher perceptions of self-efﬁcacy were associated with choosing the ﬁrst option more
often than lower capability beliefs.

Option-Generation Research Question 2: Does self-efﬁcacy predict the number of

options generated?

62

The predictive validity of self-efﬁcacy on the average number of options
generated was explored via bivariate correlations and hierarchical regression analysis.
The correlation between efﬁcacy beliefs and the number of Options generated was not
signiﬁcant, r = -.20, p = .105. However, the hierarchical regression suggested that, after
controlling for competitive playing experience and basketball knowledge, self-efﬁcacy
was a signiﬁcant, negative predictor of the number of options participants generated, ,8 =
-.35, t(66) = -2.67, p = .01, R2 = .09. These results demonstrate that higher efﬁcacy
beliefs were associated with fewer generated options.

Decision-Making Performance Hypothesis 9: Decision-making self-eﬂicacy will be
positively related to decision-making quality.

This hypothesis examined the relationship between decision-making self-efﬁcacy
and decision quality on the decision-making performance task. The bivariate correlation
between the variables (r = .36) revealed that self-efﬁcacy was positively related to
decision quality. As in the above analyses, a hierarchical regression analysis was used to
examine the relationship between efﬁcacy and decision quality, after controlling for the
effects of basketball knowledge and competitive basketball playing experience. Self-
efﬁcacy was not a signiﬁcant predictor of decision quality on the decision-making task, ,8
= .12, t(66) = 1.04, p = .301 , after controlling for the effects of competitive experience (,6
= .07, t(66) = 0.57, p = .570) and basketball knowledge (,8 = .45, t(66) = 3.69 p < .001).
Therefore, this hypothesis was not supported.

Decision-Making Performance Hypothesis [0: Self-efﬁcacy will be positively related to

decision-making speed.

63

Similar to decision quality, it was predicted that self-efﬁcacy would be
signiﬁcantly and positively related to decision-making speed. The relationship between
self-efﬁcacy and decision speed was evaluated by a hierarchical regression analysis, with
years of competitive basketball playing experience and basketball knowledge controlled
for at Step 1. According to this analysis, self-efﬁcacy beliefs were not signiﬁcantly
related to decision speed on the decision-making task, ,6 = -.21, t(66) = -l .73, p = .089,
after controlling for the effects of competitive experience (13 = .09, t(66) = 0.75, p = .455)
and basketball knowledge (,6 = -.39, t(66) = 3.08, p = .003). It is worth noting that the
self-efﬁcacy-decision speed relationship was approaching signiﬁcance (p < .10) and was
in the predicted, negative direction. Decision-making performance hypothesis 10 was not
supported.

Decision-Making Performance Hypothesis 11: Self-eﬂicacy will be positively related to
decision confidence on the decision-making performance task.

It was also hypothesized that self-efﬁcacy beliefs would be a signiﬁcant and
positive predictor of decision conﬁdence. The correlation between the two variables
supported this relationship (r = .70). Further exploration of this relationship was
examined using hierarchical regression analysis. After controlling for experience (B = -
.08, t(66) = -0.80, p = .42) and knowledge (,6 = .08, t(66) = 0.74 p = .46), self-efﬁcacy
was a signiﬁcant predictor of decision conﬁdence on the decision-making performance
task, ,6 = .69, t(66) = 6.74, p < .001. Self-efﬁcacy beliefs accounted for an additional 35%
of the variance in decision conﬁdence, F ( 1 , 66) = 45.42, p < .001. Thus, this hypothesis

was supported.

64

Decision-Making Performance Research Question 3: Do participants who have a high
tendency to use TT F make better decisions than those who have a low tendency to use
YT F ?

In order to evaluate this research question, participants were split into two
extreme groups based on their TTF score. The low TTF group was deﬁned as those
participants whose TTF score (i.e., frequency) was in the lowest one-third of all study
participants. Participants in the low TTF group (n = 21) chose the ﬁrst option between 3
and 8 times. Conversely, the high TTF group included participants whose TTF scores
were in the top one-third of all performers. Twenty-seven participants were grouped into
the high TTF category and used TTF at least 11 times out of the 13 trials. As this
question related to high and low use of TTF, data from the 21 participants in the
moderate TTF group were omitted. A one-way ANOVA was used to examine differences
between high and low TTF participants on the quality of decisions made on the decision-
making performance task. While the mean decision-making performance of high TTF
participants (M = 3.10, SD = 0.58) was slightly higher than that of the low TTF
participants (M = 2.86, SD = 0.50), the ANOVA failed to ﬁnd signiﬁcant group
differences, F(1, 47) = 2.40,p = .13.

Decision-Making Performance Research Question 4: Do participants who have a high
tendency to use TTF make faster decisions than those who have a low tendency to use
77‘ F ?

As in the previous research question, participants were split into low TTF (n = 21)

and high TT F (n = 27) groups. The decision speed of the two groups was compared using

a one-way AN OVA. Results suggested that there were signiﬁcant differences between

65

high and low TTF participants, F (1, 47) = 6.22, p = .016, Cohen's d = .73. Speciﬁcally,
high TTF participants responded .71 s faster than participants who had a low tendency to
use TTF.
Decision-Making Performance Research Question 5: Are participants who have a high
tendency to use TTF more conﬁdent in their decisions than are those who have a low
tendency to use YT F ?

This research question was evaluated by a one-way AN OVA in which use of
TTF, high or low, served as the between-subjects factor and decision conﬁdence
represented the dependent variable. The ANOVA revealed that high and low TTF
participants reported signiﬁcantly different levels of decision conﬁdence, F (1 , 47) = 9.74,
p = .003. Speciﬁcally, high TTF participants expressed higher conﬁdence (M = 8.32, SD
= 1.31) in their decisions than did low TTF participants (M = 6.99, SD = 1.66).
Additionally, this effect was quite large, Cohen's d = .92. However, an ANCOVA
indicated that decision conﬁdence did not signiﬁcantly differ between the high and low
TTF groups after controlling for gender, F(l , 47) = 3.35, p =5 .074. While not signiﬁcant

at the desired alpha level, this result was approaching signiﬁcance (p < .10).

66

CHAPTER 5

DISCUSSION

The TTF heuristic was proposed as a tool to understand how people generate
options and subsequently choose the best decision from among those options in familiar,
yet ill—deﬁned situations. While early research has provided some support for TTF, there
is much more that needs to be understood and tested about the heuristic. Therefore, the
purpose of this study was to provide a rigorous evaluation of the major propositions of
TTF and to explore the inﬂuence of self-efﬁcacy on TTF and decision-making in sport.
This chapter discusses the ﬁndings of the current study, identiﬁes practical implications
of these results, and outlines future research directions.
Take the First and Option-Generation

The results of this study support many of the predictions made by the TTF
heuristic. First, participants frequently used TTF, as they chose the ﬁrst option on over
70% of the trials. This ﬁnding is similar to previous results, which indicated that people
used TTF approximately 60% of the time (Johnson & Raab, 2003; Raab & Johnson,
2007). Likewise, the between-subjects analysis also supported participants’ high use of
TTF. The frequency with which participants used TTF was somewhat higher in the
current study than reported in previous research. One explanation for the different rates of
TTF may be related to the study participants. Participants in the current study were
undergraduate and graduate students who had at least 1 year of competitive experience.
For many of these participants, this experience may have been when they were very
young (e. g., elementary school) and they may not have participated in organized
basketball for quite some time. However, the Johnson & Raab (2003) sample comprised

current adolescent male handball players. It is possible that, as current sport participants,

67

these athletes may have viewed the test as an opportunity to improve their strategic
decision-making skills. Thus, they may have had more incentive to do well on the task
and may have taken a more critical, analytical approach to making their decisions.
However, the participants in the current sample were no longer participating in
competitive basketball and were not likely looking to use this experience as an
Opportunity to improve their skills. As a result, they may have simply been content to
choose their ﬁrst option because they assumed that it would be at least somewhat
acceptable. Based on the combined ﬁndings, it can be concluded that people often “go
with their gut” when making decisions in dynamic, time-pressured situations in sports
environments.

A second proposition Of the TTF heuristic that was supported in this study was the
inverse relationship between option quality and serial position. Based on the original
conceptualization of the heuristic, option-generation occurs in a meaningful, sequential
fashion in which better options are generated ﬁrst. The results of this study conﬁrm this
hypothesis, as earlier Options were of higher quality than later ones. In this manner, the
ﬁrst option was, on average, the best option, with the second option representing the next
best decision. However, the fourth Option, as Opposed to the ﬁfth Option, was found to be
the poorest option. Likewise, the ﬁfth option was not statistically different than any of the
other positions, including the ﬁrst. This ﬁnding is likely due the very low number of ﬁfth
options that were actually generated. Out of the 910 possible trials, participants only
generated 8 options in the ﬁfth position. With such a small number of instances, one must
be cautious when interpreting the statistical results regarding the ﬁfth Option. The ﬁnding

that option quality decreased with serial position supports Johnson and Raab’s (2003)

68

contention that options are not randomly generated. Rather, Options are generated in a
meaningful order, likely determined by the strength of the connection within the
associative network. This meaningful, sequential order of option-generation can be used
to explain why Option quality decreased according to serial position.

Another aspect of TTF that was investigated in the option-generation task was
dynamic inconsistency. According to the TTF heuristic, dynamic inconsistency, or failure
to choose the ﬁrst option, increases with the number of options that are generated. Upon
ﬁrst glance, it would appear that the results of this study suggested that there were
differences in dynamic inconsistency based on number of options participants generated.
While the overall results of the AN OVA suggested signiﬁcant differences, 112 revealed
that the effect was small. Moreover, the post hoc tests revealed that dynamic
inconsistency was only different for the ﬁrst option. In other words, when participants
generated more than one Option they were more likely to choose an option other than the
ﬁrst one than when they only came up with one choice. This ﬁnding is not surprising
considering that a person must choose the ﬁrst option when it is their only choice.
However, there were no differences in dynamic inconsistency when participants came up
with two, three, four, or ﬁve options. Thus, participants who came up with only two
options were no more or less likely to TTF than were those who generated ﬁve different
choices.

This ﬁnding is in contrast to the heuristic and previous research (Johnson & Raab,
2003; Raab & Johnson, 2007). The previous results may be different due to the way in
which the data were analyzed. Linear regression was used to analyze the results of these

previous studies. As the articles make no mention of omitting any data, it can be assumed

69

that the regression included instances in which only one option was generated. These
instances, which force participants to TTF, would have a powerful impact on the results
by presenting a somewhat misleading positive relationship between dynamic
inconsistency and number of Options. The results from the Johnson and Raab studies
could be considered misleading because the relationship may not hold true except in
cases when only one option was generated. Therefore, it is important to consider the
differences in the manner in which the data were analyzed and interpreted when
comparing the results of the current study with previous ﬁndings.

Another possible explanation for the present study’s ﬁnding is that the
participants may not have really considered the other options, but simply generated those
options because of the experimental conditions. For example, simply knowing that the
researcher would see their answers may have made participants feel pressured to come up
with several alternatives. In other words, participants may have felt that it would look bad
if they only came up with one option, so they cited other options in order to please the
. researcher. However, these additional options may have never been seriously considered
as potential choices for the best decision. Rather, participants may have already known
that the ﬁrst option was the best, but they just wrote down a few other options to give the
false impression that they were contemplating other alternatives.

According to Johnson & Raab (2003), dynamic inconsistency may be related to
one’s conﬁdence. Speciﬁcally, they state that people may be less likely to choose the ﬁrst
Option after generating subsequent ones because they lose conﬁdence that the ﬁrst one is
actually the best. While the current study did not measure conﬁdence in the ﬁrst decision,

it did examine conﬁdence in the ﬁnal decision. By extending the logic presented in

70

Johnson & Raab (2003), it could be assumed that having more options to choose from
would reduce one’s conﬁdence in his or her ﬁnal decision. Initial results indicated that
there were differences in conﬁdence based on the number of Options generated.
Participants were more conﬁdent in their ﬁnal decisions when they only generated one
decision. However, generating two, three, four, or ﬁve options made no impact on
participants’ conﬁdence. These ﬁndings combined with the smalleffect size suggest that
conﬁdence plays at least a minor role in dynamic inconsistency. However, it does not
appear to have a signiﬁcant impact beyond the second Option. Similar to the explanation
offered above, this could be because the participants did not really view the other options
as serious contenders for the best decision; therefore, those additional choices did not
affect conﬁdence in their ﬁnal decision.

A major facet of TTF is the notion that participants can make better decisions
simply by going with the ﬁrst idea that springs to mind. In other words, choose the ﬁrst
option because generating more than one choice only serves to hurt the quality of the
ﬁnal decision. As was discussed earlier in this chapter, Option 1 was better than any of
the other options, on average. However, results indicated that the average ﬁnal decision
was signiﬁcantly better than the average ﬁrst decision. This suggests that on situations
when participants did not TTF, the switch had a positive impact on decision quality. In
fact, out Of the 252 trials on which participants did not TTF, switching from the ﬁrst
Option to the ﬁnal decision proved to be beneﬁcial on 161 of those occasions. The switch
was considered beneﬁcial because it increased decision quality by switching from a poor
ﬁrst option to a good ﬁnal decision. Conversely, participants made a poor switch, from a

good ﬁrst Option to a poor ﬁnal decision, 84 times.

71

This ﬁnding contradicts previous results by Johnson & Raab (2003). These
researchers reported that participants in their study would have done better, on average, if
they had only generated one possibility. In other words, the average ﬁrst option was
better than the average ﬁnal option. Moreover, these researchers reported that participants
predominantly switched from a good ﬁrst Option to a poor ﬁnal decision. It is worth
mentioning that the Johnson and Raab study did not look at the quality of each option,
but rather at the frequency of responses rated as “appropriate”. Therefore, “appropriate”
options were more frequently cited as the ﬁrst option as compared to the ﬁnal decision.
However, this does not account for different levels of “appropriateness”. The scoring
system employed by Johnson and Raab did not differentiate between those responses that
were simply adequate and those that were very good, or even the best. Perhaps if the
researchers had employed an interval scale, not a dichotomous scoring system, they
would have found results similar to the current study.

Another explanation as to why the studies may have yielded different ﬁndings is
related to the participants. As was discussed earlier in this chapter, previous research on
TTF involved current athletes; whereas, the present sample consisted of athletes who
were no longer participating in organized, competitive basketball. Likewise, it is possible
that some participants had not played basketball in several years. This prolonged period
of inactivity in the sport could have had a negative impact on the intuitive decision-
making capabilities of study participants. In other words, the participants may have been
a little “rusty” and it took them a few tries to generate the best answer. Accordingly, the
ﬁndings of the present study suggest that while it may be best to TTF most of the time,

there are instances when it can be beneﬁcial to choose another. Option.

72

Self-Efﬁcacy and Option-Generation

The current study also examined the inﬂuence of self-efﬁcacy on various aspects
of the option-generation task. First, the relationship between self-efﬁcacy and decision
quality in the Option-generation phase was explored. Consistent with previous research on
self-efﬁcacy and decision-making (Bandura & Wood, 1989; Hepler & F eltz, 2008),
efﬁcacy beliefs were positively associated with decision quality in the option-generation
task. Accordingly, participants with higher efﬁcacy beliefs made better decisions than did
participants with lower efﬁcacy beliefs. This ﬁnding is not surprising, as the relationship
between self-efﬁcacy and performance has been consistently supported by theory and
previous research (Bandura, 1997; Moritz, Feltz, Farbach, & Mack, 2000).

Furthermore, the relationship between self-efﬁcacy and speed of Option-
generation was also investigated. In sport, superior decision-making performance can be
conceptualized as making good decisions quickly. Likewise, failure can result from
making a poor decision (e. g., taking a bad shot) or ﬁom making a good decision tOO
slowly (e. g., waiting too long to pass to an open teammate, allowing the defense to
intercept the pass). In this manner, decisions must be of high quality and high speed. As
self-efﬁcacy is purpOrtedly related to performance, and speed is a key aspect of decision-
making performance in sport, it was expected that self-efﬁcacy would predict speed of
Option-generation. However, self-efficacy was not signiﬁcantly related to the speed with
which participants generated the ﬁrst Option. One reason why self-efﬁcacy did not predict
response time may be the time-pressure utilized in the option-generation task. The
purpose of the time pressure was to serve as an incentive for participants to give their

intuitive ﬁrst responses. However, this time-pressure may have forced some participants,

73

particularly those with low efﬁcacy beliefs, to respond quicker than normal. This idea is
supported by the fact that participants responded approximately 1 s faster on the Option-
generation task than they did on the decision-making performance task. Therefore, the
time-pressure used in the option-generation task may have eliminated the inﬂuence Of
self-efﬁcacy on option- generation speed.

In addition to decision quality and option-generation speed, the study also
explored the link between self-efﬁcacy and decision conﬁdence. Intuitively, it makes
sense that people who feel more conﬁdent in their decision capabilities would ultimately
express more conﬁdence in their ﬁnal decisions. Indeed, self-efﬁcacy was positively
related to decision conﬁdence. Participants with higher efﬁcacy beliefs were signiﬁcantly
more conﬁdent than were participants with lower ability perceptions. This suggests that
general capability beliefs extend to situation-speciﬁc conﬁdence.

A particularly unique contribution of the current study was the investigation of the
relationship between self-efﬁcacy and participants’ use of TTF. This research question
was based on the assertion that efﬁcacy beliefs are related to analytic strategies (Bandura,
1997; Wood, Bandura, & Bailey, 1990) and heuristics (Wood et al., 2000). However,
there are no published studies that have examined the link between self-efﬁcacy and the
TTF heuristic. As predicted, self-efﬁcacy was a signiﬁcant and positive predictor of TTF.
Accordingly, higher efﬁcacy beliefs were associated with more ﬁequent use of TTF as
compared to lower capability beliefs. Bandura and Wood (1989) suggest that high
efficacy beliefs facilitate efﬁcient cognitive functioning by allowing people to focus their
full attention and energy on analyzing and solving problems. Conversely, people with

low efﬁcacy beliefs are often distracted ﬁ'om the task at hand by thoughts of personal

-

74

shortcomings, fear of evaluation, and visions Of failure; thereby impairing cognitive
capabilities. Results of the current study lend support for this assertion, as high efﬁcacy
individuals relied on TTF more than low efﬁcacy individuals. Some people may view
TTF as a very risky strategy because it represents intuitive decision-making, as opposed
to logical, rational, and deductive decision-making. Therefore, using TTF requires a
person to have a great deal of conﬁdence in his or her abilities and leaves no room for
second-guessing or fear of failure. In the case of TTF, it is possible that feelings of
efﬁcaciousness allow people to trust their instincts more than people with lingering self-
doubts. This ﬁnding is important because it establishes a clear, empirical link between
efﬁcacy beliefs and the use of judgmental heuristics.

The ﬁnal purpose related to self-efﬁcacy and option-generation examined the
number Of options generated. According to Bandura (1997 , p. 216), “if self-efﬁcacious
individuals ﬁnd solutions readily, they have no need to persist”. In other words, when
people who have a ﬁrm belief in their capabilities generate a desirable ﬁrst Option, there
is no reason to think of additional alternatives. Thus, it was predicted that self-efﬁcacy
would be negatively related to the number of options generated. Results of this study
supported this hypothesis, as high efﬁcacy beliefs were associated with fewer options
than were low efﬁcacy beliefs. Similar to the previous discussion, it is possible that
people with high self-efficacy beliefs simply trust their ﬁrst options more than those with
lower perceptions of efﬁcacy. Therefore, they do not feel it necessary to come up with
alternative choices because they believe that their ﬁrst Option represents a desirable
course of action.

Self-Eﬂicacy and Decision-Making Performance

75

Similar to the option-generation phase, the relationship between self-efﬁcacy and
decision quality, speed, and conﬁdence on the decision-making performance task were
investigated. Unlike the results of the option-generation task, self-efﬁcacy was not a
signiﬁcant predictor of decision quality in the decision-making performance phase, after
controlling for competitive basketball playing experience and basketball knowledge. Not
only is this inconsistent with the results in the option-generation phase, but it is also
contrary to the predictions of self-efﬁcacy theory and previous research ﬁndings. It is
important to note that self-efﬁcacy, by itself, is signiﬁcantly and positively related to
decision quality.

These results may be related to the role of basketball knowledge and competitive
playing experience, as well as the nature of the two experimental tasks. As can be seen
from the results of the statistical analyses and the raw correlations, basketball knowledge
and previous playing experience were important factors in inﬂuencing decision quality on
both the option-generation and decision-making performance tasks. Likewise, both of
these factors were signiﬁcantly correlated with self-efﬁcacy. Thus controlling for
knowledge and experience in the hierarchical multipleregression likely removed a large
amount of common variance in decision quality. This type of statistical control may have
been more inﬂuential on the decision-making performance task than on the option-
generation task due to the amount of time participants had to make their decisions. In the
Option-generation task, participants were allowed up to 45 s to generate various options
and choose among them. During this deliberation, it is likely that participants took into
consideration their perceived decision-making capabilities. In particular, 45 s is a long

time to sit and think about one's decision-making deﬁciencies and may have made them

76

more susceptible to doubt their ﬁrst top choice and pick a different option as the best
decision. So, this prolonged decision time may have allowed self-efﬁcacy beliefs to exert
a stronger inﬂuence on performance. Conversely, the decision-making performance task
required participants to make a ﬁnal decision rather quickly. On average, participants
made those decisions in less than 3 s. This may have been such a short time that self-
efﬁcacy beliefs were not able to exert a powerful enough inﬂuence that could be detected
after removing the effects of competitive playing experience and basketball knowledge.
Perhaps when quick decisions must be made, decision-making self-efﬁcacy exerts its
inﬂuence on the execution of that decision, rather than on the actual decision itself. In
other words, low self-efﬁcacy may make a person hesitate when executing their decision
(e. g., pass to Player B) whereas high ability perceptions allow a person to act without
hesitation. Thus, one should not interpret these results as self-efﬁcacy had no inﬂuence
on decision quality. Rather, that self-efﬁcacy did not exert a detectable inﬂuence when
controlling for other pertinent variables such as knowledge and playing experience.
Another possible explanation for this relationship could be related to a
discrepancy between initial ratings of decision-making self-efﬁcacy and efﬁcacy beliefs
while performing the decision-making performance task. While participants did not
receive any feedback, it is possible that their efﬁcacy beliefs changed throughout the
course of the study, particularly during the decision-making performance task. For
instance, it was not uncommon on the decision-making performance tasks for participants
to state one answer and then ask if they could change their answer. In other words, they
doubted their decision and that doubt served as a source of perceived performance

feedback. Based on their perceived performance, participants' self-efﬁcacy beliefs may

77

have changed during the course of the study. Accordingly, the initial self-efﬁcacy
measure would no longer be an accurate predictor of performance. A key factor to keep
in mind is that this same type of self-doubt probably did not exist on the option-
generation task because participants were allowed an extended period of time to make
their decisions. In fact, participants were slightly more conﬁdent in their decision on the
option-generation task (M = 7.81') than they were on the decision-making performance
task (M = 7.68). However, this difference was not signiﬁcant (p = .14).

Similarly, self-efﬁcacy was also not a signiﬁcant predictor of decision-making
speed, after controlling for the inﬂuence of playing experience and knowledge. This
ﬁnding is congruent with the results regarding the option-generation task. However, it
should be mentioned that this relationship was marginally signiﬁcant (p < .10).
Moreover, the correlation between the two variables was also signiﬁcant. Perhaps with a
larger sample size, this relationship would have been statistically signiﬁcant.

Finally, self-efﬁcacy was also found to be a positive predictor of conﬁdence in
one’s decision. This result, which is in accordance with what was found on the Option-
generation task, suggests that participants with high self-efﬁcacy are more conﬁdent in
their ﬁnal decisions than are participants with low efﬁcacy beliefs. Once again, this
supports the notion that conﬁdence in one’s abilities extends to conﬁdence in one’s
decisions. This could have important implications in sport, as it is likely that a conﬁdent
decision about what move to make next will produce quick, distinct actions. Conﬁdent,
resolute action is a key ingredient to success, because a moment of hesitation or self-
doubt can result in a disastrous outcome.

Take the First and Decision-Making Performance

78

An important aspect in evaluating the TTF heuristic is to determine if using the
rule helps improve decision-making performance. A key element to decision-making
performance is the quality of the decision. Previous research, as well as the Option-
generation phase of the current study, has supported the link between TTF and decision
quality. However, the way in which decision-making was examined in these tasks was
rather artiﬁcial. Previously, participants were asked to watch a video scenario, state the
ﬁrst intuitive option, generate other options, and ﬁnally choose one of the options as the
best decision. Additionally, these actions were performed serially, not simultaneously.
Participants may have spent many seconds or even a few minutes completing these tasks.
In real-world settings, such as those in sports, people are required to do all of these tasks
in an instant. Therefore, the decision-making performance task in the current study was
designed to be more realistic, as participants Were put in a situation and asked to make a
decision. Unlike previous ﬁndings involving Option-generation tasks, results of the
decision-making performance task found that participants who had a high frequency Of
choosing the ﬁrst option did not make better decisions than did participants who used
TTF less frequently. This ﬁnding suggests that using TTF does not produce higher
quality decisions on realistic tasks not requiring explicit option-generation.

One possible explanation for this result relates to the TTF groups themselves. The
TTF groups, which represented the one-third most frequent and least frequent users of
TTF, were created based on the results Of the option-generation task. A main assumption
was that participants would continue to follow the same pattern during the decision-
making performance task. However, participants may have changed how much they

relied on TTF. For instance, participants in one or both groups may have regressed closer

79

to the mean. In this manner, participants in the low TTF group may have became more
frequent users of TTF or those in the high TTF group may have used TTF less often. The
results of this study would have been signiﬁcantly inﬂuenced if participants changed their
use of TTF. However, based on the study design, it is not possible to determine whether
or not participants continued their same pattern of use of TTF.

While the high and low TTF groups did not differ on decision quality, there were
signiﬁcant differences on another aspect of decision-making performance: decision
speed. Participants with a high tendency to TTF made their decisions signiﬁcantly faster
than low TTF participants. Furthermore, the effect size of TTF on decision speed was
moderate-to-large (Cohen's d = .73). This ﬁnding is likely related to the number of
Options generated. Participants in the high TTF group may typically only generate one
option, as one choice is all that is really necessary. However, on trials when participants
did not TTF, then they must have generated at least one additional alternative. Thus, it
would take more time to generate multiple options and choose among them than to
generate and choose one single option. Results from the option- generation phase support
this contention, as participants in the high TTF group generated fewer options (M = 1.79)
than did participants who had a low TTF frequency (M = 2.29). This ﬁnding suggests that
using the TTF heuristic can be advantageous for decision-making by decreasing their
decision response time.

The ﬁnal research question evaluated differences in decision conﬁdence. Results
of an ANOVA suggested that high TTF participants were more conﬁdent in their
decisions than were low TT F participants. However, after controlling for gender, these

differences became only marginally signiﬁcant (p < .10). There are two possible

80

explanations for this ﬁnding. The ﬁrst is related to the number of Options generated. As
mentioned above, it is likely that participants who ﬁ'equently used TTF generated fewer
options than participants who did not use TTF very often. Generating more Options may
have made participants feel less certain in their ﬁnal decision. Differences in self-efﬁcacy
could also explain the different levels of decision conﬁdence between the two TTF
groups. Results of this study found that self-efﬁcacy has a strong relationship with
decision conﬁdence. Upon inspection of the average self-efﬁcacy beliefs of the two TTF
groups, the high TTF group was discovered to have much higher self-efﬁcacy beliefs (M
= 7.70, SD =. 1.18) than participants in the low TTF group (M = 5.79, SD = 1.58).
Consequently, group differences on decision conﬁdence may have been due to group
differences on self-efﬁcacy.
Gender Diﬂerences

Another aspect of this study that warrants discussion is the issue Of gender
differences. Preliminary analyses revealed that males and females differed on many
pertinent study variables. For instance, females scored lower on the basketball knowledge
test, had lower perceptions of decision-making self-efﬁcacy, used TTF less, and had
lower conﬁdence in their decisions than did male participants. These results may be
related to differences in experience playing and watching basketball. There were no
differences between men and women on years of competitive experience or on number of
years at various competitive levels (e.g., junior high, high school varsity). However, there
were signiﬁcant differences on years of recreational experience, F(1, 68) = 17.98, and
average number of hours spent watching basketball per week, F (1 , 68) = 6.87.

Speciﬁcally, males had more years of recreational basketball playing experience (M =

81

8.71, SD = 5.78) and watched more basketball (M = 4.34) than did females (M = 3.54, SD
= 4.25; M = 2.49; SD = 2.70). Basketball viewing habits reﬂected participants' current
behavior. Moreover, it is also likely participants' recreational basketball experience may
have represented their most current basketball playing experience. Only 22 (12 females,
10 males) of the 70 participants played high school varsity basketball. As males reported
rather extensive recreational playing experience, it is likely that they continued to play
basketball even after they stopped playing competitively. Additionally, it is important to
understand that watching basketball and playing the sport recreationally can have a
signiﬁcant inﬂuence on decision-making capabilities and ability perceptions. So, the
gender differences observed in this study may simply be related to more extensive and
more recent experiences with the sport of basketball.
Implications

A critical step in changing or improving behavior is to understand how and why
the behavior occurs. In terms of decision-making, it is important to understand how
people make decisions. In conjunction with previous research, the present study has
helped to shed light on how athletes make decisions in sport. First, this study helped
reveal that not only do people use the TTF heuristic to make decisions in dynamic, time
pressure situations, but also that TTF can be advantageous to decision-making
performance (i.e., decision speed). This ﬁnding suggests that it is important for
experienced athletes to learn to trust their gut instincts when making decisions in sport.
Coaches can play an inﬂuential role in this process by providing positive feedback,
discouraging second-guessing, and teaching athletes what decisions are best in different

situations. This last aspect, training correct intuitive responses, is a key aspect because

82

learning to trust one's gut would only be advantageous if those ﬁrst intuitive responses
were of high quality. Simulation training, including video and real-life situations, could
be used as an important teaching tool to better develop athletes’ decision-making
capabilities. This type of training could be particularly useful for athletes who play key
decision-making positions, such as quarterback, point guard, shortstop, setter, or goal
keeper. Decision training would help to strengthen the connections among high quality
options within the associative network; thereby, facilitating the generation of high quality
ﬁrst Options.

Moreover, results indicated that self-efﬁcacy plays an important role in decision-
making and in the use of TTF. Thus, it would be beneﬁcial to build the decision-making
self-efﬁcacy beliefs of athletes. A few techniques that might be particularly effective at
enhancing decision-making efﬁcacy beliefs are simulation training, guided practice, and
modeling. Simulation training, as mentioned above, could include real-life simulations or
video presentations. Real-life simulations would involve physical practice of various
situations. Many coaches commonly use simulations, such as practicing three-on-two fast
break situations, in their practices. However, more ﬁequent simulation practice involving
a more diverse array of situations could help increase athletes’ efﬁcacy beliefs. Video
simulation training typically involves game and practice footage, but could also include
cutting-edge technologies such as virtual reality or video games. In this type of
simulation training, athletes can practice making decisions in almost any type of
situation. In guided practice, a coach, or other knowledgeable authority, would walk
athletes through various situations and explain why speciﬁc decisions would be good or

bad. This type of training would not only increase self-efﬁcacy, but also help to deveIOp

83

athletes’ situation-speciﬁc knowledge and strengthen connections of the best options in
the associative network. Modeling could be a useful technique by allowing athletes to
Observe the effective decisions made by other players. For a more complete discussion on
efﬁcacy-enhancing techniques in sport, see Feltz, Short, & Sullivan (2008).
Future Research Directions 1

The ﬁndings of this study suggest that TTF is an important heuristic in sport and
that self-efﬁcacy beliefs play an inﬂuential role in TTF and decision-making. However,
TTF is a relatively new construct which has not yet been studied extensively. There are
several new directions in which future research could help enhance our understanding of
TTF and self-efﬁcacy in sport decision-making. First, future research should try to
identify the situations on which people do and do not use the TTF heuristic. According to
research, people use the heuristic 60-70% of the time. But why do people not rely on the
rule 100% of the time? Are there speciﬁc conditions or situations where people do not
use TTF? Moreover, what are the common elements of the 60-70% of the cases in which
people do TTF? This research could also advance our understanding of heuristics by
identifying the rule(s) people use when they do not TTF. Future research should try to
identify those factors which inﬂuence whether or not a person chooses the ﬁrst option.
For instance, there is some research to suggest that pressure can inﬂuence a person’s use
of speciﬁc problem-solving strategies (Beilock & DeCaro, 2007). Accordingly, different
levels of anxiety may affect a person’s tendency to TTF.

The type of task is another factor that could also be explored in the future.
Research should examine TTF in the context of open and closed, interdependent and

independent, and gross and ﬁne motor skills. This research would help researchers better

84

understand who uses TTF, how Often they use TTF, and what sports TTF is advantageous
(i.e., better and/or faster decisions). Likewise, research should also examine option-
generation and decisiOn—making in dynamic, time pressured environments outside Of
sport, such as those encountered by police ofﬁcers, military personnel, and medical
professionals.

In the future, research should attempt to explore TTF in real-world settings. Thus
far, TTF research has been conducted experimentally. However, these ﬁndings on
simulated, experimental tasks may not generalize to real-life decision-making. In other
words, just because participants use the heuristic in the lab does not mean that they use it
to make decisions on the basketball court. Field research would help determine whether
or not TTF is a salient decision-making strategy in the real-world.

Another suggestion for ﬁrture research is to include participants of varying
abilities and experience. This line of research could use novice, intermediate, and expert
performers to examine some of the main predictions of TTF, as well as other pertinent
questions. For example, research could compare participants of different experience
levels on the ﬁ'equency of TTF use. Likewise, it could also explore relationships, such as
the relationship between TTF and decision quality, decision speed, and decision
conﬁdence, and the self-efﬁcacy-TTF link. Another way to vary the participants is to
consider age and maturational factors. It is possible that TTF is related to age, cognitive
development, or physical maturation. For example, TTF may not have a positive impact
on the decision-making performance of 10-year olds, but may be useful for 14-year olds.
Perhaps this effect is not due to age, but rather to experience or cognitive and/or physical

development. Likewise, this research could shed some light how a person develops a

n.

85

tendency to TTF. Furthermore, future TTF research should not only focus on athletes, but
should also examine the decision-making of coaches.

Future studies could also expand on the current study by making minor
methodological changes. One such way would be to include an option-generation task
that did not involve any time pressure. This would help determine if participants cited the
same ﬁrst responses under conditions of time pressure as they did with no time pressure.
This would be important to help evaluate the validity of previous research (i.e., did
previous studies get accurate measures of ﬁrst options) and to help shape the
methodology of ﬁrture research studies. Another way in which an untimed option-
generation task would be useful is to compare those response times to the response times
Observed with time pressure. The current study found a nearly signiﬁcant relationship
between Option-generation speed and self-efﬁcacy; however, it was hypothesized that the
time pressure exerted undue inﬂuence on response times. Thus, the self-efﬁcacy—option-
generation speed relationship could be clariﬁed by introducing a condition with no time
pressure. Another useful modiﬁcation would be to ask participants to indicate the
position of the Option they chose on the decision-making performance task. Participants
were categorized into high and low TTF groups based on the results of the Option-
generation task. However, it is impossible to know if participants exhibited the same
tendencies on the decision-making performance task. In order to truly understand how
TTF inﬂuences performance on the more realistic decision-making performance task,
participants could be asked to list the serial position Of the option that they chose (i.e.,
ﬁrst, second, etc.).

Conclusions

86

To date, there has been very limited research on how people generate and select options
in sport. The TTF heuristic has helped researchers gain better insight into these cognitive
processes. Accordingly, the current study supported the notion that people do in fact use
the TTF heuristic to make decisions, but also that using TTF can improve decision-
making performance. Moreover, a person’s self-efﬁcacy beliefs are also related to TTF

and their decision-making performance.

87

Appendix A
Demographic and Importance Questionnaire
1. Gender (circle ONE): Female Male
2. Age:
3. Year in school (circle ONE):
First Year Sophomore Junior Senior Other

4. Total number of years of organized, competitive basketball experience (i.e., a coach,
records/standings):

5. Please indicate the number of years of organized, competitive (i.e., a coach,
records/ standings) basketball experience at EACH of the following levels:

Elementary (Grades K-6) Junior high (Grades
H.S. Junior varsity (Grades 9-12) H.S. Varsity (Grades 9-12)
H.S. travel ball (e. g. AAU) College club

Other (explain)

 

6. Please indicate the number of years of recreational basketball experience:

 

7. Primary position played (circle ONE):
point guard shooting guard forward center

8. Approximately how many hours of basketball do you watch, in person and/or on TV,
in an average week:

 

9. How important is it for you to be successful in this basketball decision-making skill
test?

0 1 2 3 4 5 6 7 8 9 10

not at all somewhat very
important important important

88

Appendix B

Basketball Knowledge Test

Circle ONE answer for each question.

1.

When a defensive player is fronting (standing in front of) an offensive low-
post player, what pass can be used to get the post player the ball?

a. chest pass c. lob pass

b. bounce pass d. don’t pass it, you should shoot it
How many seconds can an offensive player dribble in place while being
closely guarded by a defensive player?

a. 3 seconds c. 10 seconds

b. 5 seconds (I. as long as they want
In general what should an Offensive player do when he/she is double-teamed
by the defense?

a. change pivot feet 0. shoot

b. pass to the Open player d. dribble penetrate

What player is typically responsible for handling the ball and running the

offense?
a. point-guard c. shooting guard
b. forward d. center

When playing against a zone defense, offensive players should cut to the:
a. baseline c. top of the key

b. elbow d. gaps

89

6. What does the term “weak side” mean?
a. side of court with guards c. side of court away from ball
b. side of court with posts (1. side of court with ball
7. If a defensive player is overplaying (denying) the passing lane, what should
the offensive player do?
a. cut harder c. clear out
b. cut to the basket d. set a screen (pick)

8. A free throw is worth:

 

 

a. 1 point c. 3 points
b. 2 points d. a foul
9. In a “give and go” a person the ball and then
a. shoots, rebounds c. rebounds, passes
b. dribbles, shoots (1. passes, cuts

10. When posting up, players should set up above the:
a. baseline c. free throw line

b. block (1. 3 point line

90

Appendix C
Decision-Making Self-Efﬁcacy Questionnaire
Directions: In this basketball video task you will watch videos depicting offensive
situations in basketball. When the video stops, you must decide what the player with the
ball should do next. ‘

For this video test, please rate your conﬁdence in your ability to...

 

Not at all Extremely
Conﬁdent Conﬁdent

 

1. Make the right decision atthe right moment 0 l 2 3 4 5 6 7 8 9 10

 

2. Knowwhotopasstheballto 0 1 2 3 4 5 6 7 8 9 10

 

3. Make good decisions when playing against 0 1 2 ‘3 4‘ 5 6 7 8 9 10
a zone defense

 

4. Knowwhentoshoottheball 0 1 2 3 4 5 6. 7 8 9 10

 

5. Make good decisions when playing against 0 1 2 3 4 5 6 7 8 9 10
a person-to-person defense

 

 

 

 

 

6. Make decisions when the game is on the 0 l 2 3 4 5 6 7 8 9 10
line

7. Know when to drive to the basket 0 1 2 3 4 5 6 7 8 9 10

8. Make decisions quickly 0 1 2 3 4 5 6 7 8 9 10

9. Make good decisions in critical situations 0 1 2 3 4 5 6 7 8 9 10

10. Know what do to next with the ball 0 1 2 3 4 5 6 7 8 9 10

 

91

Appendix D
Example Option-Generation Score Sheet
Situation 1: 15 minutes to play in 2"“ half, score is tied 28-28.
1. Write down the ﬁrst intuitive Option that came to your mind

a.

 

2. Write down any options YOU feel are appropriate in this situation.

b.

 

C.

 

d.

 

e.

 

3. Please circle the option YOU feel is BEST in this situation.
4. Using the scale below, rate how conﬁdent are you that this is the BEST decision.
Not at all Extremely

Conﬁdent Conﬁdent
0 1 2 3 4 5 6 7 8 9 10

92

Appendix E

Example Decision-Making Score Sheet

Write down the BEST decision for each situation.

Situations 14-23: 15 minutes to play in 2"d half, score is tied 28—28.
14.

 

Using the scale below, rate how conﬁdent are you that this is the BEST decision.
Not at all Extremely

Conﬁdent Conﬁdent
0 1 2 3 4 5 6 7 8 9 10

93

Appendix F
Informed Consent Form

You are being asked to participate in a study conducted by graduate student Teri Hepler, under
the supervision of Deborah Feltz, Ph.D., from Michigan State University. The purpose of this
study is to examine the relationship between self-efﬁcacy and decision-making performance in
sport. It is believed that this study will have practical implications for athletes and coaches
regarding training practices in sport.

As part of this research, you will be asked to complete a questionnaire about your decision-
making capabilities. The questionnaire consists of 10 questions and takes approximately 2
minutes to complete. Also, you will be asked to watch 20 video clips of offensive situations in
basketball and decide upon a course of action you believe to be the best decision in each situation.
The duration of your participation in this study should last no longer than 45 minutes.

Your participation in this study will remain conﬁdential; no one except the principal investigators
will have access to these responses or to participation records. After participating, you will not be
able to be identiﬁed. At the end of the project, responses will be presented at the group level to
ensure the conﬁdentiality and anonymity of individual responses. Group-based ﬁndings will be
made available to those who are interested. Your privacy will be protected to the maximum
extent allowable by law.

Your participation in this study would be greatly appreciated. However, please know that you
may refuse to participate or withdraw from the project at any time and without penalty. Also, you
may also refuse to answer any speciﬁc questions. If you would like to participate, please sign this
form and return it. If you have any questions concerning this study, please contact Dr. Deborah
Feltz, at 517.355.4732 [dfeltz@msu.edu] or Teri Hepler at 517.896.7491 [heplerte@msu.edu].

Ifyou have any questions or concerns about your role and rights as a research participant, or
would like to register a complaint about this study, you may contact, anonymously if you wish,
the Director of MSU’s Human Research Protection Program, Dr. Peter Vasilenko, at 517-355-
2180, Fax 517-432-4503, or e-mail irb@msu.edu or regular mail at 202 Olds Hall, MSU, East
Lansing, MI 48824.

Thank you for your time and cooperation,

 

 

Dr. Deborah F eltz, Principal Investigator Date
Teri Hepler, Graduate Student Date
I , have been informed of and voluntary agree to

 

participate in the above-mentioned study.

 

Signature Date

94

Appendix G

Study Directions

General Directions

In this basketball test, you will watch a total of 26 of basketball video clips. These

clips depict various offensive situations in a high school boys basketball game. When the

video stOps, you will be asked to decide what the player with the ball should do next. You

should be very speciﬁc in your answers. For instance, if you think the player should pass

to a teammate, be sure to indicate the letter of the player he should pass to. Your

performance score will be based on the speed and quality of your decision. High quality

decisions that are made quickly will receive the best scores. Quality of decisions will be

evaluated by basketball experts, while the computer program will determine decision

speed.

A few general things to keep in mind:

assume all offensive and defensive players are of average and equal ability
(for high school boys basketball players)

there is no shot clock (or time limit to shoot) in high school basketball
carefully review the game conditions before viewing the video

You can only watch each clip ONE time, so pay close attention

Clips range from 2-25 seconds

Be speciﬁc (which direction, teammate, etc.)

95

Before we begin, please read and sign this consent form if you agree to
participate. Next, please complete the ﬁrst three pages in the packet, which includes a

short basketball quiz, a background questionnaire, and a self-appraisal inventory.

Directions for Option-Generation Task

This part of the study requires you to watch 13 of video clips, generate potential
options about what move the player with the ball should make, and then to choose the
best decision from among those options. First, you will need to review the game
conditions and then I will start the video. When the video stops, verbally state the very
first option that comes to mind. Please do this as quickly as possible, as 2 seconds after
the video stops, a loud buzzer will sound. The buzzer will continue to go off until you
give a response. However, if you respond in less than 2 seconds, the buzzer will not
sound. When you state your response, I will stop the video and record your response
time. After giving your verbal response, please record this Option on the provided score
sheet. Next, you will have 45 seconds to write down any other options you feel would be
appropriate. After generating all desired options, please circle the Option you feel
represents the best possible decision. Finally, rate how conﬁdent you feel that your
decision is in fact the best decision. In this task, you will always be a member of the team
in black jerseys. Likewise, you are the home team. Remember that your decision will be
rated on speed and quality.

Let’s go through an example, to help familiarize you with the process.

Here are the game conditions for Situation 1. Please read these over carefully, and when

you are ready, I will start the video.

96

Directions of for Decision-Making Performance Task

This part of the study requires you to watch 13 video clips and make a decision
regarding what move the player with the ball should make next. You will need to read the
game conditions before I play the video. After viewing the clip, you will, as quickly as
possible, state what you believe the player should do. Just to remind you, each decision
will beqscored based how fast you make the decision and how good the decision is. In this
task, you are the home team, indicated by the white jerseys.

Let’s do a quick example before we begin the test. Here are the game conditions
for Situation 14. Please read these over carefully, and when you are ready, I will start the

video.

97

Table 1

Means and Standard Deviations of all Independent and Dependent Variables (N = 70)

 

 

Variable Mean Standard Deviation
Competitive basketball experience 4.47 3 .33
Basketball knowledge score 6.84 2.20
Self-efﬁcacy 6.82 1.61
O-G decision quality 2.94 0.48
DMP decision quality 2.97 0.52
O-G response time 1.79 0.43
DMP response time 2.71 1.18
O-G decision conﬁdence 7.81 1.47
DMP decision conﬁdence 7.68 1.56
TTF frequency 9.40 2.14
Number of options generated 1.98 0.61
First Option quality 2.71 0.55

 

O-G: Option-generation task; DMP: Decision-making performance task; TTF: Take the
First

98

Table 2

Descriptive Statistics for all Independent and Dependent Variables and Various
Demographics by Gender (N = 70)

 

 

 

 

_Males (n = 34)_ Females ln = 36)

M SD M SD
Competitive basketball experience 4.26 3.24 4.67 3.44
Recreational basketball experience 8.71 5.77 3.54 4.25
Hours watching basketball 4.34 3.20 2.49 2.70
Importance rating 7.47 1.76 6.64 1.61
Basketball knowledge score 7.82 1.78 5.92 2.18
Self-efﬁcacy 7.44 1.12 6.23 1.79
O—G decision quality 3.08 0.42 2.80 0.49
DMP decision quality 3.11 0.46 2.85 0.54
O-G response time 1.75 0.41 1.83 0.45
DMP response time 2.47 1.26 2.94 1.06
O-G conﬁdence 8.40 0.93 7.26 1.67
DMP conﬁdence 8.30 1.12 7.10 1.71
TTF frequency 10.26 1.64 8.58 2.26
Number of options generated 1.97 0.66 1.99 0.57
First option quality 2.91 0.43 2.52 0.60

 

O-G: Option-generation task; DMP: Decision-making performance task; TTF: Take the
First

99

mo. v m genome-Aw?- oba gouge-SO... “BE 06 SEE ”P:- u—mﬁ conga-Saba mega-commmoow ”ASE Emﬁ eoumcoeoméoumo ”0-0

 

NH

tum.-

3

*mm.

em.-

2

tom. 2.-
ion. :.-
as. E.-
so.
a w

*3. ho.- ...Nm. *nm. *on.

43.- :4. 3.6... 2.- .5...
am. 8. as. we. .2.
.mm. .8.- ..s. :4. .mm.
2.- E.- 8. 49.- 8.-
an. an. .am. am.
3.- B.- S.-

em. as.

.54.

A. o m v m

45m.

*9?-

5%.

1%.

m:-

*hm.

*mm.-

.36.

*Nm.

*5».

N

om. 883:8 .95 .2

m a .- one ages .95 ..:
em. amaze haze .2
*3. base Sens can .a

No; ecumbeow 30:90 me c3852 .w

*8. 668886 0.0 .5

mo: 2:: uncommon 0-0 .0
.mm. cease 0-0 .m
*8. panacea E .4
tom. eases-eons
as. BREE _aeﬂam .N

l-- ooaotomxo weRSm zeﬁoxmmm A

 

SEE-SS uzmnemme ~38 Berton-ones he 26953 meouSoto-O

m 033.

100

Table 4

Means and Standard Deviations of Option Quality according to Serial Position

 

 

Variable Mean Standard Deviation
Option 1 (n = 910) 2.71 1.41
Option 2 (n = 579) 2.18 1.31
Option 3 (n = 253) 1.66 1.23
Option4 (n= 57) 1.19 1.01
Option 5 (n = 8) 1.33 1.24

 

101

Table 5

Post Hoc Test Comparing Option Quality according to Serial Position

 

1. Option 1 ----

2. Option 2 0.53 (.00)
3. Option 3 1.05 (.00)
4. Option 4 1.52 (.00)

5. Option 5 1.38 (.09)

0.52 (00)
0.99 (.00)

0.85 (.39)

0.46 (03)

0.32 (.94) -0.14 (1.00)

 

Mean difference (p-value)

102

 

 

Table 6

Post Hoc Test Comparing 77’ F Frequency according to Number of Options Generated

 

 

1. 1 Option ----

2. 2 Options 0.43 (.00) ----

3. 3 Options 0.45 (.00) 0.02 (.99) ----

4. 4 Options 0.40 (.00) -0.03 (.99) -0.06 (.95) ----

5. 5 Options 0.37 (.34) -0.06 (1.00) -0.08 (.99) -0.03 (1.00) ----

 

Mean difference (p-value)

103

 

Table 7

Post Hoc Test Comparing Decision Conﬁdence according to the Number of Options

 

Generated
1 2 3 4 5
l. 1 Option ----
2. 2 Options 0.78 (.00) ----
3. 3 Options 0.98 (.00) 0.20 (.72) ----
4. 4 Options 1.08 (.00) 0.30 (.81) 0.10 (1.00) ----
5. 5 Options 0.61 (.34) -0.17 (1.00) -0.37 (.94) -0.47 (.91) ----

 

Mean difference (p-value)

104

REFERENCES

Abernethy, B. & Russell, DB. (1987). Expert-novice differences in an applied
selective attention task. Journal of Sport Psychology, 9, 326-345.

Allard, F., Graham, 8., & Paarsalu, ME. (1980). Perception in sport:
Basketball. Journal of Sport Psychology, 2, 14-21.

Allard, F. & Starkes, IL. (1980). Perception in sport: Volleyball. Journal of
Sport Psychology, 2, 22-23.

Bandura, A. (1977). Self-efﬁcacy: Toward a unifying theory of behavioral
change. Psychological Review, 84, 191-215.

Bandura, A. (1997). Self-efﬁcacy: The exercise of control. New York:
Freeman.

Bandura, A. (2005). Guide for creating self-efﬁcacy scales. In F. Pajares & T. Urdan
(Eds.), Self-eﬂicacy beliefs of adolescents (pp. 307-338). Greenwich, CT:
Information Age Publishing.

Bandura, A. & Wood, RE. (1989). Effect of perceived controllability and
performance standards‘on self-regulation of complex decision-making.
Journal of Personality and Social Psychology, 56, 805-814.

Bar-Eli, M., & Tractinsky, N. (2000). Criticality of game situations and decision making
in basketball: an application of performance crisis perspective. Psychology of
Sport &.Exercise, 1, 27-29.

Beilock, S. L., & DeCaro, M. S. (2007). From poor performance to success under stress:
Working memory, strategy selection, and mathematical problem solving under
pressure. Journal of Experimental Psychology: Learning, Memory, & Cognition,
33, 983-998.

Bennis, W.M. & Pachur, T. (2006). Fast and frugal heuristics in sports. Psychology of
Sport and Exercise, 7, 611-629.

Bouffard-Bouchard, T., Parent, 8., Larivee, S. (1991). Inﬂuence of self-efﬁcacy on self-
regulation and performance among junior and senior high-school age students.
International Journal of Behavioral Development, 14, 153-164,

BrOder, A. (2000). Assessing the empirical validity of the Take the Best heuristic as a

model of human probabilistic inference. Journal of Experimental Psychology:
Learning, Memory, and Cognition, 26, 1332-1346.

105

Bryant, P. (2007). Self-regulation and decision heuristics in entrepreneurial opportunity
evaluation and exploitation. Management Decision, 45, 732-748.

Chamberlain, C.J. & Coelho, A.J. (1993). The perceptual side of action:
Decision-making in sport. In J.L. Starkes & F. Allard (Eds.), Cognitive
issues in motor expertise (pp 135-158). Amsterdam: Elsevier

Chase, MA. (2001). Children's self-efﬁcacy, motivation intentions, and
attributions in physical education and sport. Research Quarterly for
Exercise and Sport, 72, 47-54.

Clark, B. & Warren, N. (1935). Depth perception and interpupillary distance as
factors in proﬁciency in ball games. American Journal of Psychology, 47,
485-487.

ESPN (n.d.). Biggest chokes Of the last 25 years. Retrieved February 9, 2008, from
http://sports.espn.go.com/espn/espn25/story?page=listranker/25biggestchokes.

F eltz, D.L., Short, S.E., & Sullivan, P.J. (2008). Self-eﬁ‘icacy in sport. Champaigrr, IL:
Human Kinetics.

George, T. R. (1994). Self-conﬁdence and baseball performance: A causal examination
of self-efﬁcacy theory. Journal of Sport and Exercise Psychology, 16, 38 l -3 99.

Gigerenzer, G. (2004). Fast and frugal heuristics: The tools of bounded rationality. In D. J.
Koehler & N. Harvey (Eds.), Blackwell handbook of judgment and decision making
(pp. 62-88). Oxford, UK: Blackwell.

Gigerenzer, G. (2007). Gut Feelings: The intelligence of the unconscious. New York:
Viking Press. ‘

Gigerenzer, G., Czerlinski, J., & Martignon, L. (1999). How good are fast and frugal
heuristics? In R. EliO (Ed.), Common sense, reasoning, and rationality (pp. 148-
173). New York: Oxford University Press.

Gigerenzer, G., & Goldstein, D. G. (1996). Reasoning the fast and frugal way: Models of
bounded rationality. Psychological Review, 103, 650-669.

Gigerenzer, G., Todd, P.M., & ABC Research Group (1999). Simple heuristics that make
us smart. New York: Oxford University Press.

Goldstein, D.G. & Gigerenzer, G. (2002). Models of ecological rationality: The
recognition heuristic. Psychological Review, 109, 75-90.

Grehaigrre, J-F., Godbout, P., & Boutlrier, D. (2001). The teaching and
learning of decision-making in team sports. Quest, 53, 59-76.

106

Helsen, W., & Pauwels, J.M. (1988). The use of a simulator in evaluation and training of
tactical skills in soccer. In T.Reilly, A. Lees, K. Davids & W.J. Murphy (Eds.),
Science and Football (pp 493-497). London: E. & F .N. Spon Ltd.

Hepler, T.J., & Chase, MA. (2008). Relationship between decision-making, task self-
efﬁcacy, and the performance of a sport skill. Journal of Sports Sciences, 26, 603-
610.

Hepler, T.J., & Feltz, D.L. (2008). Self-efﬁcacy and decision-making
performance on a simulated baseball task. Manuscript submitted for
publication. '

Johnson, J .G., & Raab, M. (2003). Take the ﬁrst: Option-generation and resulting
choices. Organizational Behavior and Human Decision Processes, 91, 215-229.

Martignon, L., & Schnritt, M. (1999). Simplicity and robustness of fast and ﬁ'ugal
heuristics. Minds and Machines, 9, 565-593.

McLeod, P., & Dienes, Z. (1996). DO ﬁelders know where to go to catch the ball or only
how to get there? Journal of Experimental Psychology: Human Perception and
Performance, 22, 531-543.

Moritz, S.E., Feltz, D.L., Fahrbach, K.R., & Mack, DE. (2000). The relation of
self-efﬁcacy measures to sport performance: A meta-analytic review.
Research Quarterly for Exercise and Sport, 71, 280-294.

Morris, D. (1981). The Soccer Tribe. London: Jonathan Cape.

Olsen, E. (1956). Relationship between psychological capacities in college
athletics. Research Quarterly, 27, 78-89.

Raab, M., & Gigerenzer, G. (2005). Intelligence as smart heuristics. In R.J. Sternberg &
J .E. Pretz (Eds.), Cognition and Intelligence: Identifying the mechanisms of the
mind (pp. 188-207). Cambridge, MA: Cambridge University Press.

Raab, M., & Johnson, J .G. (2007). Expertise-based differences in search and option-
generation strategies. Journal of Experimental Psychology: Applied, 13, 158-170.

Schwartz, 3., Ward, A., Monterosso, J ., Lyuborrrirsky, S., White, K., & Lehman, DR.

(2002). Maximizing versus satisﬁcing: Happiness is a matter of choice. Journal of
Personality and Social Psychology, 83, 1178-1197.

Serwe, S., & Frings, C. (2006). Who will win Wimbledon? The recognition heuristic in
predicting sports events. Journal of Behavioral Decision-Making, 19, 321-332.

Shank, M.D. & Haywood, K.M. (1987). Eye movement while viewing a

107

baseball pitch. Perceptual and Motor Skills, 64, 1 191-1 197.

Simon, HA. (1956). Rational choice and the structure of the environment. Psychological
Review, 63, 129-138.

Simon, HA. (1957). A behavioral model of rational choice. Models of Man, Social and
Rational: Mathematical Essays on Rational Human Behavior in a Social Setting.
New York: Wiley.

Simon, HA. (1990). Invariants of human behavior. Annual Review of Psychology, 41, 1-
19.

Snook, B., & Cullen, RM. (2006). Recognizing National Hockey League greatness with
an ignorance-based heuristic. Canadian Journal of Experimental Psychology, 60,
33-43.

Starkes, J .L. (1987). Skill in ﬁeld hockey: The nature of the cognitive advantage. Journal
of Sport & Exercise Psychology, 9, 146-160.

Starkes, J .L. & Deakin, J. (1984). Perception in sport: A cognitive approach to
skilled performance. In W.F. Straub & J .M. Williams (Eds.), Cognitive Sport
Psychology (pp 115-128). Lansing, NY: Sport Science Association.

Starkes, J.L., & Ericsson, K.A. (Eds) (2003). Expert performance in sports. Human
Kinetics: Champaign, IL.

Tabachnick, B.G. & Fidell, LS. (2001). Using multivariate statistics (4“ ed. ). Boston:
Allyn and Bacon.

Tenenbaum, G. (2004). Decision making in sport. Encyclopedia of applied psychology,
1, 575-584.

Tenenbaum, G., Levy-Kolker, N., Sade, S.S., Lieberman, D., & Lidor, D. (1996).
Anticipation and conﬁdence of decisions related to skilled performance.
International Journal of Sport Psychology, 27, 293-307.

Tenenbaum, G., Yuval, R., Elbaz, G., Bar-Eli, M., & Weinberg, R. (1993). The
relationship between cognitive characteristics and decision-making. Canadian
Journal of Applied Physiology, 18, 48-62.

Thomas, K.T. (1994). The development of sport expertise: From Leeds to
MVP legend. Quest, 46, 199-210.

Todd, P.M., & Gigerenzer, G. (2001). Putting naturalistic decision making into the
adaptive toolbox. Journal of Behavioral Decision Making, 14, 353-384.

-

108

 

Wood, R.E., Atkins, P., & Tabemero, C. (2000). Self-efﬁcacy and strategy on complex
tasks. Applied Psychology: An International Review, 49, 430-446.

Wood, R.E., & Bandura, A. (1989). Impact of conceptions of ability on self-
regulatory mechanisms and complex decision-making. Journal of
Personality and Social Psychology, 56, 407-415.

Wood, R. E., Bandura, A., & Bailey, T. (1990). Mechanisms governing organizational
performance in complex decision-making environments. Organizational Behavior
and Human Decision Processes, 4 6, 181-201.

109

 

92

ill

3 02956

l‘lll

ll
3 129

 

 

 

 

 

 

 

 

 

 

 

I...

 

;.

..= 5":