ABSTRACT CARLTON, TROY A. Evaluating the Influence of Sport

ABSTRACT

CARLTON, TROY A. Evaluating the Influence of Sport Context Factors and Coaching Behavior on the Physical Activity Production of High School Athletes During Practice Time. (Under the direction of Dr. Michael Kanters).

Most U.S. children and adolescents do not achieve sufficient levels of moderate-to- vigorous physical (MVPA) on a daily basis. Organized sports have been promoted as a potential strategy to increase physical activity (PA) behavior of youths, however research does not consistently support that sport participation alone contributes to higher rates of PA. The purpose of this study was to examine the practice design and coaching behavior of a variety of high school sports during practices in North Carolina and its association with the PA behavior observed in athletes. The System for Observing Fitness Instruction Time (SOFIT) was used to objectively measure PA in twenty different sports. Data collection was conducted during 598 varsity sport practices in twelve schools. Participants accrued a substantial amount of PA during practice time, but PA production varied. The sports that were with inherently high amounts of running such as cross country, track & field, and soccer had high MVPA. Conversely, sport where there is more standing such as cheerleading, softball, and baseball were lower in MVPA.

A majority of the boys’ sports generated more PA than girls’ sports. The context was significant in determining how much PA occurred. Based on the regression analyses, sports that placed extra emphasis on game simulation, fitness, and skill development drills can be expected to have higher levels of MVPA. Practices where the coach was either promoting fitness or observing had higher MVPA rates than during instruction or management tasks. Regression analyses revealed that practice context (R2 = 0.31) had a greater predictive relationship to PA than sport type

(0.06) and coaching behavior (0.05). The full models explained 41% and 52% of the PA behavior for boy and girl athletes, respectively. Nearly half of the variability in PA production is explained by the sport context factors measured in this study. The results contribute to our

understanding of how different sports and their practice contexts contribute to PA during practices. Managers/coaches can improve PA rates of youths by modifying the environmental factors surrounding sport.

Evaluating the Influence of Sport Context Factors and Coaching Behavior on the Physical Activity Production of High School Athletes During Practice Time

by Troy A. Carlton

A dissertation submitted to the Graduate Faculty of North Carolina State University in partial fulfillment of the requirements for the degree of Doctor of Philosophy

Parks, Recreation, and Tourism Management

Raleigh, North Carolina

2020

APPROVED BY:

______Michael A. Kanters Jason N. Bocarro Committee Chair

______Michael B. Edwards Jonathan M. Casper ii

DEDICATION

This dissertation is dedicated to my wife Marella, sons Sawyer and Holden, and daughter Adalyn.

iii

BIOGRAPHY

Troy was born and raised in Hickory, NC with his three siblings, Ben, Jill, and Wade.

Sports played an influential role in Troy’s life, participating in year round soccer and basketball at both the interscholastic and club level. He received his undergraduate degree, with a Bachelor of

Science in Leisure/Sport Management, from Elon University. After graduating in 2009, Troy worked in the banking industry for three years before applying to graduate school. In October

2012, he received acceptance into the NC State graduate program and enrolled full-time in

January 2013 as a Research Assistant for the North Carolina Community Transformation Grant

(CTG), funded by the Department of Health and Human Services. Under his graduate appointment, Troy served as a project manager for the CTG, evaluating public school athletic facility use with community organizations through shared use agreements.

After completing his Master of Science degree, he matriculated into the Parks,

Recreation, and Tourism Management doctoral program at NC State where he accepted another appointment as the Project Manager for the Healthy Sport Index (HSI), funded by the Hospital for Special Surgery with support from the Aspen Institute. Troy traveled all over North

Carolina visiting public high schools observing the physical activity levels and coaching behavior of varsity sport coaches.

After proposing his dissertation topic, Troy and his family accepted a position as

Assistant Professor of Sport Management at Warner University, a small private college in Lake

Wales, FL. While completing his dissertation, Troy taught undergraduate courses for the sport management major and advised students majoring in both exercise science and sport management.

Troy currently lives in Salisbury, NC with his wife and three children working as an

Assistant Professor of Sport Management at Catawba College in the School of Health Sciences

& Human Performance.

ACKNOWLEDGMENTS

This project was conceptualized and commissioned by the Aspen Institute based out of

Washington, DC as part of its Project Play initiative. I wholeheartedly thank Tom Farrey and the Aspen Institute team for funding my dissertation project and entrusting me with overseeing the research done at all of our participating schools. I would like to thank all participating schools, principals, athletic directors, and coaches for allowing access to their varsity sports teams. I wish I could name you and your school specifically but for confidentiality reasons, I cannot. To the schools, please know that your willingness to open your doors to allow studies to be done on your campus is not common. You were extraordinarily supportive and hospitable towards myself and the other data collectors, thank you! To the teachers and coaches, thank you for spending countless hours after school watching sports practices (sometimes in not so great weather). You could easily have said no and spent your leisure time doing something else.

Instead, you demonstrated interest, stewardship, and care for the sports programs in your community and the students they serve by investing your time and effort in observing practices.

In 2013, I was recruited by my advisor, Dr. Michael Kanters, to become a master’s student at NC State. Who knew that six years later, I would be finishing my PhD under the guidance and mentorship from the same man. In those six years, I’ve had the privilege to witness Dr. Kanters not just as my graduate advisor but as a wonderful husband, father, colleague, and friend. The conversations I remember the most with him were not about theory, statistics, or conferences…. they were about life. The “life talks” that we had were invaluable to me and that’s probably why we worked so well together because he values relationships just as much as I do. He’s probably cringing reading this right now because he never likes to be acknowledged in any formal manner. Instead, he prefers to divert the praise and recognition to someone other than himself. This is a quality of his that goes unnoticed by many but one that I

vi admire about him. One thing I will always remember is texting him after a hard day in my doctoral studies saying something like, “I’m just going to quit and go be a long shore man”. Dr.

Kanters was an expert in talking me off the ledge, giving me sound advice on what next steps to take.

Along with Dr. Kanters, I want to thank the other members of my committee who I consistently referred to as “Murderers’ Row”. I am a big believer in having “a few good men (or women)” to work with and I had a royal flush of good men on my committee. Dr. Edwards, came to NC State as a faculty member about the same time that I did. I was immediately drawn toward his personable and likeable character. After taking a couple of his classes, I asked him.

Dr. Edwards was an invaluable resource for me in understanding both theory and statistics

(mostly statistics). Statistics has always felt to me like mental gymnastics so I would frequent his door and pick his brain about how he would analyze something. Most importantly, I want to thank Dr. Edwards for always living it out. And by that I mean

Dr. Casper….. the Denver Broncos fan. As much as I loved talking sports with him over the years, he always loved a good sports debate and would taunt me if one of my teams lost a big game (and brag if his team won). I also had to endure Super Bowl 50 in 2016 when the

Broncos defeated my beloved Carolina Panthers. Defeated is the wrong word, “crushed” is more like it. Guess who was the first person who asked me if I watched the game that following

Monday morning? Yep…. it was Casper. But in all seriousness, Dr. Casper had the ability to communicate complicated ideas in simple ways to where I could understand them. Thank you for always lending an ear to hear.

I still remember it to this day, reading a journal article by Dr. Bocarro about getting children active through playing outdoors. After reading the paper, I was hooked. I was convinced that I wanted to go to graduate school and study youth sport and recreation for the

vii purposes of creating a more active generation. Dr. Bocarro is one of the most likeable people I have ever known. If somebody does not like Dr. Bocarro, then I would seriously question that individual’s sanity. One of the things I appreciated most about Dr. Bocarro was his consistent open door policy. He very rarely would shut his office door, leaving it open for any student to come in and chat with him. As much as R1 institutions preach research agendas, Dr. Bocarro was truly devoted to his teaching and mentorship, and I took many of his methods into my first teaching appointment.

Dr. Thom McKenzie is not one of my official committee members but he might as well be. Thom wrote the book (literally) on the methodology that I wanted to use in this project and over the past six years, he has shown extreme humility in his accomplishments. Thank you Dr.

McKenzie for allowing me to be a small part of your amazing journey as a scholar.

Dr. Luis Suau wasn’t just one of my first mentors in social science research, he WAS the first mentor. As a scared little master’s student, I was pseudo-assigned to Luis to learn about what would become the most valuable piece of instrumentation in my research: systematic observation. Together, we built and refined iSOPARC and iSOFIT. From day one, he was patient with me and asked for my opinion on things. He didn’t talk down to me, even though he had the “PhD” at the end of his name, he treated me like a friend. I looked up to him. We shared the same faith and family values. It was awesome…. He was awesome! Luis was just the person I needed in my life during my years in graduate school; not a faculty member, not a family member (which are all very important), but that intellectual and emotional resource I needed when I was at a stalemate with my research. After learning of me about to have my first child, Luis personally drove a truck load of baby stuff to me in the rain. And this was valuable stuff too; like a pack n’ play and other furniture items. It was one of the nicest gestures I had ever experienced. Luis, you will never know on this side of heaven the impact you have made

viii on my life. I very rarely thanked you over the years so from the bottom of my heart: THANK

YOU.

TABLE OF CONTENTS

LIST OF TABLES ...... x LIST OF FIGURES ...... xiii

INTRODUCTION ...... 1 Why Coaching Matters ...... 9 Statement of the Problem ...... 11 Research Purpose ...... 13 Significance of the Study ...... 14 Summary ...... 18 Research Questions ...... 19 Primary research question ...... 19 Secondary research questions ...... 19 Delimitations ...... 20 Definition of Terms ...... 21

REVIEW OF LITERATURE ...... 23 Relevant Theories and Concepts ...... 23 Application of the Social Ecological Model ...... 24 Tracking Sport Participation...... 28 Role of Interscholastic Sports ...... 30 Sport and Physical Activity ...... 31 Observing physical activity in sport settings ...... 34 Measures ...... 39 Measures in youth sport participation ...... 41 Measurements for type of sport ...... 44 Direct observation ...... 45 Summary ...... 48

METHODOLOGY ...... 50 Sampling ...... 50 Instruments ...... 55 System for Observing Fitness Instruction Time ...... 55 iSOFIT® ...... 57 Observer Training and Reliability ...... 59 Data Collection Procedures ...... 60 Analysis ...... 63 PA dependent variables of interest ...... 64 Independent variables ...... 65 Inferential Statistics ...... 66 One-way analysis of means ...... 66 Regression models ...... 67 Summary ...... 70

RESULTS ...... 71 Research Questions ...... 72

Descriptive Data of SOFIT ...... 72 Overall Mean Scores by Sport ...... 74 Findings in Practice Context ...... 78 Findings in Coaching Behavior ...... 83 Regression Analysis ...... 86 Sport type models ...... 86 Practice context models ...... 89 Coaching behavior models ...... 91 Full generalized linear model ...... 93 Odds Ratios...... 99 Summary ...... 112

DISCUSSION ...... 116 Findings in the Present Study ...... 117 The relationship between physical activity and sports ...... 118 Practice design as an influential factor in physical activity ...... 123 The importance of the coach ...... 128 Gender differences ...... 131 A Compendium of Sports ...... 133 Significance of Findings ...... 136 Recommendations for Practice ...... 138 Recommendations for Future Research ...... 141 Limitations of the Present Study ...... 147 Strengths of the Present Study ...... 150 Conclusion ...... 151

REFERENCES ...... 156

APPENDICES ...... 192 Appendix A: IRB Approval Letter ...... 193 Appendix B: Sport Action Compendium ...... 194

LIST OF TABLES

Table 3.11 Sampling of North Carolina high schools ...... 51

Table 3.12 Demographic characteristics of sample high schools ...... 52

Table 3.13 Boys high school sports included in the study ...... 54

Table 3.14 Girls high school sports included in the study ...... 54

Table 3.15 Operationalization of SOFIT coaching instruction codes and sport related examples ...... 60

Table 3.16 SOFIT data collection schedule for 2017-2018 academic school year ...... 63

Table 3.17 Description of variables for analysis ...... 63

Table 4.11 Descriptives for completed observations during high school sport practices during the 2017-2018 school year ...... 73

Table 4.12 Ranking of observed high school sports based on average METs ...... 77

Table 4.13 Mean differences of METs and practice context for SOFIT observations ...... 79

Table 4.14 Duncan post hoc homogeneous subsets for practice context ...... 80

Table 4.15 Mean differences of METs and coaching behavior for SOFIT observations ...... 83

Table 4.16 Duncan post hoc homogenous subsets for coaching behavior ...... 84

Table 4.17 Model 1 – Linear regression parameter estimates of METs based on sport type...... 87

Table 4.18 Model 2 – Linear regression parameter estimates of percent practice time spent in MVPA based on sport type ...... 88

Table 4.19 Model 3 – Linear regression parameter estimates of METs based on practice context ...... 90

Table 4.20 Model 4 – Linear regression parameter estimates of percent practice time spent in MVPA based on practice context...... 91

Table 4.21 Model 5 – Linear regression parameter estimates of METs based on coaching behavior ...... 92

Table 4.22 Model 6 – Linear regression parameter estimates of percent practice in MVPA based on coaching behaviors ...... 92

xii

Table 4.23 Model 7A – Summary of regression analysis for variables predicting METS for boy athletes during high school sports practices ...... 94

Table 4.24 Model 7B – Summary of regression analysis for variables predicting METS for girl athletes during high school sports practices ...... 95

Table 4.25 Model 8A – Summary of regression analysis for variables predicting percent time in MVPA for boy athletes during high school sports practices ...... 96

Table 4.26 Model 8B – Summary of regression analysis for variables predicting percent practice time in MVPA for girl athletes during high school sports practices ...... 98

Table 4.27 Model 9 – Ordinal logistic regression results with odds ratios for the effect of sport type on physical activity for high school athletes ...... 100

Table 4.28 Model 10 – Ordinal logistic regression results with odds ratios for the effect of practice context on physical activity for high school athletes ...... 102

Table 4.29 Model 11 – Ordinal logistic regression results with odds ratios for the effect of coaching behavior on physical activity for high school athletes ...... 103

Table 4.30 Model 12 – Ordinal logistic regression results with odds ratios for the effect of sport type on percent of practice time spent in MVPA for high school athletes ...... 104

Table 4.31 Model 13 – Ordinal logistic regression results with odds ratios for the effect of practice context on percent of practice time spent in MVPA for high school athletes ...... 106

Table 4.32 Model 14 – Ordinal logistic regression results with odds ratios for the effect of coaching behavior on percent of practice time spent in MVPA for high school athletes ...... 107

Table 4.33 Model 15 – Ordinal logistic regression results with odds ratios for the effect of independent variables on PA for high school athletes ...... 107

Table 4.34 Model 16 – Ordinal logistic regression results with odds ratios for the effect of independent variables on percent of practice time spent in MVPA for high school athletes ...... 110

xiii

LIST OF FIGURES

Figure 1.11 Percentage of high school students who were physically active for 60 minutes on five or more days per week ...... 2

Figure 1.12 Percentage of high school students who were physically active for 60 minutes every day of the week ...... 3

Figure 1.13 Percentage of high school students who played on at least one sports team ...... 5

Figure 1.14 Health through sport conceptual model ...... 6

Figure 2.11 Model showing strength of reach and permanence across ecological influences ...... 25

Figure 2.12 Social-ecological model for enhancing health outcomes of youth sport participation ...... 26

Figure 3.11 Location of public NC high schools selected as study sites ...... 52

Figure 3.12 Example of iSOFIT® summary of PA observed from a practice session ...... 57

Figure 3.13 Example of iSOFIT® summary of practice context observed from a practice session ...... 58

Figure 3.14 Example of iSOFIT® summary of coaching behavior observed from a practice session ...... 58

Figure 3.15 Primary iSOFIT® counter tool ...... 62

Figure 4.11 Physical activity levels across sports for girls ...... 75

Figure 4.12 Physical activity levels across sports for boys ...... 75

Figure 4.13 Frequency and distribution of practice context measures ...... 81

Figure 4.14 Frequency and distribution of girls sport practice context measures ...... 82

Figure 4.15 Frequency and distribution of boys sport practice context measures ...... 82

Figure 4.16 Distribution of physical activity and coaching behavior ...... 84

Figure 4.17 Frequency and distribution of boys sport coaching behavior measures ...... 85

Figure 4.18 Frequency and distribution of girls sport coaching behavior measures ...... 86

xiv

Figure 5.11 Relationships between physical activity and sport ...... 139

INTRODUCTION

“This is the least physically active generation in history and we should never get comfortable with that.”

-Caitlin Morris, Nike General Manager of Global Community Impact, 2017

Participation in physical activity (PA) has been associated with numerous health benefits in children and adolescents (Janssen & LeBlanc, 2010). Current PA guidelines recommend children accrue 60 minutes of moderate-to-vigorous physical activity (MVPA) per day (U.S.

Department of Health and Human Services, 2018). Youths who meet this criteria can be expected to improve their health and well-being status in a number of different areas, including decreased cardiovascular disease risk factors (Carnethon et al., 2005), lower blood pressure

(Nielsen & Andersen, 2003), weight management (Ness et al., 2007), increased bone density

(Petit et al., 2002), muscular (Kriemler et al., 2011) and cardiovascular fitness (Fernhall &

Agiolasitis, 2008). Most notably, the depth and breadth of PA involvement during early ages has shown to be strongly associated with adopting a more active lifestyle into adulthood (Ekblom-

Bak, Ekblom, Andersson, Wallin, & Ekblom, 2018).

Despite the overwhelming evidence of the health benefits of PA, most U.S. children and adolescents are not sufficiently active (Bassett et al., 2013; Stefan et al., 2018). For instance, only

22 percent of adolescents meet the national PA guidelines (YRBSS, 2017) and one out of five high school aged youths are completely inactive as reported by the Physical Activity Council

(2019). Approximately 26 percent of American youth in high school participate in 60 minutes of

PA every day (Figure 1.12) and in North Carolina, it is even lower for females (13.9 percent) as opposed to males (30.7 percent) (YRBSS, 2017). On average, 47 percent of high school youths participate in 60 minutes of PA on at least 5 days of the week (Figure 1.11) (YRBSS, 2017).

Again, in the state of North Carolina, high school males (50.5 percent) are substantially more

2 active than females (34.2 percent). Self-report measures of PA confirms that 24.2 percent of youths meet daily PA guidelines (NSCH, 2017), with boys (28 percent) being more physically active than girls (20 percent). These numbers have been fairly consistent over the past 15 years.

Figure 1.11 Percentage of high school students who were physically active for 60 minutes on five or more days per week (YRBSS, 2017).

70 63.1 60 56.4 56.3 51.2 50 45.3

40 34.2 34.6 32.2

Percent 30

0 9th 10th 11th 12th Grade Level

Male Female

Decline in sport participation with age parallels trends in PA. In the U.S., 2016 data shows participation in school sport declined from the time children entered high school (14 years old), to the time they graduated high school (18 years old), 64 percent to 61 percent in boys and 63 percent to 52 percent in girls (Child Trends Data Bank, 2018). Trends are corroborated by the U.S. sporting goods industry which indicated a decline in participation in team sports form 12 to 26 years old (Physical Activity Council, 2019).

Given the high proportion of youths who participate in high school sport (and sport in general), coupled with the myriad of health and developmental benefits associated with sports

3 participation, sport has the potential to be a powerful health-promoting environment for adolescents. And one of the most pertinent attributes of participating in sports, regarding health outcomes, is its potential to contribute considerably to levels of MVPA in adolescents

(Guagliano et al., 2013; Wickel & Eisenmann, 2007). Compared to other activities such as P.E. classes and out-of-school leisure time, participating in sport can more effectively facilitate the achievement of youths meeting the recommended 60 minutes of PA. But certainly, all of the leisure domains in which PA can occur (i.e., active transportation, active recreation, school-based activities, and household activities) complement each other in this regard.

Figure 1.12 Percentage of high school students who were physically active for 60 minutes every day of the week (YBRSS, 2017).

39.7 40 36.7 34.5 29.8 30

22.0

Percent 20 15.2 15.9 16.4

0 9th 10th 11th 12th Grade Level

Male Female

Sport is one of the most popular leisure-time activities among young people (Geidne et al., 2013), and is often organized in the form of private sports clubs, non-profit community organizations, public recreational offerings, or interscholastic athletics at schools (DeMartelaer &

Theeboom, 2006). Consequently, governments and local municipalities are beginning to take

4 considerable interest in promoting sport, particularly youth sport, in their communities

(USDHHS, 2018). With so many children participating in highly structured, organized, and adult-led physical activities on a daily basis (National Physical Activity Plan Alliance, 2018), the role that sports can play in adolescent development is undeniable. However, discussions on how organized sport can serve as a method to facilitate more daily PA and an active lifestyle are equivocal (Demetriou & Honer, 2012; Janssen et al., 2010; McKenzie, 2019). Researchers have yet to determine the quantity or quality of high school after school sport that actually contributes to adolescents’ health (Curtner-Smith et al., 2007).

According to the National Federation of State High School Associations (NFHS, 2017), over 7.8 million high school students participated in interscholastic sports in 2015-2016 and this number has grown by one million participants every decade since the 1970s. In 2017, 58 percent of youth, ages 6 to 17 years, participated in team sports or took sports lessons after school or on weekends in the previous 12 months (Figure 1.13) (NSCH, 2017; USDHHS, 2019). An estimated 54 percent of high school students report participating in sports, and this estimate has remained relatively stable since 1999 (YRBS, 2017), so there is a clear opportunity to support youths getting more PA through sports given the growing participation rates. According to the

YRBS (2017), 60 percent of boys and 49 percent of girls in high school report participating on at least one sports team in the past year, and participation by both girls and boys decreased as grade level increased (Figure 1.13).

Figure 1.13 Percentage of high school students who played on at least one sports team (YRBSS, 2017)

70 63.9 59.2 59.5 60 56.4 55.9 49.2 50 47.0 43.8

Percent 30

0 9th 10th 11th 12th Grade Level

Male Female

The Social Ecological Model (SEM) is commonly used among community organizations and public health professionals as a guiding theoretical framework to increase the PA behavior of the U.S. population (McLeroy et al., 1988). The SEM illustrates an interactive relationship between individuals, their social network, their environment, and relevant policies – emphasizing the need to address behavior at multiple levels of influence. Ecological models of health behavior have been adapted and applied to different settings but the general SEM posits five spheres of influence: intrapersonal, interpersonal, organizational, environmental, and public policy (Mehtälä, et al., 2014; Richard et al., 2011). The model proposes that health behaviors are impacted at several different levels (Figure 1.14).

Starting at the center, individual factors such as knowledge and skills about PA are important factors influencing behavior. Moving outward, interpersonal relationships with parents, peers, and coaches also influence the PA of children and adolescents. Organizations

6 also have a role to play and in terms of adolescent sport participation, the mode (e.g., team sports, individual sports, organized but noncompetitive sports), setting (e.g., schools, clubs, neighborhood setting), and type (e.g., specific sports such as basketball and swimming) are included in this level (Eime et al., 2013). The school setting, especially through PE and interscholastic sports, plays an important role in increasing PA levels in youths (Dudley et al.,

2011). Finally, federal, state, and local policies determine funding, laws, and programs that impact all of the previous levels. One of the unique aspects concerning the delivery of sport for the purpose of increased PA production, especially among youths, is that the multiple levels of influencers touted by the SEM all play a major role in determining the effectiveness of the sport on the individual participant (Fenton et al., 2017).

Figure 1.14 Health through sport conceptual model (Eime et al., 2013)

Although youth sport may provide an ideal opportunity for youth to accumulate substantial amounts of MVPA, a large proportion of players’ time during organized youth sports

7 is spent inactive or in light intensity activity (Guagliano et al., 2013; Leek et al., 2011; Sacheck et al., 2011). For example, some studies show that children spend up to 70 percent of their time playing sport either completely inactive or minimally active (Leek et al., 2011; Wickel &

Eisenmann, 2007). While sports-playing children are generally more active, have greater energy expenditure and spend more time in MVPA than non-participants, data are limited about how their PA varies with types and levels of sports participation (Bailey et al., 2015; Silva et al., 2013).

So, while existing studies indicate that participation in youth sport offers children and adolescents the opportunity to accrue a substantial amount of MVPA (Leek et al., 2011), it seems to be the case that PA accrued during sports is not sufficient to meet recommended levels of MVPA (Gould & Carson, 2008; Ridley et al., 2018).

In sport, coaches carry considerable influence over their players (Conroy & Coatsworth,

2006). Coaches, then, may be able to increase their players’ PA levels, particularly during practice, where they are more capable of influencing PA intensity, as compared to games.

Furthermore, there has been a call to evaluate strategies for how to best increase MVPA in youth sports by also reducing the amount of sedentary behavior commonly found during practices (Kanters et al., Ridley et al., 2018; Schlechter et al., 2018). In the PA contexts most frequently experienced by high school students (e.g., youth sport, physical education (PE) classes), the social environment is largely dominated by the interpersonal styles of adults acting within these settings (i.e., the coach created social environment). Within the context of high school sport, the interpersonal behaviors of the coach are a prominent factor influencing the social environment and athlete behavior (Mageau & Vallerand, 2003). But the reality is that coaches are also concerned with goals other than maximizing PA production of athletes such as technique, strategy, and tactics. It is important to recognize that high school coaches are not solely focused on the MVPA makeup of their practices so strategies and training for coaches in

8 this area may have a ceiling on their effectiveness and even prove to be negligible (Schlechter et al., 2017).

Because of the increasing number of adolescents participating in interscholastic sports in the U.S., high schools and their student athletes have become an important target for community based interventions and policies of different kinds (Bailey et al., 2009; Priest et al.,

2008; Ridley et al., 2018; Schlechter et al., 2017). At the organizational level, the aggressive expansion of youth sport organizations, particularly in the private sector, has resulted in major fluctuations and heterogeneity in coaching strategies and accepted standards of practice (Fenton et al., 2017; Rocchi et al., 2013).

High school sports in the U.S. are fragmented with little to no continuity from program to program (Camiré, 2014). While programs are governed by the National Federation of State

High School Associations (NFHS) and also a state level governing body (e.g., North Carolina

High School Athletic Association (NCHSAA)). The philosophy and organization of each sport team is based almost entirely on the perspectives and goals of its sponsoring school principal and adult coach who manages the specific team. One of the palpable downsides to not having a cohesive interscholastic sports philosophy is the overabundance of untrained coaches due to the lack of requirements for coach education. Although it is logical that increased coach education and training would in turn, increase the capacity of a coach run a physically active practice,

Schlechter et al. (2017) discovered that this is not always the case. Coaches usually take the form of a volunteer parent who lacks the time, knowledge, and resources to learn about youth developmental principles and proper coaching protocols (Conroy & Coatsworth, 2006; Seefeldt

& Ewing, 1997). As a result, many youth sport programs consist of structures and standards imposed by adults unfamiliar with age-based developmental orientations of children (Guagliano et al., 2014). Male coaches in particular may also feel that they already understand sports and do

9 not need training to work with adolescents (Schlechter et al., 2017), especially if it is for a sport they have previous experience playing.

Why Coaching Matters

It is well-documented that coaches play an integral role in shaping youths’ performance, participation, and personal development in sports (Smith & Smoll, 2007; Vella et al., 2011).

Coaches are individuals who provide the structure that contributes to the variability in MVPA of athletes. Dudley et al. (2011) performed a large meta-analysis on the effectiveness of PE and school sport interventions targeting PA. The researchers concluded that the available evidence supports the notion that teachers and coaches in the school setting are highly influential in impacting PA outcomes for their pupils. In their review, it was clearly noted that the methods and behavior demonstrated by the instructor/coach had strong impact on outcome variable desired (e.g., PA participation). And although empirical studies examining students’ PA in PE is extant (McKenzie & Smith, 2017), others have stated that potential studies furthering research in contexts outside of the traditional PE setting is valuable (Schwamberger & Wahl-Alexander,

2016).

Furthermore, coaches who place primary emphasis on external motivations such as winning, social comparisons, and public recognition, can create a performance-focused environment (Esteban-Cornejo et al., 2015). On the opposite end of the spectrum, coaches who focus on effort, self-improvements, and intrinsic motivation create a mastery environment

(Keegan et al., 2009). There is considerable support for the notion that youth sport participants in mastery climates experience better health outcomes than those in performance laden climates

(Fenton et al., 2017; Mageau & Vallerand, 2003; Rocchi et al., 2013; Teixeira et al., 2012). It has been recommended that practitioners should attempt to create administrative and training

10 systems that promote the autonomy and competence of their frontline staff (Ryan & Deci,

2000). Sport managers must better identify the influencers on youth sport participation to understand the barriers for athletes generating more PA to inform the planning, implementation, and marketing of sport programs.

In a sports environment, the coach serves as the formal leader of the group. This is why most sports research on leadership concentrates on the interactions between the coach and athlete (Beauchamp & Eus, 2008; Carron et al., 2005; Weinberg & Gould, 2011; Williams, 2010).

For example, studies indicate that coaches can influence athletes’ performance by adopting positive interpersonal styles, such as transformational leadership behaviors or autonomy- supportive behaviors (Charbonneau et al., 2001; Gillet et al., 2010). A number of studies also suggest that coaches can influence outcomes related to participation, including self-determined motivation, sport enjoyment, and PA (Alvarez et al., 2009; Pelletier et al., 2001). There is also an emerging body of research exploring coaches’ contributions to youths’ personal development such as the acquisition of transferrable physical literacy skills (MacDonald, 2015).

A good sports leader is a coach who provides his or her athletes with adequate technical preparation, shows them support and influences their actions in order to accomplish desired outcomes. The coach is an expert whose task is to lead the athlete to reach the full extent of his or her capabilities and achieve the best results possible. Many proponents in youth sports stress the importance of the coach to teach the physical, technical, and development of skills necessary for a healthy sport experience and to fight the temptation to coach strictly to win games or tournaments. Directly related to this is the fact that the coach spends enormous amounts of time with the athlete and often develops emotional ties with athletes. Needless to say, conducting research in the area of coaching influence when sport is being delivered to athletes

11 would have significant impact on the way administrators and coaches can be trained to maximize their impact in terms of health-enhancing PA in their athletes.

In the sports realm, coaches serve as the front-line workers for carrying out the mission of the institution they serve. However, it has been estimated that 90 percent of youth sport coaches in the U.S. do not have formal training in coach education or youth development

(Petitpas et al., 2005). In the School Health Policies and Practices Survey (SHPPS, 2017), 74 percent of school districts reported that they require high school head coaches to complete a training course but a vast minority (25 percent) actually require coaches to have previous coaching experience in any sport. Therefore, a study that understands the role coaches play with regards to instructional methods used in sport practice settings will have practical significance for furthering the investigation of PA and coaching practices in a variety of youth sports at the high school level.

Statement of the Problem

Over the past decade, the Healthy People 2020 initiative (USDHHS, 2008) has fueled research addressing grand public health problems using interdisciplinary methodology, particularly those involving youths. In terms of PA, half of the report’s objectives comprise of interventions directly targeting children and adolescents in schools and other child care settings.

When attempting to promote PA in communities, it is clear that constituents are prioritizing (a) young people and (b) school settings. Since youths are deemed vulnerable populations and behavior change is best accomplished when interventions begin in early years of maturation, focusing on adolescents is understandably appropriate. And with a majority of youths aged six to eighteen spending a large proportion of their week days at school, interventions involving public schools (e.g., physical education (PE), recess, interscholastic sports) is a natural fit,

12 especially considering that schools already have the resources, supervision, and infrastructure needed to facilitate PA opportunities.

Addressing the PA problem in the U.S. has been a focal point for sport and recreation researchers over the past ten years. But there is widespread acknowledgement that a large gap exists between the science of athlete and coach development and the day-to-day reality of athletes and coaches (Gilbert et al., 2009; Martel, 2015). Numerous studies have found that sport practices are plagued with high frequencies of inactivity time (Bassett et al., 2013; Kanters et al., 2015; Ridley et al., 2018; Sacheck et al., 2011; Schlechter et al., 2018; Wickel & Eisenmann,

2007). For example, high school athletes have spent upwards of 60 percent of sports practice time being sedentary (Ridley et al., 2018). Sport practices tend to overemphasize skill and competition strategies during instructional time, thereby limiting efficient use of practice facilities and the amount of time athletes engage in MVPA.

While sport has been promoted as a PA strategy, the research does not consistently support that sport participation contributes to higher rates of PA (Schlechter et al., 2017;

Timperio & Ball, 2004). The handful of studies measuring PA levels during regularly scheduled sports practices suggest that activity levels can be relatively low and insufficient for meeting guidelines for daily MVPA (Bocarro et al., 2014; Leek et al., 2011). Additionally, most of our knowledge regarding the topic of PA and sport revolves around sport in the general sense (i.e., sport participation versus non-sport participation) (Allender et al., 2006; Eime et al., 2013;

Findlay & Coplan, 2008; Holt, 2011a; Marsh, 1993; Taliaferro et al., 2011). One group was successful in evaluating PA behavior in more than a dozen middle school intramural and interscholastic sports activities (Bocarro et al., 2014). However, there is inadequate literature that provides insight into the differences between particular sports (e.g., playing basketball versus playing lacrosse) in high school sports.

Therefore, although organized sport may create opportunities for children to be active, its full potential as a strategy to increase PA may be limited by the various contexts surrounding sport. Policies that govern the structure and delivery of sport practices, characteristics of the sporting environment, coach-athlete interactions, and the leadership (or lack thereof) of head coaches can all have major influence on the amount of PA experienced by athletes, yet empirical research is certainly deficient in these areas. The next step in understanding PA production in interscholastic sport is to perform a cross-sectional examination of athletes across a multitude of different sports and coaching methods, adding contextual knowledge to how PA occurs in this setting. This would move beyond the sport versus non-sport analysis and give greater insight into the influences that modifiable variables have on affecting change in athlete behavior.

Research Purpose

The aim of this study is to add to the literature pertaining to youth sport and PA, gain additional knowledge about the management and design of sports practices, and to use the findings from systematic observations to support previous scholarship or provide new perspectives related to theory used in understanding PA production among high school athletes.

The SEM is most useful for identifying a range of factors that influence human behavior and the environment that surrounds them. In this case, the framework is useful in identifying a multifaceted approach to affecting sport participation and it’s PA outcomes to inform the design of sports programs as to best maximize the health benefits to its participants.

As past reviews have shown, there is a growing body of literature examining the impacts of sport on youth development and PA and the influence of adult leaders in this context (Bailey et al., 2009; Fraser-Thomas et al., 2005; Gould & Carson, 2008). Generally, these reviews indicate that the positive outcomes of youth participating in sport is not automatic but is

14 dependent on a complex combination of factors. One of the essential factors to consider is the particular environment in which sport is played, because sport programs vary greatly according to their structure and the people involved in them (Gould & Carson, 2008). Through the utilization of systematic observation, the current research will attempt to unravel the complex interplay between multiple variables in an effort to further the conceptual and applied understanding of high school sport’s context in delivering healthy sport to adolescents. This study helps in the call for future studies implementing observational methods to address if current interscholastic sport programs are providing students with high rates of MVPA

(Schwamberger & Wahl-Alexander, 2016).

Specific to this study, I attempt to understand adolescent behavior in the high school sport environment at the individual (i.e., gender), interpersonal (i.e., coaching impact) and organizational (program/practice design, type of sport) levels of the SEM model. Hence, this will be one of the first studies to objectively examine a multitude of different sports at the high school level and the influence of key factors like practice design and coaching behavior on the

PA production of athletes.

Therefore, the purpose of this study is to examine sport context factors on high school athletes’ PA production during organized practice sessions in North Carolina public high schools during the 2017-2018 school year.

Significance of the Study

Sport is a setting with the potential to reach a large number of youths and have a significant public health impact (Camiré, 2014). While numerous factors impact youths’ benefits garnered from participation, the behaviors exhibited by adults, namely coaches, are particularly influential in the context of sport (Brustad et al., 2001; Garcia Bengoechea & Strean, 2007;

Keegan et al., 2010; Mageau & Vallerand, 2003; Weiss et al., 2002). In fact, the orientations and actions of coaches and program administrators have been identified as factors in youth developmental outcomes in sports (Fredricks & Eccles, 2004; Horn, 2004; Kay & Spaaij, 2011;

Smoll & Smith, 2003).

Although social influencers, organizational capacities, and program features are known to influence the efficacy of youth sport programs to facilitate healthy outcomes (Doherty &

Cousens, 2013), the empirical research to back up these claims is limited. As a result, there is a large body of literature linking sport with a variety of physical, social, and psychological outcomes but little insight on the practical implications for understanding this process. In the

PE setting, providing training to teachers on management strategies has shown to increase in- class MVPA (Lonsdale et al., 2013). But as the private youth sport sector expands along with its growth in participant numbers, it is now becoming imperative for delivery systems to be better informed as to best maximize the benefits that high school sport opportunities affords.

Sport participation has been understandably touted as an accessible vehicle for youth to experience many health benefits like PA, socialization and the like but there exists a tendency towards treating sport as a relatively homogenous activity. Combining all sports into a single homogenous pot leads to claims that are simply not true, especially concerning PA. Every sport has unique characteristics that appeal to different interests, abilities, and expectations and the impact on athletes’ health can certainly vary by sport type. In fact, the type of sport played has been cited as a determinant of what health outcomes are realized (Adler & Adler, 1998; Coakley,

2011; Côté & Fraser-Thomas, 2007; Crissey & Honea, 2006). It is also important to highlight that research on factors influencing participation has primarily focused on PA in general, rather than sport specifically.

In addition to providing unique conceptual, methodological, and managerial insights into the field of sport-based PA, this research will offer important information to practitioners and policymakers surrounding youth sport organizations. For example, the coaching behavior piece has the potential to shift the attention from competitive outcomes commonly found in youth sports today (e.g., winning, individual statistics) to the aspects that ultimately translate into positive outcomes and lifelong healthy behavior (e.g., skill development, PA) as previously described in the tracking and LTAD philosophy. This will help sport administrators, coaches, and other personnel identify key strategies in the development process of adolescents, and direct activities to influencing the structure of sport practices that contribute to desirable youth outcomes. For example, it is hypothesized that football practices are rampant with sedentary activity as players are often found to be standing around, absorbing strategy given by coaches and taking turns running routes or performing drills. If the current study supports the previous hypothesis to be true, this can inform national sporting governing bodies surrounding football as well as coaches in a myriad of other sports to reevaluate their practice curriculum.

From a managerial perspective, this research may help inform national and state governing bodies of school-sponsored sport programs on (1) PA outcomes of various high school sports, (2) the training of coaches on effective practices for the daily managing of a sport program, and (3) the education of youth sport governing bodies on how to best enforce the structure of program design through added formalization and standardization of curricula. As youth sports programs primarily utilize volunteer coaches or school teachers on a part-time basis, providing coach training may afford the skills necessary to conduct an active practice.

Currently, sport delivery is extremely fragmented as coaches are given huge amounts of leeway for how they manage a sports practice as the U.S. is one of the few countries that lacks a centralized governing agency that controls all sport delivery methods (Coakley, 2011).

It has shown historically that the more levels of the SEM that are addressed, the greater impact the strategy has on PA behavior. These factors can contribute to health outcomes in both a positive and negative way. More research is needed to understand the health impacts based on certain sport environments. Therefore, this study will help fill theoretical gaps along with enhancing practitioner knowledge for the administration and organization of interscholastic sports. It is critical to move beyond the assumption that sport participation is healthy without legitimate data to back up these claims. This study attempts to understand how sport programs can be best delivered to young athletes where it not only produces higher rates of PA, but also translates into skill development which can theoretically have lasting impacts on the health of young athletes through their lifetime as supported by the LTAD model. This, in turn, has a chance to inform sport delivery personnel (e.g., athletic directors, coaches) and interpersonal networks (e.g., parents, siblings) on what factors have the greatest impact on PA.

Ideally, this research would aid in the advisement of a more standardized delivery system similar to that of PE teachers. In a recent systematic review, Dudley and colleagues (2011) reported that a key component to the most effective interventions in schools targeting PA was the provision of professional development programs for teachers using well-designed prescribed curriculum. The curriculum was often supported with additional resources such as lesson plans and in-school consultants. I do not presume that this research will inspire massive overhaul of interscholastic sport policies across the nation, as these systems have historically been known to be multi-layered and highly complex. However, I do believe that more cross-sectional studies will aid in our theoretical understanding (and importance) of both the coach-athlete relationship and the efficient management of athletes regarding sport activities during practice time. The results of this study could provide a small and logical step in the right direction for reforming how sport and recreation managers think about the many facets surrounding interscholastic

18 sport offerings. An ancillary benefit would be to provide a foundation of information (in lieu of speculative reporting) to help policymakers form evidence-based recommendations on the future governing of the American youth sport delivery systems.

Summary

In summation, high school sport facilitates millions of active participants on an annual basis (NFHS, 2017; SHPPS, 2017). Regular PA associated with youth sport has the potential to favorably influence positive health outcomes in adolescents (Schlechter et al., 2017). For this reason, governing bodies have identified organized sports as a vehicle for youth to obtain adequate bouts of PA among many other benefits (USDHHS, 2019). Still, sport is a social institution and our understanding of the complexities and relationships between the many moving parts surrounding the context in which sport is delivered to adolescents is marginal at best (Logan et al., 2019). In particular, the role that a coach plays in the skill development and

PA behavior of high school athletes is theoretically sound (Kokko et al., 2015) but remains under researched (Schwamberger et al., 2016). Recently, there has been a call to evaluate strategies for increasing MVPA in youth sport practices (Leek et al., 2011). This study uses the

SEM as a guiding theoretical framework in conjunction with the LTAD to explore PA and its contexts in high school sport settings. The study holds important theoretical implications as it will aid in our understanding of the multi-layered sport environment and the multi- dimensionality of coaching behavior as it relates to sports and practice design in addition to clarifying the importance of sporting contexts when delivering physical activities to high school aged athletes. In practice, the aggressive expansion of youth sport organizations has resulted in major fluctuations and heterogeneity in coaching strategies and accepted standards of practice.

If PA in adolescence is to track into adulthood, a more nuanced understanding of PA within high school sport can help coaches and administrators identify the key contributors to PA

19 during practice time. There is potential, then, for coaches to be able to influence their players’

PA levels; particularly during training where coaches are better able to dictate the intensity of

PA, as compared to during a game. This research has a chance to inform sport delivery personnel and interpersonal networks on what factors have the greatest impact for creating a physically active sporting experience.

Research Questions

The focus of this dissertation is to (1) explore the high school sport environment and determine what factors contribute to PA outcomes experienced by athletes through participation, (2) examine the behavior exhibited by coaches to build our understanding of appropriate coaching methods to maximize PA outcomes and skill development in sports.

Therefore, this study will address the following research questions:

Primary research question

RQ. What factors are most strongly associated with PA behavior during high school sports practices and to what degree?

Secondary research questions

SRQ1. How physically active in terms of MVPA are athletes during high school sports practices?

SRQ2. What are the variations in MVPA across different sports for boys and girls?

SRQ3. What is the association between practice context and PA during high school sports practices and how does it vary based on gender and sport type?

SRQ4. What is the association between coaching behavior and PA during high school sports practices and how does it vary based on gender and sport type?

SRQ5. What are the interrelationships between sport type, coaching behavior, practice context, and proportion of time spent in MVPA during high school sports practices?

SRQ6. How do sport context factors combine to predict PA behavior during high school sports practices?

The knowledge gained from addressing these research questions will provide insight into future sport managerial and coaching research related to the respective concepts. Sport represents one of the main developmental contexts for children and adolescents, next to family and school (Rutten et al., 2007), and millions of children participate in sport every day

(USDHHS, 2019). In the U.S., high school sport represents an important and relevant domain for interventions seeking to promote PA during adolescence. There is an urgent need to discover methods by which, and contexts in which, the PA levels of interscholastic athletes can be increased (Curtner-Smith et al., 2007) and it remains unclear with respect to opportunity for

PA how time is spent and how coaches conduct themselves during practices and games in organized sports (Guagliano et al., 2013). This study aims to address these research gaps.

Delimitations

This study is delimited by the following conditions:

1. Only public high schools in North Carolina were included in the study.

2. Only data on the ten most popular boy sports and ten most popular girl sports in North

Carolina were collected and reported.

3. Observations were completed primarily based on the availability of trained data

collectors and were not done based on an established schedule.

4. Data collection included regular season practices only (i.e., no pre-season or playoff

practices). In other words, observers did not start until the first regular season game and

ended before the beginning of the postseason state tournaments.

5. Data collectors observed varsity team practices (i.e., no JV practices). Practices were

selected in lieu of games/competitions as athletes are anticipated to spend more time

participating in team practices than in actual games. Practices are also more inclusive

than games, involving every athlete regardless of talent level.

6. No pre- and post- test intervention and evaluation of varying practice models and

coaching strategies was conducted.

Definition of Terms

1. Physical Activity (PA): Any bodily movement produced by skeletal muscles that requires

energy expenditure – including activities undertaken while working, playing, carrying out

household chores, travelling, and engaging in recreational pursuits (World Health

Organization, 2010).

2. Moderate-to-Vigorous Physical Activity (MVPA): According to the CDC (2012),

moderate physical activity is defined as the expenditure of between 3.0 and 6.0 METs (or

3.5-7.0 kcal/min), while vigorous physical activity is defined as an expenditure of over

6.0 METs (or >7.0 kcal/min).

3. Health-enhancing Physical Activity (HEPA): any form of PA that benefits health and

functional capacity without undue harm or risk (Foster, 2000).

4. Metabolic Equivalent Task (MET): The ratio of exercise metabolic rate. One MET is

defined as the energy expenditure for sitting quietly, which, for the average adult,

approximates 3.5 ml of oxygen uptake per kilogram of body weight per minute (1.2

kcal/min for a 70-kg individual). For example, a 2-MET activity requires two times the

metabolic energy expenditure of sitting quietly (CDC, 2012).

5. Sport: all forms of activity which, through casual or organized participation, aim at

expressing or improving physical fitness and mental well-being, forming social

relationships or obtaining results in competition at all levels (Council of Europe, 2001).

6. Interscholastic Athletics: all school-sponsored sports offered at the high school level (i.e.,

ages 14 to 18). The term is sometimes used interchangeably with ‘high school athletics’.

7. Tracking: the relative stability of a characteristic, or the maintenance of relative position

or rank in a group, over time (Malina, 1996).

8. Compendium: a collection of concise but detailed information about a particular subject

usually in a sequential arrangement or numerical order.

REVIEW OF LITERATURE

The primary purpose of this study is twofold: (1) to explore the associations of high school sports practice design and PA behavior and (2) investigate the amount of influence that leadership styles have on coaching behavior and PA outcomes of athletes. Literature specific to previous academic research examined several relevant topics. The review begins with the theoretical frameworks that have aided in my understanding of managing youth sport for PA outcomes along with the specific theoretical models guiding this study. This section will summarize what I have learned from these studies and how they serve as the foundation for my current research. I will subsequently present an overview of programmatic approaches in school based PA, namely physical education and interscholastic athletics, benefits and measurements of

PA, role of public schools, previous studies on the utility of high school athletics, research on leadership and coaching models, and the applications of observational methodologies.

Relevant Theories and Concepts

While a large number of researchers and policymakers have designed interventions to increase healthy sport outcomes (Casey & Eime, 2015), the evidence on the success of interventions conducted in the sports setting are very limited, with many focused on PA in after- school programs or school-based programs (e.g., physical education) rather than organized sport organizations (Metcalf et al., 2015). Most of the empirical studies done in building theory towards a better understanding of healthy sport for youth have been outcomes-based and focused almost entirely on the individual (Holt, 2011b). For instance, there is a litany of literature that helps inform our grasp of sport participation and its influence on youth development outcomes like motivation, competence, confidence, enjoyment, PA, and others as reported in some systematic reviews (Demetriou et al., 2017; Jones et al., 2016; Whitley et al.,

2019). However, there has been less attention to programmatic features and managerial aspects of youth sport delivery, leading to a lack of clarity regarding how health outcomes are produced in the sport context.

In the following pages, I present an overview of some of the more prominent theoretical frameworks that have informed our ever-growing investigation into managing sport for health.

Specifically, I concentrate on some of the more meso-level paradigms that have influenced the management of youth sport programs for achieving goals of a “healthy sport” environment rather than theories focused on individual-level components.

Application of the Social Ecological Model

Social ecology defined is an ecological perspective that suggests that environments enhance a range of behaviors by promoting and sometimes requiring certain actions while discouraging or prohibiting other behaviors; multiple facets of the physical and social environments influence the well-being of people (Stokols, 1992). Furthermore, an ecological approach to explaining human development concerns itself with understanding the contexts and interactions between the individual and his or her surroundings. Stokols (1992) used the framework of social ecology to develop his own social ecological model under four assumptions.

First, health behavior is influenced by physical environments, social environments and personal attributes. Second, human-environment interactions occur at varying levels of combination

(individuals, families, groups, whole populations). Third, people influence their settings and the changed settings then promote either positive or negative health behaviors. The final assumption is that environments are multidimensional.

There has been a significant increase in the application of ecological models in guiding health promotion research (Cochrane & Davey, 2008; Fleury & Lee, 2006; Richard, Gauvin, &

Raine, 2011; Sallis, Owen, & Fisher, 2008). The SEM has adapted the general ecological framework to explain PA behavior and its determinants to inform multi-level interventions. In

Moos’ (1980) first SEM model, four categories of environmental factors (physical settings, organizational settings, sociocultural characteristics, and the social climate) were all contributors to people’s behavior. The interrelationships between environmental conditions and human behavior and well-being are at the core of this framework. Hence, the social ecological framework is interdisciplinary by nature and is used in fields of epidemiology, medical sociology, public health, health psychology and others (McLeroy et al., 1988). A social ecological perspective is based on a broad, overarching paradigm that bridges several different fields of research, rather than from one theory (Stokols, 1996).

Figure 2.11 Model showing strength of reach and permanence across ecological influences (Sallis, 2014)

Various conceptualizations of the levels of influences on behavior have emerged, but most factors are generally categorized under one of five domains: intrapersonal, interpersonal, organizational, physical, and policy. The basic premise being that environments and particular social factors either enable or constrain behavior by promoting or discouraging particular

26 actions. A health through sport conceptual model, proposing three primary categories of health outcomes as a result of sport participation has been developed (Eime et al., 2013). In the model, the determinants of sport participation are represented within the multiple levels of the SEM to help identify the several levels of influence: intrapersonal (e.g., gender, age); interpersonal (e.g., social connections); organizational; environmental; and policy factors. This model is useful for linking sport participation behavior and its utility in informing the planning and implementation of sports policies and programs (Rowe et al., 2013). The model suggests that different factors will influence the context of sport participation and subsequent physical, social and psychological outcomes.

The NYSS operationalizes a framework based on the SEM, which summarizes key factors that influence youth sports participation at multiple levels (Figure 2.12).

Figure 2.12 Social-ecological model for enhancing health outcomes of youth sport participation (NYSS, 2019).

HHS compiled a comprehensive list of factors related to youth sports participation and organized the factors into the levels of influence from the SEM, a widely used framework in public health. Some of these factors are intrinsic characteristics, some are processes that influence youth sports participation, and others are outcomes or behaviors that affect youth sports participation. As such, some of the factors are included at multiple levels of the framework. The individual level of the framework addresses youth directly, and the remaining levels address youth indirectly.

The interpersonal level of the framework examines factors related to adults who directly interact with and influence youth participating in sports. Adult influencers include parents and coaches. The focus is on adults because they have the ability to affect these factors through quality coaching skills and unstructured play within sport sessions. It has been recommended that adults model and encourage an active lifestyle as a way to help youth get enough daily PA

(USDHHS, 2018).

Opportunities and action items for the interpersonal level include (a) participate in training or certification programs to acquire, develop, and maintain skills for engaging youth sports participants, (b) emphasize skill development over competition and performance outcomes, (c) structure practices and games to provide all participants with more time engaged in moderate-to-vigorous PA to meet the PA guidelines, and (d) create environments that support unstructured sports play. Opportunities and action items for the organizational level include (a) providing adequate training and other necessary resources to adapt or modify sports activities to meet the needs of youth of all abilities, (b) implement the principles of the American

Development Model within youth sports programs, and (c) structure practices and games to provide all participants with more time engaged in MVPA to meet the PA guidelines.

Further testing of ecological models of PA has been recommended by many researchers as behavior change is expected to be maximized when environments and policies support healthful choices, motivate and educate people to make those choices. Lastly, ecological models are most powerful when they are behavior-specific. The strength of using a SEM framework is that more models can be developed for specific PA behaviors and population subgroups. For example, the use of instruction and motivation techniques by coaches in sport settings.

Although the SEM is used to guide this study, it is important to acknowledge that several other theories converge in identifying the important role coaches play in achievement settings such as youth sport.

Tracking Sport Participation

Organized sport provides opportunity for PA on a regular basis and in a safe environment. Children and adolescents involved in sports are more physically active than those not involved, and spend more time in MVPA (Katzmarzyk & Malina, 1998; Machado Rodrigues et al., 2012; Wickel & Eisenmann, 2007). Adolescent athletes 16-19 years for both genders had greater energy expenditure and energy expenditure in PA than non-athletes (Ribeyre et al., 2000).

Although limited to survey-based information, sport participants were also more physically active than non-participants among U.S. youths (Pfeiffer et al., 2006; Trost et al., 1997).

Participation in sport during adolescence also tends to track at higher levels than other indicators of PA (Malina, 2001a, 2001b). Can adult PA be predicted from activity or sport participation during childhood and youth? Sport club membership (and participation) tracks at a higher level than other indices of PA among adolescents and young adults (Telama et al., 1997).

Moreover, frequency of participation in sports at 14 years old (Tammelin et al., 2003), membership in sport clubs at 16 years old (Barnekow-Bergkvist et al., 2001) and sport club training and competition during adolescence (Telama et al., 2006) significantly predict PA in

29 young adults (late 20s to early 30s). The process of how participation in sport during adolescence translates into an active lifestyle in young adults needs study. An association between sport participation during adolescence and psychological readiness for PA in adulthood has been proposed (Engstrom, 1991).

Dennison, Straus, Mellits, et al. (1988) carried out a precursor of tracking studies, although their focus was on childhood determinants of adult PA patterns. In a non-concurrent prospective study, the PA levels of 453 males between 23 and 25 years old were compared with their physical fitness test scores as children and adolescents (10 to 11 year olds and 15 to 18 year olds). The researchers found that physically active adults had significantly better childhood physical fitness test scores than did inactive adults, and that physical inactivity in young adulthood was linearly related to the number of low scores on run and sit-ups tests as children.

Another early investigation, by Kuh and Cooper (1992), was more explicitly a tracking study, examining the influence of childhood activity on adult activity patterns in a national sample of more than 3,300 36 year-old adults in England. The authors reported that adults who were most active in sport had been of above average ability at sports in school than those who were less active. Studies examining the relationship between sports participation and later PA appeared a little later. For example, Finnish researchers reported that participation in sports competitions at sixteen years old predicted PA two years later (Aarnio, Winter, Peitonen, et al.,

2002), and a follow-up study with a large sample of 14 to 31 year olds found that participation in sports twice a week or more after school hours, being a member of a sports club, and a high grade in school sports at age 14 years were associated with a high level of PA at age 31

(Tammelin, Nayha, Hills, et al., 2003). A more recent study found that participation in high school sports was the best predictor of PA after age 70 (Dohle & Wansink, 2013). In addition,

Smith, Gardner, Aggio, and Hamer (2015) found that participation in sport at age 10 was highly associated with participation in PA at age forty-two.

Role of Interscholastic Sports

The settings approach to health promotion has become one of the dominant paradigms in the field, by shifting the focus from individual-based risk orientation and disease to social factors that support human health and wellbeing (Eriksson & Lindström, 2014). The application of the settings approach in schools has proved to be useful in terms of both impact and effectiveness with all age groups (Langford et al., 2015). Advocates claim that sports participation can substantially increase energy expenditure and physical fitness for those who participate, as well as contributing to wide health outcomes (Samitz et al., 2011). But as previously presented, critics have questioned these claims (Nielsen et al., 2016).

The case for the settings approach for sport as a feature of HEPA rests on a series of assumptions (Davies et al., 2016). First, large numbers of children play sport regularly. Second, the PA dose from sport is sufficient to benefit the health of the child. Third, participation in sport during childhood and youth tracks into PA later in life.

Organized youth sport is important for healthy development, growth and wellbeing

(Vella et al., 2014) and can be a useful tool for mobilizing health-promoting resources, particularly in schools (Ewing et al., 2002). Research surrounding policies affecting school-based

PA has primarily emphasized physical education (P.E.) and sport offerings (Institute of

Medicine, 2013). Existing state and national organization policies for PA in after school programming have set benchmarks for students to achieve appropriate amounts of PA, which can most readily be delivered through P.E. sessions, intramural, and varsity sports (Beets et al.,

2013). Much of the work done to manage and govern healthy sport delivery has been devoted

31 to children and adolescents; rooted in the mission of P.E. in the primary school system (Casey &

Eime, 2015), with the purpose being to provide exercise opportunities for children as well as introducing them to the many facets of health, most notably through various forms of physical activities. Many of the administrative protocols for organized sport offerings are tied directly to the original academic mission of schools. The delivery of sport to young populations has understandably been informed by developmental psychologists to align as close as possible to the educational development of students. In theory, the intellectual and physical education of youth should follow the same trajectory to result in a well-balanced and healthy young adult.

Public schools can play an essential role in providing additional opportunities for PA across the school day. Nearly 90 percent of all school-age children are enrolled in public schools

(Filardo et al., 2010). Furthermore, the growth of the U.S. population has contributed to the increase in public school enrollment and is projected to increase by at least 3 percent over the next decade (National Center for Education Statistics, 2017). As more students fill up schools, interscholastic sports used by the school and students could be the low-hanging fruit for more

MVPA to occur. This would depend on how sport is delivered to limit sedentary activity and facilitate more PA production from athletes enrolled in the sport program.

Sport and Physical Activity

A large number of studies have examined the relationship between sports participation and PA. These studies vary widely in research design, outcomes measured, and methodology used. For example, several studies have reported participation in organized sport to be associated with increased levels of PA (Mota & Esculcas, 2002; Sacheck et al., 2011), but the confidence in their findings has been undermined by the use of research methods that might be

32 inappropriate with children (e.g., self-report surveys). Consequently, there has been a more recent surge in utilizing more objective measures of activity.

Wickel and Eisenmann (2007) investigated weekday levels of MVPA in elementary and middle school boys. They found weekday MVPA to be 110 and 40 minutes per day, respectively among flag football and soccer players. They also reported levels of MVPA to be 30 minutes lower on a non-sports day compared to a sport day. Similar findings were reported by

Machado-Rodrigues et al. (2012) in their study of MVPA of 13 to 16 year old boys. Sports players accrued 114 minutes per weekday of MVPA, and an average of 97 minutes per day across the week. The researchers calculated that this equated to between 11 and 13 percent of total daily energy expenditure in organized sports which corresponds to 35 to 42 percent of the

MVPA daily energy expenditure. Data from 9 to 15 year old soccer players showed average daily

MVPA to be 122 minutes per day (Van Hoye et al., 2013). These studies have also found that the amounts of PA accrued during sport participation did not occur during non-playing days and was generally replaced with low-intensity and sedentary activities. This is consistent with other research reporting that children are unlikely to compensate for missed PA resulting from sports

(Dale et al., 2000), indicating that sport might have the potential to increase levels of PA, and also be effective in reducing bouts of sedentary behavior (Kanters et al., 2015).

Recently published studies paint a generally positive picture about the potential role of sports participation as a contribution to youth PA (Kanters et al., 2015; Guagliano et al., 2015;

Ridley et al., 2018; Schlechter et al., 2017). However, caution should be taken when interpreting findings. While sports-playing children are generally more active, have greater energy expenditure and spend more time in MVPA than non-participants, data are still limited about how their PA varies with types and levels of sports participation. Furthermore, it has been reported that some sedentary activity can be mistakenly categorized as MVPA (Kim et al., 2012).

As a consequence, levels of PA during youth sport may have been overestimated in studies involving younger participants (Leek et al., 2011). But regardless of the study’s findings, much of the research done within sport and PA have concluded that sport participation alone does not guarantee children will reach recommended daily PA levels.

In reviewing the sport and PA literature, it seems clear that different sports offer different levels of PA for different players, further emphasizing that context matters. For example, Leek et al.’s (2011) study reported levels of MVPA and the proportion of session time engaged in MVPA to be higher in soccer participants than in softball and baseball. Likewise,

Katzmarzyk and others (2001), who studied the amount and quality of PA through direct observation of 11 to 14 year old boys’ and girls’ recreational sports, found outdoor soccer to be the most robust activity, providing less sitting time than hockey. Outdoor soccer also had the greatest time spent jogging and sprinting. Additionally, indoor soccer provided less sitting time than basketball. Leek et al. (2011) also found that 7 to 10 year olds were more active than 11 to

14 year olds, and boys were more active than girls. Other studies have found that younger participants engage in higher levels of MVPA than older participants (Wickel & Eisenmann,

2007); overweight and obese participants engage in lower levels of MVPA and are inactive for more time than normal weight peers (Sacheck et al., 2011); those with access to sports facilities

(Mandic et al., 2012), and that MVPA was higher during practice sessions than games/competitions (Guagliano et al., 2013).

Other studies have investigated levels of PA during sport time specifically to determine the extent to which participation contributes to meeting MVPA targets. Guagliano et al. (2013) studied 11 to 17 year old soccer and basketball players and found that for every hour of game play or practice time, participants accumulated approximately 33 percent of the recommended

60 minutes of MVPA per day. Another study found that both boys and girls (aged 7 to 14 years)

34 engaged in approximately 45 minutes of MVPA during youth sport practice (Leek et al., 2011).

The findings also showed that 23 percent of players met the recommended levels of MVPA per day during youth sport time. Comparatively, lower levels of MVPA were reported by Wickel and Eisenmann (2007) in their study of 6 to 12 year old boys, although their findings are similar to the previously mentioned studies (49 percent).

More negatively, many of these studies also recorded large amounts of sedentary or idol activity. Some show that children spend up to 70 percent of their time playing sport engaged in either complete inactivity or light intensity activity (Leek et al., 2011) but others discussed about

50 percent of all practice time was sedentary in their particular projects (Schlechter et al., 2017;

Wickel & Eisenmann, 2007).

Observing physical activity in sport settings

SOFIT is a widely used direct observation system that uses momentary time sampling to generate data on players’ PA, lesson context, and coach behavior (McKenzie et al., 1991). To date, it is by far the most applied method for observing PA in recreation and sporting evironments. With that said, a full review of all studies using SOFIT in sport settings was conducting. The following nine peer-reviewed articles were found and annotated (Bass &

Martin, 2013; Cotton & O’Connor, 2013; Cutner-Smith et al., 2007; Cardon et al., 2004;

Guagliano et al., 2013; 2014; 2015; Peralta et al., 2014; Schwamberger & Wahl-Alexander, 2014).

Five were international studies leaving only four published articles in the U. S. citing SOFIT as the tool used to understand PA behavior in sport participants.

The purpose of the Cardon et al. (2004) study was to compare eight to twelve year old elementary school students’ PA engagement levels in swimming and nonswimming PE classes.

Swimming, both girls and boys., 8 to 12 year olds, Belgian, PE classes. A total of 234 students

35 were observed during one swimming class and one nonswimming PE class to determine PA levels. The average percentage of MVPA during lesson time was significantly higher during swimming class (52 percent) than during nonswimming PE (40 percent). Forty-one percent of the swimming lessons and 77 percent of the nonswimming PE lessons did not reach 50 percent

MVPA engagement. An important finding of the present study was that PA is higher during swimming classes than during nonswimming PE classes in elementary schools. It was still concluded that swimming and nonswimming PE classes contribute too little to PA levels in elementary school children, and that the need to increase PA in PE is high. A limited number of lessons were observed and the variety of lesson contents of the nonswimming PE classes was not evaluated.

The primary purpose of an earlier exploratory study was to determine the percentage of time in which the athletes coached by high school teachers were engaged in MVPA during sport practices (Curtner-Smith et al., 2007). There were also some ancillary purposes: determine the percentage of time allocated to fitness and knowledge content and percentage of time teachers were encouraged to be active. Participants were twenty teachers employed in twenty high schools in the U.S. 2007 (dated), girls only aged 17-18, basketball only, 60 practices. 51 percent of their time engaged in MVPA. Much of this time was spent in very active behavior (32 percent) while the remainder was spent walking (19 percent). Unfortunately, athletes also spent

42 percent of their time standing. Teachers focused on teaching the skills and strategies of basketball for 88 percent of the time (i.e., the sum of general knowledge, skill practice, and game play). Most encouraging was the fact that the teachers allocated very little time for management

(8 percent). They spent 75 percent of their time providing athletes with instruction on skills, strategies, and tactics. Conversely, they spent very little time promoting (0.52 percent) or

36 demonstrating (0.18 percent) fitness. The most important finding was that teachers engaged athletes in MVPA for 51 percent of their practice time over the course of the season.

Research out of Australia measured PA from accelerometers in only 20 sessions (10 practices and 10 games) (Guagliano et al., 2013). A total of 94 girls aged 11 to 17 participated from three different sports (i.e., basketball, netball, and soccer). Twenty sessions were observed and on average, sports contributed 18 minutes of MVPA during games and 20 minutes during practices. Large proportions of time, however, were spent in sedentary activity, on average 23 minutes during games and 18 minutes during practices. Participants accumulated 24 percent of recommended levels of PA during practices overall. Every hour of game play or practice time, girls accumulated approximately one third of the recommended 60 minutes of MVPA (Strong et al., 2005). But the study was not designed for comparison between different sports but rather to describe PA levels of three sports and to compare PA levels in games versus practices.

The systematic observation of 64 sessions included 480 rugby players under ten years old and 32 coaches (Cotton & O’Connor, 2013). The primary purpose of this Australian study was to test if a six-week intervention targeting coaching behavior could increase player activity levels and skill development opportunities using a modified SOFIT tool; PA, coach behavior, and what skills (e.g., catching, throwing). Their analytic procedure involved the data from the modified SOFIT instrument which indicated how physically active athletes were, how the coaches spent their time during training, and what skills were performed during the training session. Coaches who used the skill training tool were found to give significantly shorter pre- activity instructions and significantly shorter concurrent instructions. There were no differences between the control group of coaches and coaches using the skill training tool in the following areas: PA levels, amount and type of skills practice; and coach behavior. However, coaches who used the tool gave shorter instruction periods before and after practices.

The purpose of the (Bass & Martin, 2013) study was to examine if the number of teaching strategies used in a PE lesson increased the PA levels of students during a collegiate swim class. The participant was a lone swim teacher who agreed to be observed during a swim class that was scheduled to last 90 minutes four times a week. Five thirty minute segments of lessons over the course of two weeks were observed and recorded. There did not appear to be a strong connection between the number of teaching strategies used and the amount of time spent in MVPA. The majority (57 percent) of the observed class time was coded in the ‘standing’ category of PA. None of the observed activity was coded as vigorous. The limitations from this project were that observations occurred during the same segment of every lesson and only covered one-third of the entire instructional time, swimming was the only sport of interest, and there was only one teacher being evaluated.

The primary aim of Peralta and colleagues’ (2014) study was to investigate the effect of a community and school sport program (SCP) on urban Indigenous adolescents’ life skills and PA levels within program sessions. The study was a non-randomised pre-post test case study that involved seven to ten year old students across three secondary schools. SOFIT was used to measure MVPA and instructional time in each session over a ten week period. Observations were conducted by two trained SOFIT observers with an inter-rater agreement of more than

90% on all variables. On average, participants were engaged in MVPA for 58 percent of the school sport sessions. Skill development and implementation was a focus with half of practice lesson time spent on skills and drills and another quarter of the time on game play. The limitations gathered from this research was that it was Australian so not generalizable to the

U.S., no coaching behavior measure was taken, and focused on life skills and goal setting

(intrapersonal level) more so than the meso-level contextual criteria usually reported in SOFIT research.

Guagliano et al. (2014; 2015) was the first randomized controlled trial (RCT) conducted in an organized youth sport setting aimed at determining the short-term efficacy of coach education on player MVPA. Primarily, this two armed RCT aimed to assess whether coach education, relative to a standard-care control, can increase the proportion of time players spent in MVPA during practices over a five-day basketball program. Topics covered during the coach education sessions were: strategies to increase MVPA and decrease inactivity during practice, self-monitoring, goal-setting, and suggested target step counts/minute as done previously by other researchers (Scruggs, 2007a). The study’s findings provided support for the short-term efficacy of coach education on increasing time spent in MVPA and reducing time spent inactive without detrimental effects to players’ motivation. Players in the intervention group spent a significantly higher proportion of practice time in MVPA, and a significantly lower proportion of practice time inactive, compared to the control group. The main limitation of the study was that it only investigated the short-term efficacy of coach education on player PA in basketball (single sport), and over the duration of only five days of observing. PA was also not gathered from

SOFIT, rather, accelerometers were used (Guagliano et al., 2014). Additionally, researchers were located in Australia and only interested in female players aged 9-12 years old.

Finally, the purpose of Schwamberger & Wahl-Alexander (2016) was to determine the lesson context and children’s participation levels during a summer swimming program taught by two PE teachers. U.S., swimming only, twenty eight boys and seventeen girls (forty-five children total), P.E. teachers, 4th and 5th grade. The average percentage of MVPA engagement during lesson time was high (59 percent). Both instructors’ practices had athletes spend a majority of their time in vigorous activity (37 percent). The highest category in lesson context was skill practice with the least amount of time in knowledge content. An important finding of this study was the low percentage of time spent in management by both teachers (12 percent, 28 percent).

Measures

It is clearly established that PA is associated with health benefits and well-being (Janssen

& LeBlanc, 2010). Hence, PA is one of the most important lifestyle factors for improving public health and disease prevention in children and youth. Accurate assessment of PA and sedentary behavior is necessary to establish relationships between PA and various health outcomes as an inaccurate assessment of PA may weaken the true association between factors and outcomes.

Further, valid and reliable methods for assessing PA and sedentary behavior are essential when monitoring the effect of interventions to make cross-sectional comparisons.

Despite progress in the assessment of PA, limitations related to measurement accuracy still exist. These limitations are often amplified in youth and some are unique for this age group as youth PA is different from adults due to the nature of their PA pattern. Adolescent PA is often sporadic, including less organized activity and characterized by intermittent movements with variable intensities (Hoos et al., 2004). Additionally, the associations between subcomponents of PA and different health outcomes in youth may be weaker than in adults and therefore more difficult to establish (Janssen & LeBlanc, 2010). This has spawned consequences for the assessment of PA and for the interpretation of PA data in youth. Although many methodological considerations can be generalized between adults and youth the assessment of

PA in youth is not synonymous with that of adults and should be considered as separate populations.

Many youth sport programs do not provide the necessary amount of MVPA due to participants standing around waiting for their turn to play (Bergeron, 2007; Leek et al., 2011).

Some studies have used accelerometers to examine PA levels of engagement among sport participants (Wickel & Eisenmann, 2007), finding that average weekday MVPA to be 30 minutes

40 lower on a non-sports day compared to a sport day. Similar results have been reported internationally (Machado-Rodrigues et al., 2012; Van Hoye et al., 2013). These studies have also found the amounts of MVPA accrued during sport participation did not occur during non- playing days and was generally replaced with low-intensity and sedentary activities. This is consistent with other research indicating that children are unlikely to compensate for missed PA resulting from sports (Dale et al., 2000), suggesting that sport has the potential to increase levels of PA, and also be effective in reducing bouts of physical inactivity (Kanters et al., 2015) but only when designed and delivered correctly. But overall, the general body of evidence points to a positive effect of team sports on the physical health of youth.

There are numerous methods for assessing PA and generally these methods are divided into subjective and objective categories. Subjective methods obtain information from the individual by questionnaires, diaries, or logs and data is often collected retrospectively and depends on the individual’s cognitive function. Studies in PA have historically been based on questionnaires among adolescents and young adults (Malina et al., 2016). These methods can be cost effective and easy to obtain, but a major caveat is that self-report surveys may not be reliable or valid measures of PA as people tend to exaggerate how much PA actually occurs

(Scott et al., 2015). Objective measurement methods are not influenced by the individual’s self- assessment of PA and may be less prone to response bias. Accelerometers, inclinometers, heart rate monitoring and indirect calorimetry are all examples of objective measures for PA and highly reliable. However, the costs, both financial and time, can be extravagant in purchasing the technology and length of data collection timelines.

More recent surveys are based on accelerometry and often focus on epochs of MVPA, the intensity of activity that is most often associated with health benefits (Strong et al., 2005).

Observations derived from a week of accelerometry in a nationally representative sample of

41 youth indicated a decline in MVPA from 9 to 15 years old for boys and girls, higher levels of

MVPA in boys than in girls (Nader et al., 2008). Boys exceeded the recommended 60 minutes of MVPA at all ages except 15 years old, while girls exceeded the recommendation from 9 to 13 years when levels of MVPA fell below 60 minutes.

Measures in youth sport participation

Youth sports participation have been measured through Federal and non-Federal surveillance systems over the last few decades. Broadly, participation can be described by a variety of constructs: being on a sports team, the number of sports played, frequency of participation, and type of sport played. Another useful construct is how sports are organized

(e.g., interscholastic sports, intramural clubs, travel sports programs, and recreation leagues). A number of surveillance systems measure sports participation. These data can be used to design and implement effective programs, practices, and policies (NYSS, 2019). However, differences across systems and measures make it challenging to fully describe and understand youth sports participation and access in the U.S.

Five Federal surveillance systems collect individual-level youth sports participation data, and one (SHPPS) collects school-level data on access to sports programs. Some efforts collect only youth data, while others include all members of the household. Most individual-level systems collect information using a questionnaire, but one (ATUS) uses an interview with a one- day activity recall. Some individual-level systems use parent or caregiver reporting to collect information, while others ask youth directly, and some use a combination of the two methods.

These surveillance systems gather data on demographic variables, such as gender, race, ethnicity, age, and household income; the NSCH and the NHANES also include questions about disability. This allows for subgroup analyses and examination of differences across subgroups.

The level of detail and type of information obtained by the individual-level systems varies. The types of sports in which youth participate are captured in two systems (ATUS and

NHANES), while only one (ATUS) assess the type and time spent participating in sports. Both the NSCH and the SIPP include a yes/no question about sports participation, and the YRBS includes a question on the number of sports teams on which the individual participates. Three systems (NHANES, NSCH, and YRBS) have a question about overall PA that can include examples, such as exercise, recreation/leisure, play, and sports participation. For example, the

2016 and 2017 NSCH asked parents or caregivers to report the number of days their child exercised, played a sport, or participated in PA for at least 60 minutes per day. It separately asks about sports participation in after-school time during the last 12 months. From 2005 to 2017, the YRBS asked high school students questions about the number of days they were active for at least 60 minutes and the number of sports teams (school or community sponsored) on which they played during the past 12 months.

The only Federal surveillance system that measures youth sports at the school or community level is the SHPPS, which focuses on school systems (e.g., classroom, school, school district, state level) for kindergarten through twelfth grade. The SHPPS district-level survey includes questions that address whether the district requires or recommends that high schools provide opportunities for PA (including intramural and interscholastic sports) before and after school. It also provides information regarding whether the district has adopted policies related to coaching interscholastic sports. The SHPPS school-level questionnaire has questions that provide insight about the kinds of sports-related opportunities that schools offer students. The

SHPPS includes broad questions, such as “Does your school offer opportunities for students to participate in PA clubs or intramural sports?” and “What activity/sport is offered?”. It also has

43 more specific questions about the programs, such as whether the programs are offered to just boys or girls or to both.

The non-Federal surveillance systems found were Monitoring the Future (MTF),

National Federation of State High School Associations (NFHS), National Sporting Goods

Association (NSGA), and Physical Activity Council (PAC) Survey. Three of these systems

(MTF, NSGA, and PAC) provide public access to the methodology, but comprehensive results from the PAC and the NSGA are only available to the public for a fee. The PAC includes data from the Sports & Fitness Industry Association (SFIA), which is the primary data source for the

Aspen Institute Project Play. The systems collect data from different youth age groups. MTF

(2018) collects information directly from approximately 50,000 eighth, tenth, and twelfth grade students each year. The NSGA collects information from 34,000 youth age 7 and older. Lastly, the PAC (2019) collected information from 20,100 youth age 6 and older.

MTF asks youth about overall participation in PA but also has a separate question about participation in sports, athletics, or exercise and participation in a list of 21 competitive sports during the last 12 months (including school, community, or other organized sports). This is followed by a question about the frequency of participation in the selected sport(s). The NSGA asks about frequency of participation in a large variety (56 total) of sports and recreational activities. The PAC asks about the types of sport (from a list of 124 activities), frequency of participation for each sport in the last 12 months, participation location, and if it was the first time playing the specific sport.

Although a comprehensive picture of youth sports may be constructed by looking at data from various surveillance systems, there is a distinct lack of a single, comprehensive data source. No single system, Federal or non-Federal, measures all aspects of sports participation,

44 including who is participating (e.g., gender, race, ethnicity, age); frequency of participation; number of sports teams; time spent playing sports; type of sports; and location (e.g., school, recreational facility). NHANES measured most of these constructs from 1999 to 2006.

However, this approach was discontinued in 2007, when a new PA questionnaire was adopted.

While a number of aspects of sports participation were measured again from 2013 to 2016,

NHANES currently does not measure sports participation. Having a comprehensive understanding of the types of sports in which youth participate would enable programs to be developed that meet the needs and interests of youth. It would help identify trends, such as increases or decreases in participation rates for specific sports; identify gaps; and help support policies and programs to increase participation.

Measurements for type of sport

Data on the types of sports in which youth are participating are also limited because of the variety of data collection methods that are used. The type of sport in which youth participate is currently measured in only one Federal system (ATUS, ≥15 years old) and three non-Federal systems (MTF, NSGA, and PAC). Each system asks youth to report on the type of sport and frequency or time spent in each sport in different ways. For example, MTF asks youth to select from a predetermined list of sports. On the other hand, ATUS gathers data through interview responses. It collects information on all activities in which adolescents participated during the day before the interview. Interviewers then assign activity codes based on a classification system. The PAC asks youth to report the types of sports from a list of 124 possible options, frequency of participation in the sports, and participation location.

The use of different questions and methods to assess youth sports participation makes it challenging to identify trends and compare data across systems and time periods. YRBS (2017)

45 indicates that 54 percent of students in ninth through twelfth grade reported participating on at least one sports team during the past 12 months. The NSCH (2017) showed that 58 percent of six to seventeen year olds participated on a sports team or in sports lessons after school or on the weekend in the last 12 months. The MTF (2018) results showed that eighth, tenth, and twelfth grade students who actively participate in sports, athletics or exercising was 51 percent,

51 percent, and 43 percent respectively. While the results are similar, these examples show both the variation of how questions are asked and the corresponding participation rates they elicit.

Because of the different questions in each of these systems, it can be challenging to interpret differences in results across systems, particularly when survey questions are double barreled or not homogenous through age groupings.

Another limitation of data from existing surveillance systems is that there is little distinction between sports and other physical activities. Some of these include sports within the broad PA question, while others ask a separate sports participation question. Both of these approaches limit the opportunity to analyze how sports participation contributes to overall PA.

It is not analytically possible to determine what portion of a youth’s overall PA was sports participation when sports is only one example of the types of activities being investigated.

Direct observation

Direct observation is the most applied method for assessing PA and sedentary behavior and is by some considered the gold standard for PA assessment (Sirard & Pate, 2001). PA measurement has improved during the past two decades (Bauman et al., 2012; McKenzie et al.,

2016). One of the most accurate, comprehensive field measures of PA is direct observation

(McKenzie, 1991; Pope et al., 2002). Direct observation is a popular method for assessing PA, which requires a designated observer to document the PA behavior of subjects for a short

46 duration of time. Direct observation exceeds other measures of PA in its ability to provide detailed information on the physical and social context in which it occurs (McKenzie et al., 2010;

McKenzie & van der Mars, 2015). Several studies have adopted systematic observation in the pursuit of measuring and obtaining reliable PA data (Besenyi et al., 2013; Bocarro et al., 2009;

Bocarro et al., 2012; Carlton et al., 2017; Chung-Do et al., 2011; Cohen et al., 2007; 2011; 2013;

Floyd et al., 2011; Lafleur et al., 2013; McKenzie et al., 2006; McKenzie et al., 2010; Parra et al.,

2010; Pleson et al., 2014; Price et al., 2012; Reed et al., 2009; Shores & West, 2010; Tester &

Baker, 2009).

Some of the first systematic observation tools developed and validated were the System for Observing Play and Active Recreation in Communities (SOPARC) and the System for

Observing Play and Leisure Among Youth (SOPLAY). While the design of SOPARC was originally intended to be used in park settings, a growing number of research studies have applied this as well as SOPLAY in schools and playgrounds. The feasibility of SOPLAY was established in a large scale study of 24 diverse schools in southern California to investigate on- campus PA levels throughout the school day (McKenzie, Marshall, Sallis, & Conway, 2000).

McKenzie and colleagues discovered that only two to four percent of daily attendance were present in facilities before and after school. Although physical settings were available for play and recreation during the school day, little to no structured activities, supervision or equipment was present to encourage student PA. Through the results of systematic observation, the authors concluded that many opportunities for PA by middle school students are being wasted

(McKenzie et al., 2000).

The System for Observing Fitness Instruction Time (SOFIT) has been used extensively to study P.E. and other PA settings (e.g., recess, classroom, and sport settings). SOFIT is particularly valuable because it provides simultaneous appraisal of: (a) individual PA levels, (b)

47 activity context, and (c) teacher/instructor behavior. SOFIT has been validated with heart rate monitors (McKenzie et al., 1991) and accelerometers (McKenzie et al., 1994).

Direct observation is not without limitations. One could argue that this method is not suited for assessing activity intensity due to the subjective nature of classification and variability in the training of data collectors as they interpret physical movements. Other limitations with this method include the substantial investigator burden which makes it unsuitable for the assessment of leisure-time PA, the invasion of the individual’s privacy. And particularly with adolescents, there is the potential for subject reactivity and consequently altered behavior as a result of an coder observing a PA space. This phenomenon has been referred to as the

Hawthorne effect and is most prevalent in young ages (Adair et al., 1989).

One of the most popular scales to measure coaching behavior is the Leadership Scale for

Sports (LSS) (Chelladurai & Saleh, 1980). This instrument has been widely used and employed in sports leadership research for more than thirty years. Its reliability and validity have also been re-evaluated several times. The LSS examines five main characteristic behaviors of a coach thus making it possible to establish the degree to which the coach puts emphasis on giving the athletes detailed instructions in order to teach new techniques in a given sport. It allows to verify whether the coach is oriented toward clarifying ties and relationship between the team members and to establish whether he or she pays attention to structuring and coordinating the actions of the players (training and instruction dimension). Moreover, the questionnaire makes it possible to verify the degree to which the coach attaches importance to a positive atmosphere in the team and to establish whether his or her behavior is oriented toward concern for the well- being of the athletes (the social support dimension). When analyzing the positive feedback dimension, the LSS obtains information on the degree to which the coach reinforces the players’ behaviors by recognizing, acknowledging and rewarding good performances of his or her

48 athletes. The democratic and autocratic behaviors concern the management style of the coach

(i.e., the degree to which he or she is able to give the athletes’ opportunities to take part in making decisions relating to sports (Chelladurai, 1989).

Summary

The tendency towards treating sport as a relatively homogenous activity is evidenced in sociological writing on the relationship between sport and health, and consequently used to disqualify sport from discussions of health promotion. Waddington (2000), discusses what he calls the ’sport-health ideology’, framing sport as an inherently competitive activity which entails distinct social relations, and infused with aggressive masculinity and a relatively high tolerance of injury. There is certainly reasons to doubt the role that sport can play in health promotion.

It is clear from the literature that sports participation and associated experiences are not the only or even main determinants of PA during childhood and later in life (Cleland et al., 2009;

Hirvensalo & Lintunen, 2011). For example, children’s socioeconomic conditions can have a major effect on leisure-time PA (Mäkinen, Kestilä, Borodulin, et al., 2010b), including participation in different sports (Tammelin et al., 2003). Researchers reported upward social mobility (i.e., reaching a higher level of educational attainment) to be associated with a greater likelihood of high PA in adulthood and a greater likelihood of a high fitness level (Cleland et al.,

2009). Psychosocial factors have also been found to mediate PA, and these seem to be especially marked for females (Bailey, Wellard, & Dismore, 2004).

Qualitative research has highlighted the effect of body shape and weight management of girls’ participation (Allender, Cowburn, & Foster, 2006). Differences in the experiences and perceptions of boys and girls of early sport and PA is a persistent theme in the literature (Bailey et al., 2004; Cooky, 2009), and there seems to be little doubt that girls have traditionally been

49 disadvantaged by the sports provision made available to them. Traditionally, more boys than girls are involved in organized youth sports in childhood (Hovden & Pfister, 2006), although the

Lunn’s analysis (2010) suggests that differences are greatest in team sports, and that opportunities are more equal in individual sports.

METHODOLOGY

The purpose of this study is three-fold: (1) explore the various contexts of the most commonly offered high school sports, (2) examine the association between practice context and

PA outcomes of high school athletes during sports practices, and (3) examine coaching behavior and its association with the PA outcomes of high school athletes during sports practices. This study follows a non-experimental research design where a sample of public high schools across

North Carolina were selected to participate. A methodology of systematic observation was used to address the research questions. This chapter describes in detail the methods employed in the study. Study methods were approved by the Institutional Review Board for Human Subjects

Research at North Carolina State University (see Appendix A).

Sampling

A comprehensive list of all high schools in North Carolina was obtained from the North

Carolina High School Athletic Association (NCHSAA). Support for the project was attained from the acting Commissioner of NCHSAA with a brief follow up presentation at the 2017

North Carolina Athletic Directors Conference. High schools that offered a minimum of 18 out of the 20 varsity sports of interest by the Aspen Institute were considered eligible to be included in the study. The 20 sports were selected based on the ten most popular boys’ and girls’ programs according to the most recent participation numbers reported by the National

Federation of State High School Associations (NFHS) 2017-2018 Handbook (Tables 3.13 and

3.14). Next, schools were selected based on the following criteria:

(1) District level per pupil expenditure

(2) Percent of study body eligible for the free/reduced lunch program

(3) Percent of student body that is of racial minority

(4) Geography (to prevent regional clustering)

Table 3.11 Sampling of North Carolina high schools

Student Per Pupil Expenditure Sports School # Class Body ($) Observed1 1 1,689 8,114 4A 2 2 1,072 8,114 2A 13 3 933 9,302 2A 5 4 1,401 9,327 4A 4 5 1,851 10,557 4A 6 6 1,907 8,536 4A 16 7 1,512 8,674 4A 17 8 1,351 10,031 4A 9 9 1,398 8,516 3A 1 10 1,360 8,867 3A 2 11 1,519 8,506 4A 1 12 2,018 8,522 4A 1 Sample Average 1,465 8,922 --- 6.4 Statewide 1,017 8,888 ------Average *Note. Number of sports out of 20

Twelve public high schools in North Carolina were initially selected to participate in the study using a stratified nonrandom sampling method. The final list of study sites comprised of a statistically representative sample of all North Carolina high schools (Table 3.11). One other consideration regarding which schools would be included in the study were schools that were within close proximity (i.e., approximately 15 miles) to the residence and/or workplace of a trained data collector. Since data collection occurred on a weekly and sometimes daily basis, the schools needed to be within a reasonable driving distance from the nearest data collector. In

52 most cases, data collectors were teachers and coaches already working on site so no travel was necessary.

Figure 3.11 Location of public NC high schools selected as study sites

Finally, schools were offered a $500 VISA® gift card given to the athletic department as an incentive to participate in this study. The incentive was awarded after the completion of all data collection in June 2018. The gift card was instructed to be used toward the purchase of athletic supplies or equipment, based on the immediate needs of the school/athletic department.

Twelve public high schools were secured as the study sites for data collection (Figure 3.11).

Table 3.12 Demographic characteristics of sample high schools (NFHS, 2017-18)

School # SES1 % Black % Hispanic % White % Minority 1 29.7 11.6 11.3 72.0 28.0 2 37.5 11.6 7.8 77.1 22.9 3 24.1 2.4 4.9 86.8 13.2 4 36.1 27.8 4.5 62.6 37.5 5 52.8 42.4 27.2 24.2 75.8 6 37.6 24.3 15.9 49.6 50.4

Table 3.12 (continued).

7 51.2 55.9 5.9 33.9 66.1 8 30.2 0.9 5.7 89.0 11.0 9 45.5 41.6 14.4 40.5 59.5 10 62.9 13.9 39.6 40.8 59.2 11 43.5 3.9 5.7 85.8 14.2 12 57.5 59.3 15.6 18.6 81.4 Sample 42.4 24.6 13.2 56.7 46.1 Average Statewide 51.3 27.8 11.9 53.3 47.1 Average *Note. Percent of student population receiving free or reduced price lunch

The sports included in the investigation were selected in consultation with the NFHS annual handbook on participation numbers in the U.S. (2017). Specifically, the researchers selected the top ten most popular boys sports (Table 3.13) and the top ten most popular girls sports (Table 3.14) in terms of overall enrollment. One exception was made to the final list of twenty sports included in the study. Golf was the ninth most popular interscholastic sport in the

U.S. for boys according to the manual but was replaced by lacrosse. Golf was removed from consideration based on a number of factors. First, not every high school in North Carolina had a varsity golf team, which compromised our sampling criteria. Second, since the girls’ sport list included lacrosse, the researchers deemed it logical to also include boys’ lacrosse as the gender counterpart. And third, based on the methodology of systematic observation, it was decidedly difficult to imagine being able to apply the systematic observation protocols (discussed in a later section) adequately since golf is a unique sport that is played not within the normal boundaries of typical athletic sessions (e.g., practice structure, facilities, etc.). Therefore, the decision to substitute lacrosse for golf in the boys’ category was made.

Table 3.13 Boys high school sports included in the study (NFHS, 2017).

Sport National Participation (Rank) North Carolina Participation (Rank) Football 1,036,842 (1) 28,600 (1) Track & Field 600,097 (2) 14,663 (2) Basketball 551,373 (3) 11,063 (4) Baseball 487,097 (4) 10,959 (5) Soccer 456,362 (5) 12,298 (3) Cross Country 270,095 (6) 6,632 (7) Wrestling 245,564 (7) 8,111 (6) Tennis 158,151 (8) 3,957 (10) Swimming 138,935 (10) 3,973 (8) Lacrosse* 113,313 (14) 3,961 (9) *Note. Substituted for the ninth most popular sport, golf.

Table 3.14 Girls high school sports included in the study (NFHS, 2017).

Sport National Participation (Rank) North Carolina Participation (Rank) Track & Field 488,592 (1) 10,398 (2) Volleyball 446,583 (2) 9,565 (4) Basketball 412,407 (3) 7,913 (5) Soccer 390,482 (4) 10,254 (3) Softball 367,861 (5) 7,804 (6) Cross Country 223,518 (6) 5,250 (7) Tennis 190,768 (7) 4,316 (9) Swimming 175,594 (8) 5,062 (8) Cheerleading 162,669 (9) 20,250 (1) Lacrosse 96,904 (10) 2,556 (10)

Instruments

System for Observing Fitness Instruction Time

Given the investigation parameters illustrated in the first chapter and the background knowledge gathered from the literature review, it is reasonable to deduct how sport practice time is spent and how coaches conduct themselves concerning PA during practices in high school sport is unknown (Nelson et al., 2011). Consequently, generating data on practice context and coach behavior is best achieved through direct observation, as self-report data may be unreliable or otherwise biased (McKenzie et al., 1991). The objective measure of PA was obtained using the method of systematic observation; a widely accepted and validated technique used to assess

PA in sport and recreation settings (McKenzie et al., 2006).

As one of the many systematic observational tools available, The System for Observing

Fitness Instruction Time (SOFIT) was selected due to its unique capabilities for acquiring the necessary data for answering the research questions: sports practice design and coaching behavior (McKenzie et al., 1991). Typically, SOFIT is used for structured PA sessions such as physical education classes evaluating teacher effectiveness and is recognized as a valid research instrument for assessing student PA levels as well as the contexts of a lesson (McKenzie et al.,

1995; McKenzie et al., 1991; Pope et al., 2002). But over the past decade, the instrument has been adapted to some extent and applied in organized sports (Bass & Martin, 2013; Cotton &

O’Connor, 2013; Curtner-Smith et al., 2007; Guagliano et al., 2015; 2014; Peralta et al., 2014;

Schwamberger & Wahl-Alexander, 2016). SOFIT can easily be implemented in a sport setting as it provides a similar environment led by a coach instead of a teacher.

SOFIT is based on momentary time-sampling techniques in which student athletes are randomly selected by trained observers and coded during short interval sessions (McKenzie et

56 al., 1991). During each observation, athletes are coded based on five different levels of PA: lying down, sitting, or standing, moderate (e.g., walking), or vigorous (e.g., running, jumping, strenuous activity). The randomly selected athletes are also coded for their gender, grade level, the type of sport they are playing, the lesson context, and the coach interaction.

On a rotational basis, the PA levels, practice context, and coach behavior are coded and recorded every twenty seconds via a looped voice recording that prompted the observer to observe and record. The data collectors observes for ten seconds then has ten seconds to answer three questions regarding the athlete being watched. The first decision relates to the PA intensity of the athlete on a scale of 1 to 5. The second decision involves an observer deciding on the context of a practice every twenty seconds. The third decision involves an observer coding coach behavior every twenty seconds. Once the three decisions are made and recorded, the next observation begins and so on. One athlete is observed for twelve complete observations (approximately four minutes) before the data collector moves on to track the next randomly selected athlete. After the twelfth observation of the fourth athlete, the observer then loops back to the first athlete and begins again. An observer continues this process until the end of the sport practice.

The activity codes in SOFIT have been calibrated using heart rate monitoring (Rowe et al., 1997) and validated using accelerometers (McKenzie et al., 1994). Inter-observer reliability has been found to be high and has been successfully used to assess the relationship between PA and its contexts. The instrument has also been validated for use with high school athletes by

Rowe et al. (2004).

iSOFIT®

iSOFIT® is a new, interactive Apple® application developed with the iOS framework, free to download from the Apple® store, and is compatible for all iPad™ devices. All data were input, organized, and stored using iPad-mini™ 4’s latest app release available in lieu of the conventional way of recording observations on paper. The researchers pilot tested one of the earlier versions of the app (version 1.5.3) to gauge its utility and dependability for large scale data collection. After submitting a few recommendations for how to improve iSOFIT® to the app builder, the latest version (1.6.0) was used for the remainder of the project. The layout and main functions of iSOFIT® are similar to the original protocol and data collection follows the same procedures as SOFIT (McKenzie et al., 1991). Additional advantages and features included the streamlining of data processing, quick exportation options, automated time-keeping system, and secure data storage.

Figure 3.12 Example of iSOFIT® summary of PA observed from a practice session

Figure 3.13 Example of iSOFIT® summary of practice context observed from a practice session

Figure 3.14 Example of iSOFIT® summary of coaching behavior observed from a practice session

Observer Training and Reliability

Twenty five observers were hired and trained in the SOFIT procedures during the months of August and September 2017 followed by routine “re-trainings” during reliability sessions. The project manager trained all observers to use SOFIT using recommended guidelines (McKenzie, 2015). SOFIT coding accuracy was assessed against a pre-coded ‘gold standard’ video developed by McKenzie (2012) and the most up to date protocols manual

(McKenzie, 2015). Each observer participated in a two hour, face-to-face training with the project manager at their respective high schools using revised and updated SOFIT training materials (McKenzie, 2015); beginning in a classroom setting and then by interactive practice.

Data collectors were also issued an iPad-mini™ 4 device and trained in the use of the iSOFIT® app. Observers were required to submit one 30 minute practice observation to the project manager within 48 hours of the completion of the training to certify the comprehension of the observational protocols. The project manager reviewed the exported data from the practice sessions prior to clearing data collectors to begin observing sports practices. For this study, athlete levels of activity were coded in consultation with Cardon et al. (2004) and

Schwamberger & Wahl-Alexander (2016) who customized and expanded the SOFIT coaching behavior description to fit the sport of swimming to ensure accurate coding by data collectors.

For example, Cardon et al. (2004) listed ten of the most common swim activities with corresponding 1 to 5 PA codes to enhance higher agreement rates amongst observers.

However, the current study had twenty sports of interest, not just one. The researchers recognized the heterogeneity of various sport types and thus kept the adapted codes and provided examples as generalizable as possible to fit every sport type. Table 3.15 provides an example of how each coaching action during the practices was coded.

Table 3.15 Operationalization of SOFIT coaching instruction codes and sport related examples (Schwamberger & Wahl-Alexander, 2016) Observer Code Definition Examples Promotes fitness Prompting or encouraging fitness Remarks made that stimulates related activities athletes’ fitness activity, telling athletes to give their best effort Demonstrates Models or actively participates in Running laps with the team, fitness fitness related activity demonstrating a fitness exercise, playing in a simulated game Instructs generally Lectures, describes, prompts, or Teaching a new skill, technique, provides feedback to athletes strategy or rule, positive or related to all sport content corrective feedback Manages Manages students or environment Setting up equipment, taking roll, by engaging in non-subject matter writing notes tasks Observes Monitors entire team, group, or Watching a scrimmage, Watching individual by observing without drills take place speaking Other-task Attends to events not related to Reading a magazine, looking at his/her responsibilities to the phone, turning back to team, team talking to a parent, leaves practice area for any reason

Data Collection Procedures

Practices were selected in lieu of games/competitions as athletes are anticipated to spend more time participating in team practices than in actual games. Practices are also more inclusive than games, involving every athlete regardless of talent level. Practices may also provide more

MVPA as coaches are better able to dictate the intensity of a practice compared with games.

Also, a larger proportion of the team can participate simultaneously and in smaller groups, which can provide increased opportunities for players to participate at a higher PA intensity during practice, compared with a game (Kanters et al., 2015).

Data collection was completed in multiple phases beginning in August 2017 and ending in June 2018. Before each SOFIT observation, data collectors arrived early to the practice

61 facility and randomly selected four student athletes in attendance at the practice and made notes of any identifiable characteristics (e.g., the number on the back of the jersey or the color of the athlete’s clothes). The observer maintained a physical position that did not interfere with instructional methods or materials and was positioned primarily to the side of the field or court near the main instructional area. However, when the coach moved around to instruct athletes in different areas of the playing range, the observer moved accordingly to hear and record the coaching behaviors accurately.

Once the practice started, data collectors observed the first athlete for one, ten-second iteration followed by ten seconds for coding. During each 20 second iteration, athletes are coded based on five levels of PA: lying down, sitting, standing, walking, or vigorous; six levels of practice context: management, knowledge content, fitness, skill practice, game play, or other; and six levels of coaching instruction: promotes fitness, demonstrates fitness, instructs generally, manages, observes, or other task (Figure 3.14). After twelve iterations are completed for the first athlete (approximately four minutes), the data collector then repeats this procedure for the second athlete, then the third, and so on. At the conclusion of the twelfth iteration for the fourth athlete, the data collector then circles back to observe the first athlete again. This process is continued until the coach ends practice and dismisses the athletes. The iSOFIT® app was developed and chosen over the conventional SOFIT paper-based form due to the distinct advantages it offered during data collection. First, the app does not require an internet connection to record observations, however, observers must connect to the internet in order to export to the project manager. Data was typically exported via email within 24 to 48 hours of the conclusion of every practice. Secondly, the app has a built in timer that counts down for each interval (20 seconds) and instructs the data collector on when to observe and when to record. This unique feature was critical in allowing observers to concentrate solely on correct

62 coding of athletes rather than attempting to keep pace with a self-timer. The automatic timer on iSOFIT® (Figure 3.15) ensured more consistency and reliability in observations during pilot testing. Third, the exported data came to the project manager already organized in an Excel spreadsheet. Thus, the burden of data entry and data cleaning was all but alleviated and fast- tracked.

Figure 3.15 Primary iSOFIT® counter tool

Sports were assigned to trained data collectors based on their availability and/or comfort level with certain sports (e.g., some females requested to only observe girls’ sports and vice versa). Observers were asked to record at least ten full practices for their respective sport by the end of the regular season. A total of 30 practices were targeted for each of the twenty sports.

Sports were categorized into one of three seasons: fall, winter or spring. Each season of data collection had a pseudo-official launch and close date (Table 3.16), although some exceptions

63 were made if certain sports’ seasons were prolonged due to extenuating circumstances such as make-up games.

Table 3.16 SOFIT data collection schedule for 2017-2018 academic school year

Sport season Begin End Fall August 14 October 25 Winter October 27 March 7 Spring March 11 June 1

Analysis

All data was analyzed using the Statistical Package for Social Sciences (SPSS) Version

26.0 software. The purpose of analyzing survey responses was to determine the extent that the data collected could help predict levels of PA in high school sport practices. Independent variables in the analysis were training and instruction, democratic behavior, autocratic behavior, social support, and positive feedback. Descriptive statistics were calculated and produced, including frequencies and cross tabulations as the first step for understanding the distribution of

PA and the accompanying context that occurred in high school sports practices. A table of all variables and scales is presented in Table 3.17.

Table 3.17 Description of variables for analysis

Outcome variables (DV) Categories Scale Physical activity 5 Ordinal % session in MVPA --- Ratio METs --- Ratio MVPA/Sedentary 2 Binary

Table 3.17 (continued).

Active/Sed Practice 2 Binary Explanatory variables (IV) Categories Scale Sport type for boys 10 Nominal Sport type for girls 10 Nominal Gender of coach 2 Nominal Practice context 6 Nominal Coach behavior 6 Nominal *Note. Sports were dummy coded for inferential analysis.

PA dependent variables of interest

The primary aim of this study revolves around understanding PA behavior and its contexts during sports practices for high school athletes. Therefore, the three major dependent variables for the study are: (a) physical activity (PA), (b) percentage of practice session spent in

MVPA, and (c) estimated energy expenditure rate (METS). On the individual level, PA is the categorical variable assigned to an athlete for each observation under SOFIT protocols: lying down, sitting, standing, moderate, or vigorous. For each full practice session, the proportion of practice time that the observed athletes spent in MVPA behavior was calculated and given by the iSOFIT® app.

Finally, a widely used absolute physiological term for expressing the intensity of physical activities is multiples of resting metabolic rate or standard metabolic equivalents (METs)

(Ainsworth et al., 2011). Previous researchers have successfully utilized METs which can be calculating by multiplying each observed athletes’ PA level with an assigned energy expenditure value: 1.5 for every sedentary athlete, 3.0 for every moderate athlete and 6.0 for every vigorous athlete.

These values have been accepted and widely used in estimating the amount and level of PA based on observational data (Ainsworth et al., 2011; Cohen et al., 2013; Edwards et al., 2014).

The only qualm found in the literature using METs was that there is evidence suggesting that the commonly adopted definition of one MET overestimates resting energy expenditure (e.g., sitting down) by twenty to thirty percent in adults (Byrne et al., 2005; Savage et al., 2007). For example, a person sitting or lying down would realistically only be expending between half of one MET to one MET (Owen et al., 2010) but would be coded as expending 1.5 METs according to the accepted intervals previously presented. But public health recommendations on PA intensity are expressed as METs and moderate intensity PA refers to three to six METs (Ainsworth et al.,

2011), which aligns well with the current conversion calculation. Nevertheless, calculating

METs is helpful in data analysis as it translates a previously ordinal variable into a continuous variable. Continuous variables are ideal to work with, as they are needed to fit data to linear regression models.

Independent variables

The attempt to determine the association of PA and the sport-level and coach-level context is reliant on also measuring the independent variables which were taken by the SOFIT instrument. At the conclusion of data collection, data files from each school were merged together to form a single data set so that associations between the predictor and outcome variables could be generated in SPSS.

PA does not happen in a vacuum; there is a time, a place, and a context that accompanies PA (McKenzie et al., 1991). Other important dimensions of PA are the type or mode of activity (McKenzie, 2016). Type of PA concerns the specific PA being performed, and is usually assessed with self-report methods (Warren et al., 2010) or by direct observation (Hardy et al., 2013). The domain of PA is the context in which the activity takes place (e.g., home, school, during work, during sports). The current study establishes the setting of PA to be in the

66 after-school environment using direct observation. The sport type was the top ten boys and girls sports according to the most recent national data (NFHS, 2017). The twenty sport types are listed in Table 3.13 for boys and Table 3.14 for girls. The gender of the coach along with the practice context (i.e., management, knowledge content, fitness, skill practice, game play, other) were gathered for each observation. The final independent variable gathered from SOFIT was the coaching instruction component (i.e., promotes fitness, demonstrates fitness, instructs generally, manages, observes, or other task).

Inferential Statistics

One-way analysis of means

To test the possible relationship between the independent variables and PA levels, independent one-way analysis of variance (ANOVA) were completed using SPSS 24.0 for the categorical independent variables to test for differences in the means across the ordinal PA dependent variables, specifically for sport type. Previous work using SOFIT has successfully utilized one-way ANOVAs in order to determine whether percentages of time spent in MVPA differs or changes based on athlete activity, coach behavior, and other contexts (Cotton &

O’Connor, 2013; Curtner-Smith et al., 2007). Chi-square tests for independence and association were performed in conjunction with the one-way ANOVAs per recommendations by Cotton and O’Connor (2013). Because the data does not meet all required assumptions for a true

ANOVA to be run (i.e., normally distributed), Kruskal-Wallis tests were chosen as it is a non- parametric alternative to the one-way ANOVA. The Kruskal-Wallis test is also appropriate for ordinal level data as the primary dependent variable, and the advantage of using this type of test is that it can be used to calculate differences between two or more means (Ruxton &

Beauchamp, 2008). The disadvantage is that the only way to know where specific differences

67 exist is to follow up with a form of post-hoc analysis. Therefore, in the cases where statistically significant differences existed, pair-wise comparisons were made for each subgroup within that specific category using Duncan’s post-hoc, the conservative option as far as post-hocs are concerned. Duncan follow-ups were also used and recommended in past SOFIT studies examining PA in sports (Curtner-Smith et al., 2007). Level of significance for inferential tests was established at p < 0.05.

Regression models

Regression analysis was used to measure the effect of specific independent variables on the dependent variables involving PA behavior of athletes and PA production of sport practices.

Since the main PA variable is ordered, ordered logistic regression (OLS) was used to fit the models. OLS is the appropriate regression analysis to conduct when the dependent variables are categorical and one or more of the explanatory variables are nominal or ordinal, which is the case for the current study. OLS can also be used to test for interactions between multiple independent variables to predict the dependent variable (e.g., using practice context, coach instruction, and sport type to predict PA outcomes). The general model for OLS is written in

… the form logit [P(Yi≤j)] = αj + β1x1 + β2x2 + . Each predictor variable (xi) has its own

“measured effect” (Menard, 2010) that can be determined when other predictor variables in the model are held constant (Agresti, 2010; Menard, 2010). Additionally, a greater magnitude of the

β coefficients in the model (a greater distance from zero) indicates a stronger effect for the predictor variables (xi). Each individual explanatory variable will be assessed for its significant prediction to the dependent variables, and the combination of the explanatory variables will be evaluated as to how they predict PA behavior. The research models to be used can be written symbolically as PA is a function of sport type, practice context, and coaching behavior.

Model 1. YPhysAct f (X1SportType)

Model 2. Y%MVPA f (X1SportType)

Model 3. YPhysAct f (X1PracContext)

Model 4. Y%MVPA f (X1PracContext)

Model 5. YPhysAct f (X1CoachBehav)

Model 6. Y%MVPA f (X1CoachBehav)

Model 7. YPhysAct f (X1SportType, X2PracContext, X3CoachBehav)

Model 8. Y%MVPA f (X1SportType, X2PracContext, X3CoachBehav)

Finally, to help bolster the translation of the results to sport practitioners, odds ratios

(ORs) were reported with separate logistic regression analyses for sport type/MVPA, practice context/MVPA, and coach behavior/MVPA. Results from the ordinal logistic regression were reported with the ORs that represent the odds of an outcome occurring given particular circumstances compared to the odds of an outcome occurring in the absence of certain conditions (Fullerton, 2009).

In order to acquire the binomial dependent variables needed for the logistic regression with odds ratios, dummy variables were created for both the ordinal PA levels and ratio level percent practice time in MVPA. The original five levels of PA measured through SOFIT (i.e., lying down, sitting, standing, moderate, and vigorous) were collapsed and recoded into two categories: 0=sedentary and 1=moderate/vigorous. The purpose behind the recoding was to aid in the translation of the resulting ORs. For example, this research was not interested in understanding what influences moved an athlete moved from the “lying down” category to the

“sitting” category. Rather, a more informed analysis of what influenced an athlete to move from

69 being sedentary to MVPA behavior was of higher interest and hypothesized to result in a stronger argument within the relationship between sport context and PA.

Similarly, the percent practice time in MVPA (0-100) was recoded into two categories:

0=0.00 to 67 and 1=67.01 to 100. Again, the premise behind this new variable was to allow the logistic regression to produce more robust ORs to allow comparisons between low PA producing practices versus high PA producing practice sessions. For the purposes of this study, an “active” practice is considered that in which at least two-thirds of the total practice time was spent in MVPA and a “sedentary” practice was defined as a practice that had less than two- thirds of the practice time spent in MVPA.

Finally, the independent variable of sport type was also later dummy coded into 20 new variables for each sport using IBM SPSS 26.0’s feature “recode into different variables”. This resulted in 20 new columns in the data set where 1=[sport of interest] and 0=[every other sport]. This allowed for a single sport (e.g., girls lacrosse) to be isolated, analyzed, and compared with all other sports combined, rather than a single sport serving as the reference category by itself.

Model 9. YMVPA f (X1SportType)

Model 10. YMVPA f (X1PracContext)

Model 11. YMVPA f (X1CoachBehav)

Model 12. Y%Practice f (X1SportType)

Model 13. Y%Practice f (X1PracContext)

Model 14. Y%Practice f (X1CoachBehav)

Model 15. YMVPA f (X1PracContext, X2CoachBehav, X3SportType)

Model 16. Y%Practice f (X1PracContext, X2CoachBehav, X3SportType)

Summary

This chapter described the sampling procedure of study sites in NC and the selection of high school sports in addition to providing the quantitative approach to the study. The chapter also describes the data collection schedule and process of completing SOFIT observation. The

PA levels, session context, and coaching behavior of four randomly selected sport participants in each practice session were coded every twenty seconds throughout the observed session on a rotational basis. I conclude by presenting the process used in the analysis of the data and the ordinal logistic regression models to be used in answering the research questions.

RESULTS

The aim of this study was to further our understanding of PA in sport settings by determining what factors are positively or negatively associated with the PA experienced by high school athletes during practice time. This chapter presents the results of the direct observations of the randomly selected athletes conducted in this study, including findings related to practice design, coaching behavior, and type of sport being played. The chapter is presented in three sections. The first section includes an overview of the data gathered from SOFIT along with pertinent descriptive statistics and reliability analyses. An assembly of histograms accompany the descriptives to help illustrate the distribution of PA levels across sports. The second section reports the inferential tests that were run; namely, the hierarchical logistic regression analyses.

The regression models are accompanied with tests for independence in addition to explanations and interpretations of the statistics as it relates to the associations of practice level measures and their predictive power on PA behavior both in individual athletes and practices as a whole. The final section presents the odds ratio results from the regression analyses where the practice level independent variables gave the likelihood of either increased or decreased MVPA levels from occurring based on the practice context or coaching behavior.

The research questions used to guide this study are also included in this section. The first question is a general descriptive question on the make-up of PA in sports practices. The second and third questions revolved around context factors and its association with PA. The third question dealt with the interrelationships and collinearity of the independent variables. The fifth question was concerned with developing a prediction model for PA based on all variables of interest.

Research Questions

SRQ1. How physically active are athletes during high school sports practices?

SRQ2. What are the variations in PA across different sports for boys and girls?

SRQ3. What is the association between practice context and PA during high school sports practices and how does it vary based on gender and sport type?

SRQ4. What is the association between coaching behavior and PA during high school sports practices and how does it vary based on gender and sport type?

SRQ5. What are the interrelationships between sport type, coaching behavior, practice context, and proportion of time spent in MVPA during high school sports practices?

SRQ6. How do sport context factors combine to predict PA behavior during high school sports practices?

Descriptive Data of SOFIT

Each of the twenty high school sports were observed in their respective fall, winter and spring high school sport season dispersed within 12 public high schools in North Carolina.

Observations were conducted at 598 varsity sports practices, for a total of 696 hours of practice time, amounting to 2,408 total athletes observed. Unadjusted means, standard deviations, and percentages were calculated for participant PA levels, lesson/practice context, and coach behavior following previous SOFIT analyses (Peralta et al., 2014).

Table 4.11 displays the distribution of observations for each sport and shows that a total of 125,218 individual observations were recorded for 76 different sports teams. Boys baseball was the only sport with less than three teams included in the study. Average practice length was one hour and twenty minutes but with high variability across sports (SD = 26.4 minutes).

Overall, 43% of the time athletes were observed engaging in vigorous activity. Walking

(moderate activity) was observed 17% of the time and sedentary (lying down, sitting, standing) was observed 40% of the time. The average percent of practice time devoted to MVPA activities was 61 percent. But again, there was major fluctuations based on the sport type (SD =

21.9 percent).

Table 4.11 Descriptives for completed observations during high school sport practices during the 2017- 2018 school year Practice Percent Number of Number of Sport Teams Length (in Practice in Observations Practices minutes) MVPA Boys Sports M (SD) M (SD) Baseball 2 3338 17 72 (13.6) 43 (22.7) Basketball 4 8444 32 99 (26.0) 67 (14.9) Cross 3 2783 21 50 (12.6) 86 (7.0) Country Football 5 11886 43 101 (27.4) 64 (26.2) Lacrosse 3 8324 34 90 (29.7) 63 (14.7) Soccer 4 5449 26 74 (15.4) 70 (11.8) Swimming 5 6649 37 75 (21.3) 61 (17.6) Tennis 3 3580 21 59 (12.2) 59 (15.6) Track & 3 2618 17 56 (8.2) 87 (9.0) Field Wrestling 4 11825 46 87 (25.9) 63 (17.2) Girls Sports Basketball 4 10922 38 100 (23.1) 55 (17.1) Cheerleading 4 6432 33 71 (22.0) 28 (20.1) Cross 3 2175 14 67 (25.9) 75 (16.9) Country Lacrosse 4 4039 26 55 (12.1) 60 (19.6) Soccer 3 5258 28 58 (20.2) 71 (12.2) Softball 4 4930 27 65 (17.8) 45 (25.4)

Table 4.11 (continued).

Swimming 5 7292 40 74 (22.0) 62 (16.6) Tennis 3 5489 30 69 (16.4) 59 (16.2) Track & 5 6897 32 75 (18.5) 54 (15.2) Field Volleyball 5 6888 36 72 (19.2) 63 (31.3) TOTAL 76 125218 598 80 (26.4) 61 (21.9)

Inter-observer reliability was checked using the procedures recommended by van der

Mars (1989) using the Kappa statistic. Reliability was calculated by using strict interval by interval comparisons or in this case, athlete by athlete. Reliability percentages resulting from this check were 72.6 percent (physical activity), 72.6 percent (practice context), and 68.7 percent

(coach behavior) and thus slightly below the eighty percent level recommended by van der Mars

(1989) for claiming excellent agreement. Yet, the consistency among observers was still within the acceptable range of Fleiss, Levin, and Paik, (2003) who consider 0.70 and above as excellent agreement. According to Fleiss et al. (2003), Kappa coefficients above 0.55 is interpreted as a moderate level of agreement by most researchers.

Overall Mean Scores by Sport

A comparison across sports for the three levels of observed PA is presented in Figure

4.11 for girls’ sports and Figure 4.12 for boys’ sports. For girls, cross country and soccer had the highest percent of observed MVPA where cheerleading and softball had the highest amounts of observed sedentary behavior. For boys, cross country and track had the highest amounts of observed MVPA and baseball had the highest amounts of observed sedentary behavior.

Figure 4.11 Physical activity levels across sports for girls

MVPA Sedentary

100%

75%

50% Percent

25%

Sport

Figure 4.12 Physical activity levels across sports for boys

MVPA Sedentary

100%

75%

50% Percent

25%

Sport

Mean METs were calculated at the observational level for each of the twenty sports, comparing differences in scores for high school athletes. A one-way ANOVA was conducted to determine if PA levels were different for athletes in different sports. Among the twenty sport types included in the study (see Table 4.11), there existed a statistically significant difference in

METs across sports, F(19, 123486) = 417.04, p < 0.001). Data was normally distributed for each sport and homogeneity of variances was violated as assessed by Levene’s test of homogeneity of variances (p < 0.001). A chi-square test of independence revealed a statistically significant association between sport type and PA level (χ2 = 11420.5, p < 0.001).

Observed physical activity data were converted to standard metabolic equivalents

(METs) for each sport to compare sports based on the estimated energy expended during practices (see Table 4.12). METs can be calculated by multiplying each observed athletes’ PA level with an assigned energy expenditure value: 1.5 for every sedentary athlete, 3.0 for every moderate athlete, and 6.0 for every vigorous athlete. These values have been accepted and widely used in estimating the amount and level of PA based on observational data (Ainsworth et al., 2011; Edwards et al., 2014). Overall, cross country was found to have the highest MET level

(p < 0.05) compared to all other sports while cheerleading was significantly the lowest.

Wrestling, swimming, and soccer all had significantly similar high MET values (> 4.0). Boys cross country had the highest MVPA rates with an average MET value of 4.84. Boys track & field (4.62) and girls cross country (4.35) were second and third, respectively. Swimming and soccer for both genders were also included in the top half of the index. The sports that were observed having the lowest physical activity levels (i.e., highest amounts of sedentary behavior) were girls cheerleading (2.41), girls softball (2.76), boys baseball (3.02), girls track & field (3.23), and girls tennis (3.46). Results also indicated that overall, boys sports generated more PA during practices than girls sports.

Table 4.12 Ranking of observed high school sports based on average METs* Sport Type Average METs No. Boys Sports Cross Country*** 4.84 2794 Track & Field*** 4.62 2618 Swimming 4.11 6649 Wrestling 4.10 11825 Soccer 4.05 5462 Lacrosse** 3.82 8324 Basketball*** 3.71 8444 Football** 3.62 11894 Tennis** 3.61 3580 Baseball*** 3.02 3338 Girls Sports Cross Country*** 4.35 2187 Soccer*** 4.23 5258 Swimming 4.11 7292 Basketball* 3.51 10922 Volleyball** 3.89 6888 Lacrosse* 3.53 4039 Tennis* 3.46 5501 Track & Field*** 3.24 6909 Softball*** 2.76 4930 Cheerleading*** 2.41 6432

Note. METs indicates metabolic equivalent of task; B=boys; G=girls. ***Significantly different from all sports. **Significantly different from all but one sport. *Significantly different from all but two sports.

Results from the Duncan post hoc for homogeneous subsets confirmed that significant differences for PA production were found based on sport type and what specific sports differed

78 from each other. The top four sports (boys cross country, boys track & field, girls cross country, and girls soccer), and the bottom four sports (girls cheerleading, girls softball, boys baseball, and girls track & field) stood alone in their respective MET grouping and differed from all other nineteen sports. Four other sports (boys tennis, boys football, boys lacrosse, and girls volleyball) were different from every other sport with the exception of one. The middle section of the ranking were combined within the same subgroup (girls swimming, boys swimming, boys wrestling, girls soccer) and were found to be homogenous to each other but heterogeneous to the other sixteen sport types.

Findings in Practice Context

In addition to observing PA, observers also recorded the practice context for every interval. In other words, for each observation sample, a decision is made regarding whether the practice time is being allocated for management, knowledge content, fitness, skill practice, game play, or other. Management refers to actions performed by athletes not involved in the sport content which can include transition, break times, changing gear, or moving from one space to another. Knowledge content refers to time spent where the primary focus is on athlete acquisition of knowledge related to the sport which can include technique, strategy, rules, and teaching from coaches. Fitness refers to activities whose major purpose is to alter the physical state of the individual in terms of cardiovascular endurance, strength, or flexibility. Skill practice refers to actions where the primary goal is to strengthen, or develop skills in an applied setting.

The primary characteristic of skill practice is the targeting of a specific skill (e.g., dribbling a basketball) and the athlete repeating the drill to further refine that skill. Game play refers to activity time devoted to the application of skills in a game-like setting such as a simulated competition or scrimmage. Other refers to all other actions performed by the athlete that does

79 not fit into any of the previously discussed categories and usually involves some form of unorganized free play.

Table 4.13 Mean differences of METS and practice context for SOFIT observations (N=125,218)

Practice Context Mean SD Game Play 4.67 1.88 Fitness 4.54 1.90 Skill Practice 4.28 2.03 Other 2.18 1.27 Management 2.12 0.90 Knowledge Content 1.71 0.72 Total 3.71 2.07

Table 4.13 shows the mean METs calculated within each practice context factor. The highest context factor was Game Play (M=4.67) and the lowest was Knowledge Content

(M=1.71). Skill Practice had largest variability in observations (SD=1.03) while Knowledge

Content had the lowest (SD=0.72).

Mean METs were calculated at the observational level for each of the six practice context codes, comparing differences in scores for high school athletes. A one-way ANOVA was conducted to determine if PA levels were different for athletes in different contexts. Data were normally distributed for each code and homogeneity of variances was violated as assessed by Levene’s test of homogeneity of variances (p < 0.001). A chi-square test of independence revealed a statistically significant association between practice context and PA level (χ2 =

42560.39, p <0.001)). Among the six practice contexts included in the study (see Table 4.14), there existed a statistically significant difference in METs observed, F(6, 123499) = 8793.78, p <

0.001).

Table 4.14 Duncan post hoc homogeneous subsets for practice context

Subset Practice Context N 1 2 3 4 5 6 Knowledge Content 17027 1.71* Management 13722 2.16* Other 4554 2.18* Skill Practice 44941 4.28* Fitness 23733 4.54* Game Play 18535 4.67* *Note. p<0.05

Results from the Duncan post hoc for homogeneous subsets confirmed that significant differences for PA production were found based on the practice context factor and identified which codes differed from each other. Knowledge content, management, other, skill practice, and, game play all differed significantly from each other in terms of PA measured in high school athletes.

Athletes overall spent a majority of practice time performing skill practice activities

(36%), which are defined as activities repeated by the athlete for the purposes of targeting a specific skill. Fitness was the next most common activity (19%), followed by game play (15%), knowledge content (14%), management (11%), and other (4%) (see Figure 4.13). The girls’ sports with the highest amount of game play were soccer, softball, and tennis. The boys’ sports with the highest amount of game play were tennis, cross country, and basketball. The sports with the largest proportion of practice time devoted to fitness were cross country, swimming, and track & field for both genders.

An examination of how practice context measures were associated with observed PA levels during practices indicated practice time devoted to Game Play (4.67 MET average) was associated with higher levels of PA (see Figure 4.14 and Figure 4.15). Practice time focused on

Fitness (4.54) and Skill Practice (4.28) were both also associated with mean PA levels above 4.00

METs. Practice time focused on Other (2.18), Management (2.11), and Knowledge Content

(1.71) were associated with mean MET levels that indicated a higher likelihood of sedentary activity.

Figure 4.13 Frequency and distribution of practice context measures

MVPA Sedentary 40%

30%

20% Percent

10%

Practice Context

Figure 4.14 Frequency and distribution of girls sport practice context measures

Fitness Game Play Knowledge Management Skill Practice Other

100%

75%

50% Percent

25%

Sport

Figure 4.15 Frequency and distribution of boys sport practice context measures

Fitness Game Play Knowledge Management Skill Practice Other

100%

75%

50% Percent

25%

Sport

Findings in Coaching Behavior

Table 4.15 shows the mean METs calculated within each coaching behavior factor. The highest coaching behavior factor was promotes fitness (M=4.83) and the lowest was manages

(M=3.17). Demonstrates fitness had largest variability in observations (SD=2.12) while promotes fitness had the lowest (SD=1.78).

Table 4.15 Mean differences of METS and coaching behavior for SOFIT observations (N=125,218)

Coaching Behavior Mean SD Promotes Fitness 4.83 1.78 Observes 4.15 2.02 Demonstrates Fitness 3.91 2.12 Other-task 3.44 2.02 Instructs 3.30 2.08 Manages 3.17 1.80 Total 3.71 2.07

Mean METs were calculated at the observational level for each of the six coaching behavior codes, comparing differences in scores for high school athletes. A one-way ANOVA was conducted to determine if PA levels were different for athletes within different coaching behaviors. Among the six coaching codes included in the study (see Table 4.15), there existed a statistically significant difference in METs observed, F(6, 123499) = 1090.19, p < 0.001). Data were normally distributed for each code and homogeneity of variances was violated as assessed by Levene’s test of homogeneity of variances (p < 0.001). A chi-square test of independence revealed a statistically significant association between coaching behavior and PA level (χ2 =

9242.63, p < 0.001).

Figure 4.16 Distribution of physical activity and coaching behavior

MVPA Sedentary 50%

40%

30%

20%Percent

10%

Coaching Behavior

Table 4.16 Duncan post hoc homogeneous subsets for coaching behavior

Subset Coaching Behavior N 1 2 3 4 5 6 Manages 10740 3.17* Instructs 44290 3.30* Other-task 8660 3.44* Demonstrates Fitness 4264 3.91* Observes 47566 4.15* Promotes Fitness 4810 4.83* *Note. p<0.05

Results from the Duncan post hoc for homogeneous subsets confirmed that significant differences for PA production were found based on the coaching context factor and identified which codes differed from each other. Promotes fitness, demonstrates fitness, manages,

85 instructs, observes, and other-task all differed significantly from each other in terms of PA measured in high school athletes.

Figure 4.17 Frequency and distribution of boys sport coaching behavior measures Promotes Fit Demonstrates Fit Instructs Manages Observes Other-task

100%

75%

50% Percent

25%

Sport

Figure 4.18 Frequency and distribution of girls sport coaching behavior measures Promotes Fit Demonstrates Fit Instructs Manages Observes Other-task

100%

75%

50% Percent

25%

Sport

Regression Analysis

Univariate generalized linear modelling (GLMs) and multivariate linear regression (see

Tables 4.17 – 4.29) were used to fit models for PA during high school sport practice sessions.

Predictor variables in the model were, practice context (i.e., fitness, game play, skill practice, knowledge content, management, and other), coaching behavior (i.e., promotes fitness, demonstrates fitness, instructs, manages, observes, and other-task) for all sport types. A total of sixteen separate regression models were run due to the hierarchal nature of the data using SPSS

26.0.

Sport type models

Univariate regression models were run to understand whether the type of sport being played had any effect on the PA levels exhibited by athletes and the percent of practice time in

MVPA. Volleyball was selected as the reference category because it was the median scoring sport in terms of METs overall with ten of the other sport types performing below and nine performing above in terms of PA production. Logically, this was a conservative sport type to hold constant for the regression analyses.

Table 4.17 Model 1 - Linear regression parameter estimates of METs based on sport type 95% CI Parameter β SE t p Lower Upper Intercept 3.89 0.02 160.90 <0.001 3.84 3.94 Boys Sport Type Baseball -0.88 0.04 -20.54 <0.001 -0.96 -0.79 Basketball -0.18 0.03 -5.44 <0.001 -0.24 -0.11 Cross Country 0.95 0.05 21.02 <0.001 0.86 1.04 Football -0.26 0.03 -8.49 <0.001 -0.32 -0.20 Lacrosse -0.07 0.03 -2.26 0.024 -0.14 -0.01 Soccer 0.16 0.04 4.47 <0.001 0.09 0.23 Swimming 0.22 0.04 6.33 <0.001 0.15 0.29 Tennis -0.28 0.04 -6.82 <0.001 -0.36 -0.20 Track & Field 0.73 0.05 15.61 <0.001 0.63 0.82 Wrestling 0.20 0.03 6.39 <0.001 0.14 0.26 Girls Sport Type Basketball -0.38 0.03 -12.17 <0.001 -0.44 -0.32 Cheerleading -1.48 0.04 -42.55 <0.001 -1.55 -1.42 Cross Country 0.45 0.05 9.10 <0.001 0.35 0.55 Lacrosse -0.36 0.04 -8.93 <0.001 -0.44 -0.28 Soccer 0.33 0.04 9.04 <0.001 0.26 0.41 Softball -1.14 0.04 -29.86 <0.001 -1.21 -1.06 Swimming 0.22 0.03 6.57 <0.001 0.16 0.29

Table 4.17 (continued).

Tennis -0.43 0.04 -11.78 <0.001 -0.50 -0.36 Track & Field -0.66 0.03 -19.23 <0.001 -0.73 -0.59 Volleyball 0* . . . . . *Note. Reference category.

Sport type was found to be statistically significant across all observations in predicting

PA, F(1, 19) = 417.05, p<0.001. However, the type of sport played explained a negligible amount of the variability in PA. Sport type accounted for only 6.0% of the variation in PA for all high school athletes with an adjusted R2 = 6.0%, a low size effect (Cohen et al., 2003).

Regression coefficients and standard errors can be found above in Table 4.17.

Table 4.18 Model 2 - Linear regression parameter estimates of percent practice time spent in MVPA based on sport type

95% CI Parameter β SE t p Lower Upper Intercept 63.49 0.23 278.49 <0.001 63.04 63.93 Boys Sports Baseball -20.45 0.40 -51.26 <0.001 -21.23 -19.67 Basketball 3.74 0.31 12.17 <0.001 3.14 4.34 Cross Country 22.93 0.43 53.96 <0.001 22.10 23.76 Football 0.20 0.29 0.70 0.483 -0.36 0.76 Lacrosse -0.63 0.31 -2.05 0.040 -1.24 -0.03 Soccer 6.76 0.34 19.72 <0.001 6.09 7.44 Swimming -2.63 0.33 -8.09 <0.001 -3.27 -1.99 Tennis -4.19 0.39 -10.74 <0.001 -4.95 -3.42 Track & Field 23.98 0.43 55.20 <0.001 23.13 24.83 Wrestling -0.27 0.29 -0.95 0.345 -0.83 0.29 Girls Sports

Table 4.18 (continued).

Basketball -8.30 0.29 -28.49 <0.001 -8.87 -7.72 Cheerleading -35.07 0.33 -106.90 <0.001 -35.71 -34.43 Cross Country 11.59 0.47 24.91 <0.001 10.68 12.51 Lacrosse -3.59 0.38 -9.57 <0.001 -4.32 -2.85 Soccer 7.18 0.35 20.73 <0.001 6.50 7.86 Softball -18.63 0.35 -52.77 <0.001 -19.32 -17.94 Swimming -1.89 0.32 -5.95 <0.001 -2.51 -1.27 Tennis -4.01 0.34 -11.70 <0.001 -4.68 -3.33 Track & Field -9.33 0.32 -28.94 <0.001 -9.96 -8.70 Volleyball 0* . . . . . *Note. Reference category.

Sport type was found to be statistically significant across all observations in predicting percent practice time in MVPA, F(1, 19) = 2271.11, p<0.001. The type of sport played did not explain much of the variability of the proportion of practice time in MVPA. Sport type accounted for 25.6% of the variation of percent of practice time in MVPA for all high school sports practices with an adjusted R2 = 25.6%, still a low effect size (Cohen et al., 2003).

Regression coefficients and standard errors can be found above in Table 4.18.

Practice context models

Univariate regression models were run to understand whether the context of sport practices had any effect on the PA levels exhibited by athletes and the percent of practice time in

MVPA. The ‘other’ code was selected as the reference category because it was the median scoring practice context term with regards to athlete PA production. The MVPA rates in Other were normally distributed so logically, this was a conservative choice to hold constant for the regression analyses.

Table 4.19 Model 3 - Linear regression parameter estimates of METs based on practice context

95% CI Parameter β SE t p Lower Upper Intercept 2.18 0.03 85.32 <0.001 2.13 2.23 Practice Context Fitness 2.36 0.03 84.74 <0.001 2.31 2.42 Game Play 2.49 0.03 87.42 <0.001 2.43 2.55 Knowledge Content -0.47 0.03 -16.44 <0.001 -0.53 -0.42 Management -0.06 0.03 -2.13 0.033 -0.12 -0.01 Skill Practice 2.10 0.03 78.40 <0.001 2.05 -2.15 Other 0* . . . . . *Note. Reference category.

Practice context was found to be statistically significant across all observations in predicting PA, F(1, 5) = 10910.31, p<0.001. The regression analysis for practice context measures indicated a significant predictive relationship, accounting for 30.8% of the variation in

PA for all high school athletes with an adjusted R2 = 30.8%, p< 0.001. Regression coefficients and standard errors can be found above in Table 4.19.

Practices that focused primarily on skill practice essentially yielded an average level of

PA. The more time spent in game play and fitness significantly increased PA levels above the mean and time spent in knowledge content, management, and other activities, significantly reduce the mean level of PA within practice. Generalization of this finding is cautioned as the type of game play can vary from sport to sport. For example, softball had high amounts of game play but as showed previously, had very low PA rates compared to other sports.

Table 4.20 Model 4 - Linear regression parameter estimates of percent practice time spent in MVPA based on practice context

95% CI Parameter β SE t p Lower Upper Intercept 54.31 0.321 169.283 <0.000 53.68 54.94 Practice Context Fitness 9.13 0.350 26.065 <0.000 8.44 9.82 Game Play 10.96 0.358 30.600 <0.000 10.26 11.66 Knowledge Content 0.31 0.361 0.847 0.397 -0.40 1.01 Management 1.76 0.370 4.746 <0.000 1.03 2.48 Skill Practice 6.96 0.337 20.662 <0.000 6.30 7.62 Other 0* . . . . . *Note. Reference category.

Practice context was found to be statistically significant across all observations in predicting percent practice time in MVPA, F(1, 5) = 719.48, p<0.001. However, the context of practice sessions did not explain much of the variability of the proportion of practice time in

MVPA. Practice context accounted for 2.9% of the variation of percent of practice time in

MVPA for all high school sports practices with an adjusted R2 = 2.8%, an extremely low size effect according to Cohen and others (2003). Regression coefficients and standard errors can be found above in Table 4.20.

Coaching behavior models

Univariate regression models were run to understand whether the coaching behavior measure had any effect on the PA levels exhibited by athletes and the percent of practice time in

MVPA. The ‘other-task’ code was selected as the reference category because it was the median scoring coach behavior term with regards to athlete PA production. The MVPA rates in other-

92 task were normally distributed so logically, this was a conservative choice to hold constant for the regression analyses.

Table 4.21 Model 5 - Linear regression parameter estimates of METs based on coaching behavior

95% CI Parameter β SE t p Lower Upper Intercept 3.44 0.02 158.82 <0.001 3.40 3.49 Coaching Behavior Demonstrates Fitness 0.47 0.04 12.40 <0.001 0.39 0.54 Instructs -0.15 0.02 -6.25 <0.001 -0.20 -0.10 Manages -0.27 0.03 -9.40 <0.001 -0.33 -0.22 Observes 0.70 0.02 29.76 <0.001 0.66 0.75 Promotes Fitness 1.38 0.04 38.15 <0.001 1.31 1.46 Other-task 0* . . . . . *Note. Reference category.

Coaching behavior was found to be statistically significant across all observations in predicting PA, F(1, 5) = 1304.17, p<0.001. However, coaching behavior explained a negligible amount of the variability in PA. Coaching behavior accounted for only 5.1% of the variation in

PA for all high school athletes with an adjusted R2 = 5.1%, a low effect size according to Cohen and others (2003). Regression coefficients and standard errors can be found above in Table

4.21.

Table 4.22 Model 6 - Linear regression parameter estimates of percent practice in MVPA based on coaching behavior 95% CI Parameter β SE t p Lower Upper Intercept 53.54 0.23 228.46 <0.001 53.08 54.00

Table 4.22 (continued).

Coaching Behavior Demonstrates Fitness 6.48 0.41 15.88 <0.001 5.68 7.28 Instructs 7.89 0.26 30.79 <0.001 7.39 8.39 Manages 4.83 0.32 15.34 <0.001 4.22 5.45 Observes 7.12 0.26 27.93 <0.001 6.62 7.62 Promotes Fitness 16.01 0.39 40.82 <0.001 15.24 16.78 Other-task 0* . . . . . *Note. Reference category.

Coaching behavior was found to be statistically significant across all observations in predicting percent practice time in MVPA, F(1, 5) = 379.35, p<0.001. However, the behavior showed by coaches did not explain much of the variability of the proportion of practice time in

MVPA. Coach behavior only accounted for 1.6% of the variation of percent of practice time in

MVPA for all high school sports practices with an adjusted R2 = 1.5%, a low size effect according to Cohen and others (2003). Regression coefficients and standard errors can be found above in Table 4.22.

Full generalized linear model

Due to the hierarchal nature of the SOFIT data, several multiple regression models were run controlling in efforts to explore all possible combinations of sport context factors on PA behavior of athletes and the MVPA make up of practices. The final multiple regression models were run with all independent variables from SOFIT observations as predictors of PA (Tables

4.23 – 4.27) and percent practice time in MVPA (Tables 4.28 – 4.29). Boys’ and girls’ sports were run separately as indicated by the model titles of “A” and “B” since significant differences in PA behavior existed between the two genders.

Table 4.23 Model 7A - Summary of regression analysis for variables predicting METS for boy athletes during high school sports practices (N=125,218) 95% CI Parameter β SE t p Lower Upper Intercept 2.56 0.13 19.18 <0.001 2.30 2.82 Practice Context Fitness 2.42 0.19 12.60 <0.001 2.04 2.80 Game Play 1.28 0.28 4.57 <0.001 0.73 1.83 Knowledge Content -1.46 1.06 -1.37 0.17 -3.55 0.63 Management 0.74 0.72 1.03 0.30 -0.67 2.15 Skill Practice 1.75 0.42 4.20 <0.001 0.94 2.57 Other 0* . . . . . Coach Behavior Demonstrates Fitness -2.84 0.85 -3.34 <0.001 -4.50 -1.17 Instructs -0.05 0.29 -0.16 0.87 -0.62 0.53 Manages -2.17 2.55 -0.85 0.40 -7.18 2.83 Observes 1.19 0.31 3.84 <0.001 0.58 1.80 Promotes Fitness 0.44 0.92 0.48 0.63 -1.37 2.25 Other task 0* . . . . . Sport Type Baseball -0.20 0.33 -0.62 0.53 -0.85 0.44 Basketball 0.33 0.19 1.78 0.08 -0.03 0.70 Cross Country 0.76 0.44 1.72 0.09 -0.11 1.63 Football -0.25 0.35 -0.71 0.48 -0.93 0.44 Lacrosse -0.36 0.19 -1.89 0.06 -0.74 0.01 Soccer -0.17 0.20 -0.87 0.38 -0.55 0.21 Swimming -0.86 0.21 -4.12 <0.001 -1.27 -0.45 Tennis -1.06 0.72 -1.48 0.14 -2.47 0.35 Track & Field 0.44 1.13 0.39 0.70 -1.77 2.65

Table 4.23 (continued).

Wrestling 0* . . . . .

*Note. Reference category

The full GLM model for boys sports (Model 7A) was a good fit for the data and were statistically significant across all observations in predicting PA, F(1, 317) = 139.02, p < 0.001

(Table 4.23). All independent variables explained 41.6% of the variability in PA for all high school athletes with an adjusted R2 of 41.3%, a decent size effect according to Cohen and others

(2003). Sport context measures from SOFIT and interaction effects statistically significantly predicted METs, F(1, 183) = 9.31, p < 0.001.

Table 4.24 Model 7B - Summary of regression analysis for variables predicting METS for girl athletes during high school sports practices (N=125,218) 95% CI Parameter β SE t p Lower Upper Intercept 1.73 0.21 8.13 <0.001 1.32 2.15 Practice Context Fitness 0.78 0.21 2.88 0.004 0.25 1.31 Game Play 1.27 0.85 1.48 0.14 -0.41 2.94 Knowledge Content -1.32 1.85 -0.71 0.48 -4.94 2.31 Management 0.03 0.37 0.08 0.94 -0.69 0.75 Skill Practice 0.16 0.23 0.68 0.50 -0.30 0.61 Other 0* . . . . . Coach Behavior Demonstrates Fitness -0.23 1.03 -0.23 0.82 -2.26 1.79 Instructs -0.23 1.45 -0.16 0.87 -3.07 2.60 Manages 4.27 1.45 2.95 0.003 1.43 7.10 Observes 0.18 0.48 0.37 0.72 -0.77 1.12

Table 4.24 (continued).

Promotes Fitness 2.27 1.15 1.97 0.05 0.02 4.52 Other-task 0* . . . . . Sport Type Basketball 0.22 0.37 0.61 0.54 -0.50 0.94 Cheerleading 0.10 0.24 0.44 0.66 -0.36 0.56 Cross Country 2.20 0.26 8.30 <0.001 1.68 2.71 Lacrosse 0.43 0.52 0.83 0.41 -0.59 1.46 Soccer 0.25 0.24 1.04 0.30 -0.22 0.71 Softball -0.22 0.25 -0.87 0.39 -0.72 0.28 Swimming 0.02 0.26 0.08 0.94 -0.49 0.53 Tennis -0.07 0.25 -0.29 0.78 -0.56 0.42 Track & Field 0.18 0.24 0.76 0.45 -0.29 0.66 Volleyball 0* . . . . .

*Note. Reference category.

The full GLM model for girls sports (Model 7B) was a good fit for the data and were statistically significant across all observations in predicting PA, F(1, 352) = 179.28, p < 0.001

(Table 4.24). All independent variables explained 52.2% of the variability in PA for all high school athletes with an adjusted R2 of 51.9 percent, a strong size effect according to Cohen and others (2003). Sport context measures from SOFIT and interaction effects statistically significantly predicted METs, F(1, 201) = 12.17, p < 0.001.

Table 4.25 Model 8A - Summary of regression analysis for variables predicting percent time in MVPA for boy athletes during high school sports practices (N=125,218)

95% CI Parameter β SE t p Lower Upper Intercept 74.36 1.33 55.80 <0.001 71.75 76.98

Table 4.25 (continued).

Practice Context Fitness -6.58 1.92 -3.43 <0.001 -10.34 -2.83 Game Play -5.49 2.80 -1.96 0.05 -10.98 0.00 Knowledge Content -1.11 10.62 -0.11 0.92 -21.93 19.70 Management -23.42 7.18 -3.26 <0.001 -37.49 -9.35 Skill Practice -3.39 4.16 -0.82 0.42 -11.55 4.77 Other 0* . . . . . Coach Behavior Demonstrates Fitness -0.11 8.48 -0.01 0.99 -16.73 16.51 Instructs 0.09 2.92 0.03 0.98 -5.63 5.80 Manages -55.83 25.49 -2.19 0.03 -105.79 -5.87 Observes 6.56 3.09 2.12 0.03 0.50 12.61 Promotes Fitness -30.21 9.20 -3.28 <0.001 -48.25 -12.18 Other-task 0* . . . . . Sport Type Baseball -19.20 3.27 -5.88 <0.001 -25.60 -12.80 Basketball -14.89 1.88 -7.94 <0.001 -18.57 -11.22 Cross Country 14.37 4.42 3.25 <0.001 5.70 23.03 Football -40.70 3.48 -11.68 <0.001 -47.53 -33.87 Lacrosse -25.23 1.91 -13.21 <0.001 -28.97 -21.49 Soccer -3.28 1.95 -1.68 0.09 -7.10 0.54 Swimming -18.93 2.08 -9.09 <0.001 -23.01 -14.84 Tennis -35.66 7.18 -4.97 <0.001 -49.73 -21.60 Track & Field 10.38 11.23 0.93 0.36 -11.63 32.40 Wrestling 0* . . . . . *Note. Reference category.

The full GLM model for boys sports (Model 8A) was a good fit for the data and were statistically significant across all observations in predicting an active practice, F(1, 317) = 116.86,

98 p < 0.001 (Table 4.25). All independent variables explained 37.4 percent of the variability percent practice time in MVPA with an adjusted R2 of 37.1 percent, a decent size effect according to Cohen and others (2003). Sport context measures from SOFIT and interaction effects statistically significantly predicted METs, F(1, 183) = 21.28, p < 0.001.

Table 4.26 Model 8B - Summary of regression analysis for variables predicting percent practice time in MVPA for girl athletes during high school sports practices (N=125,218) 95% CI Parameter β SE t p Lower Upper Intercept 71.70 2.54 28.24 <0.001 66.73 76.68 Practice Context Fitness -66.46 3.22 -20.64 <0.001 -72.77 -60.15 Game Play -39.74 10.16 -3.91 <0.001 -59.64 -19.83 Knowledge Content 43.68 22.01 1.99 0.05 0.55 86.81 Management -50.90 4.37 -11.66 <0.001 -59.46 -42.35 Skill Practice -60.43 2.76 -21.91 <0.001 -65.84 -55.03 Other 0* . . . . . Coach Behavior Demonstrates Fitness -1.30 12.31 -0.11 0.916 -25.43 22.82 Instructs -23.58 17.22 -1.37 0.171 -57.33 10.17 Manages 16.72 17.22 0.97 0.332 -17.03 50.47 Observes -53.08 5.73 -9.27 <0.001 -64.31 -41.85 Promotes Fitness 58.16 13.67 4.26 <0.001 31.37 84.96 Other-task 0* . . . . . Sport Type Basketball -34.37 4.37 -7.87 <0.001 -42.93 -25.82 Cheerleading -28.48 2.80 -10.18 <0.001 -33.96 -22.99 Cross Country 2.32 3.15 0.74 0.461 -3.85 8.49 Lacrosse -23.78 6.22 -3.82 <0.001 -35.97 -11.59

Table 4.26 (continued).

Soccer -5.06 2.80 -1.81 0.071 -10.55 0.43 Softball -45.84 3.02 -15.19 <0.001 -51.75 -39.92 Swimming -31.62 3.08 -10.26 <0.001 -37.66 -25.58 Tennis -32.64 2.95 -11.07 <0.001 -38.42 -26.86 Track & Field -23.48 2.88 -8.16 <0.001 -29.12 -17.84 Volleyball 0* . . . . . *Note. Reference category.

The full GLM model for girls sports (Model 8B) was a good fit for the data and were statistically significant across all observations in predicting an active practice, F(1, 352) = 140.54, p < 0.001 (Table 4.26). All independent variables explained 46.2 percent of the variability in the percent of practice time spent in MVPA with an adjusted R2 of 45.8 percent, a strong size effect according to Cohen and others (2003). Sport context measures from SOFIT and interaction effects statistically significantly predicted METs, F(1, 201) = 17.82, p < 0.001.

Odds Ratios

Binomial logistic regression was run to attain odds ratios to predict PA outcomes based on sport type, practice context, and coaching behavior. A new binomial dummy variable was created for the PA categories previously coded as 1 to 5. The newly created variable codes for odds ratios converted these to 1=moderate or vigorous and 0=sedentary. In so doing, this helped substantiate and clarify results so that the research questions could be answered more accurately by not diluting the linear regression statistics. For example, the purpose of the study was not to understand the differences between athletes who are sitting versus those who are standing. On the contrary, the researcher was interested in knowing the indicators that influenced athletes to go from sedentary behaviors (categories 1 to 3) to MVPA movements (categories 4 to 5).

100

Therefore, the new coding system of 0 and 1 (dummy variables) were created for PA prior to running the binomial logistic regression.

Similarly, to attain odds ratios for the MVPA make up of practice sessions, a new variable was created. Again, a new binomial categorical dummy variable was created for the previous continuous variable of percent time in MVPA (0 to 100). The newly created variable codes converted this ratio number to 1=active practice and 0=sedentary practice. An active practice was defined by one where the proportion of the entire practice session where athletes were moderate or vigorously active was greater than or equal to 67 percent. Furthermore, A low activity or sedentary practice was operationalized as a practice session where less than 67 percent of practice time was made up of MVPA. Sports were also converted using a dummy variable where 1=sport of interest and 0=everything else. In this way, sport types were isolated and examined to help better translate the odds ratios more effectively for the specific sports and to help build the SAC for individual sport types (Appendix B).

Table 4.27 Model 9 - Ordinal logistic regression results with odds ratios for the effect of sport type on physical activity for high school athletes 95% CI Predictors β SE Wald x2 df p OR Lower Upper Sport Type 5881.54 19 <0.001 Boys Sports Baseball -0.88 0.04 414.45 1 <0.001 0.42 0.38 0.45 Basketball 0.13 0.03 13.83 1 <0.001 1.14 1.06 1.21 Cross 1.30 0.06 456.35 1 <0.001 3.66 3.25 4.12 Country Football 0.01 0.03 0.08 1 0.78 1.01 0.95 1.07 Lacrosse -0.05 0.03 2.09 1 0.15 0.95 0.89 1.02 Soccer 0.31 0.04 62.29 1 <0.001 1.36 1.26 1.47

101

Table 4.27 (continued).

Swimming -0.15 0.04 18.23 1 <0.001 0.86 0.80 0.92 Tennis -0.19 0.04 20.65 1 <0.001 0.83 0.76 0.90 Track & Field 1.24 0.06 408.26 1 <0.001 3.44 3.05 3.88 Wrestling -0.07 0.03 4.81 1 0.03 0.93 0.88 0.99 Girls Sports Basketball -0.38 0.03 145.28 1 <0.001 0.68 0.64 0.73 Cheerleading -1.48 0.04 1577.97 1 <0.001 0.23 0.21 0.24 Cross Country 0.55 0.06 98.00 1 <0.001 1.73 1.55 1.93 Lacrosse -0.34 0.04 68.89 1 <0.001 0.72 0.66 0.77 Soccer 0.27 0.04 46.38 1 <0.001 1.30 1.21 1.41 Softball -0.87 0.04 518.83 1 <0.001 0.42 0.39 0.45 Swimming -0.11 0.04 10.38 1 <0.001 0.89 0.84 0.96 Tennis -0.17 0.04 20.72 1 <0.001 0.84 0.79 0.91 Track & Field -0.42 0.04 147.48 1 <0.001 0.66 0.61 0.70 Volleyball 0.55 0.03 488.60 1 <0.001 1.74 Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is MVPA dummy coded as 1 = moderate or vigorous, and 0 = sedentary. *p < 0.05.

Table 4.27 shows the results of the first binomial regression for sport type only. The model was statistically significant, χ2(19)=6749.39, p<0.001. The model explained 7.1 percent

(Nagelkerke R2) of the variance in PA and correctly classified 63 percent of cases. The following boys sports were less likely to be engaged in MVPA compared to the reference category: baseball

(0.42), lacrosse (0.95), swimming (0.86), tennis (0.83), and wrestling (0.93). Athletes participating in these particular sports had reduced chances of being physically active compared to all other sports. The following boys sports were more likely to be engaged in MVPA compared to the reference category: cross country (3.66), basketball (1.14), soccer (1.36), and track (3.44).

Athletes participating in these sports had greater odds of being either moderately or vigorously

102 physically active. Boys football (p = 0.782) and lacrosse (p = 0.148) were not statistically significant. The following girls sports were less likely to be engaged in MVPA compared to the reference category: cheerleading (0.23), basketball (0.68), lacrosse (0.72), softball (0.42), swimming (0.89), tennis (0.84), and track (0.66). The following girls sports were more likely to be engaged in MVPA compared to the reference category: volleyball (1.74), cross country (1.73), and soccer (1.30).

Table 4.28 Model 10 - Ordinal logistic regression results with odds ratios for the effect of practice context on physical activity for high school athletes 95% CI Predictors β SE Wald x2 df p OR Lower Upper Practice Context 23274.78 5 <0.001 Fitness 2.25 0.04 3837.44 1 <0.001 9.48 8.83 10.18 Game Play 2.29 0.04 3734.82 1 <0.001 9.84 9.14 10.59 Knowledge Content -1.35 0.04 1056.10 1 <0.001 0.26 0.24 0.28 Management 0.35 0.04 87.86 1 <0.001 1.42 1.32 1.52 Skill Practice 1.81 0.03 2792.51 1 <0.001 6.08 5.69 6.50 Other -0.87 0.03 722.37 1 <0.001 0.42

Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is MVPA dummy coded as 1 = moderate or vigorous, and 0 = sedentary. *p < 0.05.

Table 4.28 shows the results of the binomial regression for practice context only. The model was statistically significant, χ2(5)=34095.07, p<0.001. The model explained 32.9 percent

(Nagelkerke R2) of the variance in PA and correctly classified 76.1 percent of cases. Game Play had 9.84 times higher odds of athletes exhibiting MVPA compared to other categories with

Fitness (9.48) and Skill Practice (6.08) likewise having positive ORs. However, athletes observed to be in Knowledge Content were 74 percent more likely to be sedentary.

103

Table 4.29 Model 11 - Ordinal logistic regression results with odds ratios for the effect of coaching behavior on physical activity for high school athletes

95% CI Predictors β SE Wald x2 df p OR Lower Upper Coach Behavior 6041.81 5 <0.001 Demonstrates Fitness 0.23 0.04 37.19 1 <0.001 1.26 1.17 1.36 Instructs -0.33 0.02 198.16 1 <0.001 0.72 0.68 0.75 Manages 0.11 0.03 13.15 1 <0.001 1.11 1.05 1.18 Observes 0.64 0.02 713.33 1 <0.001 1.89 1.81 1.98 Promotes Fitness 1.45 0.05 1021.90 1 <0.001 4.28 3.92 4.68 Other-task 0.25 0.02 136.96 1 <0.001 1.29 Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is MVPA dummy coded as 1 = moderate or vigorous, and 0 = sedentary. *p < 0.05.

A binomial logistic regression was performed to ascertain the effect of coaching behavior on the likelihood that high school athletes were moderate-to-vigorously active during sport practices (Table 4.29). The logistic regression model was statistically significant,

χ2(5)=6488.99, p<0.001. The model explained 7.1 percent (Nagelkerke R2) of the variance in

MVPA makeup of practice time and correctly classified 62.1 percent of cases. All six coaching behavior codes were statistically significant. Coaches promoting fitness (4.28), observing (1.89), or demonstrating fitness (1.26) had higher odds of being physically active than managing, instructing, or doing other tasks. When athletes were being given instruction from coaches, had lower odds of being moderate-to-vigorously active (0.72).

104

Table 4.30 Model 12 - Ordinal logistic regression results with odds ratios for the effect of sport type on percent of practice time spent in MVPA for high school athletes

95% CI Predictors β SE Wald x2 df p OR Lower Upper Sport Type 18674.80 19 <0.001 Boys Sports Baseball -2.40 0.06 1681.29 1 <0.001 0.09 0.08 0.10 Basketball -0.94 0.03 789.75 1 <0.001 0.39 0.37 0.42 Cross Country 20.79 761.89 0.001 1 0.98 1066470391.80 0.00 --- Football -0.01 0.03 0.12 1 0.73 0.99 0.93 1.05 Lacrosse -0.69 0.03 428.39 1 <0.001 0.50 0.47 0.54 Soccer 0.43 0.04 123.08 1 <0.001 1.53 1.42 1.65 Swimming -1.08 0.04 907.00 1 <0.001 0.34 0.32 0.37 Tennis -1.10 0.04 649.43 1 <0.001 0.33 0.31 0.36 Track & Field 3.04 0.12 685.38 1 <0.001 20.94 16.68 26.30 Wrestling -0.06 0.03 3.72 1 0.05 0.94 0.89 1.00 Girls Sports Basketball -2.04 0.04 3278.62 1 <0.001 0.13 0.12 0.14 Cheerleading -3.07 0.06 3002.31 1 <0.001 0.05 0.04 0.05 Cross Country 1.13 0.06 336.91 1 <0.001 3.09 2.74 3.48 Lacrosse -0.88 0.04 465.44 1 <0.001 0.42 0.38 0.45 Soccer 0.40 0.04 106.63 1 <0.001 1.49 1.38 1.61 Softball -2.31 0.05 2233.17 1 <0.001 0.10 0.09 0.11 Swimming -1.42 0.04 1545.41 1 <0.001 0.24 0.23 0.26 Tennis -1.22 0.04 1026.32 1 <0.001 0.29 0.27 0.32 Track & Field -1.87 0.04 2259.68 1 <0.001 0.15 0.14 0.17 Volleyball 0.42 0.03 284.52 1 <0.001 1.52 Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is % practice time in MVPA dummy coded as 1 = >67%, and 0 = <67%. *p < 0.05.

105

The logistic regression Model 12 was statistically significant, χ2(19)=31207.52, p<0.001.

The model explained 29.7 percent (Nagelkerke R2) of the variance in MVPA makeup of practice time based on sport type and correctly classified 71.2 percent of cases. Table 4.30 shows that the following boys sports practices had a significantly positive predictive relationship with active practices: cross country, soccer, and track and field. The odds of an athlete participating in an active practice is greater when participating in these sports. Conversely, the following boys sports practices had a significantly negative predictive relationship with active practices: baseball, basketball, lacrosse, swimming and tennis. The odds of an athlete participating in an active practice is reduced when participating in these sports. The following girls sports practices had a significantly positive predictive relationship with active practices: cross country, soccer, volleyball. The odds of an athlete participating in an active practice is greater when participating in these sports. Conversely, the following girls sports practices had a significantly negative predictive relationship with active practices: basketball, cheerleading lacrosse, softball, swimming, tennis, and track and field. The odds of an athlete participating in an active practice is reduced when participating in these sports.

The extremely large expected β found in boys cross country (OR=1066470391.80, p=0.978) was anticipated and typical for the non-normally distributed data and given the massive amount of MVPA that was present in this particular sport. For instance, out of the 21 practices recorded for boys cross country, there were no cases where the entire practice time registered less than 67 percent in MVPA. This was the only sport to not have at least a few practices that were below the 67 percent threshold for an active practice. Therefore, the output for boys cross country for the binary logistic regression predicting percent MVPA was unresolved.

106

Table 4.31 Model 13 - Ordinal logistic regression results with odds ratios for the effect of practice context on percent of practice time spent in MVPA for high school athletes

95% CI Predictors β SE Wald x2 df p OR Lower Upper Practice Context 3926.49 5 <0.001 Fitness 0.78 0.04 485.06 1 <0.001 2.19 2.04 2.35 Game Play 1.20 0.04 1101.96 1 <0.001 3.33 3.10 3.57 Knowledge Content 0.23 0.04 37.99 1 <0.001 1.26 1.17 1.35 Management -0.05 0.04 1.45 1 0.23 0.96 0.89 1.03 Skill Practice 0.73 0.03 448.30 1 <0.001 2.07 1.94 2.22 Other -0.95 0.03 828.12 1 <0.001 0.39 ------Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is % practice time in MVPA dummy coded as 1 = >67%, and 0 = <67%. *p < 0.05.

A binomial logistic regression was performed to ascertain the effect of practice context on the likelihood that a sport practice was a highly active practice (Table 4.31). The logistic regression model was statistically significant, χ2(5)=4125.57, p<0.001. The model explained 4.5 percent (Nagelkerke R2) of the variance in MVPA makeup of practice time and correctly classified 59.6 percent of cases. All but one context code (i.e., management) were statistically significant. Activities in which athletes were doing game play (3.33), fitness (2.19), and skill development (2.07) were associated with an increased likelihood of having an active practice.

Athletes in knowledge content (1.26) activities were still associated with high MVPA in practice time, but the discrepancy between knowledge content and the ‘management’ and ‘other’ categories was not as big as was derived from higher ORs indicated from fitness, game play, and skill practice.

107

Table 4.32 Model 14 - Ordinal logistic regression results with odds ratios for the effect of coaching behavior on percent of practice time spent in MVPA for high school athletes

95% CI Predictors β SE Wald x2 df p OR Lower Upper Coaching Behavior 1428.03 5 <0.001 Demonstrates Fitness 0.39 0.04 98.00 1 <0.001 1.48 1.37 1.60 Instructs 0.66 0.03 663.74 1 <0.001 1.93 1.84 2.03 Manages 0.46 0.03 218.86 1 <0.001 1.58 1.49 1.68 Observes 0.62 0.03 592.41 1 <0.001 1.86 1.77 1.96 Promotes Fitness 1.35 0.04 1275.78 1 <0.001 3.87 3.60 4.17 Other-task -0.91 0.02 1466.63 1 <0.001 0.40 Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is % practice time in MVPA dummy coded as 1 = >67%, and 0 = <67%. *p < 0.05.

A binomial logistic regression was performed to ascertain the effect of coaching behavior on the likelihood that a sport practice was a highly active practice (Table 4.32). The logistic regression model was statistically significant, χ2(5)=1499.35, p<0.001. The model explained 1.7 percent (Nagelkerke R2) of the variance in MVPA makeup of practice time and correctly classified 58.5 percent of cases. All six coaching behavior codes were statistically significant. Practice time where coaches were promoting fitness (3.87), instructing (1.93), or observing (1.86) was associated with an increased likelihood of having an active practice.

Demonstrating fitness (1.48) and management (1.58) activities from coaches were associated with a reduction in the likelihood of a sport practice being active.

Table 4.33 Model 15 - Ordinal logistic regression results with odds ratios for the effect of independent variables on PA for high school athletes

95% CI Predictors β SE Wald x2 df p OR Lower Upper

108

Table 4.33 (continued).

Coaching Behavior 476.21 5 <0.001 Demonstrates Fitness -0.05 0.05 1.08 1 0.29 0.95 0.87 1.04 Instructs 0.12 0.03 16.06 1 <0.001 1.13 1.07 1.20 Manages 0.56 0.04 243.32 1 <0.001 1.75 1.63 1.87 Observes 0.22 0.03 57.63 1 <0.001 1.25 1.18 1.32 Promotes Fitness 0.68 0.05 171.18 1 <0.001 1.98 1.79 2.20 Practice Context 19987.95 5 <0.001 Fitness 2.29 0.04 3229.60 1 <0.001 9.90 9.15 10.72 Game Play 2.42 0.04 3390.90 1 <0.001 11.22 10.34 12.17 Knowledge Content -1.36 0.05 835.95 1 <0.001 0.26 0.23 0.28 Management 0.22 0.04 27.35 1 <0.001 1.24 1.14 1.34 Skill Practice 1.99 0.04 2650.82 1 <0.001 7.30 6.76 7.87 Sport Type 6316.32 19 <0.001 Boys Sports Baseball -1.19 0.05 585.45 1 <0.001 0.31 0.28 0.34 Basketball 0.48 0.04 138.69 1 <0.001 1.61 1.49 1.75 Cross Country 1.76 0.07 573.77 1 <0.001 5.82 5.04 6.72 Football 0.03 0.04 0.80 1 0.37 1.03 0.96 1.11 Lacrosse 0.33 0.04 64.49 1 <0.001 1.39 1.28 1.50 Soccer 0.50 0.05 108.50 1 <0.001 1.64 1.50 1.80 Swimming 0.02 0.04 0.26 1 0.61 1.02 0.94 1.11 Tennis -0.31 0.05 37.64 1 <0.001 0.74 0.67 0.81 Track & Field 2.15 0.08 703.39 1 <0.001 8.62 7.35 10.11 Wrestling 0.04 0.04 1.09 1 0.30 1.04 0.97 1.12 Girls Sports Basketball 0.16 0.04 16.81 1 <0.001 1.17 1.09 1.26 Cheerleading -1.41 0.04 1079.77 1 <0.001 0.25 0.23 0.27 Cross Country 1.21 0.07 323.68 1 <0.001 3.34 2.93 3.81 Lacrosse -0.19 0.05 14.46 1 <0.001 0.83 0.75 0.91

109

Table 4.33 (continued).

Soccer 0.70 0.05 202.29 1 <0.001 2.02 1.83 2.22 Softball -0.90 0.04 414.75 1 <0.001 0.41 0.37 0.44 Swimming 0.35 0.04 69.70 1 <0.001 1.42 1.31 1.54 Tennis 0.02 0.05 0.24 1 0.62 1.02 0.94 1.12 Track & Field -0.01 0.04 0.09 1 0.77 0.99 0.91 1.07 Constant -1.22 0.05 609.44 1 <0.001 0.30 Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is MVPA dummy coded as 1 = moderate or vigorous, and 0 = sedentary. *p < 0.05.

A binomial logistic regression was performed to ascertain the effects of sport type, practice context, and coach behavior on the likelihood that high school athletes are physically active (Table 4.33). The logistic regression model was statistically significant, χ2(29)=41383.63, p<0.001. The model explained 39.4 percent (Nagelkerke R2) of the variance in MVPA of athletes and correctly classified 77.4 percent of cases. Of the three predictor variables, all were statistically significant. For practice design, game play had 11.22 times higher odds of a player being moderate-to-vigorously active than any other context. Skill practice (7.30) and fitness

(9.90) also was shown to have extremely high odds for producing MVPA behavior from athletes. knowledge content (0.26) was the only practice context that was associated with a lower likelihood of athletes being in MVPA. A coach promoting fitness had 1.98 times higher odds of observing an athlete doing MVPA than any other behavior with manages (1.75) and observing

(1.25) close behind in the high odds category. The boys sports of baseball (0.31) and tennis

(0.74) were less likely to be engaged in MVPA compared to the reference category. High school athletes within these sport types were associated with a reduction in the likelihood of exhibiting

MVPA behavior when controlling for practice context and coaching behavior. The following boys’ sports were more likely to be engaged in MVPA compared to the reference category:

110 basketball (1.61), cross country (5.82), lacrosse (1.39), soccer (1.64), and track & field (8.62).

High school athletes within these sport types were associated with an increased likelihood of being moderate-to-vigorously active. The following girls’ sports were less likely to be engaged in

MVPA compared to the reference category: cheerleading (0.25), lacrosse (0.83), and softball

(0.41). The following girls’ sports were more likely to be engaged in MVPA compared to the reference category: basketball (1.17), cross country (3.34), soccer (2.02), and swimming (1.42).

Table 4.34 Model 16 - Ordinal logistic regression results with odds ratios for the effect of independent variables on percent of practice time spent in MVPA for high school athletes 95% CI Predictors β SE Wald x2 df p OR Lower Upper Practice Context 3169.95 5 <0.001 Fitness 0.33 0.04 58.52 1 <0.001 1.39 1.28 1.51 Game Play 1.02 0.04 545.17 1 <0.001 2.79 2.56 3.04 Knowledge Content -0.14 0.05 9.32 1 0.002 0.87 0.80 0.95 Management -0.42 0.05 79.98 1 <0.001 0.66 0.60 0.72 Skill Practice 0.55 0.04 172.72 1 <0.001 1.73 1.60 1.88 Coaching 1527.73 5 <0.001 Behavior Demonstrates 0.07 0.05 1.76 1 0.184 1.07 0.97 1.18 Fitness Instructs 1.00 0.03 868.54 1 <0.001 2.72 2.55 2.91 Manages 1.00 0.04 633.00 1 <0.001 2.71 2.51 2.93 Observes 0.70 0.03 444.07 1 <0.001 2.01 1.88 2.14 Promotes Fitness 1.24 0.05 700.09 1 <0.001 3.45 3.15 3.78 Sport Type 16961.89 19 <0.001 Boys Sports Baseball -2.64 0.06 1813.46 1 <0.001 0.07 0.06 0.08 Basketball -1.05 0.04 888.49 1 <0.001 0.35 0.33 0.38 Cross Country 21.00 754.28 0.001 1 0.98 1315272557.29 0.00 ---

111

Table 4.34 (continued).

Football -0.11 0.03 11.98 1 0.001 0.89 0.84 0.95 Lacrosse -0.64 0.04 337.88 1 <0.001 0.53 0.49 0.56 Soccer 0.33 0.04 66.91 1 <0.001 1.40 1.29 1.51 Swimming -1.02 0.04 724.19 1 <0.001 0.36 0.34 0.39 Tennis -1.37 0.05 890.51 1 <0.001 0.26 0.23 0.28 Track & Field 3.09 0.12 641.24 1 <0.001 21.94 17.28 27.86 Wrestling -0.13 0.03 15.18 1 <0.001 0.88 0.82 0.94 Girls Sports Basketball -2.12 0.04 3238.02 1 <0.001 0.12 0.11 0.13 Cheerleading -3.05 0.06 2776.25 1 <0.001 0.05 0.04 0.05 Cross Country 1.52 0.07 551.72 1 <0.001 4.58 4.04 5.21 Lacrosse -.939 0.04 460.90 1 <0.001 0.39 0.36 0.43 Soccer 0.26 0.04 41.56 1 <0.001 1.30 1.20 1.41 Softball -2.36 0.05 2170.65 1 <0.001 0.09 0.09 0.10 Swimming -1.30 0.04 1159.78 1 <0.001 0.27 0.25 0.29 Tennis -1.33 0.04 1064.22 1 <0.001 0.27 0.24 0.29 Track & Field -1.72 0.04 1701.65 1 <0.001 0.18 0.16 0.19 Constant -0.71 0.05 194.72 1 <0.001 0.49 Note. SE = standard error; OR = odds ratio; CI = confidence interval. The dependent variable in this analysis is % practice time in MVPA dummy coded as 1 = >67%, and 0 = <67%. *p < 0.05.

Table 4.34 shows the results of the final binomial regression for all independent variables. The model was statistically significant, χ2(29)=35079.30, p<0.001. The model explained 34.0 percent (Nagelkerke R2) of the variance in percent MVPA of sports practices and correctly classified 72.1 percent of cases. Of the three predictor variables, all were statistically significant. Game play had 2.79 higher odds to be in an active practice any other practice context. Skill practice (1.73) and fitness (1.39) also exhibited favorable odds for characterizing an active practice. A coach promoting fitness had 3.45 times higher odds of having an active

112 sports practice than any other behavior. The following boys sports had negative odds ratios: baseball (0.07), basketball (0.35), football (0.89), lacrosse (0.53), swimming (0.36), tennis (0.26), and wrestling (0.88). Practices within these sport types had higher chances of being a sedentary practice when controlling for practice context and coaching behavior. The following boys sports were more likely to be engaged in MVPA compared to the reference category: cross country, soccer (1.40), and track (21.94). Practices within these sport types had higher chances of being moderate to highly active practices. The following girls’ sports were less likely to be engaged in MVPA compared to the reference category: cheerleading (0.05), basketball (0.12), lacrosse (0.39), softball (0.09), swimming (0.27), tennis (0.27), and track (0.18). The following girls sports were more likely to be engaged in MVPA compared to the reference category: cross country (4.58), and soccer (1.30).

The extremely large expected β found in boys cross country (OR=1315272557.29, p=0.978) was anticipated and typical for the non-normally distributed data and given the massive amount of MVPA that was present in this particular sport. For instance, out of the 21 practices recorded for boys cross country, there were no cases where the entire practice time registered less than 67 percent in MVPA. This was the only sport to not have at least a few practices that were below the 67 percent threshold for an active practice. Therefore, the output for boys cross country for the final binary logistic regression predicting percent MVPA was unresolved.

Summary

The findings from 125,218 observations in 598 high school sports practices yielded significant results across all twenty sports. The purpose of running a multitude of different regression models was to find the best model that predicts PA behavior in high school sports practices. Participation in sports practices has substantial contribution to PA behavior of

113 athletes. While observing the 2,408 athletes included in this study, athletes were vigorously active 43 percent of the time, moderately active 17 percent, and sedentary 40 percent of the time.

However, the amount of PA accrued during this study varied based on the type of sport being played (χ2=11420.5). Higher PA rates were found in boy sports when compared to girl sports.

The boys’ sports that had the most active practices were track & field (87%), cross country

(86%), and soccer (70%). The girls’ sports that had the most active practices were cross country

(75%), soccer (71%) and volleyball (63%). The boys sports that had the lowest rates of practice time in MVPA were baseball (43%), tennis (59%), and swimming (61%). The girls’ sports that had the lowest rates of practice time in MVPA were cheerleading (28%), softball (45%), and track & field (54%). The amount of PA accrued can vary based on the type of sport being played but type of sport is not a significant predictor of PA behavior.

The following were the odds ratios (ORs) for MVPA in male athletes gathered from the final analyses for each boys sport independently (Table 4.27) and in the full model (Table 4.33): cross country (3.66 and 5.82), track & field (3.05 and 8.62), soccer (1.36 and 1.64), basketball

(1.14 and 1.61), football (1.01 and 1.03), lacrosse (0.95 and 1.39), wrestling (0.93 and 1.04), swimming (0.86 and 1.02), tennis (0.83 and 0.74), baseball (0.42, and 0.31). Again, an OR greater

“1” indicates a higher likelihood of being physically active and an OR less than “1” specifies a sport with a lower chance of being physically active, in other words, sedentary. The following were the odds ratios (ORs) for MVPA in female athletes gathered from the final analyses for each girls sport independently (Table 4.27) and in the full model (Table 4.33): cross country

(1.73 and 3.34), soccer (1.30 and 2.02), swimming (0.89 and 1.42), tennis (0.84 and 1.02), lacrosse

(0.72 and 0.83), volleyball (1.74 and 1.00), basketball (0.68 and 1.17), track & field (0.66 and

0.99), softball (0.42 and 0.41), and cheerleading (0.23 and 0.25).

114

The most significant predictor of PA observed was the practice context measure. The amount of PA accrued varied considerably based on the practice context (χ2=42560.39).

Practices that were organized around fitness (4.54 average METs), game play (4.67), and skill practice had undoubtedly the highest levels of PA (Table 4.15 and Figure 4.13). Athletes in knowledge content (1.71) and management (2.16) had low levels of PA. Skill practice was the most frequent sport practice activity observed, making up 37 percent of all practice time. The following were the ORs for athletes in MVPA for the practice context variable (Table 4.31 and

Table 4.34): fitness (9.48 and 9.90), game play (9.84 and 11.22), skill practice (6.08 and 7.30), knowledge content (0.26 and 0.26), management (1.42 and 1.24), other (0.42 and 0.30).

Coach behavior was also significant predictor of PA, however, did not explain much of the variability in the PA observed during practices. The amount of PA accrued varied based on coaching behavior (χ2=9242.63). Practices where the coach was either promoting fitness (4.83) or observing (2.02) had the highest levels of PA (Table 4.17 and Figure 4.16). Athletes were the least active when the coach was doing instruction (3.30) or management (3.17) activities, but still moderately active nevertheless. Seventy-eight percent of all practices were made up of observing

(42 percent) and instructing (36 percent) coaching behaviors.

The boys’ sports that had the highest amounts of coach doing fitness activities or observing were basketball (56%), track & field (63%), football (56%), lacrosse (42%), and swimming (48%), and cross country (54%). The girls’ sports that had the lowest amounts of coach observing were volleyball (64%), lacrosse (56%), cross country (56%), swimming (55%), and cheerleading (50%). The following were the ORs for athletes in MVPA for the coaching behavior variable (Table 4.32 and Table 4.36): promotes fitness (4.28 and 1.98), observes (1.89 and 1.25), demonstrates fitness (1.26 and 0.95), manages (1.11 and 1.75), and instructs (0.72 and

1.13).

115

The full regression model showed that the context of sport participation (i.e., all independent variables gathered by SOFIT) accounts for approximately half of the variance in

PA. Specifically, the sport context explained 41.6 percent of the PA behavior for boy athletes observed and 52.2 percent of the PA behavior for girl athletes observed. In terms of practices themselves, the sport context explained 37.4 percent of having an active practice or not for boys and 46.2 percent for girls. Overall, practices that were organized around fitness, game play, and skill practice had the highest levels of PA and had the highest odds of facilitating moderate-to- vigorously active high school athletes and active practices in general. Practices where the coach was either promoting fitness or observing had the highest levels of PA whereas coaching behavior reflecting instruction, management, or other tasks resulted in sedentary practice activities.

116

DISCUSSION

Health advocates at the federal and state level recommend that adolescents participate in a minimum of 60 minutes of PA each day (CDC, 2015; USDHHS, 2018). Moreover, previous research and current reports suggest that high school aged students are falling short of reaching this recommendation (YBRSS, 2017) as physical education classes do not provide students with enough activity to meet the recommended MVPA levels (Cardon et al., 2004; Friedman, Belsky,

& Booth, 2003; Scruggs, 2007b). Being physically active in childhood can lead to a healthier adulthood (USDHHS, 2018), and research suggests that playing sports during childhood and adolescence can translate into being physically active in adulthood, either through sports or other activities (Perkins et al., 2004; Pfeiffer et al., 2006; Tammelin et al., 2003; Taylor et al.,

1999; Telama et al., 2006).

With this in mind, many researchers have called to discover different opportunities for youths in this age group to become physically active on a daily basis (Schwamberger &

Sinelnikov, 2015) with the possibility of modifying environmental factors to increase PA production (Schlechter et al., 2017). The SEM framework provides guidance in this area stating that as more levels of a person’s environment is addressed, the greater impact can be expected to changing human behavior. Increasing intensity levels during organized sports is speculated to being an effective PA intervention strategy if evidence could show how this could be done through managing the contexts surrounding sport participation (Ridley et al., 2018).

The primary purpose of this study was to examine PA behavior of high school athletes and to determine the impact of interscholastic sport participation and its contexts. More specifically, this study sought to determine the association between PA output observed by sport participants and the practice activities and coaching actions within ten different sports for both

117 male and female athletes. The research aims were to (a) provide insight into the differences in

PA between particular sports, (b) determine what contextual factors are most influential for

MVPA during sport practices, and (c) determine what coaching strategies are most influential for

MVPA during sports practices. In this chapter, I will discuss the findings of the systematic observations of North Carolina high school athletes during the 2017-2018 academic year at twelve public high schools. As part of this chapter, I will also discuss the significance and implications of the findings, and make recommendations for future research endeavors and offer practical advice for constituents responsible for the delivery of interscholastic sport to high school students.

Findings in the Present Study

This study is the first to provide a detailed account of a large cross-section of high school sport practices using SOFIT, an observational tool that simultaneously examines athlete

PA and contextual information. Over 125,000 observations were collected and analyzed across

76 different sports teams within the ten sport types for boys and girls playing on varsity high school sport teams amounting to 696 total hours of practice time. Athletes experienced high levels of PA overall, yet, the results showed that PA production is not automatic with mere participation but depends on a variety of factors. As an example, the amount of PA accrued can vary based on the sport and the gender of the athlete. It was intuitive to find that the sport forms that are running centric outperformed those that involved intermittent bursts of play and

PA. However, the middle 14 sports differed by such small margins that conclusions of PA based on sport type need to be tapered to some extent. Male athletes were more active than female athletes, which is not surprising based on the literature available on gender differences.

118

The most powerful determination for how much MVPA an athlete produced and whether or not a practice was considered highly active was the practice context component, at least that was the case for this study. Undeniably, knowing what the practice context was during each observation and sport practice session, was more meaningful for determining the PA output of athletes than any other sport context variable collected. Surprisingly, the behavior shown by coaches, while statistically significant, was not an overwhelming influencer for PA behavior as previously hypothesized. But the context surrounding sport participation (i.e., all independent variables) accounted for a third to nearly half of the variance found in PA levels exhibited by athletes, which is a noteworthy finding.

The discussion laid out in the following pages explores these overarching results organized by each independent variable and how this study contributes to our current knowledge for how sport participation and its accompanying contexts are associated with PA production during high school sport practices and what that means in terms of both theory and practice.

The relationship between physical activity and sports

Observation results revealed that participation in high school sports practices has substantial contribution to PA behavior of athletes overall. The average practice length for all sports was 80 minutes which is similar to the only other systematic observational study that reported practice length (Curtner-Smith et al., 2007) but fluctuated by sport types. Practice length discussed by the Curtner-Smith et al. (2007) study reported practices lasting 91 minutes in the early-season, 89 minutes in the middle season, and 72 minutes in the late-season for the sport of basketball. In this study, basketball practices were among the longest at 99 minutes for boys and 100 minutes for girls.

119

Organized youth sports participation has been claimed to be positively associated with an increased likelihood of complying with national PA guidelines (Vella et al., 2013); which recommend accumulating at least one hour of MVPA (USDHHS, 2018). These findings support this claim but with one caveat, the PA behavior exhibited by athletes is dependent on the type of sport being played. Even practices with high proportions of MVPA, did not last long enough to meet the 60 minute threshold (e.g., girls soccer at 58 minute average practice length and 71 percent MVPA practices).

As expanded upon later, sports participation can be viewed as complimentary to helping adolescents achieve adequate MVPA levels, but not the sole source. Research in PE classes through the years has garnered some of the same insight: that attaining 60 minutes of MVPA in a single session, whether it be sport, PE or otherwise, is improbable. Recently, Ridley and colleagues (2018) found that high school aged athletes only received 32 minutes of MVPA during a sample of organized sports practices. Therefore, it is more realistic to take a holistic view of a child or adolescent’s day to find that other domains where PA can occur (e.g., transportation to school, household chores, leisure-time) are also playing a role. So even though an athlete may only achieve twenty minutes of MVPA in a high school sports practice, the same athlete could still be meeting PA guidelines by being physically active elsewhere.

It was not surprising to find sports such as cross country, track and field, soccer, and basketball with the greatest odds of participants being active since those particular sports are inherently running oriented. Similarly, baseball, football, and tennis were anticipated to have low odds of being active since these particular sports are characterized by playing in intervals rather than a continuous model of playing. Meaning, baseball participants stop activity in between every pitch; football stops and starts after every play; and with tennis, there is ample sedentary time in between sets and services.

120

What was most notable was the low odds in the girls track and field practices compared to the boys track and field practices. Even though the sport types are essentially the same, this can occur due to differing styles of coaches and team practices. Also, since the sport of track and field is full of specialized athletes (e.g., vault, discus, 400-meter run), when randomly selecting athletes, there can be a great disparity from one athlete to another. It was also easy to predict that softball and cheerleading were the lowest on the OR spectrum since those sports are typically littered with sedentary activity during both games and practices. It is certainly not the intention of this study to shed negative light on those low-activity producing sports. As later mentioned in the discussion on practice context, all sports can be designed and coached in a manner where high levels of MVPA occur, no matter the sport type.

USDHHS (2010) recommend that 50 percent of PE class time should engage students in

MVPA. Although not in the physical education setting, results from this study showed that some sports consistently exceeded this threshold (e.g., lacrosse, swimming, and cross country) while others fell short (e.g., baseball, softball, and cheerleading). Given the degree to which PE is marginalized in America (Cone, 2004), it seems sensible to search for multiple contexts in which the health of youths can also be enhanced by participating in PA for appropriate time periods and at appropriate levels of intensity. The amount of time spent in MVPA has been shown to vary by the type of sport, gender of the player, style of the coach, level of competition, and whether it is a practice or a game (Katzmarzyk & Malina, 1998; Guagliano et al., 2013; Leek et al., 2011; Sacheck et al., 2011; Smith et al., 2016; Wickel & Eisenmann, 2007). Some sports, such as soccer or basketball, may include more physically active time than other sports, such as baseball or softball. However, many sports can be modified to incorporate more PA, either during practices or games, by altering the number of youths on the field or court or by adjusting drills to limit standing time as reported in another practice design study (Kanters et al., 2015).

121

It is widely accepted that childhood PA through sport contributes significantly to the quality of life of youth when you sum up all of the associated benefits. A large number of studies have examined the relationship between sports participation and PA. For this population, high school sports seemed to make a substantial contribution to the recommended levels of PA of participating athletes. These findings are comparable with the findings of earlier studies. Sacheck et al. (2011) found 33 percent of soccer games were spent in MVPA, whereas

Leek et al. (2011) examined PA levels during soccer and baseball/softball practices and found 46 percent of practice time in MVPA across sports. Other studies have looked at PA during sport time specifically to determine the extent to which participation contributes to meeting MVPA targets. One in particular found that both boys and girls engaged in approximately 45 minutes of MVPA during youth sport practice (Leek et al., 2011). The findings also showed that 23 percent of players met the recommended 60 minutes of MVPA per day during youth sport time.

However, lower levels were found in another study (Wickel & Eisenmann, 2007), although their findings are similar to those from the studies reported above in terms of proportion of time active (49 percent). Indeed, some studies have found little or no evidence for differences between team sports and individual exercise. For example, McGale et al. (2011) compared team sport intervention and exercise intervention and found no noteworthy differences between intervention groups. But overall, these results support the general body of evidence that points to a positive effect of team sports on the physical health of youth.

Wickel and Eisenmann (2007) found average weekday MVPA to be 30 minutes lower on a non-sports day compared to a sport day for adolescents playing a multitude of sports. Similar results have been reported internationally (Machado-Rodrigues et al., 2012; Van Hoye et al.,

2013). These studies have also found the amounts of MVPA accrued during sport participation did not occur during non-playing days and was generally replaced with low-intensity and

122 sedentary activities. This is consistent with other research indicating that children are unlikely to compensate for missed PA resulting from sports (Dale et al., 2000), suggesting that sport has the potential to increase levels of PA, and also be effective in reducing bouts of physical inactivity

(Kanters et al., 2015).

School and non-school sport, both formal (organized) and informal, are major contexts of PA for youths, while many youths perceive PA as equivalent with sport. Participation in sport has the potential to contribute to the objectives of PA, but formal and informal sport, and specific sports vary in intensity of activity. Sport activities can include warm-up, instruction and practice (drills, repetitions), rest, scrimmages, games, among others. Team sports such as soccer, basketball, and swimming involve reasonably continuous activity that varies in intensity during practices or games, whereas baseball and football involve intermittent activities among frequent periods of relative inactivity. Intermittent activities are also characteristic of softball and cheerleading and some events for track and field, while continuous activity is a more likely feature of the prototypical running sports.

Another notable finding was that the type of sport athletes played, alone, was not a significant predictor of PA behavior but was for the amount of MVPA in an entire practice session. Model 1 reported an R2 a paltry 6 percent for sport type predicting METs but Model 2 reported a much higher R2 of 26 percent for predicting MVPA at the practice level. But controlling for coach behavior, practice context and sport type combined to predict 42 percent

(Model 7) and 34 percent (Model 8) of the variability in PA and practice time MVPA, respectively. The binomial regression models affirmed these deductions for the type of sport explained a mere 7 percent of the variability in PA when the five PA codes were recategorized as being either sedentary or moderate-to-vigorous activity (Table 4.30) and 30 percent of the

123 variability for percent practice time in MVPA after recategorizing for active and nonactive practices (Table 4.33).

These findings indicate that although sports alone does not provide amounts of PA sufficient to meet daily recommendations, it does provide an ideal opportunity for high school athletes to be physically active and to contribute to daily MVPA by consistently participating.

Furthermore, evidence indicates that adolescents who participate in sports are more active than those who do not and are more likely to meet recommended PA guidelines, though they may not achieve maximum amounts of PA during the sport session (Nelson et al., 2011; Silva et al.,

2010). So despite some sports providing higher levels of PA on a regular basis than others, there is still value in concluding that any sport participation is better than no sport participation at all.

One of the original impetuses of this study was to confront the assertion that sport is good for us. It has long been touted that participating in sport during adolescence is a healthy endeavor, despite some evidence debunking this assumption. Based in a functionalistic viewpoint, sport and PA can be seen as positively contributing to society and benefiting participants through the development of psychological, social, and physical well-being. So while some (including myself) acknowledge the potential dangers and detriments associated with playing sports at a young age (e.g., injuries, burnout, financial/time burdens), I believe the benefits still outweigh the costs. But debating the PA value that sport adds to adolescents is fair.

Practice design as an influential factor in physical activity

PA does not occur in a vacuum; it is a multi-dimensional behavior that occurs in multiple contexts (McKenzie, 2016). Although activity is an important avenue for learning, physical and psychological development, enjoyment, and social interactions and self- understanding, it is currently viewed most often in terms of potential health benefits associated

124 with regular energy expenditure in MVPA. Proficiency in a variety of movement skills and physical fitness is important dimensions of activity. Context is an additional dimension of PA that is often overlooked (i.e., types, settings, purposes of activity) which can be done through play, PE sessions, exercise, and extracurricular sport.

One of the most important discoveries was that practice design alone was a significant predictor of PA levels observed in athletes but was not for the amount of MVPA in a practice.

Model 3 reported an R2 of 31 percent for practice context predicting METs but Model 4 reported a much lower R2 of only three percent for predicting MVPA at the practice level across all sports. The binomial regression models for the practice context explained 33 percent of the variability in PA when the five PA codes were recategorized as being either sedentary or MVPA

(Table 4.31) and five percent of the variability for percent practice time in MVPA after recategorizing for active and nonactive practices (Table 4.34). The majority of practice sessions were focused on improving athlete skill levels through explicit skill practice (37 percent), and game play (12 percent). This is in contrast with Dudley and colleagues’ (2012) research with their participants spending only five percent of available time in skill instruction and nearly double the time spent in game play (44 percent).

The major takeaway from the analyses of the practice context factor was that even when low MVPA coaching behaviors (e.g., instructs, manages, other-task) were present, athletes were still showing strong PA behavior due to the activities being designed and executed during practices. To illustrate, when looking at the interaction effects, all coaching actions had significant high MVPA rates among athletes when the practice context was also in game play.

The same was true for skill practice; all coaching behaviors were positively associated with PA rates when coupled with skill practice. Vice versa, athletes who were in knowledge content were almost always sedentary, no matter what coaching behavior was observed simultaneously.

125

Perhaps this is why the coaching behavior variable was not as powerful of an agent to sustaining high levels of PA behavior (discussed later on).

One of the more profound findings was how fitness and game play were so close with one another in the PA production of high school athletes. It is obvious to think that any athlete doing fitness activities would be tremendously physically active, but game play was not as intuitive. I think this points toward the need for more practice time devoted to the application of skills in a game or competitive setting. As previously presented, one of the characteristics of the game play code was sport activity that was ‘generally performed without major intervention from the coach’ that mimics a real game/match/meet. Hence, when athletes were scrimmaging in a game-like simulation, with little contact with the coach, they were nine times more likely to be physically active compared than any other practice context code. This affirms some of the literature that mentions the need for more “free play” in both sport and recreation. It also supports the notion that coaches can be a little more hands-off when it comes to certain practice designs, allowing players to not feel overanalyzed or critiqued while trying to perform in their sport.

An interesting note on practice design found in the literature was the discussion on the built environment as an influencer in how practices are managed and subsequently, how active participants are during practice time. In Warburton and Woods’ study (1996), the lack of activity during swimming practice was partly explained by the small size of the pool. This brings about an important question when addressing the delivery of sport in the school setting: Is there enough space for coaches to support large bouts of MVPA during practice time? Theoretically, a coach with more players in a small space, would yield less PA levels in athletes than a coach with fewer players on a bigger playing area, regardless of the practice context.

126

An effort needs to be made to use support materials appropriately, use space more efficiently, and structure practices appropriately in order to increase PA levels for high school athletes. Increasing child involvement by better use of space, active instruction behavior, and properly structured practices, including playing with small teams and avoiding that require athletes to stand in line and “wait their turn” to perform a skill drill, is needed to increase PA engagement levels (Cardon et al., 2004). But it is understandable that some coaches, particularly in the public school system, do not have the luxury of having state of the art equipment or ample space for various stations to be established that would maximize PA behavior of athletes.

Sadly, this was not an element investigated in this research but deserves to be mentioned as a possible input variable that could determine how active practices can be achieved.

While a limited body of literature has examined youth sport program designs, the few that exists confirms that the context or design of youths’ sport programs can play a significant role in the sport outcomes experienced (Fraser-Thomas et al., 2005; Guagliano et al., 2014; 2015;

Kanters et al., 2015; Leek et al., 2011; Ridley et al., 2018; Sacheck et al., 2011; Schlechter et al.,

2018). Of the few studies that have examined PA in sports, one study found that during soccer games, children were in MVPA for 33 percent of the time (Sacheck et al., 2011), and another found that approximately 46 percent of practice time was spent in MVPA (Leek et al., 2011).

However, children also spend high percentages of time inactive or in light PA (Katzmarzyk et al., 2001; Leek et al., 2011). Recent research has supported that high school athletes spend less than 50 percent of practice time engaged in MVPA during sports practices (Schlechter et al.,

2017). Although social influencers, organizational capacities, and program features are known to influence the efficacy of youth sport programs to facilitate healthy outcomes (Doherty &

Cousens, 2013), the empirical research to back up these claims is limited. As a result, there is a

127 large body of literature linking sport with a variety of physical, social, and psychological outcomes but little insight on the practical implications for understanding this process.

The many benefits of sports are much more likely to occur when quality, intentional programming is provided (President’s Council on Sports, Fitness & Nutrition, 2019). Quality sports programs create a focus on individual and team development through skill building while also recognizing the value of a well-balanced practice that includes sedentary time. This programming takes into account the age and stage of the athlete to facilitate the development of physical literacy and sports skills. Sports-based youth development programs may provide youth with a substantial amount of PA (Perkins & Borden, 2006). In fact, the newly released NYSS report (USDHHS, 2019) focuses on the benefits of sports participation and strategies to create quality sports programming that will increase the likelihood of youths obtaining these benefits.

I also want to point out that although massive amounts of sedentary activity were found during knowledge content and management activities, these components of a practice are not necessarily negative when considering the holistic development of an athlete. Physical inactivity or sedentary behavior is independent of PA and is also multi-dimensional. Although the biomedical view of PA is critical of excessive inactivity, many sedentary behaviors are highly valued such as school, studying, or reading and have benefits for young people. PA and inactivity are independent behaviors performed in a societal context, and both have high value in a sport practice. There is a need for better understanding of the meanings attached to and motivations for PA and sedentary behaviors by adolescents, particularly at the high school age.

Most high schools (91 percent) offer students opportunities to participate in at least one interscholastic sport per year (Lee et al., 2007).

128

The current study indicates that athletes were generally very active during practices, however, some sports (e.g., baseball, cheerleading) could be enhanced by implementing strategies to increase PA during all practice contexts. For example, splitting the team into different subgroups or establishing stations for the same activity could reduce sedentary time spent standing in line to participate (Kanters et al., 2015). Also, coaches could minimize time spent doing management tasks (e.g., transition periods, equipment setup). Providing coaches with more support, education, and strategies on how to design practices to maximize PA outcomes has been recommended (Schlechter et al., 2018) and are highly encouraged by the current study. Thus, there is clearly an opportunity to optimize MVPA levels and reduce inactivity and light PA during sport practices by modifying the activities themselves as well as the logistics of how to maximize the playing space.

The importance of the coach

Past sport practice studies concluded that adolescents spend large proportions of time during organized youth sports inactive or in light PA, citing coach behavior as the primary deterrent (Gualiano et al., 2013; Leek et al., 2011; Sacheck et al., 2011). Furthermore, using direct observation, Guagliano and colleagues (2013) observed that coaches spent a considerable proportion of training time managing and instructing their players, time when youths would be relatively inactive.

It was interesting to see how influential coach ‘observation’ turned out to be on the PA of athletes. When observing was taking place during a practice session, players were twice as likely to be moderate-to-vigorously active. I think this speaks to the notion that players respond positively when they know their coach is paying attention, and therefore, put forward their best effort on the field or court. In fact, some of the highest MVPA levels were found when observing was taking place with game play and skill practice simultaneously. It was when the

129 coach lost focus on the practice or athlete (e.g., managerial tasks, being distracted, showing absenteeism during practice time, etc.) that athletes were the least likely to be physically active.

Hence, coaches who were present and caring had the most active practices, regardless of the practice activities being conducted. This is known as the Hawthorne effect where subject reactivity can occur when participants know they are being watched (McCambridge et al., 2014).

Hawthorne effect studies were primarily based on managerial theories to increase worker productivity and its translation into a coaching strategy has not been evaluated to the level as perhaps it should have.

However, coach behavior alone was not a significant predictor of PA levels observed in athletes or percent of practice time in MVPA; going against one of my original hypotheses.

Model 5 reported an R2 of 5 percent for coaching behavior predicting and Model 6 reported an

R2 of only 2 percent for predicting MVPA at the practice level. The binomial regression models for the coach behavior explained 7 percent of the variability in PA when the five PA codes were recategorized as being either sedentary or moderate-to-vigorous activity (Table 4.32) and 2 percent of the variability for percent practice time in MVPA after recategorizing for active and nonactive practices (Table 4.35). But as discussed, controlling for practice context, coach behavior and sport type combined to predict 17 percent (Model 7) and 33 percent (Model 8) of the variability in PA and practice time MVPA, respectively. So taking coaching behavior within specific sports into consideration, more accurate calculations on the PA of participants were seen.

Why, exactly, some sporting experiences are positive and others negative is not well- understood. However, some empirical studies have identified some clues. The review by

Allender et al. (2006) reported that those who participated from youth to adulthood tended to recall the importance of positive influences at school from coaches in becoming and staying

130 physically active, while overly competitive classes, a lack of coach support, negative experiences at school, and peer pressure and identity conflict were major barriers to PA. McPherson, Atkins,

Cameron, et al.’s (2015) retrospective study of children’s experience of sport found that while their adult participants reported the lasting developmental benefits of participation in organized sport as children, more than half also reported negative experiences, including emotional and physical harm and sexual harassment.

In the U.S., sport is often not considered as an extension of the education curriculum, as it is in other countries, particularly in Europe. Moreover, the predominant model of American youth sport is highly conservative and focused on winning in a narrow range of team games.

Indeed, teachers are often hired because of their ability to coach a sport rather than their teaching acumen (Curtner-Smith et al., 2007). Additionally, the goals of school sport are not particularly educational. While many coaches claim that they have personal goals for their athletes’ physical health, the reality is that they are generally more concerned with providing entertainment for the local community, sending their players on to play at the college level, and that they and their teams are used as a means of increasing school funds. Nevertheless, in recent years the increased attention on youth sedentary activity and the obesity epidemic has fueled coaches to make unsubstantiated claims about the health benefits of interscholastic sport, particularly that children and youth who participate in it are physically active.

It is also likely that athletes would be relatively inactive while in management and knowledge delivery. This has recently been exhibited in a PE setting; the authors found a significant negative correlation between MVPA and time spent in management and knowledge delivery (Dudley et al., 2012). O’Connor and Cotton (2009) identified that junior coaches spend large amounts of time instructing (48 percent) and managing (13 percent) players during training sessions, often at the expense of activity time, with players being inactive for 36 percent of a

131 training session. They also revealed youth coaches consistently overestimate the amount of time they spend during a coaching session on skill development and giving instructions to athletes.

Therefore, decreasing the percentage of time coaches spend in management and knowledge delivery may be a strategy to consider in helping create an environment that provides the most opportunity for PA.

Training helps drive quality in programs and aims to increase positive cognitive, physical, and social development of youth. Quality training equips coaches with the skills they need to confidently deliver youth sports programming that includes PA behavior. Training can include healthy competition and good sporting behavior (game play), leading a practice session (coach behavior), sport skills development (skill practice), and strength and conditioning principles

(fitness).

Gender differences

Observations revealed that boy athletes were found to be more physically active than girl athletes. In terms of MET values per sport type, regardless of gender, seven of the top ten sports with the highest average METs were boys sports in addition to girls sports accounting for five of the bottom six physically active sports (Table 4.13). Even in the sport types that had gender counterparts, the boys designation (i.e., cross country, 4.84 average METs; lacrosse, 3.82; track & field, 4.62; swimming, 4.11; soccer, 4.05; basketball, 3.71; tennis, 3.61; baseball, 3.02) had higher rates of MVPA than the girls equivalent (i.e., cross country, 4.35; lacrosse, 3.53; track & field, 3.24; swimming, 4.11; soccer, 4.23; basketball, 3.51; tennis, 3.46; softball, 2.76) for the most part. As research shows, girls are less physically active than boys (Troiano et al., 2008), with the most pronounced declines in PA participation observed in adolescence.

132

Gender inequality is a persistent and resistant concern, magnified by the relatively low levels of general PA among girls. A national report noted that significant differences in rates of participation in organized sport between boys (67 percent) and girls (54 percent) (ABS, 2012), and similar patterns have been found elsewhere. For example, Bailey et al. (2004) found consistently low levels of PA among females, both relative to males, and to PA recommendations. This and other studies report a clear trend of decreasing levels of activity as girls get older, and a widening disparity between girls’ and boys’ PA behaviors. For example, one study estimated that the decline in PA during high school is 7.4 percent for girls, compared with 2.7 percent for boys (Sallis, 1993). Another suggested that female adolescents were approximately 20 percent less active than their male peers (Booth, 1997). The picture for sports participation is more positive in some (but not all) developed countries. In the U.S., there has been a significant increase in the number of girls participating in sports over recent decades

(Coakley, 2011). However, there are still many girls who do not participate because of limited opportunities, structural barriers, and gender stereotypes, and their participation do not match those of boys (Bailey et al., 2004).

Boys are, on average, more active and less sedentary than girls, though the gender difference is weakened when maturity status is controlled in adolescents (Machado-Rodrigues et al., 2010; Thompson et al., 2003), that is, girls tend to be less active and more sedentary than boys of the same age due to their advanced biological maturation. Nevertheless, age trends in activity vary among studies and with the measurement instrument and context (Curtner-Smith et al., 2007; Harwood et al., 2015; Lubans et al., 2015). Studies have found that girls accumulate less PA than boys throughout childhood (Troiano et al., 2008), and their participation in PA drops more steeply than boys during the transition into adolescence (Kim et al., 2002). As such, girls have been identified as a high priority group for PA promotion (Biddle et al., 2014).

133

From an SEM perspective in the sport setting, the individual and biological characteristics of an individual can actually negatively influence girls’ PA production. Coaches can be guilty of unfairly stereotyping boys as being competition driven and girls as socially driven. Sport sociologists have called this gender role stereotyping and have pointed out that coaches and parents are more likely to push boys to winning and performance standards than girls (Coakley, 2011). On the other hand, it is acceptable (and sometimes expected) to allow girls of the same age a more relaxed atmosphere with an emphasis on having fun and being with friends. The truth is, both boys and girls have identical needs in terms of the psychological, social, and physical benefits that sport can provide (Litt et al., 2011). Are coaches unintentionally framing their practices differently based on the gender of their athletes? It is an interesting topic that deserves further exploration.

A Compendium of Sports

There also exists a need for parents to be more adequately educated on what they and their child athlete receives in return, especially with the ever-rising costs of sport participation through popular pay-for-play philosophies. Many people view sport as an event or experience and rightly so if the context is pure spectatorship. But sport can be argued as being much more of a “product” or something to be purchased, particularly at the grassroots and community level.

Public funding for youth sport programs has reduced drastically in the last few years (Durlak,

Weissberg, & Pachan, 2010). Rather than directing money through public agencies like local parks and recreation departments, governments are increasingly relying on a variety of nonprofit and for-profit organizations to deliver youth sport in an increasingly privatized ‘public’ goods market (Isett et al., 2011). Indeed, as more and more families are being pushed out of having the ability to affordably register their child in more than one youth sport a year, parents are more likely than ever to view youth sports in terms of maximizing their return on investment (ROI).

134

Youth sports are now taking the shape of a commodity – a consumable product available for purchase. The purchase decision-making model illustrates that the decision process involves five steps: need recognition, information search, evaluation of alternatives, purchase, and post- purchase evaluation (de Mooij, 2004). It is logical to believe that policymakers, organizers, parents, and coaches revolved around youth sports would benefit from more cross-sectional research of this nature to aid in the evaluation of different sports and varying practice designs based on specific factors of interest.

An unintended outcome of evaluating various sports based on certain criteria was the creation of a data-driven compendium based on PA outcomes (Appendix B). Compendiums have been useful for practitioners in many different facets of health where communities and other constituents can offer easily digestible guides to help decision-makers (Ainsworth et al.,

2011; 2018; American College of Sports Medicine, 2019; Bocarro et al., 2014; Butte et al., 2018;

Hermann et al., 2016; Pfeiffer et al., 2018; Ridley et al., 2008; Ridley, 2018). In 2008, a compendium was published that presented MET values for 244 activities for youth (Ridley et al.,

2008). This report used MET values from a literature review of youth-specific energy costs where only 35 percent of values were taken from empirical studies and the remaining data were taken from the adult compendium (Ainsworth et al., 2011). Therefore, the updated Youth

Compendium of Physical Activities (Ainsworth et al., 2018) includes 16 activity categories with corresponding MET values for children aged 6 to 18 years old. However, most of the compendiums/indexes such as this one fail to account for the context surrounding these particular activities.

The present research is an offshoot of a larger project termed the Healthy Sport Index

(HSI) funded by the Aspen Institute as part of their Project Play initiative. The central objective of the HSI is to coordinate and design an index that systematically ranks the most commonly

135 offered high school sports based on three measures of health as a result of participation: PA, injuries, and psychosocial indicators. Since my own research only accounts for the PA component of the HSI project, I did not feel it appropriate to use the HSI moniker to describe the current study. This was especially true when taking into consideration that the term “health” encompasses a much larger number of factors than just those measured in the HSI and I do not wish to mislead readers. Alternatively, I am proposing this particular index to be tentatively referred to as the Sport Action Compendium (SAC), which more aptly characterizes the measures being used in this dissertation: PA, practice design, coaching instruction behavior demonstrated by coaches. All of these factors are behavior-driven and actionable items within the context of the athlete-coach relationship in sports.

Despite their importance in U.S. culture, youth sports have never been governed or regulated at the national level. There is no central ministry of sport or public agency that controls sports on a state-by-state basis. There has never been a national model of youth sports nor have there been institutional structures through which such a model could be developed and implemented. The goals of youth sport programs vary with the specific interests and orientations of the adults and organizations that sponsor them. Because each of these sponsors has a different mission, the sports programs they fund are likely to offer different types of experiences for children and families. This makes it difficult to draw general conclusions about what happens in organized programs and how participation affects child development, family dynamics, community integration and public health. The SAC provides the results of this study for each specific sport rather than lumping all sports together in one pot so that national leaders in sports (e.g., U.S. Lacrosse, U.S. Soccer) can be educated on what context factors were pertinent to their sport of interest and not just sport as a whole.

136

In a sense, the parents would now have the power to “diversify” their child athlete by finding complementary sports to pair with each other based on the factors measured in the

Sport Action Compendium (SAC). The SAC would be analogous to the ACSM American

Fitness Index (2019) developed in 2008 which annually ranks the 50 largest U.S. metro areas on their respective state of health and fitness to inspire community action based on scientific data.

But I imagine the SAC tool will look more similar to the Compendium of Physical Activities for adults (Ainsworth et al., 2011) and the Youth Compendium of Physical Activities which lists the major types of physical activities (including sports) with their associated estimated energy costs

(Butte et al., 2018). Both the youth and adult compendium is used globally in clinical settings to write PA recommendations and to assess energy expenditure in individuals.

Significance of Findings

The results contribute to our understanding of how sports facilitate participant achievement of recommended levels of MVPA. Additionally, the context that characterizes sport practices is significant in determining how much MVPA occurs. Sports placing extra emphasis on game simulation, physical fitness, and skill development drills can be expected to have higher levels of MVPA. As this study examined the PA intensity during sport practices, it is important to note the importance of both the frequency and duration of sport practices as a contributing factor to overall sport PA levels.

The context surrounding sport participation accounted for a third to nearly half of the variance found in PA levels exhibited by athletes. For boys sports, the full models were able to predict 41 percent (Model 9A) and 37 percent (Model 10A) of the variability found in PA observed in athletes and percent practice time in MVPA, respectively. These numbers were a little higher in the full models for girls sports, predicting 52 percent (Model 9B) and 46 percent

(Model 10B) of the variability found in PA observed in athletes and percent practice time in

137

MVPA, respectively. The full binomial regression models helped explained 39 percent of PA levels in athletes (Table 4.36) and 34 percent of the MVPA make up of practices (Table 4.37).

Thus, there are clearly opportunities to increase MVPA given the sizable variances explained, particularly during practices in high school sports.

As the SEM postulates, PA behavior can be more permanent and reach a broader number of people with the more levels addressed in the model. Even better results can be anticipated when strategies become multidimensional within the same space and time. This study supports this notion by illustrating that when the individual (i.e., boy or girl high school athletes), interpersonal (i.e., coaching influence), and organizational (i.e., sport form and practice design) levels of the SEM are all modified to some extent in interscholastic athletics, athlete PA production of athletes can most certainly be increased during practice sessions.

It would be a mistake to assume that the benefits that have been found to be associated with PA in general necessarily accrue to sport. It is also an error to extrapolate concerns with an extreme and under representative form of sport to all sport. In both of these cases, the accuracy of claims are empirical matters, requiring empirical research. A large part of the rationale for policy attention on youth PA is the widely held belief that the early years of life provide the foundation of later activity (Bailey et al., 2015). This presumption is given added support by the finding that past exercise behavior is a reliable predictor of current activity status (Smith,

Gardner, Aggio, et al., 2015), and urgency by the fact that PA tends to decrease as individuals move into adulthood (Kwan, Cairney, Faulkner, & Pullenayegum, 2012). There are clear benefits of participating in sport during high school, from increased psychomotor competence and improved physical health to team works skills and new friendships (Bailey, Hillman, Arent, et al., 2013), but the potential long-term benefits related to continued engagement in PA as an adult could be the most rewarding benefit of all, at least from the perspective of public health.

138

In light of the high costs on inactivity and sedentary behaviors, childhood PA promotion offers a promise of a relatively cheap preventive strategy for a range of non-communicable diseases

(Matheson, Klügl, Engebretsen, et al., 2013).

With the inclusion of SOFIT in this study, not only does it provide the first glimpse of how time is spent and coach behavior during these practice sessions, it may also assist in identifying opportunities to increase MVPA, particularly during practice. Furthermore, support should be provided to coaches in an effort to increase MVPA and decrease sedentary time in organized sports, without interfering with fundamental learning opportunities and skill development that occur in organized sports. Since there are a bevy of untrained coaches in charge of delivering sport to high school students, providing substantial practice ideas and proven strategies for facilitating an active practice would yield positive results. Also, coaches may begin to reimagine how an efficient practice can look like for particular sports and how best to maximize the few after school time they have with athletes to ensure high MVPA practices can occur consistently.

Recommendations for Practice

Children participate in sport for a number of reasons, such as fun and enjoyment, the satisfaction of learning new skills, and the pleasure of being with family and friends. In many cases, the enhancement of the health is not a significant motivating factor driving PA (Cope et al., 2013). Health-enhancing Physical Activity (HEPA) has been defined as “any form of PA that benefits health and functional capacity without undue harm or risk” (Foster, 2000) but sport does not necessarily qualify as HEPA. Nor does it follow that people always engage in such activity primarily with health in mind, and a great deal of PA is incidental to other tasks. For

139 example, people walk to work or bike to school. Any health-related benefits of participation in sport are incidental to these other values.

Figure 5.11 Relationships between physical activity and sport (Bailey, 2018).

But treating sports as homogeneous, as these writers implicitly do, leads to claims that are simply not true. To extrapolate concerns with a context associated with a tiny minority of sports players to the whole is both misleading and unhelpful. Every sport has unique characteristics that appeal to different interests, abilities, and expectations; their impact on players’ health is enormously varied. Consider the twenty sports examined in this study. All forms of sport carry different levels of PA, practice activities, and coaching behavior. Moreover, there are vastly different levels of exertion and resulting health outcomes from these activities.

Mäkinen, Borodulin, Tammelin, et al., (2010a) have argued that physically active individuals are more likely to stay physically active from adolescence to adulthood, especially among those who participate in several different sports after school hours. This claim has been offered support by a number of longitudinal studies (Aarnio, Winter, Peltonen, et al., 2002;

Tammelin, Näyhä, Hills, et al., 2003; Telama et al., 2013). So possibly a better question to ask instead of “which sport attributes to more PA?” is “how many different sports that youth participate in translates into lifelong healthy PA habits?”

140

According to the most recent School Health Policies and Practices Survey (SHPPS), a majority of school districts (74 percent) have policies requiring that high school head coaches complete a coach training course (SHPPS, 2017). But a vast minority (25 percent) of school districts actually require coaches to have previous coaching experience in any sport. There is widespread acknowledgement that a large gap exists between the science of athlete and coach development and the day-today reality of athletes and coaches (Martel, 2015; Gilbert et al.,

2009). Sport practices tend to overemphasize skill and competition strategies during instructional time, thereby limiting efficient use of practice facilities and the amount of time athletes engage in MVPA. Therefore, although organized sport may create opportunities for children to be active, its full potential as a strategy to increase PA may be limited by policies that govern the structure and delivery of sport practices.

Finally, as urban areas become more densely populated, finding spaces to accommodate the growing numbers of youth athletes is becoming more difficult. The national governing body of ice hockey in the United States, USA Hockey, has taken strides to address these limitations with the development and implementation of its American Development Model (ADM), redesigning their practice structure to best utilize facilities and increases the number of children participating in high levels of vigorous PA. USA Hockey funded a study to examine whether the new national policy for sport practice improves PA behavior without compromising skill development (Kanters et al., 2015). Two sports leagues with contrasting sport practice models

(i.e., traditional vs. ADM) were evaluated over the course of one year. Data were collected using momentary time sampling techniques and the system for observing fitness instruction time

(SOFIT). While MVPA levels did not differ between the practice types, the ADM practices provided overall higher PA and energy expenditure output than traditional practices. It was concluded that sport practices could ultimately be structured in a way to successfully facilitate

141 high levels of PA for more participants without compromising attention given to skill development and instruction.

Recommendations for Future Research

Before this research study was carried out, it became apparent through reviewing the literature base that PA in high school sport (and youth sports in general) is a highly under researched area. Perhaps this is due to several assumptions that revolve around sport participation being obviously linked to higher PA in young athletes or the difficulty to collect valid and reliable data within adolescent populations. While a large number of researchers and policymakers have designed interventions to increase healthy sport outcomes (Casey & Eime,

2015), the evidence on the success of interventions conducted in the sports setting are very limited, with many focused on PA in after-school programs or school-based programs (e.g., physical education) rather than organized sport organizations (Metcalf et al., 2015). Most of the empirical studies done in building theory towards a better understanding of healthy sport for youths have been outcomes-based and focused almost entirely on the individual (Holt, 2011).

For instance, there is a litany of literature that helps inform our grasp of sport participation and its influence on youth development outcomes like motivation, competence, confidence, enjoyment, PA, and others (Jones et al., 2016). However, there has been less attention to programmatic features and managerial aspects of youth sport delivery, leading to a lack of clarity regarding how health outcomes are produced in the sport context.

One of the main topics consistently identified while conducting the literature review surrounding coaching influence on athlete behavior was on motivation. Positive sports experiences and positive outcomes during childhood and adolescence may support the development of intrinsic motivation (Larson, 2000), which could lead to increased motivation to

142 be physically active later in life (Kilpatrick, Hebert, & Bartholomew, 2005). At the moment, these are only conceivable hypotheses. Further research will be needed to properly understand the mechanisms why sport does and does not affect PA engagement during childhood and in later life.

SDT gives us a platform to understand the psychosocial conditions through which family and sport environments can influence competitive athletes, specifically for youth.

Significant adults such as parents and coaches are frequently interacting with adolescents to instruct, guide, and support their efforts in attempting to manage the various requirements of their daily living. There is good evidence for the value of SDT in providing our understanding of PA behavior and motivations and how it manifests itself within youth populations. One of the clearest findings has been the beneficial role of developing autonomous self-regulation. In a systematic review, there was consistent showing that autonomous regulation predict exercise participation across a range of samples and settings (Teixeira et al., 2012). There is also evidence to suggest that a motivational profile is important to sustain PA behaviors over time (Eime et al.,

2013). In the context of sport, studies have found a positive link between perceived autonomy support from coaches and athletes’ motivation (Amorose & Anderson-Butcher, 2007, 2016;

Gillet et al., 2010, Gagne et al., 2003). While these studies help establish links between youth sport athletes’ motivations and the degree to which adults in sport settings are autonomy- supportive, many questions still remain due to the limitations in this line of research (Amorose et al., 2016).

An objective specific to youth sports will be added to Healthy People 2030 to “increase the proportion of children and adolescents, ages 6 to 17 years, who participate on a sports team or take sports lessons after school or on weekends”. Prioritizing youth sports as a national objective in Healthy People 2030 enables policy makers to monitor progress toward improved

143 participation. USDHHS (2018) has also mentioned the need to track youth sports participation as a complement to these objectives. Effective coaching may extend beyond the improvement on players’ PA levels, thus elements of effective coaching should also be considered in the future

(e.g., education, skills, tactics) (Guagliano et al., 2015).

Tracking studies, which are longitudinal studies examining the tendency of individuals to maintain behaviors over time (Malina, 1996), have an obvious appeal within the context of PA research. These studies are relatively rare as they generally require a large financial investment and a long-term commitment of both participants and researchers. In addition to the emergence of new measurement methods and changing conceptions of PA, a host of confounding factors can occur in repeated measures (Hirvensalo & Lintunen, 2011). Although tracking may provide useful insight in the relationship between youth sport and adherence to sport, the concept of tracking is debated. First, tracking studies need longitudinal data covering years from youth to adulthood. Such data is almost impossible to get, limited in scope, and mostly unrepresentative of larger populations beyond the observed sample. This means that the available results in the literature tend not to be generalizable let alone translated into policy programs or youth sport practices. Knowledge based on tracking data so far only offers exploratory and indicative insights into the quest for lifelong adherence to sport and PA.

A second weakness of tracking is conceptual in its nature. Essentially, tracking basically observes input and output data, but leaves what happens over or during the observed period of time out. Features of youth sport such as type, frequency, intensity, or context of participation are input information. Data of the same sample many years later, on involvement in sport and

PA as adults, provides the output. Tracking studies relationships between the input and output side but what happens in between remains mostly unknown. Tracking studies do not reveal information on life phase processes related to age, work, study, leisure, living situation, and so

144 forth. Although these factors may have a significant impact on the continuity or drop out of participation in sports, they are not included in tracking studies on youth sport.

On average, level of PA tends to reach a peak around 12 to 14 years old, and subsequently declines (Malina et al., 2016). PA during childhood seems to predict physically active lifestyles in adolescence and adulthood (Anderssen et al., 2005). However, the tracking from sports participation to adult PA is less understood (Kjønniksen, Anderssen, & Wold,

2009). Also, the tracking of PA during transitions from childhood to adolescence and adulthood has been shown to be quite low, while the tracking of PA during adult life is relatively high

(Telama, Yang, Leskinen, et al., 2013). This is not surprising as education, environment, family status, and other factors have been found to influence later activity (Tammelin, Näyhä, Hills, et al., 2003; Cleland, Ball, Magnussen, et al., 2009). Limited evidence suggests that participation in sport during adolescence is reasonably predictive of PA in young adulthood. Thus, the process of how participation in sport during adolescence translates into an active lifestyle in young adults needs elaboration.

As one of the members of the systematic observation research fraternity, a common symptom experienced by data collectors is something called “observer drift”, which in short, describes an observer’s tendency over time to become less and less accurate in coding individuals during observations. In short, observer drift can be anticipated for data collectors the further away the original training occurred. The conventional solution to this would be to schedule re-trainings for all observers in essence to recalibrate them back to coding accurately and consistently for the duration of the project. The current study experienced a similar phenomenon which I can most aptly describe as “Observer Fatigue” – an offshoot of the term

“observer drift”. Observer fatigue describes the urge of an observer to actively anticipate coding an individual differently as a refreshing change of pace. I believe this was partly caused by the

145 extreme lengths that some sport practices lasted; a handful lasting on upwards of two and a half hours or more. I can imagine that this has been noted in PE studies using SOFIT as PE sessions in primary and secondary schools do not last longer than one hour.

Through the standard reliability checks that I performed during the athletic season(s), I both experienced this feeling myself as a secondary observer and had anecdotal evidence reported to me by the primary observers (i.e., the coaches on site). For example, if a data collector was observing an athlete that happened to be sitting on the sidelines, consistently coding that athlete as “sitting” for countless minutes would seem monotonous and cumbersome to the trained observer. So much so that when, after several minutes of being sedentary, the athlete of interest suddenly stands up and begins walking, the observer would feel inclined to code that athlete as being “vigorous” since there was such a drastic change of activity. This is of course inaccurate coding as the athlete is moderately active, not vigorous. However, the biases of data collectors wanting athletes to be vigorously active, particularly if they knew the student themselves (e.g., student-teacher relationship), may have resulted in certain cases of over reporting vigorous activity simply due to observer fatigue. This is certainly not backed up by the data but simply a hypothesis on my part as a result of being in the field that deserved mentioning. Future investigators using SOFIT in sport settings should take note and continuously check in on their data collectors for reliability purposes, particularly when sport games or practice sessions are lasting a long time.

Anyone involved in sports participation research knows that not all sports are the same and can vary drastically regarding how they are designed and what coaching strategies are necessary to achieve results, and the kind of sport being played. In fact, the results from this study supported these notions to be true to some degree. And yet, we systematically deployed the same SOFIT protocol across all sport types. In a perfect world, I would have much rather

146 had modified SOFIT to fit each individual sport’s unique characteristics. The training materials that were used to teach data collectors the coding protocols would have benefited from a more systematic approach to customizing the materials for specific sporting environments. For example, a skill practice session in basketball may look differently than an athlete performing skill practice activities in lacrosse that possibly produces more MVPA. The SOFIT instrument would be entirely tailored to observing basketball practices, not just sport in a generic sense. To illustrate, the game play code would specify the most common examples (i.e., full-court scrimmage, half-court set, in-bounds plays, etc.). The skill practice code could include things such as boxing out drills, stationery dribbling, shell defense drill. These are all a frequent characteristic of high school basketball practices and I believe the inter-rater reliability would have been strengthened and the findings more generalizable if sport-specific training would have been adopted. Therefore, I highly endorse researchers to explore creating a new sport specific observation tool through validation to better capture the nuance of sports game and practices that are not commonly found in physical education settings.

There is precedent to customizing the SOFIT codes to fit a particular sport setting. In the Bass and Martin (2013) swimming project, four activity level codes were developed; resting/sitting, standing, light swimming/treading water, and vigorous swimming. The lying down code was eliminated because that level of activity was not anticipated in the setting (Bass

& Martin, 2013). Future researchers looking to utilize SOFIT in sport should consider the uniqueness of each sport of interest and create coding practices that include more clarity in what practice and coaching activities should be coded in the contextual categories.

To that end, youth sports participation is measured in a variety of ways and, to my knowledge, no single system measures all aspects of sports participation, including who is participating (e.g., sex, race, ethnicity, age); frequency of participation; number of sports played;

147 time spent playing sports; type(s) of sports; and location of participation (e.g., school, recreational facility). No one can argue that these variables, just to start, are perceived to be invaluable in understanding the environment surrounding PA and sports. And yet, no tool has been conceptualized or validated that covers these elements short of a self-report survey. I believe that a systematic observational instrument designed for the collection of PA in sports could make its way into the hands of researchers eventually, and that is my charge to current researchers in this field.

As described in the methodology, prior application of the SOFIT protocols in organized youth sport settings have operationalized the context and coaching codes to fit the unique characteristics of the specific sport (Cardon et al., 2004; Schwamberger & Wahl-Alexander,

2016). I hypothesize that if each sport type were given a customized set of codes rather than generic ones, the inter-rater reliability results would have been improved to levels exceeding the

80 percent recommendation by van der Mars (1989). For example, after training observers with common examples found in swimming lessons, the authors reported a 94 percent, 96 percent, and 96 percent agreement for PA, lesson context, and teacher behavior, respectively.

Limitations of the Present Study

Before discussing the strengths and implications of this research to PA in high school sport, there are several limitations that must be acknowledged. First, the observation data was collected from the beginning of the fall semester sports season to the end of the spring semester sports season, which covers the traditional American high school academic year. It is common in interscholastic sports teams to experience significant coaching turnover from year to year.

This is study was very coach and program-centric, future studies that are able to capture at least two years of data would be able to alleviate any seasonal differences based on team rosters and

148 coaching changes. Second, this study was restricted to North Carolina high schools only. The findings can loosely be translated to other regions of the country but was cross-sectional only within the state of NC and sport teams. Third, not all high school sports were included. It is safe to say that the top twenty high school sports in terms of participation numbers is different in every state. For example, other nationally popular sport programs (e.g., golf, boys volleyball, ice hockey, water polo) are rarely offered in NC schools and hence, were not observed. Finally, only 12 schools were considered in this study, which limits the power of our analysis.

Specifically, the practice design factors are somewhat nullified by the limited sample size at this level of analysis. Forthcoming studies that are able to capture a larger cross section of high schools in different regions of the country and a larger quantity of teams within the same sport might uncover more generalizable conclusions.

Motor proficiency and level of physical fitness have not been included among correlates of activity, although the concept of physical literacy is emerging in the literature (Logan & Cuff,

2019; MacDonald, 2015; Tremblay & Lloyd, 2010). Physical literacy is the ability to move with competence and confidence in a variety of physical activities in multiple environments that benefit the healthy development of the whole person (SHAPE America, 2019; Mandigo et al.,

2009). It primarily includes the development of basic movement skills, as well as sports skills, and can also provide youth with the ability, confidence, and desire to be physically active for life.

Physical literacy and sports skills development (e.g. throwing, catching, running) can be enhanced when youths try and learn a variety of sports (Logan & Cuff, 2019, SHAPE, 2019).

Research shows that without the development of basic movement skills, it is unlikely that children will experience sports success or develop movement competence, which are both important for long-term motivation and participation (Gould & Walker, 2019). Adults play an important role in the development of youths’ physical literacy, both inside and outside of the

149 sports environment. Fundamental movement skills and sport skills do not simply occur with participation in sports or PA; they must be taught (Canadian Sport Centers, 2016). When youth begin to play sports, the development of physical literacy helps to support their ability and desire to continue participating. Research also indicates that even the perception of low physical competence is an important factor in attrition (Crane & Temple, 2015). There was a logical gap in providing information on skill development within this study that examining what specific movement skills were being performed in each sport (e.g., catching a baseball), would have been great information to know to help in the translation of results to specific sport genres.

In the Cotton & O’Connor (2013) study, they coupled SOFIT with a Training

Evaluation Tool (TET) (O’Connor & Cotton, 2010) that was designed to help a coach review the number of times a player performed a particular skill during a specific activity in a training session (i.e., number of times a player tackles during a tackling drill). O’Connor and Cotton

(2009) identified that junior coaches spend large amounts of time instructing (48 percent) and managing (13 percent) players during training sessions, often at the expense of activity time, with players being inactive for 36 percent of a training session. They also revealed youth coaches consistently overestimate the amount of time they spend during a coaching session on skill development and giving instructions to athletes.

It would also be ill-advised of me make conclusions on PA in sports without recognizing that participation in youth sports is influenced by an equally diverse range of biological, behavioral, cultural, and psycho-social factors; all of which were not investigated in this study.

More positive perceptions of the self, in particular body image and athletic competence, peer and familial support, and coach encouragement have been noted as important predictors of sport participation in girls (Biddle et al., 2005a). Parental engagement in exercise is also associated with increased sport involvement in youth (Cleland et al., 2005). Conversely, relative

150 age and maturity timing have also been documented as important predictors of participation in a number of sports, in particular those in sports where a specific physique or body size affords an athletic advantage. With regard to sociocultural factors, socioeconomic status and ethnicity have also been shown to impact youth engagement in sport, with certain groups being more or less likely to engage in specific sports, and sport in general (Biddle et al., 2005b).

Strengths of the Present Study

Despite these limitations, this study used objective measures that allowed for a rigorous description of the PA levels that high school athletes achieved during sports practices with some of the highest participation rates in the U.S. (NFHS, 2017). This was one of the first studies to evaluate PA across a variety high school sport offerings and its association with practice context and the interactions with coaches. To the best of my knowledge no study has evaluated PA behavior in this many high school sports teams within the same timeframe. This was, indeed, a youth sports participation study that provided a robust cross-section sample of athletes and coaches that is rarely achieved in social science. Through all of the literature, the number of schools, sport types, and observations completed previously was nowhere close to the quantity that was included in this project. The time frame of which observations were gathered was also a major strength. As a large research project backed by the Aspen Institute, researchers were given a full year to collect data and even more time to organize and analyze the data set. Typical sport and PA studies were not this lengthy. Not to a fault of their own, but given the strong results of the study, the extended time frame gives merit and legitimacy to the results, no matter how intuitive they may have been.

151

Conclusion

In summary, PA levels of 2,408 participants in twenty different high school sports were measured during organized sports practices. Findings indicate the amount of PA that athletes accrue and energy expended during practice time is high and aligns well with prior research but varies from sport to sport. The sports that are characterized with high amounts of running like cross country, track & field, and soccer had the highest levels of MVPA, which was expected.

Conversely, cheerleading, softball, and baseball were the lowest scoring sports with regards to

MVPA. It was also found that a majority of the boys sports had higher MET values (generating more PA) than girls sports. The results add clarity to our understanding of which sports facilitate more participant PA during practices. Additionally, the context that characterizes sport practices is significant in determining how much PA is produced by athletes. Most of the sports under investigation that place extra emphasis on game simulation, fitness, and skill development drills can be expected to have higher levels of MVPA. Based on how practices are delivered to athletes, we can accurately predict how much PA athletes can expect to achieve within a given sport season assuming length and frequency of practices are recorded.

This study provided insight on PA characteristics, lesson context, and leader/coach behavior (through SOFIT). Sport has unique characteristics that suggest it can play a valuable role in the promotion of PA, during teenage years when sport is highly attractive to this age group. The findings from this study demonstrate that sport can make a useful, if not sufficient role in helping high school athletes reach recommended PA levels. The effect this has on later

PA is more ambiguous, as mediating factors seem to play more of a role. However, a plausible case can be made for a generally positive influence of sporting context and coaches on PA outcomes. First, high school sports participation does provide substantial PA output for athletes based on the results of this study. Second, sport may provide opportunities for children to

152 acquire the skills and confidence that will facilitate participation in physical activities later in life

(Bailey et al., 2010). Third, adult coaches can provide environments and opportunities for school sport activities, which can positively influence a student athletes’ PA behavior (USDHHS,

2012). Coaches have a tremendous influence over the quality of a youth’s sporting experience

(National Physical Activity Plan Alliance, 2016). Quality coaches improve participation rates, health behaviors, and outcomes (Aspen Institute, 2015). Additionally, coaches who are well trained and equipped understand best practices in sport and physical skill development. Quality coaches provide developmental and age-appropriate activities and skills training to participants and understand how feedback and reinforcement can shape behavior (Gould & Carson, 2008;

Theokas, Danish, Hodge, Hekel, & Forneris, 2008).

One thing does seem clear: sports participation does not automatically result in long- lasting physical benefits. Simply put, participation alone is unlikely to be enough and requires significant attention to be given to the management of the context surrounding sport. It is possible to disagree with critics like Coalter (2007), who have questioned the empirical base of the various physical, physiological, psychological, and sociological outcomes claimed for sport, while accepting that there remains ‘a widespread lack of even more valuable evidence relating to sufficient conditions: the type of sports, and the conditions under which the potential to achieve the desired outcomes is maximized’. As Svoboda (1994) stated, the presumed positive outcomes of sport are ‘only a possibility’, and a direct linear affect between simple participation and effect cannot be assumed. This is certainly understandable as there is ample evidence to suggest that participation in sport and PA can result in negative as well as positive outcomes (Bailey et al.,

2015).

Although participation in youth sport can potentially promote PA, it is best not to take the relationship as a “given”; it can be difficult to achieve and can only be realized in association

153 with a series of conducive change mechanisms as this study found. Also, I do not presume that this research will radically improve these problems as they have historically been known to be multi-layered and highly complex. However, I do believe that the results provide a small and logical step in the right direction for reforming how leaders of high school athletics think about the many facets surrounding athletes playing sports.

The evidence base in the literature is much less well-developed than that of outcomes, however, it seems clear that these outcomes are mediated by a host of contextual factors, including the sport type itself, the ways in which sport is delivered, managed, and coached. If sport is going to be leveraged for wider policy concerns, then it is likely that there is intentionality in the design of the programs so that sports are deliberately structured and implemented to achieve the desired health outcomes such as PA production.

Regular PA, including participation in sport, has the potential to enhance a variety of indicators of health in youth. Given the available data, there is need for further study of the type and amount of PA associated with health benefits during adolescence. There is also a need to evaluate the influence of external factors like coaches on PA of athletes in accordance with social ecological perspectives. One area of promise within the SEM sphere of organizational systems and promoted by governing bodies like the U.S. Department of Health and Human

Services (2018) is the use of organized sports. A multitude of youth sport offerings (some not included in this research) have been identified as being capable and attractive vehicles for youths to obtain adequate bouts of PA, specifically in the after-school environment (Cardon et al., 2012;

Gortmaker et al., 2012; Lai et al., 2014). Although high school sport participation was observed to provide a considerable proportion of the recommended amounts of MVPA, there may be potential for improvement in the contribution that sport participation makes to daily MVPA.

154

This information on organized sports can be used as a platform on which to inform policies and to develop strategies to increase interscholastic PA levels through sport practices.

Many adults are involved in youth sports in some capacity (e.g., coach, parent, referee, spectator). As a result, adults can influence the youth sport experience through their actions and behaviors at multiple levels. Coaches determine how much PA occurs during practices and how much emphasis is placed on skill development, learning strategy, or game play. Additionally, they have influence over day-to-day experience of youths in sports programming and shaping health behaviors. However, very few studies have published on the PA of high school athletes

(both male and female) across a variety of sports (Ridley et al., 2018). A better understanding of youths activity levels is needed if successful intervention strategies for high school sports are to be designed and implemented.

The Physical Activity Guidelines for Americans (2018) highlights many health benefits associated with PA for youth. Benefits can be obtained regardless of gender, age, ability, or current fitness level. Compared to their less active peers, youths who engage in regular PA have improved bone health, weight status, cardiorespiratory and muscular fitness, cardiometabolic health, cognitive function, and a reduced risk of depression (USDHHS, 2018). In short, youths who are physically active tend to be healthier, have less body fat, and live more active lifestyles.

While there is some uncertainty in drawing a direct connection between PA and sports, it is clear that too few youths are meeting the PA guidelines and are therefore, missing out on the important health benefits. It is a reasonable expectation that the health benefits of PA can come from participating in sports. But regardless of the sport being played, so much hinges on the training and leadership ability of coaches in charge of conducting sports practices on a daily basis. Coaches are the delivery mechanism for quality sport programming at the high school

155 level, and thus, need to be trained in the art of long-term athlete development so that PA benefits can be sustained over the lifetime of the athlete.

Although high school sports are governed by State level associations which dictate the duration of the sport season and the number and duration of practices, those can vary dramatically from sport to sport and even from high school to high school. Finally, given the influential role of practice context in determining athlete PA, it is important to emphasize that coaches are responsible for making most of the contextual decisions. This highlights the importance of coaches being trained and certified so that sport environments can achieve their full potential for participant benefits.

156

REFERENCES

Aarnio, M., Winter, T., Peitonen, J., Kujala, U. M., & Kaprio, J. (2002). Stability of leisure-time

physical activity during adolescence – A longitudinal study among 16-, 17- and 18-year-

old Finnish youth. Scandinavian Journal of Medicine & Science in Sports, 12(3), 179-185.

Adair, J. G., Sharpe, D., & Huynh, C. L. (1989). Placebo, Hawthorne, and other artefact

controls: Researchers’ opinions and practices. Journal of Experimental Education, 57, 335-

341.

Adler, P., & Adler, P. (2002). Do university lawyers and the police define research values? In W.

C. Van Den Hoonard (Ed.), Walking the tightrope: Ethical issues for qualitative researchers (pp.

34-42). Toronto, Ontario, Canada: University of Toronto Press.

Agans, J. P., Säfvenbom, R., Davis, J. L., Bowers, E. P., & Lerner, R. M. (2013). Positive

movement experiences. Advances in Child Development and Behavior, 45, 261-286.

Agresti, A. (2010). Analysis of Ordinal Categorical Data. Hoboken, NJ: John Wiley & Sons.

Ainsworth, B. E., Haskell, W. L., Herrmann, S. D., Meckes, N., et al. (2011). 2011 compendium

of physical activities: A second update of codes and MET values. Medicine & Science in

Sports & Exercise, 1575-1581.

Ainsworth, B. E., Watson, K. B., Ridley, K., Pfeiffer, K. A., Herrmann, S. D., Crouter, S. E.,

McMurray, R. G., Butte, N. F., et al. (2018). Utility of the youth compendium of physical

activities. Research Quarterly for Exercise and Sport, 89(3), 273-281.

Allender, S., Cowburn, G., & Foster, C. (2006). Understanding participation in sport and

physical activity among children and adults: A review of qualitative studies. Health

Education Research, 21(6), 826-835.

157

Amorose, A. J., & Anderson-Butcher, D. (2007). Autonomy-supportive coaching and self-

determined motivation in high school and college athletes: A test of self-determination

theory. Psychology of Sport & Exercise, 8, 654-670.

Amorose, A. J., & Anderson-Butcher, D. (2016). Exploring the independent and interactive

effects of autonomy-supportive and controlling coaching behaviors on adolescent

athletes’ motivation for sport. Sport, Exercise, and Performance Psychology, 4, 206-218.

Anderssen, N., Wold, B., & Torsheim, T. (2005). Tracking of physical activity in adolescence.

Research Quarterly for Exercise and Sport, 76(2), 119-129.

ACSM American Fitness Index: Health and community fitness status of the 50 largest

metropolitan areas. (2017). Retrieved from http://www.americanfitnessindex.org/wp-

content/uploads/2014/02/ACSM-AFI-Report-2017.pdf

Aspen Institute (2019, January 17). Healthy Sport Index. Retrieved from

www.healthysportindex.com.

Bailey, R. P. (2006). Physical education and sport in schools: A review of benefits and outcomes.

Journal of School Health, 76(8), 397-401.

Bailey, R. P. (2016). Sport, physical activity and educational achievement-towards an explanatory

model. Sport in Society, 20(7), 768-788.

Bailey, R. (2018). Is sport good for us? Understanding youth sport as a public health setting. In

D. Parnell & P. Krustrup (Eds.), Sport and Health: Exploring the Current State of Play.

Routledge: New York.

Bailey, R., Armour, K., & Kirk, D. (2009). The educational benefits claimed for physical

education and school sport: An academic review. Research Papers in Education, 24(1), 1-27.

Bailey, R. P., Collins, D., Ford, P., MacNamara, A., Toms, M., & Pearce, G. (2010). Participant

development in sport: An academic review. Leeds: Sports Coach UK.

158

Bailey, R., P., Cope, E., & Parnell, D. (2015). Realising the benefits of sports and physical

activity: The human capital model. Retos, 28, 147-154.

Bailey, R. P., Hillman, C., Arent, S., & Petitpas, A. (2013). Physical activity: An underestimated

investment in human capital? Journal of Physical Activity and Health, 10, 289-308.

Bailey, R. P., Wellard, I., & Dismore, H. (2004). Girls’ participation in physical activities & sports:

Benefits, patterns, influences & ways forward. Geneva: World Health Organization.

Balyi, I., Cardinal, C., Higgs, C., Norris, S., & Way, R. (2005). Canadian sport for life: Long-term

athlete development. Vancouver, BC: Canadian Sport Centres.

Balyi, I., & Way, R. (1995). Long-term planning of athlete development: The training to train

phase. B.C. Coach, 2-10.

Barnekow-Bergkvist, M., Hedberg, G., Janlert, U., & Jansson, E. (2001). Adolescent

determinants of cardiovascular risk factors in adult men and women. Scandinavian Journal

of Public Health, 29, 208-217.

Bass, D., & Martin, E. H. (2013). Influence of different teaching strategies on physical activity

levels in a collegiate swim class. Perspectives in Learning, 14(1), 4-9.

Bassett, D. R., Fitzhugh, E. C., Heath, G. W., et al. (2013). Estimated energy expenditures for

school-based policies and active living. American Journal of Preventive Medicine, 44(2), 108-

113.

Bean, C.N., Fortier, M., Post, C., & Chima, K. (2014). Understanding how organized youth sport

may be harming individual players within the family unit: A literature review. International

Journal of Environmental Research & Public Health, 11, 10226-10268.

Beets, M.W., Webster, C., Saunders, R., & Huberty, J.L. (2013). Translating policies into practice:

A framework to prevent childhood obesity in afterschool programs. Health Promotion

Practice, 14(2), 228-237.

159

Berg, B., Warner, S., & Das, B. (2015). What about sport? A public health perspective on leisure-

time physical activity. Sport Management Review, 18, 20-31.

Biddle, S.J., Braithwaite, R., & Pearson, N. (2014). The effectiveness of interventions to increase

physical activity among young girls: A meta-analysis. Preventive Medicine, 62, 119-131.

Biddle, S., Coalter, F., O’Donovan, T., Macbeth, J., Nevill, M., & Whitehead, S. (2005a).

Increasing demand for sport and physical activity by girls. Research Report no. 100.

Edinburgh: SportScotland. Retrieved from www.sportscotland.org.uk.

Biddle, S. J., Whitehead, S. H., O’Donovan, T. M., & Nevill, M. E. (2005b). Correlates of

participation in physical activity for adolescent girls: A systematic review of recent

literature. Journal of Physical Activity and Health, 2, 421-432.

Bocarro, J. N., Floyd, M., Moore, R., Baran, P., Danninger, T., Smith, W., & Cosco, N. (2009).

Adaptation of the system for observing physical activity and recreation in communities

(SOPARC) to assess age groupings of children. Journal of Physical Activity & Health, 6,

699-707.

Bocarro, J. N., Kanters, M. A., Cerin, E., Floyd, M. F., Casper, J. M., Suau, L. J., McKenzie, T. L.

(2012). School sport policy and school-based physical activity environments and their

association with observed physical activity in middle school children. Health & Place, 18,

31-38.

Bocarro, J. N., Kanters, M.A., Edwards, M. B., Casper, J. M., & McKenzie, T. L. (2014).

Prioritizing school intramural and interscholastic programs based on observed physical

activity. Journal of Health Promotion, 28(3), S65-71.

Booth, M. (1997). The NSW schools fitness & physical activity survey, 1997. New South Wales

Public Health Bulletin, 8(5), 35-36.

Brown, R. (2001). Group processes: Dynamics within and between groups. Oxford, UK: Blackwell

160

Publishing.

Brown, T. D., & Holland, B. V. (2005). Student physical activity and lesson context during

physical education. ACHPER Healthy Lifestyles Journal, 52, 17-23.

Brustad, R. J., Babkes, M. L., & Smith, A. L. (2001). Youth in sport: Psychological

considerations. In R. N. Singer, H. A. Hausenblas, & C. M. Janelle (Eds.), Handbook of

sport psychology (2nd ed., pp. 604-635). New York: John Wiley & Sons.

Byrne, N. M., Hills, A. P., Hunter, G. R., Weinsier, R. L., & Schutz, Y. (2005). Metabolic

equivalent: One size does not fit all. Journal of Applied Physiology, 99(3), 1112-1119.

Camiré, M. (2014). Youth development in North American high school sport: Review and

recommendations. Quest, 66(4), 495-511.

Canadian Sport Centres. (2016). Developing physical literacy: A guide for parents of children

ages 0 to 12. Canadian sport for life physical literacy.

Cardinal, B., Yan, Z., & Cardinal, M. (2013). Negative experiences in physical education and

sport: How much do they affect physical activity participation later in life? Journal of

Physical Education, Recreation & Dance, 84(3), 49-53.

Cardon, G. M., Van Acker, R., Seghers, J., De Martelaer, K., Haerens, L. L., & De

Bourdeaudhuij, M. (2012). Physical activity promotion in schools: Which strategies do

schools (not) implement and which socioecological factors are associated with

implementation? Health Education Research, 27(3), 470-483.

Cardon, G., Verstraete, S., De Clerq, D., & Bourdeaudhuij, I. D. (2004). Physical activity levels

in elementary-school physical education: A comparison of swimming and nonswimming

classes. Journal of Teaching in Physical Education, 23, 252-263.

161

Carnethon, M. R., Gulati, M., & Greenland, P. (2005). Prevalence and cardiovascular disease

correlates of low cardiorespiratory fitness in adolescents and adults. JAMA, 294(23),

2981-2988.

Carron, A. V., Hausenblas, H. A., & Eys, M. A. (2005). Group dynamics in sport (3rd ed.).

Morgantown, WV: Fitness Information Technology.

Carson, V., Rinaldi, R., Torrance, B., et al. (2014). Vigorous physical activity and longitudinal

associations with cardiometabolic risk factors in youth. International Journal of Obesity,

38(1), 16-21.

Casey, M., & Eime, R. (2015). Sport and health. In M. T. Bowers and M. A. Dixon (Eds.) Sport

management: An exploration of the field and its value (pp. 61-93). Urbana, IL: Sagamore.

CDC - Centers for Disease Control and Prevention. (2012). General physical activities defined by level

of intensity. Retrieved from

http://www.cdc.gov/nccdphp/dnpa/physical/pdf/PA_Intensity_table_2_1.pdf

CDC - Centers for Disease Control and Prevention. (2015). How much physical activity do children

need? Retrieved from http://www.cdc.gov/physicalactivity/basics/children.

CDC - Centers for Disease Control and Prevention. (2018). Youth risk behavior surveillance –

United States, 2017. Morbidity and Mortality Weekly Report, 67(8). Atlanta, GA: United

States Department of Health and Human Services. Retrieved from

https://www.cdc.gov/healthyyouth/data/yrbs/.

CDC - Centers for Disease Control & Prevention. (2000). Promoting better health for young people

through physical activity and sports. Atlanta, GA: United States Department of Health and

Human Services.

162

Charbonneau, D., Barling, J., & Kelloway, E. K. (2001). Transformational leadership and sports

performance: The mediating role of intrinsic motivation. Journal of Applied Social

Psychology, 31: 1521-1534.

Chelladurai, P. (1989). Manual for the Leadership Scale for Sports. Unpublished manuscript,

University of Western Ontario, London, Canada.

Chelladurai, P. (1999). Human resource management in sport and recreation. Champaign, IL: Human

Kinetics.

Chelladurai, P., & Saleh, S. D. (1980). Dimensions of leader behavior in sports: Development of

a leadership scale. Journal of Sport Psychology, 2, 34-45.

Child Trends Data Bank. (2018). Participation in school athletics. Washington, DC. Retrieved

from https://www.childtrends.org/indicators/participation-in-school-athletics.

Cleland, V. J., Ball, K., Magnussen, C., Dwyer, T., & Venn, A. (2009). Socioeconomic position

and the tracking of physical activity and cardiorespiratory fitness from childhood to

adulthood. American Journal of Epidemiology, 170, 1069-1077.

Cleland, V., Venn, A., Fryer, J., Dwyer, T., & Blizzard, L. (2005). Parental exercise is associated

with Australian children’s extracurricular sports participation and cardiorespiratory

fitness: A cross-sectional study. International Journal of Behavioural Nutrition and Physical

Activity, 2, 3.

Coakley, J. (2016). Positive youth development through sport: Myths, beliefs, and realities. In N.

L. Holt (Ed.), Positive youth development through sport (pp. 21-33) (2nd ed.). New York:

Routledge.

Coakley, J. (2011). Youth sports: What counts as positive development? Journal of Sport & Social

Issues, 35(3), 306-324.

Coalter, F. (2007). A wider social role for sport: Who’s keeping the score? London: Routledge.

163

Cochrane, T., & Davey, R. C. (2008). Increasing uptake of physical activity: A social ecological

approach. The Journal of the Royal Society for the Promotion of Health, 128(1), 31-40.

Cohen, J., Cohen, P., West, S. G., & Aiken, L. S. (2003). Applied multiple regression/correlation

analysis for the behavioral sciences (3rd ed.). Mahwah, NJ: Lawrence Erlbaum Associates.

Cohen, D. A., Lapham, S., Evenson, K. R., Williamson, S., Golinelli, D., Ward, P., Hillier, A.,

McKenzie, T. L. (2013). Use of neighborhood parks: Does socio-economic status

matter? A four-city study. Public Health, 127(4), 325-32.

Cohen, D. A., McKenzie, T. L., Sehgal, A., Williamson, S., Golinelli, D., & Lurie, N. (2007).

Contribution of public parks to physical activity. American Journal of Public Health, 97(3),

509-14.

Cohen, D. A., Setodji, C., Evenson, K. R., Ward, P., Lapham, S., Hillier, A., & McKenzie, T. L.

(2011). How much observation is enough? Refining the administration of SOPARC.

Journal of Physical Activity & Health 8(8), 1117-23.

Cone, S. L. (2004). Pay me now or pay me later: Ten years later have we seen any change?.

Journal of Teaching in Physical Education, 23, 271-281.

Conroy, D. E., Coatsworth, J. D. (2006). Coach training as a strategy for promoting youth social

development. Sport Psychology, 20(2), 281-292.

Cooky, C. (2009). ‘Girls just aren’t interested’: the social construction of interest in girls’ sport.

Sociological Perspectives, 52(2), 259-283.

Cope, E., Bailey, R. P., & Pearce, G. (2013). Why do children take part in, and remain involved

in sport? A literature review and discussion of implications for sports coaches.

International Journal of Coaching Science, 7(1), 56-75.

164

Côté J., Baker J., & Abernethy B. (2007). Practice and play in the development of sport

expertise. In, G. Tenenbaum & R. C. Eklund (Eds.), Handbook of sport psychology (3rd

ed.)(p. 184-202). New York: Wiley.

Côté J., & Abernathy, B. (2012). A developmental approach to sport expertise. In S. Murphy

(Ed.), The Oxford Handbook of Sport and Performance Psychology (pp. 435-447). New York:

Oxford University Press.

Côté J., & Fraser-Thomas, J. (2007). Play, practice, and athlete development. In D. Farrow, J.

Baker, & C. MacMahon (Eds.), Developing Sport Expertise: Researchers and Coaches Put Theory

Into Practice (pp. 17-28). Routledge: New York, NY.

Côté, J., & Vierimaa, M. (2014). The developmental model of sport participation: 15 years after

its first conceptualization. Science and Sports, 29(Suppl), S63-S69.

Côté, J., Yardley, J., Hay, J., Sedgwick, W., & Baker, J. (1999). An exploratory examination of the

Coaching Behavior Scale for Sport, AVANTE, 5, 82-92.

Cotton, W., & O’Connor, D. (2013). The effect of a training evaluation tool on youth coaches.

In H. Nunome, B. Drust, & B. Dawson (Eds.), Science and football VII: The proceedings of the

seventh world congress on science and football. Chapter 61, 385-390. New York, NY: Routledge.

Council of Europe. (2001). The European sports charter (revised). Retrieved from

www.sportdevelopment.info/index.php?option=com_content&view.

Crawford, D. (2009). The future of sport in Australia. Commonwealth of Australia, Canberra.

Crane, J., & Temple, V. (2015). A systematic review of dropout from organized sport among

children and youth. European Physical Education Review, 21(1), 114-131.

Crissey, S. R., & Honea, J. C. (2006). The relationship between athletic participation and

perceptions of body size and weight control in adolescent girls: The role of sport type.

Sociology of Sport Journal, 23(3), 248-272.

165

Curtner-Smith, M. D., Sofo, S., Chouinard, J., & Wallace, S. (2007). Health-promoting physical

activity and extra-curricular sport. European Physical Education Review, 13(2), 131-144.

Dale, D., Corbin, C.B., and Dale, K.S. (2000). Restricting opportunities to be active during

school time: Do children compensate by increasing physical activity levels after school?

Research Quarterly for Exercise and Sport, 71(3), 240-248.

Davies, L., Taylor, P. Ramchandani, G., & Christy, E. (2016). Social return on investment in sport: A

participation-wide model for England summary report. Sheffield: Sheffield Hallam University.

DeMartelaer, K., & Theeboom, M. (2006). Physical education and youth sport. In The Handbook

of Physical Education. Thousand Oaks, CA. London: Sage. Dooris, M. (2004). Joining up

settings for health: A valuable investment for strategic partnerships? Crit Public Health, 14,

37-49.

Demetriou, Y., Gillison, F., & McKenzie, T. L. (2017). After-school physical activity

interventions on child and adolescent physical activity and health: A review of reviews.

Advances in Physical Education, 7, 191-215.

Demetriou, Y., & Höner, O. (2012). Physical activity interventions in the school setting: A

systematic review. Psychology of Sport and Exercise, 13, 186-196. de Mooij, M.K. (2004). Consumer behavior and culture: Consequences for global marketing and advertising.

Thousand Oaks, CA: Sage.

Dennison, B. A., Straus, J. H., Mellits, E. D., & Charney, E. (1988). Childhood physical fitness

tests: Predictor of adult physical activity levels? Pediatrics, 82, 324-330.

Dentro, K. N., Beals, K., Crouter, S. E., Eisenmann, J. C., McKenzie, T. L., Pate, R. R., Saelens,

B. E., Sisson, S. B., Spruijt, D., Sothern, M. S., & Katzmarzyk, P. T. (2014). The 2014

United States report card on physical activity for children and youth. Columbia, SC:

National Physical Activity Plan.

166

DiFiori, J. P., Benjamin, H. J., Brenner, J., Gregory, A., Jayanthi, N., Landry, G. L., Luke, A.

(2014). Overuse injuries and burnout in youth sports: A position statement from the

American Medical Society for Sports Medicine. Clinical Journal of Sport Medicine, 24(1), 3-

20.

Dobbins, M., Husson, H., DeCorby, K., & LaRocca, R. L. (2013). School-based physical activity

programs for promoting physical activity and fitness in children and adolescents aged 6-

18. Cochrane Database of Systematic Reviews, 2(2).

Doherty, A., & Cousens, L. (2013). Introduction to the special issue on community sport. Journal

of Sport Management, 27(6), 419-421.

Dohle, S., & Wansink, B. (2013). Fit in 50 years: Participation in high school sports best predicts

one’s physical activity after age 70. BioMed Central Public Health, 13, 1100.

Dudley, D.A., Okely, A.D., Cotton, W.G., et al. (2012). Physical activity levels and movement

skill instruction in secondary school physical education. Journal of Science and Medicine in

Sport, 15, 231-237.

Dudley, D., Okely, A., Pearson, P., & Cotton, W. (2011). A systematic review of the

effectiveness of physical education and school sport interventions targeting physical

activity, movement skills, and enjoyment of physical activity. European Physical Education

Review, 17(3), 353-378.

Durlak, J. A., Weissberg, R. P., Pachan, M. (2010). A meta-analysis of after-school programs that

seek to promote personal and social skills in children and adolescents. American Journal of

Community Psychology, 45, 294-309.

Dworkin, J., & Larson, R. (2006). Adolescents’ negative experiences in organized youth

activities. Journal of Youth Development, 1(3), 1-19.

167

Edwards, M. B., Kanters, M. A., & Bocarro, J. N. (2014). Policy changes to implement

intramural sports in North Carolina middle schools: Simulated effects on sports

participation rates and physical activity intensity, 2008-2009. Preventing Chronic Disease,

11(1).

Edwards, M.B., Kanters, M.A., Bocarro, J.N. (2011). Opportunities for extracurricular physical

activity in North Carolina middle schools. Journal of Physical Activity and Health, 8, 597-605.

Eime, R. M., Young, J. A., Harvey, J. T., Charity, M. J., & Payne, W. R. (2013). A systematic

review of the psychological and social benefits of participation in sport for children and

adolescents: Informing development of a conceptual model of health through sport.

International Journal of Behavioral Nutrition and Physical Activity, 10(98), 1-21.

Ekblom-Bak, E., Ekblom, Ö., Andersson, G., Wallin, P., & Ekblom, B. (2018). Physical

education and leisure-time physical activity in youth are both important for adulthood

activity, physical performance, and health. Journal of Physical Activity and Health, 15, 661-

670.

Engstrom, L. M. (1991). Exercise adherence in sport for all from youth to adulthood. In P. Oja

& R. Telama (Eds.), Sport For All. (pp. 473-483). Amsterdam: Elsevier Press.

Eriksson, M., & Lindström, B. (2014). The salutogenic framework for well-being: Implications

for public policy. In J. H. Timo & J. Michaelson (Eds.), Well-Being and Beyond: Broadening

the Public and Policy Discourse (pp. 68-97). Cheltenham: Edward Elgar Publishing.

Esteban-Cornejo, I., Tejero-Gonzalez, C. M., Sallis, J. F., & Veiga, O. L. (2015). Physical activity

and cognition in adolescents: A systematic review. Journal of Science and Medicine in Sport,

18, 534-539.

168

Ewing, M., Gano-Overway, L., Branta, C., & Seefeldt, V. (2002). The role of sports in youth

development. In M. Gatz, M. Messner, & S. Ball-Rokeach (Eds.), Paradoxes of Youth and

Sport (pp. 31-47). Albany, NY: State University of New York Press.

Fairclough, S. J., & Stratton, G. (2005). Physical activity levels in middle and high school physical

education: A review. Pediatric Exercise Science, 17, 217-218.

Fairclough, S. J., & Stratton, G. (2006). A review of physical activity levels during elementary

school physical education. Journal of Teaching in Physical Education, 25, 239-257.

Fenton, S.A., Duda, J. L., Appleton, P. R., & Barrett, T. G. (2017). Empowering youth sport

environments: Implications for daily moderate-to-vigorous physical activity and

adiposity. Journal of Sport and Health Science, 6, 423-433.

Fernhall, B., & Agiovlasitis, S. (2008). Arterial function in youth: Window into cardiovascular

risk. Journal of Applied Physiology, 105, 325-333.

Findlay, L. C., & Coplan, R. J. (2008). Come out and play: Shyness in childhood and the benefits

of organized sports participation. Canadian Journal of Behavioral Science, 40(3), 153-161.

Fleiss, L., Levin, B., Paik, M. C. (2003). The measurement of interrater agreement. In Statistical

methods for rates and proportions, (3rd Eds.) (pp. 598-626). John Wiley & Sons, Inc.:

Hoboken, NJ.

Fleury, J., & Lee, S. M. (2006). The social ecological model and physical activity in African

American women. American Journal of Community Psychology, 37(1-2), 129-140.

Foster, C. (2000). Guidelines for health-enhancing physical activity promotion programmes. The

European Network for the Promotion of Health-Enhancing Physical Activity. University of

Oxford: England. Retrieved from

http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.198.1540&rep=rep1&type=

pdf.

169

Fraser-Thomas, J. L., Côté, J. E., & Deakin, J. (2005). Youth sport programs: An avenue to

foster positive youth development. Physical Education & Sport Pedagogy, 10(1), 19-40.

Fraser-Thomas, J., & Côté, J. (2009). Understanding adolescents’ positive and negative

developmental experiences in sport. Sport Psychologist, 23(1), 3-23.

Fredricks, J. A., & Eccles, J. S. (2004). Parental influences on youth involvement in sports. In M.

R. Weiss (Ed.), Developmental sport and exercise psychology: A lifespan perspective. (pp. 145-164).

Morgantown, WV: Fitness Information Technology.

Friedman, S. L., Belsky, J., & Booth, C. (2003). Frequency and intensity of activity of third-grade

children in physical education. Archives of Pediatric & Adolescent Medicine, 157, 185-190.

Fullerton, A. S. (2009). A conceptual framework for ordered logistic regression models.

Sociological Methods & Research, 38(2), 306-347.

Garcia Bengoechea, E., & Strean, W. B. (2007). On the interpersonal context of adolescents’

sport motivation. Psychology of Sport and Exercise, 8(2), 195-217.

Geidne, S., Quennerstedt, M., & Eriksson, C. (2013). The youth sports club as a health-

promoting setting: An integrative review of research. Scandinavian Journal of Public Health,

41(3), 269-283.

Gilbert, W., Gallimore, R., & Trudel, P. (2009). A learning community approach to coach

development in youth sport. Journal of Coaching Education, 2(2), 1-21.

Gortmaker, S. L., Lee, R. M., Mozaffarian, R. S., Sobol, A. M., Nelson, T. F., Roth, B. A., &

Wiecha, J. L. (2012). Effect of an after-school intervention on increases in children’s

physical activity. Medicine & Science in Sports & Exercise, 44(3), 450-457.

Goudas, M., Dermitzaki, I., Leondari, A., et al. (2005). The effectiveness of teaching a life skills

program in a physical education context. European Journal of Psychology of Education, 21,

429-438.

170

Gould, D., & Carson, S. (2008). Life skills development through sport: Current status and future

directions. International Review of Sport and Exercise Psychology, 1(1), 58-78.

Gould, D., & Walker, L. (2019). Youth sport: unique developmental needs of youth athletes for

preventing dropouts. In M. H. Anshel, T.A. Petrie, & J. A. Steinfeldt (Eds.), APA

Handbook of Sport and Exercise Psychology, (pp. 151-177). American Psychological

Association: Washington, DC.

Guagliano, J. M., Lonsdale, C., Kolt, G. S., Rosenkranz, R. R., & George, E. S. (2015).

Increasing girls’ physical activity during a short-term organized youth sport basketball

program: A randomized controlled trial. Journal of Science and Medicine in Sport, 18(4), 412-

417.

Guagliano, J. M., Lonsdale, C., Kolt, G. S., & Rosenkranz, R. R. (2014). Increasing girls’ physical

activity during an organized youth sport basketball program: A randomized controlled

trial protocol. BMC Public Health, 14(383), 1-11.

Guagliano, J. M., Rosenkranz, R. R., & Kolt, G. S. (2013). Girls’ physical activity levels during

organized sports in Australia. Medicine & Science in Sports & Exercise, 45(1), 116-122.

Habel, M. A., Dittus, P. J., De Rosa, C. J., Chung, E. Q., & Kerndt, P. R. (2010). Daily

participation in sports and students’ sexual activity. Perspectives on Sexual and Reproductive

Health, 42(4), 244-250.

Hallal, P. C., Andersen, L. B., Bull, F. C., Guthold, R., Haskell, W., & Ekelund, U. (2012). Global

physical activity levels: Surveillance progress, pitfalls, & prospects. The Lancet, 380(9838),

247-257.

Hardy, L. L., Hills, A. P., Timperio, A., Cliff, D., Lubans, D., Morgan, P. J., Taylor, B. J., &

Brown, H. (2013). A hitchhiker’s guide to assessing sedentary behavior among young

people: Deciding what method to use. Journal of Science and Medicine in Sport, 16, 28-35.

171

Harwood, C. G., Keegan, R. J., Smith, J. M., & Raine, A. S. (2015). A systematic review of the

intrapersonal correlates of motivational climate perceptions in sport and physical activity.

Psychology of Sport and Exercise, 18, 9-25.

Hay, J., Maximova, K., Durksen, A., et al. (2012). Physical activity intensity and cardiometabolic

risk in youth. Archives of Pediatrics & Adolescent Medicine, 166(11), 1022-1029.

Herrmann, S. D., & Pfeiffer, K. A. (2016). New data for an updated youth energy expenditure

compendium: An introduction. Journal of Physical Activity and Health, 13(Suppl1), S1-S2.

Hirvensalo, M., & Lintunen, T. (2011). Life-course perspective for physical activity and sports

participation. European Review of Aging and Physical Activity, 8, 13-22.

Holt, N. L., & Knight, C. J. (2014). Participation in youth sport. In Parenting in Youth Sport: From

Research to Practice (pp. 16-27). New York: Routledge.

Holt, N. L. (2011a). Positive Youth Development through Sport. New York: Routledge.

Holt, N. L. (2011b). Sport and positive youth development. In Stafford, I. (Ed.) Coaching Children

in Sport, 256-266. London: Routledge.

Hoos, M. B., Kuipers, H., Gerver, W. J., & Westerterp, K. R. (2004). Physical activity pattern of

children assessed by triaxial accelerometry. European Journal of Clinical Nutrition, 58, 1425-

1428.

Holt, N. L., Sehn, Z. L., Spence, J. C., Newton, A. S., & Ball, G. D. C. (2012). Physical education

and sport programs at an inner city school: Exploring possibilities for positive youth

development. Physical Education and Sport Pedagogy, 17, 97-113.

Horn, T.S. (2002). Coaching effectiveness in the sport domain. In T. S. Horn (Ed.), Advances in

sport psychology (2nd Ed.), (pp. 319-354). Champaign, IL: Human Kinetics.

172

Horn, T.S. (2004). Developmental perspectives on self-perceptions in children and adolescents.

In M.R. Weiss (Ed.), Developmental Sport and Exercise Psychology: A Lifespan Perspective (pp.

165-196). Morgantown, WV: Fitness Information Technology.

Hovden, J., & Pfister, G. (2006). Gender, power and sports. Nordic Journal of Women’s Studies,

14(1), 4-11.

Isett, K. R., Mergel, I. A., LeRoux, K., MIschen, P. A., & Rethemeyer, K. R. (2011). Networks in

public administration scholarship: Understanding where we are and where we need to

go. Journal of Public Administration Research and Theory, 21(Suppl 1), i157-i173.

Janssen, I., & Leblanc, A. G. (2010). Systematic review of the health benefits of physical activity

and fitness in school-aged children and youth. International Journal of Behavior Nutrition and

Physical Activity, 7, 40.

Jayanthi, N., Pinkham, C., Dugas, L., Patrick, B., & LaBella, C. (2013). Sports specialization in

young athletes evidence-based recommendations. Sports Health, 5, 251-257.

Jones, G. J., Edwards, M. B., Bocarro, J. N., Bunds, K. S., & Smith, J. W. (2016). An integrative

review of sport-based youth development literature. Sport in Society, 20(1), 161-179.

Kanters, M., McKenzie, T., Edwards, M., Bocarro, J., Mahar, M., & Hodge, C. (2015). Youth

sport practice model gets more kids active with more time practicing skills. Retos, 28,

173-177.

Katzmarzyk, P., Denstel, K. D., Beals, K., Carlson, J., Crouter, S. E., McKenzie, T. L., et al.

(2018). Results from the United States 2018 report card on physical activity for children

and youth. Journal of Human Movement Studies, 15(Suppl 2), S422-S424.

Katzmarzyk, P. T., & Malina, R. M. (1998). Contributions of organized sports participation to

estimated daily energy expenditure in youth. Pediatric Exercise Science, 10, 378-386.

173

Katzmarzyk, P., Walker, P., & Malina, R. (2001). A time-motion study of organized youth sports.

Journal of Human Movement Studies, 40, 325-334.

Kay, T., & Spaaij, R. (2011). The mediating effects of family on sport in international

development contexts. International Review for the Sociology of Sport, 47(1), 77-94.

Kilpatrick, M., Hebert, E., & Bartholomew, J. (2005). College students’ motivation for physical

activity: Differentiating men’s and women’s motives for sport participation and exercise.

Journal of American College Health, 54(2), 87-94.

Kim, Y., Beets, M. W., & Welk, G. J. (2012). Everything you wanted to know about selecting the

‘right’ Actigraph accelerometer cut-points for youth, but …: A systematic review. Journal

of Science and Medicine in Sport, 15(4), 311-321.

Kimm, S.Y., Glynn, N.W., Kriska, A.M., et al. (2002). Decline in physical activity in black girls

and white girls during adolescence. New England Journal of Medicine, 347(10), 709-715.

Kjønniksen, L., Anderssen, N., & Wold, B. (2009). Organized youth sport as a predictor of

physical activity in adulthood. Scandinavian Journal of Medicine & Science in Sports, 19(5),

646-654.

Kohl, H. W., Craig, C. L., Lambert, E. V., Inoue, S., Alkandari, J. R., Leetongin, G., Kahlmeier,

S., & Lancet Physical Activity Series Working Group. (2012). The pandemic of physical

inactivity: Global action for public health. The Lancet, 380(9838), 294-305.

Kriemler, S., Meyer, U., Martin, E., van Sluijs, E. M., Andersen, L. B., & Martin, B. W. (2011).

Effect of school-based interventions on physical activity and fitness in children and

adolescents: A review of reviews and systematic update. British Journal of Sports Medicine,

45, 886-895.

Kuh, D. J., & Cooper, C. (1992). Physical activity at 36 years: Patterns and childhood predictors

in a longitudinal study. Journal of Epidemiology and Community Health, 46(2), 114-119.

174

Kwan, M., Cairney, J., Faulkner, G., & Pullenayegum, E. (2012). Physical activity and other

health-risk behaviors during the transition into early adulthood: A longitudinal cohort

study. American Journal of Preventive Medicine, 42(1), 14-20.

Lai, S. K., Costigan, S. A., Morgan, P. J., Lubans, D. R., Stodden, D. F., Salmon, J., & Barnett, L.

M. (2014). Do school-based interventions focusing on physical activity, fitness, or

fundamental movement skill competency produce a sustained impact in these outcomes

in children and adolescents? A systematic review on follow-up studies. Sports Medicine,

44(1), 67-79.

Langford, R., Bonell, C., Jones, H., & Campbell, R. (2015). Obesity prevention and the health

promoting schools framework: Essential components and barriers to success.

International Journal of Behavioral Nutrition and Physical Activity, 12(1), 15.

Larson, R. W. (2000). Toward a psychology of positive youth development. The American

Psychologist, 55(1), 170-183.

Lee, I. M., Shiroma, E. J., Lobelo, F., Puska, P., Blair, S. N., Katzmarzyk, P. T., & Lancet

Physical Activity Series Working Group. (2012). Effect of physical inactivity on major

non-communicable diseases worldwide: An analysis of burden of disease & life

expectancy. The Lancet, 380(9838), 219-229.

Leek, D., Carlson, J. A., Cain, K. L., Henrichon, S., Rosenberg, D., Patrick, K., & Sallis, J. F.

(2011). Physical activity during youth sports practices. Archives of Pediatric and Adolescent

Medicine, 165(4), 294-299.

Linver, M. R., Roth, J. L., & Brooks-Gunn, J. (2009). Patterns of adolescents’ participation in

organized activities: Are sports best when combined with other activities? Developmental

Psychology, 45, 354-367.

175

Litt, D. M., Iannotti, R. J., & Wang, J. (2011). Motivations for adolescent physical activity. Journal

of Physical Activity and Health, 8, 220-226.

Logan, K., & Cuff, S. (2019). Organized sports for children, preadolescents, and adolescents.

Pediatrics, 143(6), 1-20.

Lonsdale, C., Rosenkranz, R. R., Peralta, L. R., Bennie, A., Fahey, P., & Lubans, D. R. (2013). A

systematic review and meta-analysis of interventions designed to increase moderate-to-

vigorous physical activity in school physical education lessons. Preventive Medicine, 56, 152-

161.

Lonsdale, C., Sabiston, C.M., Raedeke, T. D., Ha, A. S. C., & Sum, R. K. W. (2009). Self-

determined motivation and students’ physical activity during structured physical

education lessons and free choice periods. Preventive Medicine, 48(1), 69-73.

Lunn, P. D. (2010). The sports and exercise life-course: A survival analysis of recall data from

Ireland. Social Science & Medicine, 70(5), 711-719.

MacDonald, L. C. (2015). Moving high school students toward physical literacy. Journal of Physical

Education, Recreation, & Dance, 86(7), 23-27.

Machado-Rodrigues, A. M., Coelho-e-Silva, M. J., Mota, J., Santos, R. M., Cumming, S., &

Malina, R. M. (2012). Physical activity and energy expenditure in adolescent male sport

participants and non-participants aged 13-16 years. Journal of Physical Activity and Health,

9(5), 626-633.

Machado-Rodrigues, A. M., Coelho-e-Silva, M. J., Mota, J., Cumming, S. P., Shearer, L. B.,

Neville, H., & Malina, R. M. (2010). Confounding effects of biologic maturation on sex

differences in physical activity and sedentary behavior in adolescents. Pediatric Exercise

Science, 22, 442-453.

176

Mageau, G. A., & Vallerand, R. J. (2003). The coach-athlete relationship: A motivational model.

Journal of Sports Sciences, 21, 883-904.

Mäkinen, T. E., Borodulin, K., Tammelin, T. H., Rahkonen, O., Laatikainen, T., & Prättälä, R.

(2010a). The effects of adolescence sports and exercise on adulthood leisure-time

physical activity in educational groups. International Journal of Behavioral Nutrition and

Physical Activity, 7(27).

Mäkinen, T., Kestilä, L., Borodulin, K., Martelin, T., Rahkonen, O., & Prättälä, R. (2010b).

Effects of childhood socioeconomic conditions on educational differences in leisure-

time physical activity. The European Journal of Public Health, 20(3), 346-353.

Malina, R. M. (2001a). Physical activity and fitness: Pathways from childhood to adulthood.

American Journal of Human Biology, 13, 162-172.

Malina, R. M. (2001b). Tracking of physical activity across the lifespan. Research Digest: President’s

Council on Physical Fitness and Sports, series 3, no 14.

Malina, R. M. (1996). Tracking physical activity and fitness across the lifespan. Research Quarterly

for Exercise and Sport, 67, 48-57.

Mandic, S., Bengoechea, E., Stevens, E. de la Barra, S., & Skidmore, P. (2012). Getting kids

active by participating in sport and doing it more often: Focusing on what matters.

International Journal of Behavioral Nutrition and Physical Activity, 9(1), 86.

Mandigo, J., Francis, N., Lodewyk, K., & Lopez, R. (2009). Physical literacy for educators.

Physical & Health Education Journal, 75, 27-30.

Marsh, H. (1993). The effects of participation in sport during the last two years of high school.

Sociology of Sport Journal, 10, 18-43.

Martel, K. (2015). USA Hockey’s American Development Model: Changing the coaching and

player development paradigm. International Sport Coaching Journal, 2, 39-39.

177

Matheson, G., Klügl, M., Engebretsen, L., Bendiksen, F., Blair, S. N., Börjesson, M., Budgett, R.,

Derman, W., Erdener, U., Ioannidis, J. P., & Ljungqvist, A. (2013). Prevention and

management of non- communicable disease: the IOC consensus statement, Lausanne

2013. British Journal of Sports Medicine, 47, 1003-1011.

McCambridge, J., Witton, J., & Elbourne, D. R. (2014). Systematic review of the Hawthorne

effect: New concepts are needed to study research participation effects. Journal of Clinical

Epidemiology, 67(3), 267-277.

McGale, N., McArdle, S., & Gaffney, P. (2011). Exploring the effectiveness of an integrated

exercise/CBT intervention for young men’s mental health. British Journal of Health

Psychology, 16(3), 457-471.

McKenzie, T. L. (2019). Physical activity within school contexts: The bigger bang theory.

Kinesiology Review, 8, 48-53.

McKenzie, T. L. (2016). Context matters: Systematic observation of place-based physical activity.

Research Quarterly for Exercise and Sport, 87(4), 334-341.

McKenzie, T. L. (2015). System for observing fitness instruction time: Generic description and

procedures manual. Retrieved from

https://www.researchgate.net/publication/318276946_SOFIT_System_for_Observing_

Fitness_Instruction_Time_Description_and_Procedures_Manual_Generic_Version

McKenzie, T. L. (2009). SOFIT: System for observing fitness instruction time. San Diego State

University: San Diego, CA.

McKenzie, T. L., Catellier, D. J., Conway, T., et al. (2006). Girls’ activity levels and lesson

contexts in middle school PE: TAAG baseline. Medicine & Science in Sports & Exercise,

38(7), 1229-35.

178

McKenzie, T. L., Crespo, N. C., Baquero, B., & Elder, J. P. (2010). Leisure-time physical activity

in elementary schools: Analysis of contextual conditions. Journal of School Health, 80(10),

470-477.

McKenzie, T. L., Feldman, H., Woods, S.E., Romero, K.A., Dahlstrom, V., & Stone, E.J. (1995).

Children’s activity levels and lesson context during third-grade physical education.

Research Quarterly for Exercise and Sport, 66, 184-193.

McKenzie, T. L., Sallis, J.F., & Armstrong, C.A. (1994). Association between direct observation

and accelerometer measures of children’s physical activity during physical education and

recess. Medicine Science in Sports Exercise, 26, S143.

McKenzie, T. L., Sallis, J. F., & Nader, P. R. (1991). SOFIT: System for observing fitness

instruction time. Journal of Teaching in Physical Education, 11, 195-205.

McKenzie, T. L., & van der Mars, H. (2015). Top 10 research questions related to assessing

physical activity and its contexts using systematic observation. Research Quarterly for

Exercise and Sport, 86(1), 13-29.

McKenzie, T. L., & Smith, N. J. (2017). Studies of physical education in the United States using

SOFIT: A review. Research Quarterly for Exercise and Sport, 88(4), 492-502.

McKenzie, T. L., Marshall, S. J., Sallis, J. F., & Conway, T. L. (2000). Student activity levels,

lesson context, and teacher behavior during middle school physical education. Research

Quarterly for Exercise & Sport, 71, 249-259.

McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on

health promotion programs. Health Education Quarterly, 15(4), 351-377.

McPherson, L., Atkins, P., Cameron, N., Long, M., Nicholson, M., & Morris, M. E. (2015).

Children’s experience of sport: What do we really know? Australian Social Work, 1-12.

179

Mears, R., & Jago, R. (2016). Effectiveness of after-school interventions at increasing moderate-

to-vigorous physical activity levels in 5- to 18-year olds: A systematic review and meta-

analysis. British Journal of Sports Medicine, 50, 1315-1324.

Mehtälä, M. A., Sääkslahti, A. K., Inkinen, M. E., & Poskiparta, M. E. (2014). A socio-ecological

approach to physical activity interventions in childcare: A systematic review. International

Journal of Behavioral Nutrition and Physical Activity, 11.

Menard, S. (2010). Logistic Regression: From Introductory to Advanced Concepts and Applications. Los

Angeles, CA: Sage.

Metcalf, B.S., Hosking, J., Jeffery, A.N., Henley, W.E., & Wilkin, T.J. (2015). Exploring the

adolescent fall in physical activity: A 10-yr cohort study (EarlyBird 41). Medicine & Science

in Sports & Exercise, 47(10), 2084-2092.

Michelini, E. (2015). Disqualification of sport in health-related promotion of physical activity: A

global social phenomenon? European Journal for Sport and Society, 12(3), 257-280.

Michelini, E., & Thiel, A. (2013). The acceptance of ‘sport’ in the communication of the health

system. A sociological analysis. European Journal for Sport and Society, 10(4), 325-344.

Mota, J., & Esculcas, C. (2002). Leisure-time physical activity behavior: Structured and

unstructured choices according to sex, age, and level of physical activity. International

Journal of Behavioral Medicine, 9(2), 111-121.

Nader, P. R., Bradley, R. H., Houts, R. M., McRitchie, S. L., O’Brien, M. (2008). Moderate-to-

vigorous physical activity from ages 9 to 15 years. Journal of American Medical Association,

300(3), 295-305.

National Academy of Sciences (2019). Implementing strategies to enhance public health surveillance of

physical activity in the United States. Washington, DC: National Academies of Sciences,

Engineering, and Medicine.

180

National Federation of State High School Associations (NFHS). (2017). 2017 Handbook.

Retrieved from http://www.nfhs.org/ParticipationStatistics/PDF/2016-

17_Participation_Survey_Results.pdf

National Physical Activity Plan Alliance. (2018). The 2018 United States report card on physical activity

for children and youth. Washington, DC.

Nelson, T.F., Stovitz, S.D., Thomas, M., LaVoi, N.M., Bauer, K.W., & Neumark-Sztainer, D.

(2011). Do youth sports prevent pediatric obesity? A systematic review and commentary.

Cur Sports Med Rep, 10(6), 360-370.

Ness, A. R., Leary, S. D., Mattocks, C., Blair, S. N., Reilly, J. J., Wells, J., Ingle, S., Tilling, K.,

Smith, G. D., & Riddoch, C. (2007). Objectively measured physical activity and fat mass

in a large cohort of children. PLoS Medicine, 4(3), e97.

Nielsen, G. A., & Andersen, L. B. (2003). The association between high blood pressure, physical

fitness, and body mass index in adolescents. Preventive Medicine, 36(2), 229-234.

Nielsen, G., Bugge, A., & Andersen, L. B. (2016). The influence of club football on children’s

daily physical activity. Soccer & Society, 17(2), 246-258.

O’Connor, D., & Cotton, W. (2009). Community Junior Sport Coaching: Final Report.

Commissioned by the NSW Department of Sport and Recreation and the Australian

Sports Commission.

O’Connor, D., & Cotton, W. (2010). Community Junior Sport Coaching: Part 2 Final Report.

Commissioned by the NSW Department of Sport and Recreation and the Australian

Sports Commission.

O’Connor, D., & Cotton, W. (2012). Community Junior Sports Coaching. Available from:

http://www.dsr.nsw.gov.au/assets/pubs/industry/community_junior_sport_coaching_r

eport.pdf.

181

Orlick, T. D. (1974). Sport participation—A process of shaping behavior. Human Factors, 16(5),

558, 561.

Orlick, T. D. (1973). Children’s sport—A revolution is coming. Canadian Association of Health,

Physical Education, and Recreation Journal: 12-14.

Owen, N., Healy, G. N., Matthews, C. E., & Dunstan, D. W. (2010). Too much sitting: The

population health science of sedentary behavior. Exercise and Sport Sciences Reviews, 38,

105-113.

Pate, R. R., Davis, M. G., Robinson, T. N., et al. (2006). Promoting physical activity in children

and youth: A leadership role for schools: A scientific statement from the American Heart

Association Council on Nutrition, physical activity, and Metabolism in Collaboration

with the Councils on cardiovascular disease in youth and cardiovascular nursing.

Circulation, 114, 1214-1224.

Peralta, L.R., O’Connor, D., Cotton, W.G., & Bennie, A. (2014). The effects of a community

and school sport-based program on urban indigenous adolescents’ life skills and physical

activity levels: The SCP case study, Health, 6, 2469-2480.

Perkins, D. F., Jacobs, J. E., Barber, B. L., & Eccles, J. S. (2004). Childhood and adolescent

sports participation as predictors of participation in sports and physical fitness activities

during young adulthood. Youth & Society, 35, 523-529.

Petit, M. A., McKay, H. A., MacKelvie, K. J., Heinonen, A., Khan, K. M., & Beck, T. J. (2002).

A randomized school-based jumping intervention confers site and maturity-specific

benefits on bone structural properties in girls: A hip structural analysis study. Journal of

Bone Mineral Research, 17(3), 363-372.

182

Petitpas, A. J., Cornelius, A. E., Van Raalte, J. L., & Jones, T. (2005). A framework for planning

youth sport programs that foster psychosocial development. The Sport Psychologist, 19, 63-

80.

Pfeiffer, K. A., Dowda, M., Dishman, R. K., McIver, K. L., Sirard, J. R., Ward, D. S., & Pate, R.

R. (2006). Sport participation and physical activity in adolescent females across a four

year period. Journal of Adolescent Health, 39, 523-529.

Physical Activity Council (2019). 2019 Physical Activity Council’s overview report on U.S. participation.

Jupiter, FL: Sports Marketing Surveys USA.

Pope, R.P., Coleman, K.J., Gonzalez, E.C., Barron, F., & Heath, E.M. (2002). Validity of a

revised System for Observing Fitness Instruction Time (SOFIT). Pediatric Exercise Science,

14, 135-146.

Priest, N., Armstrong, R., Doyle, J., et al. (2008). Policy interventions implemented through

sporting organisations for promoting health behaviour change. Cochrane Database of

Systematic Reviews.

Project Play. (2014). espnW/Aspen Institute Project Play survey of parents on youth sport

issues. Washington, DC: Aspen Institute. Available online at:

http://espn.go.com/espnw/w-in-action/article/11675649/parents-concern-grows-kids-

participation-sports.

Ribeyre, J., Fellmann, N., Montaurier, C., Delaitre, M., Vernet, J., Coudert, J., & Vermorel, M.

(2000). Daily energy expenditure and its main components as measured by whole-body

indirect calorimetry in athletic and non-athletic adolescents. British Journal of Nutrition, 83,

355-362.

183

Richard, L., Gauvin, L., & Raine, K. (2011). Ecological models revisited: Their uses and

evolution in health promotion over two decades. Annual Review of Public Health, 32, 307-

326.

Ridley, K. (2013). Youth compendium of energy expenditures update: Systematic review. Retrieved from

https://www.nccor.org/tools-youthcompendium/downloads/

Ridley, K., Ainsworth, B. E., Olds, T. S. (2008). Development of a compendium of energy

expenditures for youth. International Journal of Behavioral Nutrition and Physical Activity, 5, 1-

Ridley, K., Zabeen, S., & Lunnay, B. K. (2018). Children’s physical activity levels during

organized sports practices. Journal of Science and Medicine in Sport, 21, 930-934.

Rocchi, M. A., Pelletier, L. G., & Couture, A. L. (2013). Determinants of coach motivation and

autonomy supportive coaching behaviours. Psychology of Sport & Exercise, 14, 852-859.

Rowe, P. J., Schuldheisz, J. M., & van der Mars, H. (1997). Measuring physical activity in physical

education: Validation of the SOFIT direct observation for use with first to eighth grade

students. Pediatric Exercise Science, 9, 136-149.

Rowe, P., van der Mars, H., Schuldheisz, J.M., & Fox, S. (2004). Measuring physical activity in

physical education: Validating SOFIT for use with high school students. Journal of

Teaching in Physical Education, 23, 235-251.

Rowe, K., Shilbury, D., Ferkins, L., & Hinckson, E. (2013). Sport development and physical

activity promotion: An integrated model to enhance collaboration and understanding.

Sport Management Review, 16(3), 364-377.

Rutten, E. A., Stams, G. J. J., Biesta, G. J., Schuengel, C., Dirks, E., & Hoeksma, J. B. (2007).

The contribution of organized youth sport to antisocial and prosocial behavior in

adolescent athletes. Journal of Youth and Adolescence, 36(3), 255-264.

184

Ruxton, G. D., & Beauchamp, G. (2008). Some suggestions about appropriate use of the

Kruskal-Wallis test. Animal Behaviour, 76(3), 1083-1087.

Sacheck, J. M., Nelson, T., Ficker, L., Kafka, T., Kuder, J., & Economos, C.D. (2011). Physical

activity during soccer and its contribution to physical activity recommendations in

normal weight and overweight children. Pediatric Exercise Science, 23(2), 281-292.

Sallis, J. F. (2014, March 9). Active living research 101. 2014 Active Living Research Conference.

Presentation conducted from Robert Wood Johnson Foundation, San Diego, CA.

Retrieved from http://activelivingresearch.org/sites/default/files/2014_ALR101_Sallis-

Dill.pdf

Sallis, J. F., Owen, N., & Fisher, E. B. (2008). Ecological models of health behavior. In K.

Glanz, B. K. Rimer, & K. Viswanath (Eds.), Health behavior and health education: Theory,

research, and practice (4th ed., pp. 465-486). San Francisco, CA: Jossey-Bass.

Sallis, J. F., Prochaska, J. J., & Taylor, W. C. (2000). A review of correlates of physical activity of

children and adolescents. Medicine and Science in Sports and Exercise, 32, 963-975.

Sallis, J. F. (1993). Epidemiology of physical activity and fitness in children and adolescents.

Critical Review in Food Science and Nutrition, 33, 403-408.

Samitz, G., Egger, M., & Zwahlen, M. (2011). Domains of physical activity and all-cause

mortality: Systematic review and dose-response meta-analysis of cohort studies.

International Journal Epidemiology, 40(5), 1382-1400.

Sasaki, J. E., Howe, C. A., John, D., Hickey, A., Steeves, J., Conger, S., Lyden, K., Kozey-Keadle,

S., Burkart, S., Alhassan, S., & Bassett Jr., D. (2016). Energy expenditure for 70 activities

in children and adolescents. Journal of Physical Activity and Health, 13(6 Suppl 1), S24-S28.

185

Savage, P. D., Toth, M. J., & Ades, P. A. (2007). A re-examination of the metabolic equivalent

concept in individuals with coronary heart disease. Journal of Cardiopulmonary Rehabilitation

Prevention, 27(3), 143-148.

Scanlan, T., Babkes, M., & Scanlan, L. (2005). Participation in sport: A developmental glimpse at

emotion. In J. Mahoney, J. Eccles, & R. Larson (Eds.), Organized Activities as Contexts of

Development (pp. 275-309). Mahwah, NJ: Erlbaum.

Schlechter, C. R., Guagliano, J. M., Rosenkranz, R. R., Milliken, G. A., Dzewaltowski, D. A.

(2018). Physical activity patterns across time-segmented youth sport flag football

practice. BMC Public Health, 18, 226.

Schlechter, C. R., Rosenkranz, R. R., Milliken, G. A., & Dzewaltowski, D. A. (2017). Physical

activity levels during youth sport practice: Does coach training or experience have an

influence? Journal of Sports Sciences, 35(1), 22-28.

Schwamberger, B., & Sinelnikov, O. (2015). Connecting physical education to out-of-school

physical activity through sport education. Journal of Physical Education, Recreation, and Dance,

86(9), 39-44.

Schwamberger, B., & Wahl-Alexander, Z. (2016). Children’s activity levels and lesson context

during summer swim instruction. Physical Educator, 73(4), 733-744.

Scott, J. J., Morgan, P. J., Plotnikoff, R. C., & Lubans, D. R. (2015). Reliability and validity of a

single-item physical activity measure for adolescents. Journal of Paediatrics and Child Health,

51, 787-793.

Scruggs, P.W. (2007a). Middle school physical education physical activity quantification: A

pedometer steps/min guideline. Research Quarterly for Exercise & Sport, 78(4), 284-292.

Scruggs, P.W. (2007b). Quantifying activity time via pedometers in fifth- and sixth-grade physical

education. Journal of Physical Activity and Health, 4, 215-227.

186

Silva, G., Andersen, L. B., Aires, L., Mota, J., Oliveira, J., & Ribeiro, J. C. (2013). Associations

between sports participation, levels of moderate to vigorous physical activity and

cardiorespiratory fitness in children and adolescents. Journal of Sports Sciences, 31(12),

1359-1367.

Silva, P., Sousa, M., Aires, L., et al. (2010). Physical activity patterns in Portuguese adolescents:

The contribution of extracurricular sports. European Physical Education Review, 16, 171-181.

Sirard, J. R., & Pate, R. R. (2001). Physical activity assessment in children and adolescents. Sports

Medicine, 31, 439-454.

Skille, E. A., & Solbakken, T. (2014). The relationship between adolescent sport participation

and lifelong participation in physical activity in Norway. A Critical Analysis. Scandinavian

Sport Studies Forum, 5, 25-45.

Smith, L., Gardner, B., Aggio, D., & Hamer, M. (2015). Association between participation in

outdoor play and sport at 10 years old with physical activity in adulthood. Preventive

Medicine, 74, 31-35.

Smith, R. E., & Smoll, F. L. (2007). Social-cognitive approach to coaching behaviours. In S.

Jowett, and D. Lavellee (eds.) Social Psychology in Sport (pp. 75-90). Champaign Illinois:

Human Kinetics.

Smith, R., Smoll, F. L., & Cumming, S. P. (2007). Effects of a motivational climate intervention

for coaches on young athletes’ sport performance anxiety. Journal of Sport and Exercise

Psychology, 29(1), 39-59.

Smith, N., Tessier, D., Tzioumakis, Y., Fabra, P., Quested, E., Appleton, P., Sarrazin, P.,

Papaioannou, A., Balaguer, I., Duda, J. L. (2016). The relationship between observed and

perceived assessments of the coach-created motivational environment and links to

athlete motivation. Psychology of Sport and Exercise, 23, 51-63.

187

Smoll, F.L., & Smith, R.E. (2003). Enhancing coaching effectiveness in youth sports: Theory,

research, and intervention. In R.M. Malina & M.A. Clark (Eds.), Youth Sports: Perspectives

for a New Century (pp. 227-239). Monterey, CA: Coaches Choice.

Sport for Development and Peace International Working Group (2008). Harnessing the Power of

Sport for Development and Peace. Toronto, ON: Right to Play.

Stefan, L., Misigoj-Durakovic, M., Devrnja, A., Podnar, Hrvoje, Petric, V., & Soric, M. (2018).

Tracking of physical activity, sport participation, and sedentary behaviors over four years

of high school. Sustainability, 10, 3104.

Stokols, D. (1992). Establishing and maintaining healthy environments: Toward a social ecology

of health promotion. The American Psychologist, 47(1), 6-22.

Stokols, D. (1996). The social ecology of health promotion: Implications for research and

practice. American Journal of Health Promotion, 10(4), 247-251.

Strong, W. B., Malina, R. M., Blimkie, C. J., Daniels, S. R., Dishman, R. K., Gutin, B., et al.

(2005). Evidence based physical activity for school youth. Journal of Pediatrics, 146, 732-

737.

Sullivan, P. J., & Kent, A. (2003). Coaching efficacy as a predictor of leadership style in

intercollegiate athletics. Journal of Applied Sport Psychology, 15, 1-11.

Svoboda, B. (1994). Sport and physical activity as a socialization environment: Scientific Review Part 1.

Strasbourg: Council of Europe.

Taliaferro, L. A., Eisenberg, M. E., Johnson, K. E., Nelson, T. F., & Neumark-Sztainer, D.

(2011). Sport participation by children and parents from low-income families. Psychology of

Sport & Exercise, 12, 490-499.

Tappe, M.K., & Burgeson, C.R. (2004). Physical education: A cornerstone for physically active

lifestyles. Journal of Teaching in Physical Education, 23, 281-299.

188

Tammelin, T., Näyhä, S., Hills, A. P., & Järvelin, M. R. (2003). Adolescent participation in sports

and adult physical activity. American Journal of Preventive Medicine, 24, 22-28.

Teixeira, P. J., Carraca, E. V., Markland, D., Silva, M. N., & Ryan, R. M. (2012). Exercise,

physical activity, and self-determination theory: A systematic review. International Journal of

Behavioral Nutrition and Physical Activity, 9(78), 1-30.

Telama, R., Laakso, L., Yang, X., & Vikari, J. (1997). Physical activity in childhood and

adolescents as predictor of physical activity in young adulthood. American Journal of

Preventive Medicine, 13, 317-323.

Telama, R., Yang, X., Hirvensalo, M., & Raitakari, O. (2006). Participation in organized youth

sport as a predictor of adult physical activity: A 21-year longitudinal study. Pediatric

Exercise Science, 17, 76-88.

Telama, R., Yang, X., Leskinen, E., Kankaanpaa, A., Hirvensalo, M., Tamelin, T., Viikari, J. S., &

Raitakari, O. T. (2013). Tracking of physical activity from early childhood through youth

into adulthood. Medicine and Science in Sports and Exercise, 46(5), 955-962.

The Child and Adolescent Health Measurement Initiative (2017). 2017 National Survey of

Children’s Health (NSCH) data query. Data Resource Center for Child & Adolescent

Health (databased online). Baltimore, MD: Johns Hopkins Bloomberg School of Public

Health. www.childhealthdata.org. Access October 10, 2019.

Thompson, A. M., Baxter-Jones, A. D. G., Mirwald, R. L., & Bailey, D. A. (2003). Comparison

of physical activity in male and female children: Does maturation matter? Medicine and

Science in Sports and Exercise, 35, 1684-1690.

Tremblay, M., & Lloyd, M. (2010). Physical literacy measurement: The missing piece. Physical and

Health Education, (Spring), 26-30.

189

Troiano, R. P., Berrigan, D., & Dodd, K. W., Masse, L. C., Tilert, T., & McDowell, M. (2008).

Physical activity in the United States measured by accelerometer. Medicine & Science in

Sports & Exercise, 40(1), 181-188.

Trost, S. G., Pate, R. R., Saunders, R. P., Ward, D. S., Dowda, M., & Felton, G. (1997). A

prospective study of the determinants of physical activity in rural fifth-grade children.

Preventive Medicine, 26, 257-263.

U.S. Department of Health & Human Services (USDHHS). (2008). Healthy People 2020

objective topic areas and page numbers and summary of objectives. In Healthy People

2020 (pp. 263-272). Retrieved from www.healthypeople.gov

U.S. Department of Health & Human Services (USDHHS), Centers for Disease Control and

Prevention, National Centre for Chronic Disease Prevention and Health Promotion,

Division of Adolescent and School Health. (2010). Strategies to improve the quality of

physical education. United States Government, Washington DC.

U.S. Department of Health and Human Services (USDHHS). (2012). Physical activity guidelines

for American’s midcourse report: Strategies to increase physical activity among youth.

Washington, DC. Retrieved from

https://health.gov/paguidelines/2008/midcourse/pag-mid-course-report-final.pdf.

U.S. Department of Health and Human Services (USDHHS). (2018). Physical activity guidelines

for Americans: 2nd edition. Washington, DC. Retrieved from

https://health.gov/paguidelines/second-edition/.

U.S. Department of Health and Human Services (USDHHS). (2019). National youth sports

strategy. Washington, DC. Retrieved from https://health.gov/paguidelines/youth-

sports-strategy/pdf/National_Youth_Sports_Strategy.pdf

190 van der Horst, K., Paw, M. J., Twisk, J. W., & van Mechelen, W. (2007). A brief review of

correlates of physical activity and sedentariness in youth. Medicine and Science in Sports and

Exercise, 39, 1241-1250. van der Mars, H. (1989). Observer reliability: Issues and procedures. In P.W. Darst, D.B.

Zakrajsek, & V.H. Mancini (Eds.), Analyzing physical education and sport instruction (2nd ed.,

pp. 53-80). Champaign, IL: Human Kinetics.

Van Hoye, A. I., Fenton, S., Krommidas, C., Heuze, J. P., Quested, E., et al. (2013). Associations

between sports participation, adiposity, obesity-related health behaviors in Australian

adolescents. International Journal of Behavioral Nutrition and Physical Activity, 10(1), 1.

Vella, S. A., Cliff, D. P., Magee, C. A., Okely, A. D. (2014). Sports participation and parent-

reported health-related quality of life in children: Longitudinal associations. Journal of

Pediatrics, 164(6), 1469-1474.

Vella, S. A., Cliff, D. P., Okely, A. D., et al. (2013). Associations between sports participation,

adiposity and obesity-related health behaviors in Australian adolescents. International

Journal of Behavioral Nutrition Physical, 10(1), 113.

Vella, S., Oades, L. G., & Crowe, T. P. (2011). The role of the coach in facilitating positive youth

development: Moving from theory to practice. Journal of Applied Sport Psychology, 23: 33-48.

Waddington, I. (2000). Sport and health: A sociological perspective. In J. Coakley & E. Dunning

(Eds.), Handbook of Sports Studies (pp. 408-421). London: Sage.

Warburton, P., & Woods, J. (1996). Observation of children’s physical activity levels during

primary school physical education lessons. European Journal of Physical Education, 1, 56-65.

Warren, J. M., Ekelund, U., Besson, H., Mezzani, A., Geladas, N., & Vanhees, L. (2010).

Assessment of physical activity – A review of methodologies with reference to

epidemiological research: A report of the exercise physiology section of the European

191

Association of Cardiovascular Prevention and Rehabilitation. European Journal of

Cardiovascular Prevention and Rehabilitation, 17, 127-139.

Weinberg, R. S., & Gould, D. (2011). Foundations of sport and exercise psychology. Champaign, IL:

Human Kinetics.

Weiss, M. R., & Ferrer-Caja, E. (2002). Motivational orientations and sport behavior. In T. S.

Horn (Ed.), Advances in sport psychology (2nd Ed.), (pp. 101-183). Champaign, IL: Human

Kinetics.

Whitley, M. A., Massey, W. V., Camire, M., Boutet, M., & Borbee, A. (2019). Sport-based youth

development interventions in the United States: A systematic review. BMC Public Health,

19(89), 1-20.

Wickel, E. E., & Eisenmann, J. C. (2007). Contribution of youth sport to total daily physical

activity among 6-to 12-yr-old boys. Medicine and Science in Sports and Exercise, 39(9), 1493-

1500.

Williams, J. (2010). Applied sport psychology: Personal growth to peak performance. New York: McGraw-

Hill.

World Health Organization. (2003). Health and development through physical activity and sport. Geneva,

Switzerland: WHO Document Production Services.

World Health Organization. (2010). Global recommendations on physical activity for health. Geneva:

World Health Organization.

192

APPENDICES

193

Appendix A: IRB Approval Letter

194

Appendix B: Sport Action Compendium

Boys Sports

Cross Track & Baseball Basketball Football Lacrosse Soccer Swimming Tennis Wrestling Country Field Athlete PA 3.02 3.71 4.84 3.62 3.82 4.05 4.11 3.61 4.62 4.10

R2 41.6 37.3 57.1 18.9 48.5 31.5 57.0 41.0 44.8 39.6

Fitness 5.10 3.56 5.27 3.96 4.62 4.62 4.77 4.55 5.15 4.39

Game Play 2.09 4.99 5.98 4.43 5.41 5.17 5.75 3.28 5.61 4.41

Knowledge 1.58 1.65 1.88 1.79 1.61 1.97 1.57 1.69 2.32 1.67

Mgmt 2.03 2.22 2.64 2.73 2.15 2.31 1.82 1.63 2.82 1.81

Skill 3.26 4.21 5.34 3.95 4.45 4.44 5.52 4.83 5.18 4.84

Other 2.12 2.59 2.54 2.34 2.22 2.22 1.77 2.37 3.60 2.73

Dem Fit 1.76 4.03 5.83 4.24 3.17 4.34 1.94 2.64 1.50 3.16

Instructs 1.91 3.31 2.28 3.17 3.53 4.07 3.13 3.99 3.41 4.21

Manages 3.37 2.33 4.40 3.46 2.81 3.71 3.61 2.59 2.86 2.03

Observes 3.42 4.16 4.97 3.90 4.19 4.40 5.01 3.76 5.52 4.12

Prom Fit 2.28 4.45 4.82 4.43 4.94 4.56 5.64 2.40 4.79 4.58

Other-task 3.24 2.82 5.76 2.72 3.83 2.76 3.26 2.68 4.68 3.77

Girls Sports

Cross Track & Basketball Cheerleading Lacrosse Soccer Softball Swimming Tennis Volleyball Country Field Athlete PA 3.51 2.41 4.35 3.53 4.23 2.76 4.11 3.46 3.24 3.89

R2 43.7 46.5 52.5 39.2 54.3 21.4 82.7 36.0 44.5 39.8

Fitness 4.60 3.61 5.23 3.50 4.50 3.05 5.68 4.20 4.22 3.91

Game Play 5.18 5.00 5.73 4.45 5.34 2.96 5.95 4.16 5.25 5.06

Knowledge 1.73 1.59 1.55 1.55 1.65 1.70 1.59 2.07 1.56 1.65

Mgmt 2.18 2.04 2.31 1.93 2.48 2.05 1.70 2.19 1.99 2.09

Skill 4.21 1.96 4.83 4.04 5.19 3.11 5.82 3.94 4.40 4.38

Other 2.48 1.75 4.41 2.01 2.03 1.59 1.68 1.69 1.86 1.82

Dem Fit 3.84 4.46 5.53 3.28 3.00 2.41 3.44 2.82 4.16 4.98

Instructs 2.94 1.85 2.10 2.86 3.79 2.50 2.82 3.84 2.31 2.85

Manages 2.59 1.99 2.86 3.83 2.84 2.00 2.36 2.40 3.88 3.62

Observes 4.24 2.65 4.75 3.94 4.96 3.34 4.95 3.47 3.41 4.34

Prom Fit 4.95 4.10 4.83 3.93 5.53 3.00 5.68 4.47 4.41 5.43

Other-task 2.58 3.01 5.39 3.79 2.37 1.91 4.49 2.40 3.37 1.99