Coaching behaviors of successful high school girls' basketball coaches

The primary purpose of this study was to systematically examine the behaviors of a successful female and male high school coach of girls' basketball teams and compare to collegiate coach Pat Summitt. A secondary purpose of the study was to examine the correlation between coaches' behaviors and athlete expectancy status. Throughout the season both coaches' verbal and nonverbal behaviors were video recorded over the course of eight practices. A total of 553 minutes consisting of 3,141 of a female coach's practice behaviors and a total of 590 minutes consisting of 3,678 of the male coach's practice behaviors were coded using the Arizona State University of Observation Instrument. Both the female coach and male coach provided instruction more often (female 35.5%, n = 1114, male 29.2%, n = 1072) then any coaching behavior when the categories of preinstruction, concurrent instruction, and postinstruction were combined. Results indicated that the female coach exhibited management (27.9%) more than any other behavior category. The highest category for the male coach was praise (22.1%). The praise to scold ratio for male and female coaches exceeded a 2:1 ratio with male coach exhibiting more praise behaviors and the female coach exhibited less scold behaviors than the male. A Pearson Product Correlation suggested that both coaches' perceptions of athletes' expectancy remained consistent from the beginning to the end of the season. Contrary to predictions, a MANOVA revealed no differences in the quantity or quality of the coaching behaviors that both the female and male coach directed toward high and low expectancy players in the given categories. The implications from this data suggest that successful coaches treated athletes the same in terms of the 13 coaching behaviors and that coaches spend the majority of their time using instruction.

COACHING BEHAVIORS OF SUCCESSFUL HIGH SCHOOL GIRLS' BASKETBALL COACHES by Tedi L. Searle A thesis submitted to the faculty of The University of Utah in partial fulfillment of the requirements for the degree of Master of Science Department of Exercise and Sport Science The University of Utah May 2012 Copyright © Tedi L Searle 2012 All Rights Reserved Th e Uni v e r s i t y o f Ut a h Gr a dua t e S cho o l STATEMENT OF THESIS APPROVAL The thesis of Tedi L Searle has been approved by the following supervisory committee members: James Hannon , Chair Feb. 15, 2012 Date Approved Barry Shultz , Member Feb. 15, 2012 Date Approved Nicole Miller , Member Feb. 16, 2012 Date Approved and by Barry Shultz , Chair of the Department of Exercise and Sport Science and by Charles A. Wight, Dean of The Graduate School. ABSTRACT The primary purpose of this study was to systematically examine the behaviors of a successful female and male high school coach of girls' basketball teams and compare to collegiate coach Pat Summitt. A secondary purpose of the study was to examine the correlation between coaches' behaviors and athlete expectancy status. Throughout the season both coaches' verbal and nonverbal behaviors were video recorded over the course of eight practices. A total of 553 minutes consisting of 3,141 of a female coach's practice behaviors and a total of 590 minutes consisting of 3,678 of the male coach's practice behaviors were coded using the Arizona State University of Observation Instrument. Both the female coach and male coach provided instruction more often (female 35.5%, n = 1114, male 29.2%, n = 1072) then any coaching behavior when the categories of preinstruction, concurrent instruction, and postinstruction were combined. Results indicated that the female coach exhibited management (27.9%) more than any other behavior category. The highest category for the male coach was praise (22.1%). The praise to scold ratio for male and female coaches exceeded a 2:1 ratio with male coach exhibiting more praise behaviors and the female coach exhibited less scold behaviors than the male. A Pearson Product Correlation suggested that both coaches' perceptions of athletes' expectancy remained consistent from the beginning to the end of the season. Contrary to predictions, a MANOVA revealed no differences in the quantity or quality of the coaching behaviors that both the female and male coach directed toward high and low expectancy players in the given categories. The implications from this data suggest that successful coaches treated athletes the same in terms of the 13 coaching behaviors and that coaches spend the majority of their time using instruction. iv TABLE OF CONTENTS Chapter Page ABSTRACT .................................................................................................................. iii LIST OF TABLES ....................................................................................................... vii CHAPTERS 1 - INTRODUCTION ..................................................................................................... 1 Purposes of the Study.......................................................................................... 7 Questions and Research Hypotheses .................................................................. 8 Delimitations ....................................................................................................... 9 Limitations ........................................................................................................ 10 Assumptions ...................................................................................................... 10 Definition of Terms........................................................................................... 11 2 - REVIEW OF LITERATURE .................................................................................. 13 Coaching Behaviors .......................................................................................... 13 Studies on Expert/Successful Coaches ............................................................ 19 The Effect of Gender on Coaching Behaviors .................................................. 31 Coaching Behaviors and Expectancy Theory ................................................... 34 Summary ........................................................................................................... 41 3 - METHODS .............................................................................................................. 44 Participants ........................................................................................................ 44 Measures ........................................................................................................... 47 Procedures ......................................................................................................... 49 Data Analyses ................................................................................................... 52 4 - RESULTS AND DISCUSSION .............................................................................. 56 Results ............................................................................................................... 57 Discussion ......................................................................................................... 71 5-SUMMARY, FINDINGS, CONCLUSIONS, AND RECOMMENDATIONS FOR FUTURE RESEARCH .................................................................................... 90 Summary ........................................................................................................... 90 Findings............................................................................................................. 91 Conclusions ....................................................................................................... 92 Limitations ........................................................................................................ 93 Recommendations for Future Research ............................................................ 93 Recommendations for Coaches......................................................................... 96 Appendices A - COACH INTERVIEW .............................................................................. 98 B - ATHLETE INFORMATION ................................................................... 100 C - DEFINITIONS, CATEGORIES, AND CODING FORMS OF ARIZONA STATE UNIVERSITY OBSERVATION INSTRUMENT ....................................................................... 102 D - MODIFIED EXPECTANCY RATING SCALE ..................................... 105 REFERENCES ........................................................................................................... 107 vi LIST OF TABLES Table Page 1. Coaching Behavior Totals, Percentages, and RPM for the Female and Male Coach ........................................................................................................................ 58 2. Totals for Male and Female Coaching Behaviors Toward Individuals versus Team for Each Category ........................................................................................... 59 3. Correlations Between Pre- and Post MERS, Coach Rank, and Average Playing Time for the Male and Female Coach ......................................................... 62 4. Coaching Behaviors for Male and Female Coach towards High and Low Expectancy Athletes.................................................................................................. 64 5. Coaching Behaviors for Female Coach towards High and Low and Other Expectancy Athletes................................................................................................. 66 6. Shapiro-Wilks of Female and Male Coach ............................................................... 70 7. Summary Comparison of Coaching Behaviors: High School Coaches and Collegiate Coaches.................................................................................................... 77 CHAPTER 1 INTRODUCTION Regardless of sport or level of competition, successful coaches are often easily identified but rarely understood. Why are some coaches consistently successful? What factors contribute to coaches creating successful athletes and teams with a mixture of ability? Clearly, in the case of sport the end is easily discerned whereas the means to that end are usually unknown or speculated upon. More often attention is directed toward the game behaviors and results rather than the hundreds of practice hours and processes that preceded success. Although many see coaches' behaviors on game day, the behaviors that happen on a more consistent basis during practices may be more predictive of athlete and team success. Because coaching behaviors during the workout are directly related to the quality of the workout, it is logical to assume that the practice behaviors of the coach also directly influence the quality of the athletic performance in competition (Cratty, 1989). In general, coaches aim to maximize the performance of athletes, and some coaches are known for their ability to achieve this desired outcome and be successful. Due to these known results, the study of successful coaches has become a focal point of coaching research. The word "success" has been given various meanings in regards to coaching. In order for researchers to study a successful coach and discover their behaviors and processes, it is a prerequisite to establish what a successful coach is, thus leading to defining the concept of success relative to sport. Hansen, Gilbert, and Hamel 2 (2003) described successful coaches in terms of coaching experience, career winning percentage, and awards won. Gallimore and Tharp (2004) qualified John Wooden as a successful coach because of the 10 National Collegiate Athletic Association (NCAA) championships his University of California Los Angeles (UCLA) basketball teams won with varying degrees of team talent. Pat Summitt has also been recognized as a successful coach since becoming the winningest coach in NCAA Division I basketball history (1035-196). Claxton (1988) defined successful coaches as having career winning records of 70% or higher. Clearly, the criteria used to determine success in coaching has been defined in a multiplicity of ways. However, in today's world, teams and coaches are not considered successful unless they win and many coaches will state they are trying to win, thus the meaning of success usually has been related to win-loss records and championships/titles won (Becker & Wrisberg, 2008; Lacy & Goldston, 1990; Selby, 2009). Although a specific recipe for building winning teams remains unknown, findings do suggest that coaches' behaviors influence and contribute to team performance success (Côté, Salmela, Trudel, Baria, & Russell, 1995; Horn, 2008), and "the ultimate failure to win must rest with the coach who creates the program and directs the game plan" (Wooden, 2004, p. 114). Throughout history, some coaches have been recognized for their unique ability to achieve success. Therefore, it could be deemed necessary to examine the behaviors of experienced, successful coaches to fully understand their behaviors, which may have impacted their strategies that led them to success. Lacy and Darst (1985) also suggested that to have a greater understanding of the strategies employed it would seem beneficial if coaching research were focused on the behaviors of 3 winning coaches because coaching behaviors can be related to outcomes in the form of wins and losses. John Wooden, named the greatest college coach of the 20th Century by the Entertainment and Sports Programming Network (ESPN), received attention from the general public and researchers for his coaching successes by consistently producing winning teams. Tharp and Gallimore (1976) conducted a study of John Wooden, then basketball coach of the UCLA Bruins, with use of a 10-category system they devised for systematic observation in a teaching/coaching setting. Their purpose was simple: research the practices of a master teacher whose credentials and accomplishments warranted a claim of exemplary practice to generate new hypotheses and investigative avenues (Lacy & Darst, 1985). The researchers observed John Wooden for 30 hours during 15-practice sessions spread over the 1974-1975 season. After analyzing the 2,326 coded coaching behaviors of Wooden, the investigators reported that more than half (50.3%) of his behaviors were in an instructional nature. More recently, Pat Summitt, women's basketball coach at the University of Tennessee, became the coach with the most wins in NCAA Division I basketball history and represents another coaching success with 34 winning seasons, 26 conference championships, and 8 national titles. Through the use of the Arizona State University Observation System (ASUOI), researchers were able to identify and compare her coaching behaviors to those of John Wooden and other successful coaches. Consistent with previous research on successful coaches, Coach Summitt provided instruction more frequently (48%) than any other coaching behavior (Becker & Wrisberg, 2008; Bloom, Crumpton, & Anderson, 1999; Kahan, 1999; Lacy & Darst, 1985; Segrave & Ciancio, 4 1990; Solomon, Striegel et al., 1996; Tharp & Gallimore, 1976). Virtually all of the research that has been conducted about coaching in sport within the last three decades has been motivated by a desire to identify the particular behaviors and leadership styles that are most effective (Chelladurai, 1984; Martin, Jackson, Richardson, & Weiller, 1999; Mondello & Janelle, 2001; Riemer & Chelladurai, 1998; Smith, Smoll, & Hunt, 1977). Although some systematic observation research has been conducted in an attempt to identify behaviors, additional investigation is needed to gain a greater and more in-depth understanding of this topic. Cusion and Jones (2001) stated that the emergence and expansion of the use of descriptive analytical techniques has led to the start of a coaching science (Mesquita, Sobrinho, Rosado, Pereira, & Milistetd, 2008). Based on a review of the coaching literature, Gilbert and Trudel (2004) claimed that coaching, as in other fields, requires descriptive studies for the basic understanding and accumulation of knowledge. Use of systematic observation of coaching behaviors would allow for objective analysis and comparisons of coaching behaviors to be made between successful male and female coaches working with athletes at various competitive levels. In addition, Cote' and colleagues (1995) stressed the need to examine the behaviors of successful coaches on a deeper level as a result of the growing coaching profession. The focus has been more on the behaviors of coaches rather than on the knowledge behind the behaviors and instructional psychology (Gallimore & Tharp, 2004). Additional scientific analysis of the factors effecting successful coaching behaviors could enhance the field of pedagogy and sport psychology. Although there has 5 been a recent increase in research dedicated to coaching behaviors, a limited amount of literature exists on the function of gender and expectancy theory within coaching. One factor that studies have attributed specifically to coaches' behaviors has been the coaches' gender. Early investigations were predominantly completed on male coaches, which limited available understanding of coaching behaviors. Eitzen and Pratt (1989) conducted a descriptive study of female basketball teams to investigate gender differences in coaching philosophy. They asserted that male and female coaches may differ behaviorally to some extent as a result of their gender, but their findings indicated that regardless of gender, the coaches' philosophies' remained similar in most areas. Lacy and Goldston (1990) found only slight differences in analyzing the behaviors of male and female coaches through the use of systematic observation. In examining the knowledge of high-performance gymnastic coaches, Cote et al. (1995) found that the differences between male and female coaches were neither practically nor statistically significant, but were largely representative of individual styles rather than being a result of gender. Although these studies and other coaching research have shed valuable light on coaching behaviors and gender, the existing literature is still substantially lacking. Another factor that has been suggested as contributing to coaching behaviors but has inadequate research to date has been whether coaching behaviors are influenced by the coaches' perceptions of the players' athletic abilities and skills. This relationship between coach expectations and athlete performance, a theoretical framework known as Expectancy Theory, has been described in the sport psychology literature as a four-step process (Horn, Lox, & Labrador, 2006; Solomon, 2001). In step one, the coach assesses the athlete's ability and skill, then establishes expectations for performance based on 6 three types of impression cues: personal, performance, and psychological. In step two, the coaches' expectations influence how they behave toward the athlete. This is usually observed through the quantity and quality of feedback. In step three, the athlete becomes aware of the coach's treatment, which can then subsequently affect the athlete's own self-perceptions and behaviors. In step four, the athlete's performance conforms to the coach's original expectations (e.g., high expectancy athletes typically outperform their low expectancy counterparts). These performance outcomes reinforce the coach's belief that their initial assessment of the athlete's ability was accurate. Although limited research has directly examined coaching behaviors within the framework of Expectancy Theory, a number of studies have suggested that athletes are clearly influenced by their coach's feedback (Allen & Howe, 1998; Amorose & Smith, 2003; Black & Weiss, 1992; Bloom et al., 1999; Markland & Martinek, 1988; Summers, 1991; Tharp & Gallimore, 1976). Feedback has been shown to be one motivational tool to help athletes strive for a specific goal (Magill, 1994; Smith, 2001). For example, athletes are more likely to experience feelings of success and competence when they are provided with encouragement and instruction than when they are repeatedly criticized (Black & Weiss, 1992; Smith, 2001). Likewise, research has revealed that youth athletes preferred coaches who responded with feedback to mistakes using encouragement and technical instruction (Smith, 2001; Smith & Smoll, 1990; Smoll, Smith, & Curtis, 1978; Smith, Zane, Smoll, & Coppel, 1983). Results of one study indicated that female and male varsity athletes who performed on coactive, mixed, and interactive sport teams preferred feedback from coaches that was supportive and instructional, as opposed to nonresponses or negative responses (Kravig, 2003). 7 In an effort to help improve the pedagogical practice of coaching, it seems beneficial for both coaches and researchers to examine the behaviors and factors that contribute to successful coaches. If researchers are to aid coaches in the practice of effective coaching techniques, more must be learned about behaviors and the factors that relate to coaches' exhibition of desirable behaviors found through systematic observation. Like any profession, coaches can adapt their behaviors to become more effective by gaining an awareness of their behavioral patterns (Lacy & Darst, 1985). Essentially, more effective coaching may result from coaching research (Claxton, 1988). "Although much research has been completed on teaching behaviors and student activities in the physical education environment, only a limited number of studies have focused on the athletic setting" (Lacy & Martin, 1994, p. 95). Purposes of the Study The first purpose of this study was to identify instructional and non-instructional coaching behaviors of a successful male and female head coach of high school girls' basketball teams. The second purpose of this study was to investigate the theory of Expectancy (Horn, 1984) by relating coaches' expectancy status (high vs. low) of athletes to coaching behaviors for successful high school coaches in basketball team practice settings. This study replicated the purposes and investigation methods of a study on Pat Summitt, a successful collegiate female head coach of the women's basketball team at the University of Tennessee (Becker & Wrisberg, 2008). 8 Research Questions and Hypotheses This study attempted to answer the following research questions: Research Question 1: How frequently do successful coaches of female high school basketball teams use instructional coaching behaviors in team practice settings? Hypothesis 1: Coaches will provide Instruction (preinstruction, concurrent instruction, postinstruction) at a higher rate than any other coaching behavior. Research Question 2: How does gender relate to coaching behaviors for successful girls' basketball coaches at high school team practice settings? Hypothesis 2: There will be no significant differences between male and female coaches in terms of coaching behaviors at high school basketball practice settings. Research Question 3: How does a high school female and male coach differ from a female college coach, specifically Pat Summitt, in terms of coaching behaviors in basketball practice settings? Hypothesis 3: It is hypothesized that coaching behaviors of an instructional category demonstrated by the high school coaches will be slightly lower than those demonstrated by collegiate coach, Pat Summitt. Specifically, instruction will be the most prominent coaching behavior but demonstrated at lower frequency than demonstrated by Pat Summitt. It is also hypothesized that coaching behaviors from the noninstructional categories will be demonstrated at a higher frequency than those demonstrated by Pat Summitt. Research Question 4: How does the expectancy status of athletes compare from the beginning of the basketball season to the end of the season for coaches at the high school level? 9 Hypothesis 4: Based on the results of previous expectancy research, it is hypothesized that both coaches' perceptions of their players' skills and abilities will remain consistent over the course of the season. Research Question 5: How does expectancy perceptions influence coaching behaviors, specifically feedback, of successful coaches in a high school practice setting? Hypothesis 5: Coaches will provide differential treatment in the form of feedback based on perceived athlete expectancy. High expectancy athletes at the high school level will receive a greater quantity and quality of feedback than lower expectancy athletes. Research Question 6: Do female and male coaches differ in terms of their treatment of low and high expectancy athletes? Hypothesis 6: There will be no significant difference between genders in terms of differential treatment of athletes for successful basketball coaches at team practice. Delimitations This study was delimited as follows: 1. The sample population selected for this study was delimited to coaches of female high school basketball teams in the state of Utah. Female, as opposed to male basketball teams, were selected based upon the protocol of a previous study conducted with a collegiate female basketball team with Coach Pat Summitt. The sample does represent the nature of high school basketball teams and can be generalized, in this respect, only to coaches of female high school basketball teams. 10 2. Participants were delimited to coaches at a 5A high school. 3. Participants were observed during drills, half-court work, and full-court work, including observations of warm-ups, conditioning, and interpersonal interactions during practices. Limitations The limitations of this study were as follows: 1. The ability to generalize to the entire population of coaches of various age ranges from youth to professional is limited. 2. The results of this study may not be generalized to coaches of male athletes. 3. This study was limited to an analysis of descriptive data, not to attach any evaluative judgments. Any qualitative conclusions based on this research were merely speculative. 4. Behaviors are limited to those exhibited in a team sport setting, specifically basketball, and may not apply to coaches of individual or other team sports. Assumptions The following assumptions served as the basis of conduct for this study: 1. It is assumed that the participants were representative of population of successful high school male and female coaches. 2. It is assumed that the participants did not alter their normal behavior due to the presence of video and voice recording equipment. 3. It is assumed that all participants understood and answered all questionnaires as honestly and accurately as they could. 11 4. It is assumed that sample practice behaviors are representative of all practice behaviors. Definition of Terms For the purpose of this study, the following terms had special meaning and were defined: Behaviors: "Observable verbal and nonverbal actions usually measured by commonly accepted standards" (Dictionary.com, retrieved March 30, 2011 from http://dictionary.reference.com/browse/behaviors). Coactive: Sport teams with low interdependence, which denoted tasks performed by members of the team require little interaction among athletes for success (Goldman, Stockbauer, & McAuliffe, 1977). Coactive sports include bowling, golf, and wrestling. Feedback: "Any procedures used to inform a learner whether an instructional response is right or wrong" (Kepner, 1991, p. 141). Interactive: Sport teams with high interdependence, which signifies that the tasks performed by members of the team require considerable interaction among athletes for success (Cox, 1990). Interactive sports are those such as basketball, soccer, and volleyball. Leadership Style: Leadership has been defined as the attempt to influence the behavior of an individual or group (Hersey & Blanchard, 1982). Leadership styles in sport refers to the actual, preferred, or type of coaching behaviors that produce the desired performance outcomes of athletes (Chelladurai, & Carron, 1983; Chelladurai & Saleh, 1980). 12 Sport Pedagogy: "The scientific study of teaching and coaching, and the content of what was taught by those teachers and coaches" (Siedentop, 1990, p. 274). CHAPTER 2 LITERATURE REVIEW In this chapter, the relevant literature concerning the growing field of coaching science and the factors that have been applied to behaviors of coaching were reviewed. The following review of literature is topically structured to include information to date in several areas. The areas are related to behaviors demonstrated by successful coaches during practices and the factors affecting those behaviors. Thus, this review examines relevant research on coaching behaviors, studies on expert and successful coaches, gender effects on coaching behaviors, and coaching and Expectancy Theory. The review began by establishing what has been done in the coaching field as it pertains to pedagogy and sport psychology. Coaching Behaviors Coaching an athletic team at any level has often been considered a teaching experience (Selby, 2009). Coaches, like teachers, spend considerable amounts of time with the challenge of conveying knowledge by teaching physical skills and strategies, motivating effort, correcting errors, and developing confidence. Indeed like a teacher, instruction has been considered to be the most significant aspect of a coach's role (Tinning, 1982). Several studies have offered interesting implications for the understanding of the coaches' role as a teacher. Cote' and colleagues (1995) explained in 14 their qualitative study on gymnastic coaches, "like teachers, the coach's job is to transmit and transform a collective body of knowledge and skills on a given subject in order to help athletes acquire and use that knowledge in various situations" (p. 66). In a recent study (Selby, 2009) one coach described his role in the following manner: "A coach is just a teacher, and your responsibility is to teach the youngsters under your supervision how to take and execute to the best of their born ability…." (p. 81). John Wooden was widely regarded as the greatest teacher of basketball at his retirement following the 1975 season. In the notable research on Wooden, Gallimore and Tharp (2004) and Tharp and Gallimore (1976) described their purpose as an investigation of the practices of a master teacher. During this investigation for educational purposes they realized that Wooden took advantage of practically all situations as teaching opportunities. Wooden described running a practice session as the same as teaching an English class… "I knew a detailed plan was necessary in teaching English, but it took a while before I understood the same thing was necessary in sports. Otherwise, you waste an enormous amount of time, effort, and talent" (Wooden, 1997, p. 132). He felt that it was the teaching in practice that was more valuable than the games and winning (Gallimore & Tharp, 2004). Therefore, being a teacher was one of the defining roles of a coach, thereby making them responsible for transmitting to the athlete what to do, how to do it, and how to do it well (Hodges & Frank, 2002; Selby, 2009). Because the act of coaching has often been studied with respect to the act of teaching, it has focused primarily on behaviors (Bloom et al., 1999; Claxton, 1988; Lacy & Darst, 1985; Tharp & Gallimore, 1976). More and Franks (1996) suggested that the discovery of how coaches facilitated learning for athletes, the central principle of a 15 coach's instructional behavior, could be discovered through analysis of coaching behaviors. Most of the investigations of coaching behaviors have been descriptive in nature through the use of systematic observation developed from the educational field. The use of systematic observation procedures has arguably contributed more to the understanding of teacher effectiveness than any other single pedagogical development (Darst, Mancini, & Zakrasjek, 1983). Consequently, many of the observation systems developed for teaching effectiveness have been used to learn how to develop and improve coaching techniques. Lacy and Darst (1985) suggested that when coaches have an awareness of their own behavioral habits, they could adapt those behaviors to become more effective coaches. In order to more clearly identify, understand, and analyze effective coaching behaviors, systematic observation instruments for collecting objective and quantifiable data on coach behaviors were developed (Bloom et al., 1999; Darst et al., 1983). Bloom and colleagues (1999) interpreted systematic observation as a method that "allows a trained person following stated guidelines and procedures to observe, record, and analyze interactions with the assurance that others viewing the same sequence of events would agree with his or her recorded data" (p. 157). This system has been accepted and used throughout education including the field of physical education and more recently coaching to objectively observe behaviors (Bloom et al., 1999; Claxton, 1988; Lacy & Darst, 1985). Several researchers have developed models and instruments to measure the influence of coaching behaviors and leadership styles on athletes (Chelladurai & Saleh, 1978, 1980; Martin & Barnes, 1999; Smith et al., 1977). Despite some recent criticism in favor of more qualitative approaches to investigate coaching behaviors, systematic 16 observation still has a very important role to play in developing representative evidence-based guidelines to good practice (Potrac et al., 2007). There are six steps in the process of systematic observation (Van der Mars, 1989). First, the context and participants for observation must be chosen. Second, definitions must be generated and agreed upon to describe observed behavior. Third, the researcher must select the most appropriate observational method and/or tool. Fourth, training procedures must be employed among the observers to establish adequate interobserver reliability. Fifth, observations of events are recorded. Finally, the observation data obtained were analyzed via statistical procedures (ASUOI, Lacy & Darst, 1984). One of the earlier and most notable systematic observations came from the investigation by Tharp and Gillmore in 1976, which has been used as the basis for numerous coaching and teaching behavior research examinations using systematic observation (Bloom et al., 1999; Claxton, 1988; Lacy & Darst, 1984). Tharp and Gallimore (1976) intended to improve educational research, so they chose to systemically observe the master teacher of basketball, John Wooden. They used a conventional approach to classroom research, establishing categories that captured events and behaviors and then refined them to the point that two people would independently assign the same behavior to the same category (Gallimore & Tharp, 2004). They then sat on the bleachers in Wooden's classroom and designed a system for coding his acts of teaching. The system devised for use, included 10 categories encompassing Wooden's behaviors in addition to an uncodable category. These categories consisted of the following: (a) instructions, (b) hustles, (c) modeling-positive, (d) modeling-negative, (e) praises, (f) 17 scolds, (g) nonverbal punishment, (h) nonverbal reinstruction, (i) scold reinstruction, (j) other, and (k) uncodable. Coaching behaviors have been repeatedly evaluated using modified versions of Tharp and Gallimore's (1976) 10-category system for systematic observation in a teaching/coaching setting (Claxton, 1988; Lacy & Darst, 1985). Langsdorf (1979) developed a modified version of the Tharp and Gallimore instrument (1976) with the addition of two categories to objectively observe the behaviors of Frank Kush, head football coach at Arizona State University. Langsdorf's observational instrument included the addition of two descriptive categories to expand the means for summarizing and interpreting the data by different segments of practice (Darst, Zakrajsek, & Mancini, 1989). Model (1983) completed a similar systematic observation study on six losing high school football coaches. Dodds and Rife (1981) completed a descriptive-analytic study on a highly successful women's field hockey coach using an instrument based on the work of Tharp and Gallimore. Lacy and Darst (1985) used an 11-category instrument using a modified form of Tharp and Gallimore's (1976) 10-category system based on the Langsdorf instrument to research the behaviors of a group of 10 winning high school head football coaches. They did not include an uncodable category; however, they added two new categories called use of first name (using the first name or nickname when speaking directly to a player) and management (verbal statements related to organizational details of a practice session not referring to strategies or fundamentals of the game). From the empirical research derived from Tharp and Gallimore's (1976) observation instrument of Wooden, Lacy and Darst (1985) developed the Arizona State 18 University Observation Instrument, a systematic observation instrument designed specifically for practice settings versus games (ASUOI; Kahan, 1999; Lacy & Darst, 1985). Because informative statements concerning the skills and strategies of the particular sport are crucial to effective teaching in the athletic environment, the ASUOI was expanded and modified to several behavior categories (instructional) to create a more sensitive tool capable of collecting more specific data on coaching behaviors. It included 13 behavioral categories representing three general types of behaviors: instructional (pre-instruction, concurrent instruction, postinstruction, questioning, manual manipulation, positive modeling, negative modeling), noninstructional (hustle, praise, scold, management, other), and dual codes (statements that include the recipient's name). Because the ASUOI categories were specifically defined and obviously related to coaching behaviors, face validity was apparent. Because a rationale existed for the selection of the behavior categories and those behaviors were representative of coaching behaviors as supported by previous research, the instrument also possessed content validity. Interobserver reliability was established using both event and interval recording procedures. Whether event or interval recording was used, data derived from the ASUOI provided quantitative information concerning behaviors exhibited during the observation period. Since the development of the ASUOI, numerous investigations have been conducted using the instrument. Rupert and Buschner (1989) used ASUOI to compare the teaching behaviors in physical education with the coaching behaviors of nine teacher/baseball coaches. Claxton (1988) used the ASUOI to investigate high school tennis coaches who were considered more and less successful. In his investigation, 19 Claxton used the ASUOI to observe nine coaches, using 14 categories instead of 10, dividing the category of instruction into preinstruction, concurrent Instruction, and postinstruction. Also the categories of questioning, manual manipulation, first name, management, and silence were added. Additionally, he investigated tennis coaches and suggested that different sports might require slightly different categories of behaviors. Regardless of the slight variations, through the utilization of the ASUOI, studies have endeavored to compare and contrast the instructional profiles of various successful coaches of different sports. Furthermore, in order to establish a meaningful database of coaching behaviors in a variety of sporting situations, it has been necessary for similar methodologies and instrumentation to be used. If different instruments were used to collect data, behavioral categories and coding techniques could be too dissimilar to make meaningful comparisons and valid conclusions (Lacy & Goldston, 1990). Although researchers have developed a variety of systematic observation instruments to measure coaching behaviors it has been beneficial to study coaches using similar systematic observation systems such as the ASUOI (Kravig, 2003; Lacy & Darst, 1984; Smith et al., 1977; Tharp & Gallimore, 1976). Studies on Expert/Successful Coaches The question has often been asked of how success has been determined and what makes a successful coach. In reference to coaching, success has been subjective according to various criteria that have been used from building character in athletes to winning games. Some coaches have been referred to as being successful because of high athlete graduation rates from school, or the athletes having a positive experience, or athletes having learned life-skills such as work ethic and teamwork, or the athletes 20 attained above average grade point averages. Chen, Jensen, and Mann (2004) offered that success be viewed with regard to coach's sport knowledge, leadership, communication skills, and managerial ability. Coaches and others have offered the importance of developing athletes instead of focusing on winning, thus having athletes who can play the game and also contribute in society (Selby, 2009). A coach offered the opinion in a recent study that when a team reaches their potential that a coach could still be very successful as a coach and not win a championship (Selby, 2009). Perhaps helpful to determining how to define a successful coach, would be to examine the differences between the terms of good, great, and successful. When examining simple dictionary definitions of the following good, great, and successful, it was more apparent how a successful coach may be defined. Good was defined as satisfactory in quality, quantity, or degree and morally excellent (Dictonary.com, 2011). This definition may well describe a coach who teaches the athletes moral principles, and life-skills along with sport skills. Whereas great was defined as unusual or considerable in degree, power, and a person who has achieved importance or distinction in a field, which would describe coaches who have become extremely popular for numerous reasons usually due to either exceptional athlete performances and/or conduct. Again according to a dictionary definition success refers to someone who has succeeded or gained a favorable or desired outcome, it also has referred to someone who has gained wealth, favor, or eminence. From these terms it appeared that a successful coach would be someone who can achieve the desired outcome in sports, which by many standards, though not limited to, would be winning. 21 Although character building of athletes has been expressed as an important element of being a successful coach, when the day has concluded the coaches viewed as successful have been referred to those who win games and championships (Gallimore, & Tharp, 2004; Selby, 2009). This perspective has been observed and understood even within the coaching profession. One coach did not agree that success should have meant winning, but expressed that was the way success has been determined (Selby, 2009): Well, in this job, at this point in time, success is winning national championships and having your players improve to the level for, basically what they want to achieve with their talents. I think that it's a little bit of a problem in college sports these days because success is not really defined by having all your kids grow up, graduate from college, become better people, and blah, blah, blah. I mean, those things are supposed to happen, but no one hires you based on those things happening, and you certainly get fired even if all those things happen, but you don't win. So I think that's a real dilemma in college sports really, quite frankly. All those things are supposed to happen as well as doing a certain amount of winning. (p. 110) Another coach described his analysis of success as the following (Selby, 2009), "Unfortunately, we are all held hostage to winning. I mean, you know, you look at a coach's winning percentage, and championships, and different things like that" (p. 111). In the research, successful coaches have also been defined in a multiplicity of ways. Claxton (1988) defined successful coaches as having career winning records of 70% or higher. Lacy and Goldston (1990) referred to past winning percentage and years of experience of the coaches to determine success. In 2003, Hansen and colleagues described successful coaches in terms of coaching experience, career wins, and awards won. Gallimore and Tharp (2004) depicted John Wooden as successful due to his 10 NCAA championships. Pat Summitt has been expressed to be successful due to becoming the winningest NCAA Division I coach in history (Becker & Wrisberg, 2008). Despite various definitions both in and outside of research, the predominant view of 22 success has dealt with winning, and in today's world most coaches will claim they are trying to do so (Becker & Wrisberg, 2008; Selby, 2009). Successful coaches collectively involve various personalities, behaviors, and instructional leadership styles (Mondello & Janelle, 2001). Bobby Knight was known for his autocratic style (Robinson & Miller, 2003) and John Wooden for his focus on character building (Tharp & Gallimore, 1976). In addition, Pat Summit, who coaches the women's basketball team at the University of Tennessee, demonstrates to her players that she cares about them; however, she still remains a tough disciplinarian (Wrisberg, 1990). Although books, magazines, documentaries, and interviews provide insights into the philosophies and techniques of many great and successful coaches, research studies that systematically examine successful coaches' behaviors and the factors related to those behaviors have been less common. A variety of research methods, coaching populations, and variables have been examined to help identify differences between successful and less successful coaches. A limited amount of research has been conducted on successful coaches including college basketball coach Jerry Tarkanian (Bloom et al., 1999); college football coach Frank Kush (Langsdorf, 1979); John Wooden (Tharp & Gallimore, 1976); and Pat Summitt (Wrisberg, 1990) to observe and empirically evaluate. Coaching and teaching continue to be such complex processes that many different research perspectives have been and will be necessary to fully comprehend them (Gallimore & Tharp, 2004). One of the landmark studies in coaching research was performed to investigate the coaching methods of legendary John Wooden, who led his UCLA men's basketball team to an unprecedented 10 Division I basketball championships in a 12-year period (Tharp & Gallimore, 1976). This noteworthy research was among the first to report 23 systematic observational data through the use of the Tharp and Gallimore's 10-category systematic observation system known as the Coaching Behavior Recording Form (Darst et al., 1989). Tharp and Gallimore (1976) gathered their data by sitting in the front row of bleachers at Wooden's practices, allowing them to see and hear most of his verbal exchanges. Over the course of 30 hours of 15 practice sessions 2,326 of Wooden's behaviors were coded, which revealed that more than half (50.3%) of Coach Wooden's behavior was in the instruction category, which was defined as verbal statements about what to do or how to do it. He also used hustles (12%), praise (7%), and scolds (6%). It was also noted that Wooden rarely used scolds and most of his negative statements were followed up with instruction. In addition, nearly all of Wooden's statements were brief (shorter than 20 seconds in duration). In 2004 the same researchers reflected and reanalyzed their previous study and had a personal interview with Wooden to gain a different perspective of his coaching techniques. They noted that the "Hustles" were used to intensify the learning environment and also was a method of management. They found that they were able to better understand his immense planning context of every detail of every practice, which made possible his concise behavior, including the way he instructed his players in a brief and descriptive fashion (directed at what the player should do rather than merely describing a player's action). They also discovered why the category of praise (7%) might be considered low because John Wooden considered positive coaching behaviors to come in the form of instruction rather than praise (Gallimore & Tharp, 2004). Later, Williams (1978) employed a modified version of the Tharp and Gallimore instrument to systematically observe a successful high school head basketball coach, 24 during which he observed practices twice a week throughout the season and gathered data using event-recording procedures. The results were compared to those of Tharp and Gallimore's study on John Wooden. The findings indicated that the high school coach emphasized instruction, as did Wooden, but used praise (25% of all behaviors) much more frequently than Coach Wooden did. Williams suggested that the differences in maturity, skill level, and motivation of players at the high school and college levels could possibly explain the variance in the use of praise (Lacy & Darst, 1985). In 1979, Langsdorf followed up with a similar descriptive study of coaching behaviors through observation of Frank Kush, the head football coach at Arizona State University. Results of the study showed a variety of Kush's behaviors coded during 18 spring football practices. Of Kush's behaviors, 36% were in the instruction category and hustles, scold/reinstruction, and praise being the next highest occurring behaviors. In addition, 12% of behaviors were in the scold/instruction category (Landsdorf, 1979). Lacy and Darst (1985) observed the behaviors of winning high school head football coaches using a modified form of Tharp and Gallimore's (1976) 10-category system. They did not include an Uncodable category; however, they added two new categories; Use of First Name (using the first name or nickname when speaking directly to a player) and Management (verbal statements related to organizational details of a practice session not referring to strategies or fundamentals of the game). The researchers recorded the behaviors of 10 coaches in three phases over the course of one season. Each coach was observed once during pre-season, early season, and late season. A quantitative analysis was conducted using an analysis of variance with repeated measures, followed by a post hoc Tukey's test. They discovered that coaches used a more intense teaching 25 style during the first part of the year while focusing on basics and fundamentals. Praise was used over twice as much as scold, reinforcing the researchers' opinion that coaches can accomplish more using positive interactions rather than negative interactions. In addition, this study also supported the idea that informal feedback is necessary to be an effective coach or teacher (Lacy & Darst, 1985). Claxton (1988) used the ASUOI to conduct a study to describe and analyze systematically the coaching behaviors of more and less successful high school boys' tennis coaches during practice sessions. Successful coaches were categorized as having career winning percentages of at least 70%, and less successful coaches were labeled as having records under 50%. In Claxton's (1988) research, each coach was observed for three 10-minute periods three times throughout the season (pre-season, mid-season, and late season). The data were recorded according to specific events and the time each behavior lasted. Using a Mann-Whitney Test, the researcher analyzed the event recording data, and the different groups were compared for each behavior using the percentages of all intervals for interval recording analysis. Analysis of the data showed that the more successful coaches asked a significantly greater number of questions (form of instruction) of their players than did the less successful coaches. Of the 4,031 events recorded, preinstruction, concurrent instruction, and postinstruction combined to account for 20.1% of all behaviors, making Instruction the most observed category. Claxton concluded that the less successful coaches used more instruction and praise, but questioned less than successful coaches. He also indicated that the more successful coaches displayed more Silence and Management behaviors, but spent less time on them than the less successful coaches. 26 Lacy and Goldston (1990) conducted a study with the purpose to analyze the behaviors of five male and five female varsity high-school girls' basketball coaches during preseason and in-season practice sessions. Each coach was observed three times during the preseason phase and three times in the in-season phase. Systematic observation data were collected using event recording with 13 behavior categories of the ASUOI. Coaching behaviors were generally consistent across both phases of the season for both genders. The dominant behavior of the coaches observed was verbal instruction. Verbal behavior categories of an instructional nature (about 49.6 %) dominated the observed practice sessions in both phases of the season. The highest individual behavior category for the female coaches was postinstruction (21.3 %) whereas the male subjects employed concurrent instruction (21.8 %) the most often. A second major function was giving encouragement by using the praise and hustle categories to motivate athletes to maintain and intensify their efforts. These two behavior categories totaled 18.5% of all behaviors used by the coaches. These "encouragement" behaviors were exhibited slightly more by the female coaches (20.1 %) than their male counterparts (17.1 %). Both male and female coaches used over twice as many praise behaviors as scold behaviors. The third major area that accounted for an important part of the coaching behaviors observed was that of management. Management accounted for 15.3% of all behaviors for the total season. As with the study conducted by Lacy and Darst (1985) this study indicated a correlation between positive verbal feedback and coaching success. Lacy and Martin's (1994) purpose was to examine starter/nonstarter motor-skill engagement (MSE) and coaching behaviors in different segments of preseason practices in collegiate women's volleyball. The participants were athletes and coaches of eight 27 volleyball teams. Segments of the practice were defined and coded as a warm-up, skill work, scrimmage, or conditioning. Coaching behaviors were coded with interval recording procedures (5-second observe, 1-second code) using an expanded version of the Arizona State University Observation Instrument (ASUOI). They discovered that slightly more behaviors were directed to individuals (26%) than the group (22.6%). Pre-instruction was primarily directed to the group, which demonstrated the typical teaching strategy of explanation and demonstration prior to activity. The only other categories directed more to the group than to an individual were management and hustle. Bloom et al. (1999) studied Fresno State men's basketball coach, Jerry Tarkanian, using systematic observation. They adapted the observation system utilized by Tharp and Gallimore (1976) to create the Revised Coaching Behavior Recording Form. The changes to the original set of categories included Technical Instruction, Tactical Instruction, General Instruction, and Humor. Two researchers conducted a pretest, which consisted of three 2-hour observations, prior to the behavior being formally recorded. Coach Tarkanian did not know he was being observed at the time of the study. During the recording procedure, two observers sat independently of each other in an unobtrusive place. Every observed behavior exhibited by the coach was recorded on the coding sheet. Data were gathered throughout the whole regular season during practices only. The analysis of data involved combining the information from both observers and reaching an average percentage. In addition, the coach and his assistant were both interviewed at the end of the season to confirm that the concluding results were congruent with Coach Tarkanian's actual methods. They also had a chance to ask any questions about the study. In discussing the results, the researchers noted the importance of dividing the 28 Instruction behaviors into three separate categories for this study. After calculating the total number of statements, the observers found Tactical Instruction was used the most, which was 29% of the coded behaviors. They concluded that elite level coaches have a tendency to emphasize the cognitive or tactical aspects of their sport during practice. They also indicated that Tarkanian, like Wooden, did not use physical or negative punishment. Butcher (2003) examined the leadership styles of female collegiate field hockey coaches. Sport researchers have generally accepted the idea that the personalities and leadership styles of coaches influence their teams and individual athletes (Terry, 1984). Research in the sport leadership area has been conducted under the general assumption that the type of leadership behavior exhibited by coaches will have a significant impact on individual athletes and teams (Amorose & Horn, 2001; Martin, Dale, & Jackson, 2001; Vealey, Armstrong, Comar, & Greenleaf, 1998). This study examined veteran and new coaches leadership behaviors of Training and Instruction, Autocratic, Social Support, and Positive Reinforcement. The new coaches exhibited more training and instruction, social support, and positive reinforcement than veteran coaches. In addition, the athletes rated the new coaches significantly higher in displaying the democratic leadership style whereas they rated the veteran coaches as more autocratic. When comparing the leadership styles of successful and unsuccessful coaches, a significant difference was found in training and instruction, social support, and positive reinforcement. Successful coaches used more positive reinforcement and social support, and unsuccessful coaches used training and instruction more. Overall, training and 29 instruction and positive reinforcement were found to be the behaviors exhibited by most coaches regardless of coaching experience. Mesquita and colleagues (2008) conducted systematic observation using the ASUOI of male youth volleyball coaches. They found the combined categories of instruction (35.9%) accounted for the majority of recorded behaviors. However, they showed a lower use of instructional and praise behaviors compared with results of elite-level professional coaches. The data revealed a ratio of 3:1 of praise to scold respectively, demonstrating the use of positive reinforcement as a teaching tool. Additionally, the findings on the use of first names, questioning, and negative/positive modeling, important strategies used in teaching to promote active learning, were very limited in use. Recently Becker and Wrisberg (2008) observed Pat Summitt throughout the 2004-05 season for a total of 504 minutes consisting of 3,296 coded practice behaviors. Few coaching researchers have used comparable research strategies to contrast a highly successful collegiate women's basketball coach with one of the all-time great men's college basketball coach John Wooden. The purpose was to systematically examine Summitt's practice behaviors comparing them to the results of the 30+ year old study of John Wooden and to observe any differential treatment in the form of feedback to athletes. Consistent with previous research on successful coaches, Summitt provided instruction more frequently (48%) than any other coaching behavior (Bloom et al., 1999; Kahan, 1999; Lacy & Darst, 1985; Segrave & Ciancio, 1990; Solomon, Striegel et al., 1996; Tharp & Gallimore, 1976). The most common form of instruction that Summitt provided during practices was concurrent instruction. As players executed various tasks, she frequently provided them with technical and tactical information. Another interesting 30 aspect of Summitt's instructional behavior was the higher frequency of preinstruction (before action) and lower frequency of postinstruction directed toward the team. In the study, nearly half (45%) of Summitt's statements were directed toward individual players. The second most frequent type of feedback Coach Summitt provided during practice sessions was praise (15%). She directed hustle statements toward the team as a whole more often than toward individual athletes. Throughout the course of drills, she utilized hustle statements, which comprised 11% of her overall coaching behaviors, which was similar to John Wooden. Therefore, the researchers concluded that one factor that might have contributed to the success of Summitt was the careful planning of intense, game-like practices. Feedback clearly has been an important component of the coaching process. Despite slight differences and independent of study design or population, the current body of literature gained from systematic observation suggests that successful coaches provide greater amounts of overall feedback in practice than less successful coaches (Becker & Wrisberg, 2008; Bloom et al., 1999; Butcher, 2003; Lacy & Goldston, 1985; Mesquita et al., 2008; Markland & Martinek, 1988; Tharp & Gallimore, 1976). Furthermore, the results of these investigations have demonstrated the category of ‘instruction' as one of the most observed behaviors of feedback for both male and female coaches. Moreover, as Black and Weiss (1992) claim, effective coaches tend to cultivate a more positive environment through the use of praise for their players than their less effective counterparts. Potrac and colleagues (2007) verified the use of praise by top-level coaches as an instructional strategy to enhance self-efficacy and confidence levels of players, and as a valuable tool for reinforcing desired athlete behavior. 31 The Effect of Gender on Coaching Behaviors If coaching educators are to aid coaches in the practice of effective coaching techniques, more must be learned about the factors that relate to coaches' exhibition of desirable behaviors. Of these factors, gender may play a role in the behaviors of coaches. Males may place more value on certain aspects of coaching more than females and vice versa. Little research has been conducted regarding the differences in actual behaviors of male and female coaches. In addition, relating to female athletes and relating to male athletes may require different methods of coaching. Eitzen and Pratt (1989) focused on differences in the coaching philosophy of both male and female coaches of female basketball teams. They used questionnaires to collect the data, which specified five areas of coaching philosophy: (a) the coach's role in the overall development of athletes, (b) conditions believed essential to maximize team performance, (c) team rules used, (d) use of sport aphorisms, and (e) the expectations of the athletes. Six hundred questionnaires were sent to randomly selected high school girls basketball teams, and 250 were used in the study. The significant differences found were minimal. Only one comparison in the overall development of the athletes was statistically different. The female coaches placed a greater importance on helping the athlete develop a positive self-image. In the second area, coaches were found to value the same conditions essential to maximize performance, regardless of gender. In team rules, females tended to have rules focusing on academic performance, practice the day before a game, good sportsmanship, and profanity; male coaches were more likely to include curfews and enforce rules on punctuality relative to practices. In the few 32 differences that existed concerning sport aphorisms, the females placed greater importance on the statements than the male coaches. A rather unique characteristic of competitive girls' basketball has been that both males and females serve as coaches. Using systematic observation, Lacy and Goldston (1985) analyzed the behavior of male and female coaches in high school girls' basketball. The participants included five female and five male coaches in Dallas-Fort Worth, Texas. Each coach was observed three times during the preseason and three times during the in-season, using the ASUOI. Coaching behaviors were generally consistent across both phases of the season for both genders. When analyzed separately, male coaches were found to use concurrent instruction the most, and females exhibited more post instruction. The biggest difference during this part of the season was observed in the Management category, which was used more by the female coaches. When the data were analyzed across the whole season, concurrent instruction was used most frequently and post instruction was ranked second. The greatest disparities between genders were the categories of postinstruction, preinstruction, and management. It was also interesting to note the variability of the genders across the phases in the use of first name category. The female subjects exhibited this behavior at a .97 RPM in the preseason and the male coaches used first names from .50 RPM in the preseason. During the in season phase, the first name usage was almost identical for both genders. Although there were slight variations in behaviors, the gender differences of displayed behaviors were not considered significant. In the studies of Dubois (1981) and Millard (1990), the male coaches gave more technical instruction and less encouragement than did the female coaches. However, the 33 gender differences noted in those studies might have been confounded by other related factors. As an example, Sherman and Hassan (1986) reported that high-experience coaches gave more technical instruction than did low-experience coaches. Indeed, males in the Millard study had over twice the years experience coaching as did the females and they were significantly older. Thus, differences found by gender may have reflected differences in experience and/or age of the coaches. Another factor that appeared to be gender related was the past athletic participation of coaches. In an extensive study of high school coaches in Oregon, Sisley and Capel (1986) found a greater percentage of the males than females had been varsity collegiate athletes. In summary, although differences have been found in the frequency with which male and female coaches engage in technical instruction and encouragement, years experience coaching, age, and past athletic participation have emerged as factors that could be contributing to gender behavior differences of coaches. Yet, the degree to which gender contributed to coaches' behaviors remained unclear. In another study, Pratt and Eitzen (1989) examined the differences in philosophies between male coaches of male and female basketball teams. Their purpose was to detect any differences in the coaching philosophies of male coaches with athletes of different genders. Questionnaires, which focused on five specific areas of the coaches' beliefs and behaviors, were sent to the head coaches for boys and girls' basketball. In the overall development of the athlete, significant differences existed in six of the eight items on the questionnaire. Male coaches of male teams attempted to influence the athlete's life outside of sport much more than male coaches of female teams. Few differences were found in the conditions essential to maximize performance. For example, male coaches of 34 female teams tended to focus on the behaviors during practice. In comparing the coaches' expectations of their athletes, the coaches of male teams place more importance on the athletes' self-discipline and control, and on regarding the team higher than themselves then coaches of female teams. This study also found gender to not be a significant factor and also suggested that differences in coaching of male and female athletes have been due to personality and teaching style differences rather than as a result of gender. Coaching Behaviors and Expectancy Theory Actions happen because of the expectations people hold for themselves and those of others (Jones, Armour, & Potrac, 2002). The investigations of expectations affecting actions has developed into a theoretical framework known as Expectancy Theory that evolved from the work of the 20th-century sociologist Robert Merton (1948) who coined the synonymous term, the self-fulfilling prophecy. Merton utilized this concept to explain various sociological phenomena. He found that when something is expected to happen, the initial behavior actually caused the behavior to occur. The classic work in education of Rosenthal and Jacobson (1968), Pygmalion in the Classroom, studied teachers' differential treatment of students. It demonstrated that children who were expected to be high achievers improved significantly more then those who were expected to achieve less. In a later study, Rosenthal (1974) determined that individuals regarded as high achievers were awarded more beneficial treatment than those considered low achievers in four ways. Termed the Four Factor Theory, Rosenthal (1974) showed that high expectancy students were issued superior quality feedback in greater amounts, were afforded a warmer socio-emotional climate, were offered more opportunities for input 35 (i.e., more challenging tasks), and were given additional output opportunities (i.e., given more time to respond to questions). This phenomenon of teachers treating students of high ability/expectancy differently than those of lower ability/expectancy became known as Expectancy Theory (Cousineau & Luke, 1990; Horn, 1984; Markland & Marinek, 1988; Sinclair & Vealey, 1989). These studies prompted researchers to further investigate the effects of expectations of educators towards students and later of coaches towards athletes. Expectancy Theory suggests that through a series of phases coaches directly influence and impact athlete ability, skill, and performance. Current literature characterizes the relationship between coach expectations and athlete performance as a 4- step process (Horn, 1984; Horn, Lox, & Labrador, 2006; Solomon, 2001). In the first step, the coach develops expectations for athletes based on three types of impression cues: personal (i.e., race, gender, body size), performance (i.e., coordination, speed, agility), and psychological (i.e., confidence, motivation, anxiety). In the second step, the coach's expectations are communicated to the athlete in verbal and nonverbal ways in how s/he behaves toward the athlete (Horn et al., 2001; Martinek, 1989; Sinclair & Vealey, 1989; Solomon, 2001; Solomon & Kosmitzki, 1996). The third step includes the athlete becoming aware of the coach's treatment and it consequently affecting the athlete's own self-perceptions and behaviors. In the fourth step, the athlete responds to this treatment and her/his behavior conforms to the original expectation, thus completing the expectancy cycle. This serves to reinforce to coaches that their original judgment of the athlete's ability and skill was accurate. For the purposes of this review the expectancy cycle is not dictated by how players were rated against other players in the general public, 36 athletes' perspectives, or their actual performance. Rather all the steps, specifically differential treatment, are based upon the assessments and perspectives of the coaches. A major limitation of previous expectancy research has been the lack of accurate understanding and assessment of coach expectations of athlete ability and skill. In the past, this theory and the vast majority of research on expectancy relied primarily on the assumption that coaches' used impressions of physical ability to assess athlete expectancy level. Expectancy Theory suggested that two primary categories of information were utilized for assessment: personal and performance cues. Until more recently, psychological cues were not included in expectancy research and not commonly used in expectancy measuring tools. Initially, researchers utilized a rank-order method to distinguish between the coach's perceptions of high and low ability athletes (Sinclair & Vealey, 1989; Solomon, DiMarco, Ohlson, & Reece, 1998; Solomon & Kosmitzki, 1996; Solomon Striegel et al., 1996). This method required coaches to hierarchically rank athletes from most to the least skilled, resulting in rankings that were supposedly primarily based on athletes' physical skills and abilities (Solomon, 2001). Those who have worked in any athletic capacity know coaches use more information for assessment beyond athletes' physical abilities and skills. Coaches use a multitude of sources for information, as well as physical impressions, to assess the abilities and skills of their athletes when determining expectations. To address the issue Solomon (2001) conducted research that confirmed that head coaches use psychological sources in order to assess overall athletic ability and skill, specifically the psychological impression cue of confidence was included along 37 with performance impressions, which extended Expectancy Theory to include psychological cues as potential sources of expectancy information. Due to this research, the Expectancy Rating Scale (ERS) was created to overcome some of the limitations of the rank-order method (Solomon, 2001). Additional research has continued to support the use of psychological cues by coaches and the development of tools to measure those cues (Solomon, 2010). Interviews conducted with 18 coaches of both individual and team sports revealed that coaches used a multitude of factors and impressions to judge ability and skill including but not limited to the traditional physical ability/performance (Solomon & Rhea, 2008). From the results of this study, the Solomon Expectancy Sources Scale (SESS; Solomon, 2003), a 30-item survey, was developed to provide a more comprehensive report of the sources of information used by coaches to evaluate athletes. Further qualitative investigations using the SESS and other measures have supported the concept of psychological cues being an important factor contributing to coaches' perceptions of athlete expectancy (Solomon, 2010; Solomon & Rhea, 2008). In an investigation of 70 Division I head basketball coaches, Work Ethic, Receptivity to Coaching, Willingness to Learn, Love of Sport, Willingness to Listen, and Competitiveness were the six characteristics that emerged from the SESS as the predominant sources of information coaches used when assessing athletes (Becker & Solomon, 2005). Another investigation on soccer coaches found that the coaches perceived athletes attitudes to be a function of categorizing athletes as high or low expectancy (Wilson, Cushion, & Stephen, 2006). The previous investigation found that coaches believed that attitude was an important element to developing the conditions 38 necessary for advancement of an athlete. This could illustrate the magnitude of the importance of knowing and understanding use of psychological cues by coaches. In a qualitative investigation of five male and female Division I head coaches of various sports with at least 5 years of experience, they identified the characteristics they sought in athletes, especially when recruiting. Although all the coaches in the study identified ability, at least three discussed other qualities they used to establish athletes as successful for their teams, specifically the characteristic of competitiveness and work ethic (Selby, 2009). These findings have assisted researchers in understanding and measuring how coaches determine expectancy status of athletes. Furthermore, understanding the first step of evaluation has been important because coaches' initial impression of an athlete is likely to remain inflexible over the course of time, even when new information regarding athlete ability becomes available (Solomon, Golden, Ciapponi, & Martin, 1998; Solomon & Kosmitzki, 1996; Solomon & Rhea, 2008; Wilson et al., 2006). Simply stated, first impressions were prevailing and rarely did coaches re-evaluate their initial expectation of athlete achievement. Two studies (Solomon & Kosmitzki, 1996; Solomon, Golden, et al., 1998) demonstrated that coach expectations of athlete's ability tended to be inflexible whereas the coach's perceived potential for athletic improvement was changeable. Researchers examining soccer coaches found that the coaches were willing to articulate to athletes how they could change the coaches' perceptions, however, the majority of the coaches admitted that they usually did not, if ever, change their expectancy status of an athlete (Wilson et al., 2006). In the study of Pat Summitt, a consistent athlete expectancy assessment of her athletes was demonstrated from the beginning to the end of a basketball season (Becker 39 & Wrisberg, 2008). Thus, as with the investigation by Rejeski, Darracott, and Hutslar (1979), it appears that once coaches identify athletes as either high or low expectancy, they tend to retain those perceptions. This inflexibility of coach expectations demonstrated how important the initial evaluation could be on an athlete's development and future performance. This inflexibility has been important specifically when the second step in the expectancy cycle was considered, specifically how coaches' expectations had the potential to influence coaches' behavior and athletes' perceptions of their own abilities and skills. Rejeski et al. (1979) extended the notion from education to a sport setting that coaches' feedback correlated to the perceived skill level of athletes. In their study, they coded coaching behaviors directed at high and low skilled athletes in a youth sports basketball league. Consistent with findings reported in the classroom, the high ability children were reinforced more than low-ability children. It was also found that coaches gave more general technical information to low ability than to the high ability children. This differential display of behavior became known as differential treatment and generally emerged through the form of quality and quantity of feedback given to athletes (Mavi & Sharpe, 2000). According to Solomon and Kosmitzki (1996), this process of "feedback refers to the condition whereby coaches offer varying amounts of information to athletes based on perceptions of ability" (p. 165). Continued research using similar methods has suggested that high expectancy athletes received a greater quantity and quality of feedback than their low expectancy teammates (Sinclair & Vealey, 1989; Solomon, DiMarco, et al., 1998; Solomon & Kosmitzki, 1996; Solomon & Rhea, 2008; Solomon, Striegal et al., 1996). 40 This differential treatment caused by coaches' expectations, can potentially affect athletes permanently depending on how coaches' perceptions and treatment are demonstrated. Because coaches tend to have inflexible perceptions of athletes, it is possible that they may communicate an enduring high or low expectancy status to the athletes (Solomon & Rhea, 2008). As a result, coaches could continue in the behaviors they displayed toward those athletes. This may result in high expectancy athletes consistently receiving a greater quantity of relevant feedback than the low expectancy athletes (Solomon, Golden et al., 1998). Athletes' perceptions of differential patterns of feedback from coaches depending on expectancy status have also been documented (Sinclair & Vealey, 1989). Athletes' feelings of competence appear to be related to the amount of praise and instruction they received from their coach in response to successful performance attempts (Allen & Howe, 1998; Black & Weiss, 1992). Athletes interpreting the expectations of their coach as low may be negatively affected because their enthusiasm for participation and future performance decreases. Conversely, if an athlete interprets the expectations of the coach to be high, their enthusiasm and performance may be heightened. In essence, researchers have suggested that athlete performance hinges on the continued conveyed expectations of the coach (Martinek, 1981; Martinek, Crowe, & Rejeski, 1982). It is important to note that some coaches' behavior has not been congruent with their expectations, thereby demonstrating that they do not engage in the expectancy process. This was found when studying coach Pat Summitt who was found to give both high and low expectancy athletes similar amounts of quantity and quality of feedback (Becker & Wrisberg, 2008). Instead, she distributed an equitable proportion of feedback 41 to both high and low expectancy players, this finding was not consistent with previous expectancy literature (Becker & Wrisberg, 2008; Lacy & Martin, 1994; Markland & Martinek, 1988; Sinclair & Vealey, 1989; Solomon, DiMarco et al., 1998). Results from Smith et al. (1977) indicated that athletes responded most favorably to coaches who engaged in higher percentages of supportive and instructional behaviors. Furthermore, expectations may serve to reinforce an athletes' competence if the expectation is perceived, interpreted, and adopted. These findings are particularly important in light of the research suggesting that athletes with higher levels of confidence are more likely to succeed than those lower in confidence (Gould, Guinan, Greenleaf, & Chung, 2002; Weinberg, Grove, & Jackson, 1992). Knowing that differential treatment relates to coach expectations and potentially affects athlete performance, it reinforces the importance of coaches' being aware of the expectancy cycle and their behaviors. Summary Considering the ever-increasing rates of sport participation it is imperative to investigate the behaviors that may contribute to successful coaching outcomes. Because coaching is often referred to as a teaching experience, systematic observations of coaching behaviors have been developed and adapted from previous research in education. To improve the quality and efficacy of behaviors specific to coaches, specific observational tools have been created and employed throughout the literature. This research has provided relevant information that can be used by both researchers and coaches to improve coaching science. An investigation of past research has revealed a tremendous amount of information relative to coaching. From this extensive body of literature, the following themes have received significant attention: behavior trends 42 exhibited by successful coaches, the potential effects of gender on coaching behaviors, and the impact of Expectancy Theory on coaches' actions. Through a variety of systematic observational studies, a multitude of coaching behaviors have been categorically identified. Some of the more prevalent behaviors included instruction, praise, and hustle. The literature suggested various types of instruction as the dominant behavior of successful coaches. Positive feedback, such as praise, has also frequently been observed as a common coaching strategy. Additionally, highly successful coaches employ highly organized and structured practice plans. Although differences have been observed relative to coaching experience, evidence suggests that these behaviors seem to exist among coaches regardless of the type of sport or the gender of the coach. With increasing levels of sport participation for both males and females, the role of gender in coaching has been investigated. Surprisingly, the few significant differences that have been observed relative to gender have been mild. The literature suggests that regardless of the coaches' or athletes' gender, similar coaching behaviors are observed. Specifically, instruction remains the most prevalent strategy for both male and female coaches. Another major tenant of coaching science has been the role of Expectancy Theory. Expectancy Theory suggests that a coach's expectation of an athlete's ability will influence the behavior and subsequently the outcome for both parties'. Athletes who are perceived as more skilled, and therefore expected to perform better, may receive more instruction and feedback from their coaches than their less skilled teammates. This phenomenon tends to remain salient throughout the course of the coach athlete 43 relationship. Sadly, this behavior could prove detrimental to successful outcomes for coaches and athletes alike. Interestingly, these findings may be related to the level of experience and expertise of the coach and therefore not universally practiced. Although a great deal of scholarship has been dedicated to coaching science, it seems beneficial to study coaches using a similar observational instrument. The ASUIO was used in this study to systematically record and then compare the behaviors of successful high school girl's basketball coaches to a recent study of a highly successful collegiate women's basketball coach. These findings were also compared to similar studies using the same observational instrument. In addition to coaching behaviors the current study evaluated the effects of gender and Expectancy Theory to improve both the theory and practice of coaching science. CHAPTER 3 METHODS The primary purpose of this study was to identify instructional and non-instructional coaching behaviors of a successful male and female head coach of high school girls' basketball teams. The second purpose of this study was to investigate the theory of expectancy (Horn, 1984) by relating coaches' expectancy status (high vs. low) of athletes to coaching behaviors, quantity and quality of feedback, for successful high school coaches in basketball team practice settings. This chapter described the participant selection criteria, instrumentation, methodological procedures, and statistical analyses that were used in this study. Participants Participants for this study were one male and one female successful high school varsity girls' basketball head coaches in the mountain west area of the United States. Both participants were head coaches at high schools classified as 5A, the classification for schools with the largest enrollment. The participants were chosen due to the qualifications that would most closely match the accomplishments of the successful collegiate coach, Pat Summitt. Before participating in the research study completed in 2005, Pat Summitt had completed 30+ years as a head coach, accumulated an .839 winning percentage (852-167), won 26 tournament and conference championships, and 8 45 NCAA titles. Likewise, the participants in this study were selected based on criteria that closely matched Pat Summitt's success including quality programs, total career wins, state tournament appearances, state championships, years of coaching experience, and willingness to participate in the study. Success and inclusion criteria for coaches included the following as a head coach: (a) coaching at 5A high school, (b) total career wins at or above 300, (c) 10 or more regional championships, (d) 10 or more state tournament appearances, (e) 2 or more state championships, (f) 15 or more years of coaching experience, and (g) willingness to participate. The female coach had 17 years as head coach at her current school with 4 years previous coaching experience. At the current high school she had coached with a winning percentage of .754 (295-96) with teams winning 9 regional championships and 2 state titles. She had also coached teams to 17 state tournament appearances with teams going to the finals 4 times, placing 2nd twice. For the current season her team had gone 19-5 in their league, also winning the regional championship title. From all the teams combined, 33 students have continued on to play basketball at a collegiate level. The female coach's team consisted of 10 varsity athletes with three seniors and no freshman. She did not consider any junior varsity athletes to be varsity players even though some of the junior varsity athletes were to have limited varsity game playing time, especially when a varsity player sustained an injury. The female coach used only her designated 10 varsity players and 6 junior varsity players during games. The five starters received the majority of the playing time with the other athletes substituting sparingly, regardless of the competition. Of note one junior varsity athlete received more playing time then one of the varsity athletes. A typical practice was highly structured and 46 consisted of 2+ hours and a couple of Saturday practices. Her practices consisted of game specific drills, usually timed. Practices often included junior varsity athletes versus varsity athletes for scrimmages. Practices were run using one other varsity assistant coach and two junior varsity assistant coaches. The freshman team practiced separately, although the freshman coach would often assistant with varsity practices. The female coach also used graduate students and professors from the local university to assist with monthly sport psychology consulting during practices. Periodically throughout the season, auxiliary help would be brought in during practices for specialized training including but not limited to strength training, mental preparation, and technique of shooting. The male coach had 21 years as head coach plus coaching experience for 19 years at a previous high school. He also became the second winningest basketball coach in the state. He had a .831 winning percentage (409-83) with a 17-5 season record. His teams had won 13 regional championships and 4 state tournament titles. His teams had also made 21 consecutive state tournament appearances and made it to the final four 12 times. From all these teams, 52 athletes have continued on to play basketball at various collegiate levels. The male coach considered 18 athletes to be part of the varsity team including three seniors and two freshmen. These 18 athletes included all athletes who played junior varsity. Of the 18 athletes, only 5 played strictly on varsity, whereas the remaining 13 received playing time on varsity and junior varsity. The five starters received the majority of the game playing time, however, the male coach would frequently substitute a multiplicity of athletes depending on the competition. The junior varsity and varsity 47 consistently practiced together and were separated for learning into positions versus varsity playing level then combined for team scrimmaging. Practices were 2+ hours daily and every Saturday except on game days, which usually consisted of two days of the week. The practices were efficient and organized, using basic skill drills and game play. The two varsity assistant coaches were also the junior varsity coaches. The two assistant coaches taught and called the majority of the offense plays during practice and games. There were athletic trainers who also assisted with athlete injuries; however, the auxiliary assistance was limited from athletic trainers and other outside sources. Measures Demographic Information Background information, including coaching experience and accomplishments, on both coaches was obtained through personal interviews (see Appendix A). Background information about the athletes was obtained from coaches (see Appendix B). Information provided about the athletes included their grade in school, uniform numbers, and playing time. Arizona State University Observation Instrument The Arizona State University Observation Instrument (ASUOI; Lacy & Darst, 1984) was used to assess coaching behaviors (see Appendix C). ASUOI was used in order to analyze coaching behaviors with the same observation instrument used in many previous studies, thus making meaningful comparisons and valid conclusions (Darst & Goldston, 1990). It is a widely used observational instrument in coaching research and was created specifically to examine coaching behaviors during practice sessions (Kahan, 48 1999). The behavioral categories of the ASUOI are based on the conceptual rationale that satisfy the criteria for both content and face validity (Lacy & Darst, 1984). The use of this observational system yielded the total number of times a particular behavior occurred during a practice session. It is comprised of 13 behavioral categories representing three general types of behaviors: instructional (preinstruction, concurrent instruction, postinstruction, questioning, manual manipulation, positive modeling, negative modeling), non-instructional (hustle, praise, scold, management, other), and dual codes (statements that include the recipient's name). For the purposes of the current study and to examine expectancy theory, the category of dual codes represented statements that were specifically directed toward individual players. The coaches' use of the particular player's first name with the statements fitting into another category was coded as both first name and the behavior category, then used to determine the quantity and quality of feedback provided to that particular player. It should be noted that coaches' behaviors toward individual players was established during the coding process through the coaches' use of the athletes' first name, as well as coaches' physical proximity, eye contact, gestures, and use of gaining conscious attention of the athletes. Modified Expectancy Rating Scale (MERS) A major limitation of previous expectancy research has been the lack of a complete assessment of coach expectations of athlete skill and ability, specifically the ability to measure coaches' use of psychological cues. Initially, researchers used a rank-order method to distinguish between the coach's perceptions of high and low expectancy athletes (Sinclair & Vealey, 1989; Solomon, DiMarco et al., 1998; Solomon & Kosmitzki, 1996; Solomon, Striegel et al., 1996). However, it was left to the coaches to 49 define skill when ranking athletes. Solomon (1993) created the Expectancy Rating Scale (ERS) in an attempt to scientifically define skill level. Unlike the rank-order method, the ERS is a 5-item instrument that enables coaches to rate athletes independently of one another and give equal ratings to athletes with similar skills and abilities. Like the rank-order method, the ERS limitation was its lack of evaluation of other characteristics (i.e., psychological skills/abilities) that coaches use when evaluating athletes, because its primary emphasis was an evaluation of physical abilities (Solomon, 2001). More recently, Solomon (2003) created the Solomon Expectancy Sources Scale (SESS) to determine the most common characteristics that coaches use to evaluate athlete skill and ability. This 30-item instrument was used to assess the degree of importance coaches placed on various physical and psychological characteristics. The investigation results suggested psychological cues play an important part of coaches' decisions concerning expectancy (Solomon, 2003, 2010; Solomon & Rhea; 2008). More recently, Becker and Wrisberg adapted the ERS by adding three items to create the Modified Expectancy Rating Scale (MERS; Becker & Wrisberg, 2008; Solomon, 2003; Solomon & Rhea, 2008). The MERS consists of 8-items (see Appendix D for MERS) measuring both physical and psychological skills and abilities used by coaches to establish expectancy status of athletes (Becker & Wrisberg, 2008). Content validity for the MERS was established by obtaining feedback and consensus from three experts in the field of sport psychology in a previous study (Becker & Wrisberg, 2008). The MERS provided a reliable tool to assess coaches' assessment of athlete skill and ability in a more comprehensive manner (Becker & Wrisberg, 2008). 50 Procedures After examining the past 5 years of the Utah 5A girls' basketball state tournament, select coaches were contacted for further interviewing. Before the start of the girl's basketball season, head girls' high school basketball coaches were contacted and interviewed via email and phone to determine if they fit the criteria for the study. One male and one female coach were invited to participate and then a meeting was scheduled to discuss the purposes and procedures of the study (i.e., only practices were videotaped). The coaches were provided a description of the study and informed consent was obtained. Institutional and school district approval were obtained to conduct the investigation and informed consent statements were read and signed by each coach. Athletes and their parents were also be given assent/permission forms with an opt-out option. Parts of the observed practices were defined as particular segments. The definitions of these segments were: (a) warm-up: Any activity not using a basketball during the practice, such as stretching, to prepare for the workout or practice; (b) skill work: Any drill the purpose of which was to develop a particular skill or group of skills necessary for game play; (c) scrimmage: Five players versus five players in a game-like condition; and, (d) conditioning: Activity with the sole purpose of increasing the level of physical fitness in such areas as strength, endurance, or flexibility. Over the course of the season, a total of 553 minutes of practice with the female coach and 590 minutes of practice with the male coach (30 minutes to 2 hours per practice) were video recorded at one to 2-week intervals depending on practice schedules (games, weekends, school vacations, High School Activities Association regulations). 51 Observations took place during typical practices and started when practice began and continued throughout the practice (i.e., warm-ups, drills, half-court work, full-court work). To maximize the viewing perspective, the video camera was positioned courtside or in the bleachers. This allowed the researcher to track the coaches as they moved from one end of the court to the other. During all of the recorded sessions, coaches wore a wireless microphone to ensure that all verbal communication was acquired. Coaches were allowed to turn off the microphone when talking to assistant coaches, parents, administration, or athletes concerning topics that may be considered confidential (deaths in the family, counseling, etc.) and not pertaining to coaching. The input receptor for the wireless microphone was attached to the video camera. Therefore, all of the audio and visual data were simultaneously recorded onto the same digital videotape. After the second week of practices, both coaches completed the MERS on all athletes that were selected for the varsity team and placed them in a sealed envelope. Two weeks after the final game of the season, coaches completed the MERS a second time for each athlete based on their perceptions of the athletes' overall skill and ability. Coaches also rank-ordered all varsity team members according to the rank-order method after the finish of the season. Both coaches placed these evaluations in a sealed envelope. Total playing time for all varsity athletes was also obtained from coaches as one of the four components (preseason MERS, postseason MERS, average playing time, rank-order) to determine high and low expectancy athletes. Once all the data were collected, two or three observers watched the video footage and coded the coaches' behaviors using the ASUOI (Lacy & Darst, 1984). Observers were trained by completing a manual that was specifically designed for researchers using 52 this instrument (Solomon & Reece, 1995). Event recording provided a frequency count of defined behaviors as observers coded each discrete behavior as it occurred. The same consensus building technique utilized by Becker and Wrisberg (2008) was employed to determine the coding of all coaching behaviors. That process consisted of pausing the videotape after each statement and observers independently coding the feedback. If they agreed on the coding category, the data were entered. If they did not agree, they viewed the segment again until there was consensus. No feedback statements were coded until consensus was attained. To minimize possible experimenter bias, coaches' MERS ratings (both pre- and postseason) for each player were not viewed until all coding was completed. Data Analysis Because this study was a field-based investigation, it was not possible to standardize practice sessions. Thus, the length of practices, the selection of which segments would be included in the practices, and the length of the segments were out of the control of the researcher. Statistical analysis for all data in this study was conducted on a personal computer using SPSS version 15.0 (SPSS, Chicago, IL, USA). All data were entered into an Excel spreadsheet and checked by a research assistant to verify the correctness of data input prior to being imported into SPSS. Coaching Behaviors Data were coded and quantified for each behavior category for both the team and individual athletes. Percentages and rate per minute (RPM) for each behavior category was calculated and totaled. RPM was calculated by dividing the total of each category by 53 the total number of minutes observed. Percentages were calculated by dividing the frequency of each independent behavior category by the total frequency of all the categories combined, then that number was multiplied by 100. It should be emphasized that the resultant figure represented a percentage of total behaviors, not percentage of time. By definition, the use of first name had to accompany another behavior, therefore, to calculate the percentage of each behavioral category, the use of first name was excluded and the percentage of this category was considered separately. If first name was analyzed as an independent category it would have decreased the values of other behaviors and the true percentage of these behaviors would have been distorted (Lacy & Goldston, 1990). Designation of Consistent Expectancy Status A Pearson Product Moment correlation was used to determine if the coaches' perceptions of their athletes would remain consistent over the course of the season. If there were a high and a statistically significant correlation between the coaches' preseason and postseason MERS ratings of athletes, there would be support for the hypothesis that coaches' perceptions remained relatively consistent over the season. Designation of High and Low Expectancy Athletes Coaches' expectancy assessments (preseason MERS scores, postseason MERS scores, ranking) along with each player's average amount of playing time for the entire season was analyzed to see if average playing time, ranking, and MERS assessments predicted the same athletes as low expectancy and high expectancy. Specifically, a Pearson Product Moment correlation was calculated to determine the relationship 54 between all four measures (preseason MERS, postseason MERS, ranking, and average playing time). Because the number of athletes varied on each team, the scores for each of the four measures were then converted to ranks and combined to establish a composite expectancy score for each athlete. Lower values indicated a higher expectation. High and low expectancy was determined by partitioning athletes into halved samples according to their composite expectancy score by designating athletes scoring above the 50th percentile as high expectancy (n = 5, n = 9) and athletes scoring below the 50th percentile as low expectancy (n = 5, n = 9). Three different types of scores were computed for each athlete for the purpose of analyzing quantity and quality of feedback to athletes. The first frequency score was an individual behavior score, which was calculated by dividing the number of individual behaviors received from the coach by the total number of individual behaviors given to all the athletes on that particular team. The second frequency score was an overall received behavior score, which was calculated by dividing the number of individual behaviors received from the coach by the total number of individual and group behaviors given to the teammates and the team in general. Thirdly, for the purposes of analysis, all frequency counts for each category for each athlete were converted to proportion scores. Eleven behavior type scores were calculated representing 11 of the 13 behavior categories used in the ASUOI, first name and other were excluded. Each feedback type was converted to a proportion by dividing the number of individual received behaviors by the athlete in that category by the total number of individual behaviors received by that athlete. For example, an athlete's score 55 on postinstruction represents the proportion of postinstruction that athlete received in relation to all of the individual behaviors she received from the coach. A multivariate analysis of variance (MANOVA) was conducted to analyze whether differences existed in both quantity and quality of feedback provided to high and low expectancy players for each category represented on the ASUOI. Independent variables used for the MANOVA were the players' expectancy status (high, low) and the dependent variables were the behavioral categories represented on the ASUOI. CHAPTER 4 RESULTS AND DISCUSSION The purpose of this study was to identify coaching behaviors of a successful male and female head coach of high school girls' basketball teams and determine if the factors of gender and expectancy affected the coaches' behaviors. The participants in this study included one male and one female successful high school head basketball coach of girls' basketball teams. The remaining part of this chapter clarifies the results that were found in the current study and presents a discussion of those results. Results Primary Aim and Hypothesis The primary aim of this study was to determine the instructional and non-instructional coaching behaviors of a successful male and female head coach of high school girls' basketball teams. It was hypothesized that both the male and female coach would provide instruction (preinstruction, concurrent instruction, postinstruction) at a higher rate than any other coaching behavior. Also, there would be no significant differences in terms of coaching behaviors between male and female coaches at high school basketball practice settings. It was hypothesized that coaching behaviors demonstrated by the high school coaches would be similar to those demonstrated by collegiate coach, Pat Summitt. 57 Secondary Aim and Hypothesis The secondary aim of this study was to investigate how the theory of expectancy (Horn, 1984) correlated to coaches' behaviors due to their expectancy status (high vs. low) of athletes. Based on the results of previous expectancy research, it was hypothesized that both coaches' perceptions of their players' skills and abilities would remain consistent over the course of the season. Coaches would provide differential treatment in the form of feedback based on perceived athlete expectancy. High expectancy athletes at the high school level would receive a greater quantity and quality of feedback than lower expectancy athletes. There would be no significant difference between genders in terms of differential treatment of athletes for successful basketball coaches at team practice. Throughout the 2009-2010 season at approximately two-week intervals a total of 553 minutes of practice time was observed, revealing 3,052 practice behaviors for the female coach and a total of 590 minutes consisting of 3,610 practice behaviors were coded for the male coach. Table 1 provides a summary of both coaches' frequencies of behavior towards combined team and individual basketball athletes including the total number of behaviors observed, percentage, and RPM for each defined category of the ASUOI. Table 2 provides a breakdown of the behaviors from Table 1 into both coaches' behaviors toward individual basketball athletes and the team. When examining Table 1, the behavior used most often by both coaches in the instructional categories was concurrent instruction (female 13.1%, male 12.9%) followed closely by postinstruction (12.1%) by the male coach and preinstruction (11.4%) for the female coach. Interestingly, the two coaches' overall most exhibited behavior was in the 58 Coaching Behavior Totals, Percentages, and RPM for the Female and Male Coach RPM Male .27 .79 .74 .32 .06 .12 .02 .74 1.35 .51 .73 .47 6.23 1.21 Confidence Interval 3.83-5.17 11.81-13.99 11.04-13.16 4.48-5.92 .68-1.32 1.5-2.34 .13-.47 10.94-13.06 20.75-23.45 7.40-9.20 10.85-12.95 6.83-8.57 18.50-21.10 % Male 4.5% 12.9% 12.1% 5.2% 1.0% 1.9% .4% 12.0% 22.1% 8.3% 11.9% 7.7% 100% 19.8% Total Male 161 466 438 189 35 69 13 434 797 301 429 278 3610 714 RPM Female .63 .72 .59 .47 .03 .11 .02 .36 .64 .11 2.41 .27 5.51 1.84 Confidence Interval 10.27-12.53 11.92-14.39 9.69-11.81 7.51-9.49 .33-.87 1.42-2.38 .18-.62 5.63-7.37 10.47-12.73 1.50-2.50 26.41-29.59 4.13-5.67 31.53-34.87 % Female 11.4% 13.1% 10.7% 8.5% .6% 1.9% .4% Non-Instructional Behaviors Non-instructional 6.5% 11.6% 2.0% 28.0% 4.9% 100% 33.2% Total Female Instructional Behaviors Instructional 349 399 327 258 18 59 13 199 355 62 855 150 3052 1016 Table 1 Behavior Pre-Inst. Con. Inst. Post-Inst. Question Phy. Assist + Model - Model Hustle Praise Scold Manage Other Total 1st Name 59 Totals for Male and Female Coaching Behaviors Toward Individuals versus Team for Each Category % Ind. Beh. 24.2% 29.2% 70.1% 68.8% 100% 49.3% 61.5% 17.5% 47.2% 56.5% 23.8% 65.8% Note. % to Individual represents the percentage of feedback for the given category of behaviors to individuals versus team (e.g., [108/(241+108)]*100 = 30.9%). % Male 2.4% 8.5% 19.2% 8.1% 2.2% 2.1% .5% 4.8% 23.6% 10.7% 6.4% 11.8% 44.21% Indiv. Male 39 136 307 130 35 34 8 76 376 170 102 183 1596 % Male 6.1% 16.4% 6.5% 2.9% 0 1.7% .2% 17.8% 20.9% 6.5% 16.2% 4.7% 55.79% Team Male 122 330 131 59 0 35 5 358 421 131 327 95 2014 % Ind. Beh. 30.9% 48.1% 66.7% 41.9% 100% 55.9% 61.5% 22.6% 58.0% 60.0% 43.3% 37.3% % Female 7.7% 13.7% 15.6% 7.7% 1.2% 2.4% .6% 3.2% 14.7% 2.6% 26.4% 4.0% 45.84% Indiv. Female 108 192 218 108 18 33 8 45 206 37 370 56 1399 % Female 14.6% 12.5% 6.6% 9.1% 0 1.6% .3% 9.3% 9.0% 1.5% 29.3% 6.2% 54.16% Team Female 241 207 109 150 0 26 5 154 149 25 485 102 1653 Table 2 Behavior Pre-Inst. Con. Inst. Post-Inst. Question Phy. Assist + Model - Model Hustle Praise Scold Manage Other Total 60 non-instructional category. Of the noninstructional behaviors and overall, the female coach exhibited management (28.0%) more than any other behavior category. The highest category for the male coach was praise (22.1%), also a noninstructional behavior. The praise to scold ratio for the male was approximately 3:1 with the female exhibiting a higher ratio of praise to scold of 6:1. The greatest differences in behavior categories were use of first name (female 33.2% to male 19.8%), preinstruction (female 11.1% to male 4.5%), hustle (female 6.5% to male 12.0%), praise (female 11.6% to male 22.1%), and management (female 28.0% to male 11.9%). The total RPM for all behaviors was similar for both coaches, the male coach showing a slightly higher rate (6.23) then the female coach (5.51). Inspection of Table 1 shows that both the female coach and male coach provided instruction more often (female 35.2%, n = 1114, male 29.5%, n = 1065) than any coaching behavior when the categories of preinstruction, concurrent instruction, and postinstruction were combined. Coaching behaviors were distinguished between behaviors that were directed toward the team and individual players in Table 2 from the totals in Table 1. Results revealed that both coaches were almost identical in their feedback behaviors to the team and individuals overall with 54.16 % (n = 1653) of the female coach's total behaviors were toward the team (including small groups) and the male coach directed 55.79% (n = 2014) towards the team. Likewise the female coach directed 45.84% of behaviors toward individual athletes, as the male coach was similar with 44.21% As Table 2 illustrates, the female coach's most frequent behavior to the team (29.3%) and individuals was management (26.4%). The male coach's most frequent behavior to the team (20.9%) and individuals (23.6%) was praise. The female coach 61 demonstrated a higher frequency of preinstruction (n = 241) towards the team than to individuals (n = 108). The opposite was true for postinstruction whereby more statements were directed toward individual players (n = 218) than toward the team (n = 109). Also the female and male coach directed hustle statements toward the team (female n = 154, male n = 358) more often than individuals (female, n = 45, male, n = 76). Of note, the lowest coaching behavior categories for both team and individuals were physical assistance (n < 35), sometimes referred to as manual manipulation, and negative modeling (n < 13). Expectancy Patterns The second purpose of the study was to examine coaches' differential patterns of behavior to high and low expectancy athletes. Each one of the coaches' expectancy assessments (preseason MERS, postseason MERS, and postseason ranks) along with each player's average playing time for the entire season were analyzed. In Table 3 the correlation between the assessments was calculated using a Pearson Product Moment correlation. The most significant correlation existed between coaches' rankings and average playing time (female r = .851, p = .005; male r = .777, p = .005). To assess the consistency of coaches' expectations, the Pearson Product Moment correlation was also used to compare coaches rankings of their players using the preseason MERS and postseason MERS. The results of this correlation analysis indicated there was consistency of expectancy from the beginning to the end of the season for the female (r = .758) and male coach (r = .706) with p < .001. These correlations were likely underestimated due to using a Pearson Product Correlation versus an Intraclass Correlation, which would have been a more appropriate test to run on the pre- and post 62 Table 3 Correlation Between Pre- and Post MERS, Coach Rank, and Average Playing Time for the Male and Female Coach *p < .05, **p < .01 Pre MERS Post- MERS Playing Time Coach rank Pre MERS Female _ .758* .339 .349 Male .706* .069 .448 Post- MERS Female _ .569** .630* Male .481** .723* Playing Time Female Male _ .851** .777** 63 MERS, though not suited for testing the use of average playing time and coaches rank. The Pearson Product Moment Correlation was also used to be able to make comparisons with the study of Pat Summitt, which used the same statistical analysis. Composite expectancy scores were established for each athlete. Final composite scores ranged from 4 to 40 for the female coach and 4 to 72 for the athletes of the male coach. The top 50% of the scores represented the high expectancy athletes whereas the bottom 50% represented the lower expectancy athletes, respectively, for each coach. Halved samples were used to replicate previous research and also due to the small sample size of the teams running statistical analysis would have been difficult using a smaller portion of each team. The mean for the high expectancy group for the female coach was 14.8 (SD = 6.46) and for the male coach 24.6 (SD = 10.97), whereas the mean for the low expectancy group for the female coach was 29.8 (SD = 4.32) and the mean for the male coach was 52 (SD = 8.89). Expectancy and Quality of Coaching Behaviors A review of the raw data between the groups was done as noted in Table 4. The totals of the high and low expectancy athletes combined for each category do not match the individual totals on Table 2 due to other athletes participating in practice (i.e., freshman athletes) being included in Table 2 and the totals for Table 4 were strictly limited to high/low athletes. Totals without the other category were created because the category of other consisted of behaviors that were not of a teaching nature. Statistical analyses for homogeneity of group variance were run on the total without the other category athlete scores of the high and low expectancy groups of both the female and male coach. The female coach's low expectancy group mean frequency of 64 Table 4 Coaching Behaviors for Male and Female Coach towards High and Low Expectancy Athletes Athlete Expectancy Female Male Coaching Behavior High (5) Low (5) High (9) Low (9) Pre-Inst. 61 32 19 19 Con.-Inst. 65 67 53 58 Post-Inst. 66 70 149 125 Question 44 29 52 59 Phy. Assist. 4 7 9 16 + Model 12 11 9 16 - Model 1 5 4 3 Hustle 13 15 32 36 Praise 94 64 176 169 Scold 16 13 64 71 Management 152 121 56 33 Other 20 18 111 53 Total 548 452 734 658 Total w/o other 528 434 623 605 Note. Coaching Behavior totals for high/low athletes differed from individual totals on Table 1 due to other athletes not ranked as high or low receiving communications during practices being included in Table 1. 65 coaching behaviors was 91.0 (SD = 37.202) and the high expectancy group mean was 115.4 (SD=13.939), F (1,8) = 1.886, p = .207. The Shapiro-Wilk for the low expectancy group was .90 and for the high expectancy group .79, so the groups were normally distributed. For the male coach the low expectancy group mean was 69.33 (SD = 41.39) and the high expectancy mean was 69.33 (SD = 27.56), F (1, 16), p = 1.0. The Shapiro- Wilk for the low expectancy group was .93 and for the high expectancy group was .92, respectively. In all behavior categories except management and other, the male coach displayed approximately the same amount of behaviors to both high and low expectancy athletes. Of note, the other category had a large difference between high (n = 111) and low (n = 53) expectancy athletes from the male coach, however, this category did not count towards teaching feedback. When comparing the male coach to the female coach, some differences were noticed between the frequencies of the behaviors in the categories of pre-instruction, post instruction, hustle, praise, scold, management, and other. For the female coach, the behaviors were recorded for an additional group of athletes, as viewed in Table 5, which were unrated as high or low expectancy, but practiced consistently with those athletes grouped as such. In addition, these six unrated athletes had varsity game playing time, and in fact, one of the unrated athletes had more average playing time then a few of the lower ability athletes, however, was not considered high or low, or as being on varsity, or ranked by the coach. The female coach provided approximately equal amounts in the behavior categories of concurrent instruction, physical assistance, positive/negative modeling, hustle, and other to all three groups of athletes. When considering the unrated athletes, the female coach gave pre- 66 Table 5 Coaching Behaviors for Female Coach towards High and Low and Other Expectancy Athletes A t h l e t e Expectancy Female Coaching Behavior High (5) Low (5) Unrated (6) Pre-Inst. 61 32 15 Con.-Inst. 65 67 60 Post-Inst. 66 70 82 Question 44 29 35 Phy. Assist. 4 7 7 + Model 12 11 10 - Model 1 5 2 Hustle 13 15 17 Praise 94 64 48 Scold 16 13 8 Management 152 121 91 Other 20 18 18 Total 548 452 393 Total w/o other 528 434 375 Note. Unrated athletes were athletes that had limited varsity playing time and were part of varsity practices, but were not rated by coach as being high/low on varsity. 67 instruction from the least to greatest amount from the unrated (n = 15) to low (n = 32) to high (n = 61) ability and skilled athletes. This trend was also seen in the categories of questioning, praise, scold, and management. The only category where the unrated players received a higher amount of behaviors was in the category of postinstruction (high n = 66, low n = 70, unrated n = 82). Overall total coach behaviors also displayed high expectancy athletes receiving more behaviors then the low expectancy athletes, then the unrated athletes receiving slightly less then the low expectancy. When comparing each group, the high expectancy received 39%, the low expectancy received 32%, and the unranked group received 28% of the total overall individual behaviors (n = 1393) for these three groups. Statistical analysis was similar to the statistical processes completed during the study of Pat Summitt in order to make comparisons. However, in the Summitt study instruction was grouped together instead of being broken down into pre-, concurrent, and postinstruction. However, for this statistical analysis pre-, concurrent, and postinstruction were analyzed separately to attain a specific analysis of each instructional behavior of the coaches towards the athletes. A category of Total without Other was created because this category had little impact on the actual teaching of basketball. The various MANOVA's were run using the total minus the other category. The rationale used was that other was any behavior coded that did not apply to the teaching of basketball (i.e., team items, school topics, family situations, etc.), therefore this category seemed inappropriate when examining quantity and quality of feedback to athletes. Two MANOVA's were conducted to analyze the differences between high and low expectancy athletes in relation to the quantity of behaviors each group received from 68 their coach. The dependent variables included in the analyses were individual behavior score and overall received behavior scores. The individual behavior score was calculated by dividing the number of individual behaviors received from the coach by the total number of individual behaviors given to all the athletes on that particular team. The overall received behavior score was calculated by dividing the number of individual behaviors received from the coach by the total number of individual and group behaviors given to teammates and the team in general. For the female coach no significant multivariate effect was found for differences in quantity of behavior between expectancy groups, defined by split samples, for frequency of individual behaviors and overall behaviors received from coaches. For the halved sample (n = 5), analyses of the female coach indicated that high expectancy athletes (M = .107, SD = .018) received the same individual feedback behaviors (based on individual behavior received score) as low expectancy athletes (M = .089, SD = .042). Also high expectancy athletes (M = .043, SD = .008) received the same quantity of individual feedback (based on overall behavior received) as low expectancy athletes (M = .034, SD = .015). No significant multivariate effect was found for differences in quantity of behaviors between expectancy groups, also defined by the split samples, for frequency of individual behaviors and overall behaviors received from the male coach. For the halved sample (n = 9), analyses of the male coach indicated that high expectancy athletes (M = .055, SD = .021) received the same individual feedback behaviors (based on individual behavior received score) as low expectancy athletes (M = .055, SD = .033). Also high expectancy athletes (M = .022, SD = .008) received the same quantity of individual feedback (based 69 on overall behavior received) as low expectancy athletes (M = .022, SD = .013). Tests of normality were conducted to assess differences between high and low expectancy athletes in relation to the type or quality of coaching behaviors athletes received as illustrated in Table 6. All of the type scores representing the qualitative behavior categories from the ASUOI were included as dependent variables except the variables for both coaches of manual manipulation, positive modeling, and negative modeling because they did not appear to be contributing variables due to low frequency counts (n < 20). The category of other was not used due to its lack of contributing to actual feedback. A Shapiro-Wilk test was conducted to assess the normality of the remaining eight behavior categories for both coaches. Categories with a Shapiro-Wilk statistic value greater than .90 were considered normally distributed (Becker & Wrisberg, 2008). Homogeneity of variance was found among three categories for the female coach and 6 for the male coach. For the female coach with 10 athletes, the behavioral categories of post instruction, questioning, praise, scold, and management fell below the .90 range of normal distribution. The MANOVA revealed no significant differences in the quality of feedback in the behavioral categories of preinstruction, concurrent instruction, and hustle. For the male coach with 18 athletes, the behavior categories of preinstruction, concurrent instruction, and postinstruction fell below the value of .90. The MANOVA revealed no significant differences in the quality of feedback between the expectancy groups for the categories of questioning, hustle, praise, hustle, scold, and management. 70 Table 6 Shapiro-Wilks of Female and Male Coach Tests of Normality Female Male Expect Statistic Sig. Expect Statistic Sig. PreInstruction Low .980 .932 PreInstruction Low .864 .106 High .902 .423 High .942 .602 Concurrent Instruction Low .922 .546 Concurrent Instruction Low .892 .208 High .931 .606 High .933 .514 Post- Instruction Low .831 .141 Post- Instruction Low .733 .003 High .888 .348 High .860 .095 Questioning Low .896 .389 Questioning Low .982 .975 High .906 .443 High .957 .770 Hustle Low .907 .449 Hustle Low .984 .982 High .949 .727 High .889 .196 Praise Low .784 .060 Praise Low .965 .851 High .891 .362 High .944 .623 Scold Low .766 .041 Scold Low .947 .658 High .958 .793 High .944 .624 Manage Low .978 .925 Manage Low .962 .820 High .885 .333 High .942 .606 Note. Expect stands for Expectancy group 71 Discussion The present study contributed to the existing literature by providing descriptive data pertaining to the behaviors of two high school girls' basketball coaches and analysis of factors contributing to those behaviors such as gender and coach expectancy. The methodology used in this study (as with the study previously conducted on Pat Summitt) allowed for a more in-depth analysis of behaviors and added to the growing database of descriptive information concerning practice behaviors of successful high school coaches. In the past research with Wooden, Summitt, and many other successful coaches, instruction was frequently researched and found to be the chief behavior of coaches (Becker & Wrisberg, 2008; Lacy & Goldston, 1990; Tharp & Gillmore, 1976). A dominant function of the coaches in this study was also giving instruction. Over the entire season for the period sampled, the behavioral categories classified as instructional (preinstruction, concurrent instruction, postinstruction, questioning, modeling, physical assistance) accounted for 46.6% for the female and 38.0% for the male total coaches' behaviors. Across both genders, the functions of instruction (pre-, concurrent, and postinstruction) accounted for a large percentage of the behaviors (female 35.2% to male 29.5%). These behaviors accounted for approximately one-third of all behaviors for both genders. Both coaches exhibited a higher frequency of preinstruction for team (female 14.6%, male 6.1%) compared to in