Upper elementary school students' attitudes and perceptions toward physical education
TABLE OF CONTENTS CHAPTER 1: Introduction 1 Purpose of the study 4 Rationale 4 Research questions 5 Dissertation organization 5 CHAPTER II: Development and validation of scores of an instrument to measure fourth and fifth grade student attitude toward physical education 6 Method 9 Elicitation study and instrument question development 10 Pilot study 11 Pilot questionnaire data analysis 12 Pilot content validity study 14 Validity study 14 Validity study data analysis 16 Content validity study 15 Results 16 Content validity study 21 Discussion 21 CHAPTER III: Upper elementary school student attitudes toward physical education...26 Method 30 Entry into schools 30 Instrumentation 31 i
Questionnaire administration 32 Data analysis 33 Results 33 Four Factor Model 34 Two Factor Model 35 Discussion 36 CHAPTER IV: Upper elementary school student perceptions of physical education: High attitude and moderate and low attitude student perspectives 42 Method 44 Pilot Study 45 Participant 45 Question development 45 Final Study 46 Setting 46 Participants 46 Observations 47 Interviews 47 Interview and observation data analysis 48 Trustworthiness 49 Findings 50 Physical education class was "fun" 51 Students were excited to go to class 52 Physical education was their favorite class 53 ii
The teacher influences the fun level 54 Class activities were important 57 Moderate to low attitude students love team problem solving 59 Fitness activities are controversial 61 Discussion 64 REFERENCES 69 APPENDIX A: Literature review 87 APPENDIX B: Elicitation study 109 APPENDIX C: Pilot study administrator explanation letter 110 APPENDIX D: Pilot study consent forms I l l APPENDIX E: Pilot study questionnaire 119 APPENDIX F: Pilot study demographic information sheet 123 APPENDIX G: Pilot study questionnaire administration protocol 124 APPENDIX H: Content validity first questionnaire 125 APPENDIX I: Content validity second questionnaire 129 APPENDIX J: Validation study questionnaire 133 APPENDIX K: Validation study administrator letter 136 APPENDIX L: Validation study administrator consent form 137 APPENDIX M: Validation study questionnaire permission slip 139 APPENDIX N: Validity study protocol-researcher 144 APPENDIX O: Validity study protocol-others 145 APPENDIX P: Validity study demographic information sheet 147 APPENDIX Q: Final validity study questionnaire 148 iii
APPENDIX R: Content validity study questionnaire 150 APPENDIX S: Interview consent form 153 APPENDIX T: Interview protocol 157 IV
LIST OF TABLES 2.1 Validity study factor loadings two factor model 18 2.2 Validity study factor loadings four factor model 20 2.3 Fit statistics and Cronbach's alpha values for each model 21 3.1 Four factor model descriptive statistics 34 3.2 Two factor model descriptive statistics 36 v
ACKNOWLEDGEMENTS Thank you to everyone whose energy and support made this achievement possible. First I would like to thank my parents, Frank and Sue Phillips, for your constant love and for giving me the foundation to my education. Secondly, a sincere thank you to Steve Silverman. You have guided me every step of the way to success, line-by-line, word-by-word. Your dedication, knowledge and hard work are greatly appreciated and will always be remembered. Thank you to my dissertation committee. Especially Carol Garber for all our talks and Ellie Drago-Severson for all your guidance. The support of the doctoral seminar group at Teachers College played an important role over the past four years, thank you for always being there. A special thank you to Eve Bernstein, who has been there to lend a helping hand all the way. To my friend Jodi Rigotti, thank you for the cheering and support that keeps on coming! To Marilyn Rabinowitz who helped inspire me to go for a doctorate, thank you. To Joyce Unrig who never let me lose sight of my goal, thank you. To Joe King, thank you for your statistical wizardry. Finally, thank you to Geoffrey Lenat. You have helped me celebrate each step along the way. You have shown amazing patience and kindness, and have been incredibly supportive in every way. Your love and support mean the world to vi
1 CHAPTER 1 INTRODUCTION Attitudes are an integral part of our everyday lives. Our personal attitudes impact us constantly in an array of different ways and in many facets of our lives. One of the ways attitude impacts our lives is related to our attitudes toward physical activity. Our physical activity behavior, whether or we choose to go to the gym or go for a jog, is impacted by our attitude. These behaviors and attitudes are strongly influenced by our experience in physical education. Physical inactivity currently is a growing issue among America's children and is one of the contributing factors to obesity, which is a major concern because in American society today there is an obesity epidemic among our youth (Flegal, Carroll, Ogden, & Curtin, 2010; U.S. Department of Health and Human Services, 1996). Obesity is associated with significant health problems for children and is an important early risk factor for adult morbidity and mortality (Krebs, 2003). One of the major combatants of obesity is physical activity, which should be encouraged among children and adolescents based largely on the assumption that the behavior will become part of the person's life and carry into adulthood (Kohl & Hobbs, 1998). It also has been suggested that if a student has a positive experience in physical education and develops a positive attitude about physical activity, motivation to engage in physical activity outside of school is likely to occur (Solmon & Lee, 1996). Conversely, students who show unfavorable attitudes toward physical education may refrain from indulging in physical activity outside of school (Carlson, 1995; Ennis, 1996; Portman, 1995; Robinson, 1990).
2 Attitude is defined as the degree to which a person likes or does not like something (Ajzen & Fishbein, 1980). Allport (1968, p. 59) described attitude as being "the most distinctive and indispensable concept in contemporary American psychology." A person's attitude toward any given object can have either a positive or negative impact. This attitude has the ability to strongly influence a person's behavior. The idea, that attitude guides behavior, is grounded in the Theory of Reasoned Action. This theory states that a person's personal belief systems influence a person's attitude, which ultimately determines a person's behavior (Fishbein & Ajzen, 1975). This theory manifests itself in physical education class as students enter the gymnasium with predisposed attitudes toward physical education, both positive and negative, which will influence their physical activity levels. While students do enter the gymnasium with predisposed attitudes, these attitudes can be altered over time (Ajzen & Fishbein, 1980). A positive experience with a situation once perceived to be negative has the ability to alter a person's attitude to a more positive one (Zimbardo & Leippe, 1991). The opposite is true, a negative experience with a situation once perceived to be positive could also cause a negative change in attitude. Changing a person's attitude, while possible, is quite challenging. Attitudes are impacted by personal belief systems. These belief systems are formed at a young age and once these belief systems take hold they become more difficult to alter as attitudes and behavior patterns form (Allport, 1935). Reaching students at a young age can be done through elementary school physical education classes and this interaction can have a substantial impact on the formation of student attitudes toward both physical education and physical activity.
3 Little is known about elementary school students' attitudes toward physical education, however a great deal is known about middle and high school students' attitudes. Middle and high school student attitudes are influenced by age, grade level, and skill level (Dickenson & Sparkes, 1988; Haladyna & Thomas, 1979; Lirgg, 1993; Montalvo & Silverman, 2008; Saunders, 1979; Stewart, Green, & Huelskamp, 1991; Subramaniam & Silverman, 2007; Tannehill & Zakrajzek, 1993; Wersch, Trew, & Turner, 1992). It has been noted that overall student attitude decreases with age and that girls' attitudes decline more steeply beginning in middle school than boys' attitudes (Dickenson & Sparkes, 1988; Montalvo & Silverman, 2009; Saunders, 1979; Subramaniam & Silverman, 2007; Wersch et al., 1992). In addition, student skill level has been suggested to have a relationship with student attitude (Graham, 1987; Grant, Ballard, & Gynn, 1989; Rikard, 1992; Silverman, 1985, 1993, 2009; Silverman, Subramaniam, & Woods 1998; Silverman, Tyson, & Krampitz, 1993; Silverman, Woods, Subramaniam, 1999; Sweeting & Rink, 1999). It is relatively unknown if these same variables and factors influence elementary school student attitudes. Studying student perceptions has taught us about the student's perspective in physical education. These perceptions have helped inform the teaching and learning process in physical education (Bernstein, Phillips, & Silverman, in press; Cothran, Kulinna, & Garrahy, 2003; Graham, 1995; Hopple & Graham, 1995; Portman, 1995; Ravizza & Stratton, 2007; Subramaniam & Silverman, 2002). These studies of student perceptions in physical education have provided insights and reasons to improve physical education instruction, with attention toward student attitudes toward physical education. While information about middle school student perceptions is available, little is known
4 about upper elementary school perceptions. Investigating upper elementary school students' perceptions about physical education extends the research from middle schools to an age group where attitudes and perceptions may influence future views of physical education and physical activity. Purpose of the Study There are two primary purposes of this study. The first purpose is to develop an instrument that will assess upper elementary school student attitude toward physical education. The second purpose is to explore and begin to understand upper elementary school student attitudes and perceptions toward physical education using the dual component view of attitude, which encompasses affective and cognitive factors (Bagozzi & Burnkrant, 1979; Oppenheim, 1992; Subramaniam & Silverman, 2007). Rationale This research can add to the body of literature related to the Theory of Reasoned Action. Presently, a vast majority of the literature regarding student attitudes toward physical education is grounded in and supports this theory. There is little research that has been done on the basis of the Theory of Reasoned Action with elementary school students. This study can contribute to the literature with another perspective of students' attitudes and perceptions for younger children. Equipped with a basic understanding of middle school student attitudes toward physical education and a head start on understanding high school student attitudes, we can continue to create a more comprehensive understanding of student attitude toward physical education throughout a student's entire K-12 career with the addition of a baseline understanding of upper elementary school student attitude. Given this
5 information we can begin to create an understanding of overall student attitude toward physical education and start to explore the variables and contextual factors that influence student attitude and test various interventions. Eventually this body of research can be used to inform teachers and influence the curricula that impact the learning environment. By understanding the factors that are impacting their attitude additional research can be done in a more specific nature to hone in on which aspects of these factors impact learning. That information may then be shared with university professors, pedagogy students and teachers to help improve instruction and learning within physical education. Research Questions 1. Can an instrument be developed to measure student attitude in upper elementary school students? 2. What are the attitudes of 4th graders towards physical education? 3. What are the attitudes of 5th graders towards physical education? 4. What is the relationship between gender and student attitude towards physical education in upper elementary school students? 5. What are students with positive attitudes towards physical education perceptions about the physical education? 6. What are students with negative attitudes towards physical education perceptions about the physical education? Dissertation Organization The above questions were answered in three related studies that follow in chapter II (question 1), chapter III (questions 2, 3, and 4) and chapter IV (questions 5 and 6). This dissertation uses a journal article format and each chapter represents a separate article that will be submitted for publication. A complete literature review is in Appendix A.
6 CHAPTER II DEVELOPMENT AND VALIDATION OF SCORES OF AN INSTRUMENT TO MEASURE FOURTH AND FIFTH GRADE STUDENT ATTITUDE TOWARD PHYSICAL EDUCATION Attitudes are an integral part of our everyday lives. Our personal attitudes impact us constantly in an array of different ways. One of the ways attitude impacts our lives is related to our attitudes toward physical activity. Our physical activity behavior, whether or not we choose to go to the gym or go for a run, is impacted by attitude. These behaviors and attitudes are strongly influenced by our experience in physical education. Research on student attitudes toward physical education is important because physical inactivity is increasing among America's children. Physical inactivity is one of the contributing factors to obesity. This is a major concern because there is an obesity epidemic among American youth (Flegal et al., 2010; USDHHS, 1996) and obesity is associated with significant health problems for children and is an important early risk factor for adult morbidity and mortality (Krebs, 2003). The relationship to education exists because it has been suggested that if a student has a positive experience in physical education and develops a positive attitude about physical activity, motivation to engage in physical activity outside of school is likely to occur (Solmon & Lee, 1996; Wallhead & Buckworth, 2004). Conversely, students who show unfavorable attitudes toward physical education may refrain from indulging in physical activity outside of school (Carlson, 1995; Ennis, 1996; Portman, 1995; Robinson, 1990). Attitude is defined as the degree to which a person likes or does not like something (Ajzen & Fishbein, 1980). Allport (1968, p.56) described attitude as being "the most
7 distinctive and indispensable concept in contemporary American psychology." A person's attitude toward any given object can have either a positive or negative impact. This attitude has the ability to strongly influence a person's behavior. Little is known about elementary school students' attitudes toward physical education. A great deal, however, is known about middle and high school students' attitudes. Currently, there is an instrument measuring middle school student attitude toward physical education developed by Silverman and Subramaniam (2000). Montalvo and Silverman (2008) showed the scores for this measurement also are reliable and valid for urban high school students. There is, however, currently no measure of student attitude toward physical education for students in fifth grade or earlier. If we have an instrument to assess fourth and fifth grade students we can begin to gain an insight into what influences their attitudes toward physical education. This study is grounded in both attitude theory and psychometric theory. The Theory of Reasoned Action guides much of the attitude research in physical education and states that a person's belief system guides their attitude, which ultimately impacts their behavior. The elementary school age gap in the research on physical education is crucial because research suggests that belief systems take hold at a young age and become increasingly difficult to change (Ajzen & Fishbein, 1980). There are a number of psychometric reasons for needing to validate scores of an instrument used in different age groups. An instrument is not transferable to different contexts unless its scores have temporal generalizabihty and ecological generalizabihty. Temporal generalizabihty refers to the need for age and developmental levels to be taken into consideration when constructing the validity of an instrument (Emmerich, 1964,
8 1968; Joreskog & Sorbom, 1979; Kagan, 1971; Messick, 1989; Tucker, 1966). Ecological generalizabihty refers to creating an instrument that can span across age groups, gender, socio-economic groups and geographical locations a (i.e. urban, rural or suburban) (Bracht & Glass, 1968; Messick, 1989). Temporal generalizabihty addresses whether or not the participant's age and development levels are appropriate for the instrument. Currently, there is an instrument available that can examine middle and high school students attitudes toward physical education (Montalvo & Silverman, 2008; Subramaniam and Silverman, 2000). While middle and high school students have been successfully evaluated by the same instrument, this may not be true for younger students. Therefore, it is necessary to examine and, if necessary, adapt the original instrument to determine if the scores are appropriate to fit an elementary school age group and developmental level. Ecological generalizabihty addresses whether or not an instrument is able to span across groups. This ensures that the participants' environmental conditions, gender or age are not going to impact their ability to participate in the study or to alter their responses to the instrument (Bracht & Glass, 1968; Messick, 1989). Altering the original instrument will help ensure that the scores will fit the appropriate audience and will span across gender and fourth and fifth grade students, which will satisfy ecological generalizabihty. Students who are 10 years of age or older do not have any major limitations in participating in quantitative research methodologies (McKay, Halperin, Schwartz, & Sharma, 1994; McKenna, 2001; Rebok et al., 1997). The most common form of quantitative methodology employed in education research on elementary school students'
9 attitudes has been the use of questionnaires. Specifically, Likert-type scale questionnaires have been frequently used for students in upper elementary school, students eight through twelve years old (Andre, Whigham, Hendrickson, & Chambers, 1999; Digelidid, Kamtsios, & Theodorakis, 2007; Garner & Garfinkel, 1979; Maloney, McGuire, & Daniels, 1988; Sainsbury & Schagen, 2004; Shariff & Yasin, 2005; Twist, Gnaldi, Schagen, & Morrison, 2004; Twist, Sainsbury, Woodthorpe, & Whetton, 2003). The purpose of this validation study was to develop an instrument with reliable and valid scores that can assess student attitude toward physical education. Specifically, this instrument was tailored to students in fourth and fifth grade. Having the results from a study done with an instrument that produces reliable and valid scores for upper elementary school students, can start to fill a gap in the research on student attitude toward physical education. Method Development and validation of the scores for an instrument to measure fourth and fifth grade student attitude toward physical education required a multi-phase design including (a) an elicitation study and instrument question development; (b) a pilot study; (c) a pilot study content validity study; (d) a validity study assessment of construct validity and reliability; and (e) a content validity study. Before the elicitation study and item generation, an extensive literature review was conducted to determine whether or not an attitude instrument to assess physical education attitude existed (see Appendix A for literature review). Subramaniam and Silverman (2000) created an attitude instrument, which was selected as a basis for this instrument. It was chosen because the instrument's scores are both reliable and valid in
10 delineating students with high and low attitudes toward physical education in middle school and in high school (Montalvo & Silverman, 2008). Subramaniam and Silverman's twenty-item questionnaire is based on a five point Likert-type scale ranging from 1 (strongly disagree) to 5 (strongly agree), with 8 of 20 items negatively written. The instrument contains two constructs: enjoyment (affect) and usefulness (cognition) each with two contextual sub-factors: the physical education teacher and curriculum. Elicitation Study and Instrument Question Development The first step in altering the questionnaire to fit the appropriate age group was an elicitation study (see Appendix B for elicitation study questions). In this study the students were asked to write in their own words what they liked and disliked about physical education. This study was conducted with 157 fourth and fifth graders in the summer during a sports camp. Once language was ascertained from these students the items were rewritten using their words. The elicitation study suggested physical activity and exercise are important to upper elementary school students and another sub-factor, exercise, was added. This decision, to add the sub-factor of exercise, was determined due to the constant discussion of exercise and sweating noted throughout the 157 elicitation responses. The two constructs of the questionnaire: enjoyment and usefulness initially had three sub-factors each; physical education teacher, curriculum, and exercise, rather than two from the previous work by Subramaniam & Silverman (2000). Eight items were added to the original questionnaire and included wording such as sweating and exercise, using four negatively written items; all worded according to the language used during the elicitation study.
11 Once the questionnaire language was altered and the eight exercise items were added, another step was taken to ensure the usability of the instrument with fourth and fifth grade students. The questionnaires were sent to one elementary school literacy specialist and 15 fourth and fifth grade language arts teachers for feedback. Suggestions were provided and minor alterations were made to the questions. One such alteration for example, was the change in the use of the word "games". The word "games" was changed to "activities" because it was discovered that fourth and fifth grade students do not differentiate between games such as Capture the Flag and other activities such as Project Adventure. Pilot Study Students at four schools took part in the pilot study, all of which are located in suburban areas. The administrators were sent a letter explaining the study from the researcher, along with a copy of the participants' rights and assent form and the instrument, and once these schools agreed to participate the physical education teachers were contacted (see Appendix C for the pilot study administrator explanation letter, Appendix D for the pilot study consent forms and Appendix E for the pilot study questionnaire). A total of 493 participants rights and assent forms were sent to the schools and 373 assent slips were returned. Out of the 367 completed questionnaires, 303 were viable, the other 64 had missing responses. School one had 93 student responses, school two had 128 student responses, school three had 76 student responses and school four had six student responses. There were 141 males and 162 females with 178 fourth graders and 125 fifth graders.
12 Arrangements were made to administer the demographic information sheets (see Appendix F for the pilot study demographic information sheet) and questionnaires. The same instructions for administering the questionnaire were followed each time (see Appendix G for pilot study questionnaire administration protocol). Extra pencils and pens were provided for any students who had forgotten to bring one to class. The students were spread out around the space provided and asked not to talk during the administration of the questionnaire. Students were told what the purpose of the questionnaire was and how to take it, they were reminded that this is an anonymous questionnaire, assured that their teachers would not know who wrote what on any questionnaires and asked to be as honest as possible with their responses. Students placed their completed questionnaires directly into a manila folder facedown so as to remind the students that their anonymity would not be broken. It took approximately 20 minutes for students to complete the questionnaires, including instruction and collection. Pilot questionnaire data analysis. The questionnaire responses were initially entered into an Excel spreadsheet and categorized by school, gender, grade and each item response. Reverse coding was completed to ensure that the negative phrasing did not cause an inaccurate reflection of the students' attitude scores when calculated by summing the scores. These data were then uploaded into SPSS and principal component analysis (PCA) and exploratory factor analysis (EFA) were completed. This information and a scree plot were used for diagnostic purposes. Data were then analyzed using SAS. First a correlation matrix was created. This matrix then was used to run confirmatory factor analysis using structured equation modeling to examine score construct validity and determine if the data fit the model. Fit
13 statistics were examined to determine the Goodness of Fit Index (GFI), the Adjusted Goodness of Fit Index (AGFI), the Root Square Mean Error of Approximation (RSMEA) and the Root Mean Square Residual (RMR). The G-C alpha was examined to determine internal consistency and overall score and factor reliability. The data were analyzed using six factor, four factor and two factor models. Questions were moved between factors as guided by the theoretical model and their factor loadings. The best of each of these factor models were evaluated. The best six factor model actually was the original model created with the questions kept in their original factors. This model included the six sub-factors of (a) cognitive- teacher, (b) cognitive-curriculum, (c) cognitive-exercise, (d) affect-teacher, (e) affect- curriculum and (f) affect-exercise. The fit statistics were at a moderate level (GFI = 0.74, AGFI = 0.698, RMR = 0.687, RMSEA = 0.0884, Cronbach's Alpha cognitive- teacher = 0.77, cognitive-curriculum = 0.673, cognitive-exercise = 0.534, affect-teacher = 0.795, affect-curriculum = 0.814 and affect-exercise = 0.774). The data were then run using a four factor model. As in the six factor model, questions were shifted around as per their factor loadings and theoretical model. The best four factor model fit statistics were discovered when all the exercise questions were removed, except for two affect- exercise questions which were moved to the affect- teacher factor. These fit statistics were also moderate (GFI = 0.77, AGFI = 0.72, RMR = 0.688, RMSEA = 0.097, Cronbach's Alpha cognitive-teacher = 0.77, cognitive- curriculum = 0.673, affect-teacher = 0.85, affect-curriculum = 0.814 and overall Cronbach's Alpha = 0.997).
14 Throughout the entire process the factor loadings yielded results that strongly encouraged running a two factor model using only affect and cognition. First a two factor model was run using all the questions. Following this, another two factor model was run dropping all of the exercise questions. Finally a two factor model was run dropping all but two of the exercise questions, the two that were previously noted above as being moved from the affect-exercise sub-factor to the affect-teacher sub-factor. This last model had the highest fit statistics of all the models (GFI = 0.7718, AGFI = 0.7225, RMR = 0.0696, RMSEA = 0.0969, Cronbach's cognitive = 0.851, affect = 0.906 and overall Cronbach's Alpha = 0.997). Results of the pilot study analysis show that the each of the models that were run have the potential to fit the data. Since the final two factor model, which had the best fit statistics, was only marginally better than the original six factor model fit statistics, the questionnaire employed for the validity study initially kept a six factor model. In an effort to maintain this model the questionnaire was slightly altered based upon the guidance of the fit statistics. New questions, based on the wording from the elicitation study, were added to the cognitive-curriculum sub-factor as well as the cognitive-exercise sub-factor. The two questions that fit the content of both the affect-teacher and affect- exercise sub-factors were removed because they did not fit the theoretical model. Pilot Content Validity Study This newly altered questionnaire was presented to a group of physical education doctoral students (JV= 8) to be reviewed for content validity. Each person was asked to read each item and assign it to one of the models categories (see Appendix H for the content validity first questionnaire). The results of this round of content validity were
15 reviewed and based on these results, slight alterations were made in language and questions were added or dropped to strengthen sub-factors. This new questionnaire was then presented to the same group (see Appendix I for the content validity second questionnaire) and the results were reviewed. Minor alterations were made and the study questionnaire was created (see Appendix J for the validation study questionnaire). Validity Study Invitations to participate in the study were sent to school administrators throughout the tri-state area, California, Connecticut and North Carolina (see Appendix K for administrator letter and Appendix L for administrator consent form). After approval by the Institutional Review Board and once board of education permission at each of the schools was granted, permission slips were sent to 2467 fourth and fifth grade students from 13 school districts encompassing 17 different schools and five different states. Students were sent home from school with permission slips inviting them to participate (see Appendix M for the permission slip and assent form). Students (JV= 1452) returned permission slips and there were 1344 viable questionnaires. The questionnaires were administered using the same protocol as for the pilot study (see Appendix N for the validity study protocol for the researcher, Appendix O for the validity study protocol for others, Appendix P for the validity study demographic information sheet). For locations that were not within commuting distance for the researcher, the questionnaire was administered by someone at the school. An effort was made to have someone other than the physical education teacher administer the questionnaire and in four out of the six cases this occurred.