THE RELATIONSHIP BETWEEN PARTICIPATION IN CAMPUS RECREATION
PROGRAMS AND COLLEGE STUDENT ACADEMIC SUCCESS
By
Jessica E. Kiss

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Kinesiology-Doctor of Philosophy
2017

	

	

ABSTRACT
THE RELATIONSHIP BETWEEN PARTICIPATION IN CAMPUS RECREATION
PROGRAMS AND COLLEGE STUDENT ACADEMIC SUCCESS
By
Jessica E. Kiss
The academic success of undergraduate students is necessary for degree attainment and
fulfilling career goals. Universities recognize factors that affect academic achievement and
promote strategies that support satisfactory grades, progression through degree programs, and
graduation for students. It is essential to determine predictors of success for all students but also
just as important to determine factors that may promote success for those who traditionally
struggle academically like first generation students and freshmen. Participation in campus
recreation programs may encourage academic accomplishments. The relationship between
campus recreation and student success has been studied for decades, although consistent results
have not been reported. Therefore, the purpose of this study was to determine the relationship
between participation in specific campus recreation programs and college student academic
success.
Our sample consisted of male and female undergraduate students who participated in club
sports, intramural sports, and group fitness classes at a mid-sized university in the Midwest. A
nonparticipant group of students was also created to compare to the participants. Demographic
data and academic success variables were populated for participants and nonparticipants (e.g.
grade point average, credits completed, and one-year and two-year retention). Propensity score
matching was used to match the participants with nonparticipants based on covariate information
provided by the Registrar. Grade point average and credits completed were compared between
the total sample of participants and nonparticipants, first generation students, and based on the

	

	

number of programs participated in by the students. One-year and two-year retention were
determined for the freshmen sample. Two-sample tests for proportions (z-tests) were used to
compare participants’ and nonparticipants’ GPA, credits completed, one-year retention, and twoyear retention. Effect sizes were calculated to determine the magnitude of effect in significant
findings between participation in campus recreation programs and academic outcomes. ANOVA
was used to compare amount of participation and GPA and credits completed. Linear trend
analyses were conducted to determine significant differences between the groups.
Overall, participants of campus recreation programs completed significantly more credits
per year and obtained a significantly higher GPA per year. First generation participants also
completed significantly more credits per year compared to first generation nonparticipants.
Participants in one campus recreation program per year completed significantly more credits than
nonparticipants. Participants in one campus recreation program and participants in two or more
campus recreation programs both obtained a significantly higher GPA when compared to
nonparticipants. First generation participants in one campus recreation program per year earned a
significantly higher GPA when compared with first generation nonparticipants. Freshmen
campus recreation participants had a significantly higher one-year retention rate and two-year
retention rate. First generation freshmen participants also had a significantly higher one-year
retention rate and two-year retention rate.
The findings of this study should be of interest to students, campus recreation
professionals, and university higher administration. Participation in campus recreation may
possibly influence academic success in all students, including those who are known to struggle
academically. Suggesting participation in campus recreation programs may be a helpful strategy
to use for undergraduate students seeking academic success.

	

	

Copyright by
JESSICA E. KISS
2017

	

	

ACKNOWLEDGEMENTS

First and foremost, I would like to thank my advisor, Dr. Jim Pivarnik. I spent more time
with you than probably most graduate students and I thank you for your patience and guidance
throughout these past six years. You did not have to agree to work with me while I was
employed full time, but you did, and I will forever be grateful. I thank Dr. Kim Maier for helping
me with statistical analyses that were new to me. I have known you as the “go to” statistician at
MSU and I thank you for taking time to help me with my dissertation. I would like to
acknowledge Dr. Rick McNeil for meeting with me so early on in my career as a Ph.D. student
to chat about campus recreation. I love this field and it has opened up so many opportunities for
me. Without you, I may not have found my love for campus recreation research. Dr. Jeremy
Knous, you have been a great friend, coworker, and dissertation committee member. All of this
started with us putting on hardhats, walking through the Health and Human Services Building (at
SVSU) that was under construction during my interview. Thank you for being there for me
throughout my career and graduate program. I would like to thank Dr. Nick Wagner (Director of
Institutional Research at Saginaw Valley State University) for teaching me all there is to know
about Propensity Score Matching and being willing to meet with my committee when we had
questions. You spent a great deal of time on this project and I hope you know how grateful I am
for your expertise. Finally, I would like to thank my campus recreation friends, Aaron Mowen
and Andrea Hamlin for being on board with this project. Thank you for sharing your
participation data with me. I hope the results of this dissertation help you and the campus
recreation field as much as they have helped me achieve my degree.

v	
	

	

To my friends and family in Ohio and Michigan, thank you for always understanding that
“I have work to do.” You have been supportive throughout this whole process and I couldn’t get
to where I’m at without you. Deanna Reynolds-Griffin, John Lowry, Andrea Hamlin, and
Lyndee Zimpfer: we have the world’s best group message. Thank you for bringing laughter to
my life on a daily basis. Thank you for bringing me care packages during comps, presenting at
professional conferences with me (Three’s Company), and being there for me when I just needed
a friend. You are the absolute best.
To my furbabies, Maddox and Josie, I love you. You always understand when I need to
work and have spent countless hours lying next to me while I write. When you insist you need to
go for a run, you give me a reason to take a break and spend some time with you. You are truly a
human’s best friend.
Finally, I want to acknowledge and thank my husband, Shawn. You have only known
“grad school Jess.” You have been amazing all of these years taking on a lot more than your fair
share of the work just so I could progress through my program. If the only way we could spend
time together was you sitting next to me on the couch while I worked, you didn't complain and
always saw the positive. I’m convinced you have earned this degree too. I will forever love you
for being so patient with me. Thank you.

vi	
	

	

TABLE OF CONTENTS

LIST OF TABLES ......................................................................................................................... ix
LIST OF FIGURES ....................................................................................................................... xi
CHAPTER ONE: INTRODUCTION ..............................................................................................1
CAMPUS RECREATION AND ACADEMIC SUCCESS ................................................3
FIRST GENERATION STUDENTS ..................................................................................9
FRESHMEN STUDENTS .................................................................................................12
RESEARCH AIMS............................................................................................................16
ORGANIZATION OF THE DISSERTATION.................................................................17
REFERENCES ..................................................................................................................18
CHAPTER TWO: LITERATURE REVIEW ................................................................................24
INTRODUCTION .............................................................................................................24
THE RELATIONSHIP BETWEEN CAMPUS RECREATION PARTICIPATION
AND GRADE POINT AVERAGE (GPA)........................................................................24
SUMMARY ...........................................................................................................29
THE RELATIONSHIP BETWEEN CAMPUS RECREATION PARTICIPATION
AND RETENTION ...........................................................................................................30
SUMMARY ...........................................................................................................37
THE RELATIONSHIP BETWEEN CAMPUS RECREATION PARTICIPATION
AND MULTIPLE ACADEMIC ACHIEVEMENT OUTCOME MEASURES ...............37
SUMMARY ...........................................................................................................43
FIRST GENERATION STUDENTS ................................................................................44
SUMMARY ...........................................................................................................51
FRESHMEN STUDENTS .................................................................................................51
SUMMARY ...........................................................................................................56
SOPHOMORE STUDENTS .............................................................................................56
SUMMARY ...........................................................................................................61
OVERALL SUMMARY ...................................................................................................61
REFERENCES ..................................................................................................................64
CHAPTERS THREE, FOUR, AND FIVE: METHODS ..............................................................70
ORGANIZATION OF THE METHODS ..........................................................................70
METHODS: CHAPTER THREE ......................................................................................71
METHODS: CHAPTER FOUR ........................................................................................71
METHODS: CHAPTER FIVE ..........................................................................................71
STUDY POPULATION AND RECRUITMENT .............................................................72
Demographic Variables .........................................................................................72
ACADEMIC SUCCESS OUTCOMES .............................................................................73
CAMPUS RECREATION PARTICIPATION .................................................................73
Club Sport Participation .........................................................................................73

vii	
	

	

Intramural Sport Participation ..............................................................................74
Group Fitness Class Participation .........................................................................74
Non-Participants .....................................................................................................74
Procedure................................................................................................................74
STATISTICAL ANALYSES ............................................................................................75
REFERENCES ..................................................................................................................77
CHAPTER SIX: RESULTS ..........................................................................................................79
SPECIFIC AIM #1: CREDITS COMPLETED (TOTAL SAMPLE) ...............................79
SPECIFIC AIM #1: GPA (TOTAL SAMPLE) .................................................................83
SPECIFIC AIM #2: CREDITS COMPLETED (FIRST GENERATION) .......................87
SPECIFIC AIM #2: GPA (FIRST GENERATION) .........................................................92
SPECIFIC AIM #3: CREDITS COMPLETED (TOTAL SAMPLE-AMOUNT OF
PARTICIPATION) ............................................................................................................96
SPECIFIC AIM #3: GPA (TOTAL SAMPLE-AMOUNT OF PARTICIPATION) .......103
SPECIFIC AIM #4: CREDITS COMPLETED (FIRST GENERATION-AMOUNT
OF PARTICIPATION) ....................................................................................................110
SPECIFIC AIM #4: GPA (FIRST GENERATION-AMOUNT OF
PARTICIPATION) ..........................................................................................................116
SPECIFIC AIM #5: ONE-YEAR & TWO-YEAR RETENTION (TOTAL SAMPLEFRESHMEN STATUS) ...................................................................................................123
SPECIFIC AIM #6: ONE-YEAR & TWO-YEAR RETENTION (FIRST
GENERATION-FRESHMEN STATUS)........................................................................129
OVERALL SUMMARY .................................................................................................134
REFERENCES ................................................................................................................136
CHAPTER SEVEN: DISCUSSION ............................................................................................138
CAMPUS RECREATION AND ACADEMIC SUCCESS ............................................138
CREDITS COMPLETED ................................................................................................139
GPA .................................................................................................................................142
ONE-YEAR & TWO-YEAR RETENTION ...................................................................145
FIRST GENERATION STUDENTS ..............................................................................147
FRESHMEN STUDENTS ...............................................................................................152
LIMITATIONS OF THE STUDY...................................................................................155
STRENGTHS ..................................................................................................................157
RECOMMENDATIONS FOR FUTURE RESEARCH..................................................159
CONCLUSIONS .............................................................................................................160
REFERENCES ................................................................................................................162

viii	
	

	

LIST OF TABLES

Table 1.0. Propensity Score Matching Sample Sizes ....................................................................79
Table 1.1. Two-Sample Test for Proportions (z-test) for Participants and Nonparticipants:
Credits Completed .........................................................................................................................83
Table 1.2. Propensity Score Matching Sample Sizes ....................................................................83
Table 1.3. Two-Sample Test for Proportions (z-test) for Participants and Nonparticipants:
GPA ...............................................................................................................................................87
Table 2.0. Propensity Score Matching Sample Sizes ....................................................................88
Table 2.1. Two-Sample Test for Proportions (z-test) for First Generation Participants and
First Generation Nonparticipants: Credits Completed ...................................................................91
Table 2.2. Propensity Score Matching Sample Sizes ....................................................................92
Table 2.3. Two-Sample Test for Proportions (z-test) for First Generation Participants and
First Generation Nonparticipants: GPA .........................................................................................96
Table 3.0. Propensity Score Matching Sample Sizes ....................................................................96
Table 3.1. Descriptives for Participants and Nonparticipants: Credits Completed .....................100
Table 3.2. Test of Homogeneity of Variances .............................................................................101
Table 3.3. ANOVA Results .........................................................................................................101
Table 3.4. Robust Tests of Equality of Means .............................................................................101
Table 3.5. Post Hoc Tests ............................................................................................................102
Table 3.6. ANOVA Table for Linear Trend Analysis .................................................................102
Table 3.7. Propensity Score Matching Sample Sizes ..................................................................104
Table 3.8. Descriptives for Participants and Nonparticipants: GPA ...........................................107
Table 3.9. Test of Homogeneity of Variances .............................................................................108
Table 3.9.1. ANOVA Results ......................................................................................................108

ix	
	

	

Table 3.9.2. Post Hoc Tests .........................................................................................................109
Table 3.9.3. ANOVA Table for Linear Trend Analysis ..............................................................109
Table 4.0. Propensity Score Matching Sample Sizes ..................................................................110
Table 4.1. Descriptives for First Generation Participants and First Generation
Nonparticipants: Credits Completed ............................................................................................114
Table 4.2. Test of Homogeneity of Variances .............................................................................115
Table 4.3. ANOVA Results .........................................................................................................115
Table 4.4. ANOVA Table for Linear Trend Analysis .................................................................115
Table 4.5. Propensity Score Matching Sample Sizes ..................................................................116
Table 4.6. Descriptives for First Generation Participants and First Generation
Nonparticipants: GPA ..................................................................................................................120
Table 4.7. Test of Homogeneity of Variances .............................................................................121
Table 4.8. ANOVA Results .........................................................................................................121
Table 4.9. Post Hoc Tests ............................................................................................................122
Table 4.9.1. ANOVA Table for Linear Trend Analysis ..............................................................122
Table 5.0. Propensity Score Matching Sample Sizes ..................................................................124
Table 5.1. Two-Sample Test for Proportions (z-test) for Freshmen Participants and
Freshmen Nonparticipants: 1-Year Retention .............................................................................128
Table 5.2. Two-Sample Test for Proportions (z-test) for Freshmen Participants and
Freshmen Nonparticipants: 2-Year Retention .............................................................................129
Table 6.0. Propensity Score Matching Sample Sizes ..................................................................129
Table 6.1. Two-Sample Test for Proportions (z-test) for First Generation Freshmen
Participants and First Generation Freshmen Nonparticipants: 1-Year Retention ........................133
Table 6.2. Two-Sample Test for Proportions (z-test) for First Generation Freshmen
Participants and First Generation Freshmen Nonparticipants: 2-Year Retention ........................134

x	
	

	

LIST OF FIGURES

Figure 1.0. Distribution of Propensity Scores................................................................................80
Figure 1.1. Standardized Difference in Means ..............................................................................80
Figure 1.2. Histogram of Propensity Scores Before and After Matching ......................................81
Figure 1.3. Histogram of Standardized Differences Before and After Matching ..........................82
Figure 1.4. Distribution of Propensity Scores................................................................................84
Figure 1.5. Standardized Difference in Means ..............................................................................85
Figure 1.6. Histogram of Propensity Scores Before and After Matching ......................................85
Figure 1.7. Histogram of Standardized Differences Before and After Matching ..........................86
Figure 2.0. Distribution of Propensity Scores................................................................................88
Figure 2.1. Standardized Difference in Means ..............................................................................89
Figure 2.2. Histogram of Propensity Scores Before and After Matching ......................................90
Figure 2.3. Histogram of Standardized Differences Before and After Matching ..........................90
Figure 2.4: Distribution of Propensity Scores................................................................................93
Figure 2.5. Standardized Difference in Means ..............................................................................93
Figure 2.6. Histogram of Propensity Scores Before and After Matching ......................................94
Figure 2.7. Histogram of Standardized Differences Before and After Matching ..........................95
Figure 3.0. Distribution of Propensity Scores................................................................................97
Figure 3.1. Standardized Difference in Means ..............................................................................98
Figure 3.2. Histogram of Propensity Scores Before and After Matching ......................................99
Figure 3.3. Histogram of Standardized Differences Before and After Matching ..........................99
Figure 3.4. Means Plot for Linear Trend Analysis ......................................................................103

xi	
	

	

Figure 3.5. Distribution of Propensity Scores..............................................................................104
Figure 3.6. Standardized Difference in Means ............................................................................105
Figure 3.7. Histogram of Propensity Scores Before and After Matching ....................................106
Figure 3.8. Histogram of Standardized Differences Before and After Matching ........................106
Figure 3.9. Means Plot for Linear Trend Analysis ......................................................................110
Figure 4.0. Distribution of Propensity Scores..............................................................................111
Figure 4.1. Standardized Difference in Means ............................................................................112
Figure 4.2. Histogram of Propensity Scores Before and After Matching ....................................113
Figure 4.3: Histogram of Standardized Differences Before and After Matching ........................113
Figure 4.4. Means Plot for Linear Trend Analysis ......................................................................116
Figure 4.5. Distribution of Propensity Scores..............................................................................117
Figure 4.6. Standardized Difference in Means ............................................................................118
Figure 4.7. Histogram of Propensity Scores Before and After Matching ....................................119
Figure 4.8. Histogram of Standardized Differences Before and After Matching ........................119
Figure 4.9. Means Plot for Linear Trend Analysis ......................................................................123
Figure 5.0. Distribution of Propensity Scores..............................................................................125
Figure 5.1. Standardized Difference in Means ............................................................................125
Figure 5.2. Histogram of Propensity Scores Before and After Matching ....................................126
Figure 5.3. Histogram of Standardized Differences Before and After Matching ........................127
Figure 6.0. Distribution of Propensity Scores..............................................................................130
Figure 6.1. Standardized Difference in Means ............................................................................131
Figure 6.2. Histogram of Propensity Scores Before and After Matching ....................................132
Figure 6.3. Histogram of Standardized Differences Before and After Matching ........................132

xii	
	

	

CHAPTER ONE: INTRODUCTION

The academic success of undergraduate college students is essential for degree attainment
and fulfilling career goals. Acceptable grades in coursework lead to the progression through
degree programs, persistence from semester to semester, and finally, graduation. To many,
degree attainment reflects academic success (Astin & Oseguera, 2012). At a university level,
academic success outcomes are crucial in the operation of the institution. In order for universities
to accomplish institutional goals and operate most effectively, students must be enrolled in
classes continuously, if possible. To this end, universities strive to retain as many students as
possible until they graduate. Unfortunately, enrollment and retention at universities across the
nation has declined causing concerns for department budgets and a demand for programs thought
to improve academic performance, persistence, and graduation rates to increase student retention
(Kampf & Teske, 2013). Consequently, many universities have created centers for academic
success, offer free tutoring, and provide services in hopes of improving academic student success
measures. Furthermore, nonacademic university programs and activities such as faith based
student organizations, leadership development programs, and campus recreation programs are
available as an indirect means for improving academic success (Stier, Schneider, Kampf, Haines
& Wilding, 2005). Involvement in programs such as intramural sports and club sports have been
linked with academic success measures like GPA, credits completed, and retention (Becker,
Cooper, Atkins & Martin, 2009; McElveen & Rossow, 2014). Vincent Tinto’s (1975) student
integration model illustrates the value of such programs to the overall success of college
students. Tinto (1975) suggests that the more students are integrated into their college
communities, the less likely they are to drop out of college. Tinto (1975) states that when

1	
	

	

students do not do well or drop out of school, it is often because they have not been involved in
the campus community. More recently, Swail (2004) noted that it is essential for students to
connect to the campus culture early on in their academic careers by participating in student
organizations and engaging in social traditions. By doing so, students may build friendships with
peers, gain mentors, and connect with faculty members (Swail, 2004). This type of involvement
allows for student integration to the university and may positively influence their institutional
commitment and retention (Swail, 2004). To determine the effectiveness of campus programs,
many universities compare themselves to their peer institutions by analyzing academic outcomes
such as 1st-2nd year retention, retention from freshmen year to graduation, and graduation within
six years (Becker et al., 2009). Because of the great emphasis placed on student success at the
university level, researchers have sought to determine factors that affect academic achievement.
To predict success, the likelihood students will succeed in relationship to variables such
as background, amount of financial aid they are awarded, and if they are first generation (first in
their families to attend college) have been studied (Paulsen & St. John, 1997, 2002; Choy, 2001;
Ishitani, 2003). Additionally, researchers have also surveyed students regarding the effect that
participating in specific on campus programs; like campus recreation programs, has on student
outcomes such as enjoyment, sense of belonging, and building friendships (Henchy, 2013;
Miller, 2011; Moffitt, 2010). Although these success outcomes are not considered academic
outcomes, they may indirectly affect, or modify the effect of other variables on academic
outcome measures (Hall, 2006; Miller, 2011). Astin (1984) suggests the importance of
extracurricular activities for student success in his Theory of Involvement. Astin (1984) explains
that the amount of learning students can gain is proportional to the quality and quantity of their
campus involvement with academic as well as nonacademic programs. Involvement in

2	
	

	

nonacademic programs like campus recreation programs has been a popular entity on campuses
to study due to the magnitude of benefits reported (e.g. stress reduction, holistic wellness,
physical strength, and academic performance) (Lower, Turner & Petersen, 2013). Although
numerous academic and nonacademic benefits of campus recreation participation are cited, some
specific recreation programs have either not been studied sufficiently or have employed different
research methods which have created questionable results (Danbert, Pivarnik, McNeil &
Washington, 2014). Consequently, the issues suggested by Danbert et al. (2014) have been a
particular concern for campus recreation programs and have warranted continued research to
confidently suggest the benefits of campus recreation participation (Forrester, 2015; Todd,
Czyszczon, Carr, & Pratt, 2009).
CAMPUS RECREATION AND ACADEMIC SUCCESS
Student participation in campus recreation programs and academic success has been
studied for several decades dating back to the 1930’s (Washke, 1939). Early research by Washke
(1939) focused on participation in intramural sports, although since then, countless campus
recreation programs have become available and have been successful in drawing student interest
(e.g. group exercise classes, aquatics programs, student wellness programs). Considering each
program, participation may take on many meanings and could prove difficult to reconcile when
attempting to generalize the success of all campus recreation programs. While this issue may
play a role in mixed results over time, promising findings encourage continued research on the
topic of campus recreation program participation and academic success (Todd et al., 2009;
Forrester, 2015). Positive findings are especially important due to budget cuts and campus
recreation departments’ need to justify their existence. The relationship between participating in
campus recreation programs and academic outcomes relates well with university goals and may

3	
	

	

rationalize the continuation of campus recreation programs (Danbert et al., 2014; Lindsey &
Sessoms, 2006; Kampf & Teske, 2013). Consequently, several researchers have studied the
relationships between campus recreation programs and GPA, 1st-2nd year retention, retention
from freshmen year to graduation, credits completed, and graduation rates (Danbert et al., 2014;
Lindsey & Sessoms, 2006; Kampf & Teske, 2013; Miller, 2011). To determine if fitness center
membership at a large sized university in the Midwest United States was related to student
academic success outcomes in first time degree seeking freshmen, Danbert et al. (2014) obtained
membership data from the university’s campus recreation center and academic data from the
university Registrar. The researchers were able to obtain high school GPA, cumulative college
GPA, college credits completed, one year and two year retention, and class standing. College
GPA, college credits completed, one year and two year retention for the members were
compared to the same outcomes for nonmembers. Danbert et al. (2014) found that campus
recreation members completed more credits and had higher GPAs than nonmembers. After the
first year, members were more likely to be retained in school compared to nonmembers. The
same was true after the second year. Kampf & Teske (2013) believe the need for more campus
recreation research is due to the fact that students are not enrolling in college at the rates they
previously did and universities are having trouble retaining the students who are enrolled.
Acknowledging this issue with recruitment and retention, the investigators decided to evaluate
the engagement-persistence relationship in campus recreation programs at a large, four-year,
residential university. Kampf & Teske (2013) obtained club sport, campus recreation student
employment, and student recreation center entry count data from the university’s Division of
Student Affairs Department and demographic data, precollege data, latest term one-year college
GPA, and Fall-to-Fall retention status of students from the Office of Institutional Research.

4	
	

	

Using a regression model, the researchers found that students participating in club sports were
2.22 times more likely to enroll the following academic year as compared to nonparticipants.
Additionally, entry counts into the campus recreation center were correlated with high school
GPA and latest one-year college GPA. Students who used the campus recreation center were also
more likely to be retained. In addition, the amount that students used the recreation center was
also related to retention.
Forrester (2015) noted that previous research has provided the benefits of participating in
campus recreation programs, but multi-institutional research is lacking. Therefore, Forrester
(2015) investigated the relationship between participating in campus recreation sports and
retention using data from a large multi-institutional sample. The 2013 NASPA Assessment and
Knowledge Consortium was distributed to students at 38 universities across the United States
(n=66,970). Of the total sample who completed the survey (n=33,522), approximately 75% of
respondents indicated they used campus recreation sports facilities, programs and services. For
students who participated in campus recreation sports programs, 66.6% indicated campus
recreation programs and 73.9% indicated campus recreation facilities were important in their
decision to stay at their university.
Although campus recreation participation has been found to be a very beneficial exposure
for campus communities, most researchers study students at universities that maintain a high
percentage of Caucasian students and therefore may miss out on studying minority populations
(Washke, 1939; Danbert et al., 2014; Lindsey & Sessoms, 2006; Kampf & Teske, 2013; Miller,
2011). Determining exposures for academic success for these populations is especially important
considering academic success measures (e.g. graduation rates, retention, GPA) are obtained at a
lower rate as compared to majority populations at predominately White universities (Thayer,

5	
	

	

2000; Attinasi, 1989; Rendon, Jalomo, & Nora, 2000; Tierney, 1992; King, 2002; Bowen & Bok,
1998). Lindsey, Sessoms & Willis (2009) suggest that campus recreation research on African
American students attending historically black colleges and universities is lacking, so they
assessed the impact of campus recreation programs on student recruitment and retention for
African American students at a small, private, historically black university in the Southeast
United States. Undergraduate students were asked to complete the National Intramural
Recreational Sports Association’s Quality and Importance of Recreational Services Survey to
report on how important recreational sports facilities and programs were to their college
experiences. When asked about the importance of recreational sports when choosing a university,
47% reported recreational sports were important/very important. Even more students (50%)
reported the availability of recreational sports to be important/very important in deciding to stay
at that university.
While numerous benefits of campus recreation participation have been discovered over
the past eight decades, not all studies support the link between campus recreation participation
and academic success (Hall, 2006; Zizzi, Ayers, Watson, & Keeler, 2004; McElveen & Rossow,
2014). Taking a qualitative approach to determining the relationship between campus recreation
participation and retention, Hall (2006) interviewed eight campus recreation participants from a
small, private, liberal arts college on the West Coast of the United States. Hall developed seven
questions to discuss with each participant. These questions focused on physical activity since
childhood, types of programs students participate in at the university, and role the campus
recreation programs play in continuing education at the university. Hall (2006) found that the
general common themes discussed by participants included friendship, socialization, and being
physically active. The central theme was sense of community. When students were asked “What

6	
	

	

role, if any, has your participation in the campus recreation program played in continuing your
education at the university?” only three out of the eight participants mentioned campus
recreation programs had an influence on their persistence. Furthermore, these three participants
clarified that they continued at the university only in part due to participation in campus
recreation programs. In addition, Zizzi et al. (2004) did not find a positive relationship between
campus recreation participation and academic success. The researchers surveyed 655 students
from a large, mid-Atlantic university to determine demographics and perceived physical activity
rates among student recreation center users and non-users. Although academic success was not a
main study purpose, the investigators analyzed and described it in the study results. Student
recreation center users were asked to complete a survey that was developed for this research
study. Non-users completed a survey that was constructed slightly differently so that they were
not being asked campus recreation participation specific questions. To determine academic
success measures, students were asked to report their college GPA. Results showed there was not
a significant difference in GPA between student campus recreation center users and non-users.
McElveen & Rossow (2014) performed a more focused study designed to determine if there was
a relationship between intramural sports participation and academic success. The researchers
tracked first time in college freshmen to determine if participation in intramural sports had a
relationship with GPA and retention at a small, private university located in the Southeast United
States. Participation in intramural sports was tracked during the Fall 2010 and Spring 2011
semesters. The researchers also obtained demographic data, GPA for both semesters, and
retention from Fall 2010 through Fall 2011. To determine if frequency of participation made a
difference in GPA, the researchers divided participants into the following categories: 1) nonusers, 2) moderate participation (1-3 sports per semester), or 3) heavy participation (4 or more

7	
	

	

sports per semester). Although retention from Fall 2010 to Spring 2011 and Fall 2010 to Fall
2011 was higher in the first time in college students who participated in intramural sports, GPA
did not differ significantly among intramural sports participant groups.
It would likely be of benefit to university campus recreation directors to collect
participation data for the benefit of their campus recreation programs. Although several
researchers have found academic and non-academic benefits of campus recreation programs,
university administrators may believe campus recreation’s role in student success is ancillary,
and some believe campus recreation is a “perk” rather than an entity promoting success (Danbert
et al., 2014; Braunsein & McGrath, 1997; Brandon, 2010). Although campus recreation
participation has been linked to outcome measures such as sense of belonging, improved selfesteem, group bonding, stress reduction, fun, enjoyment and happiness, improved fitness, ability
to build friendships, GPA, retention, credits completed, and higher graduation rates, consistent
agreeable findings would be helpful in supporting the causal argument (Lower et al., 2013).
Overall, different sized and types of university campus recreation centers have been
studied, specific programs within campus recreation have been tracked, academic success
measures have been related to campus recreation participation through survey or by Registrar
records, and specific populations of students have been studied to determine the affect campus
recreation has on student academic success (Danbert et al., 2014; Lindsey, 2012; McElveen &
Rossow, 2014; Lindsey et al., 2009; Kampf & Teske, 2013; Lindsey & Sessoms, 2006; Lower et
al., 2013). Suggestions for continued research in campus recreation are plentiful. Broad
suggestions include continuing to conduct research in campus recreation to provide justification
to campus administrators to fund campus recreation programs. More specifically, analyzing
multiple programs within campus recreation centers may provide a great deal of knowledge

8	
	

	

about each individual program which may also provide campus recreation directors with
justification to maintain programs (Miller, 2011; Lindsey et al., 2009; Moffitt, 2010; Todd et al.,
2009). Additionally, longitudinal campus recreation participation data would help researchers
study this important exposure variable and outcome measures such as academic performance
over time (Henchy, 2013). Finally, it has been suggested that since universities across the United
States are diverse, it is necessary to study a variety of student populations to determine whether
campus recreation participation has the same relationship to academic success for all (Henchy,
2013; Lindsey & Sessoms, 2006; Lindsey et al., 2009; McElveen & Rossow, 2014). Researchers
have previously found the value of studying populations like African American and
Latino/Latina students’ and their participation in campus recreation (Lindsey & Sessoms, 2006;
Lindsey et al., 2009), but first generation student campus recreation participation has been
studied less.
FIRST GENERATION STUDENTS
Several subgroups of college students are studied to determine probability of their
academic success (Lohfink & Paulsen, 2005; Lindsey et al., 2009). Students from different
racial/ethnic backgrounds, with varying financial aid statuses, different high school GPAs and
ACTs, and who participate in certain on campus programs are just a few groups studied (Lindsey
& Sessoms, 2006; Lindsey et al., 2009; Lohfink & Paulsen, 2005; Thayer, 2000). Another
subgroup that may fall under several of the above categories is first generation students. These
students may have a more difficult transition into college and lower retention rates compared to
students whose family culture includes college graduation (Choy, 2001; London, 1989, 1996;
Nunez & Cuccaro-Alamin, 1998; Ishitani, 2003).

9	
	

	

Although there is not a standard definition for first generation college students, there are
several suggestions as to how a university should define them (Lohfink & Paulsen, 2005).
Thayer (2000) suggests a student can be defined a first generation student if neither of his or her
parents has earned a Bachelor’s degree. Ishitani (2003) agrees with Thayer but adds that first
generation students could also include those whose parents started college but never attained a
degree. Pascarella, Wolniak, Pierson & Terenzini (2003) suggest that first generation students
are students whose parents have never attended college. McConnell (2000) specifies that a first
generation student is an undergraduate student whose parents have no college experience.
London (1989) narrows the definition of first generation student down to include only students
who are the first in their families to go to college. Finally, Inkelas, Daver, Vogt & Leonard
(2007) define a first generation student as a college student whose parents have a high school
education or less and did not begin a postsecondary degree. Regardless of the precise definition,
researchers who study first generation students agree that students who fall into this category are
disadvantaged when compared to students whose parents went to college and obtained degrees
(Darling & Smith, 2007; Lohfink & Paulsen, 2005; Thayer, 2000).
Terenzini et al. (1996) reports that first generation students they studied were different
than traditional students on many levels. For example, when students completed pre-college
admissions forms, first generation students differed from traditional students on 14 of the 37
characteristics surveyed. Some questions that were answered differently by first generation
students related to race/ethnicity, family income, gender, age, if they had children, and degree
aspirations (Terenzini et al., 1996). Pascarella et al. (2004) studied students at multiple
universities to obtain results about the differences between first generation and traditional
students. They found that first generation students enrolled in and earned fewer credit hours,

10	
	

	

worked more hours, more likely lived off campus, participated in less extracurricular activities
outside of the classroom, had fewer non-academic interactions with peers, and earned lower
grades (Pascarella et al., 2004). Additionally, Choy (2001), Richardson and Skinner (1992), and
Terenzini et al. (1996) found that first generation students are less prepared academically, have
lower math, reading, and critical thinking skills, and tend to attend high schools with less
rigorous curricula. Choy (2001) also found that first generation students were more likely to
leave a 4-year institution before their second year when compared to traditional students.
Moreover, first generation students are less likely to return to college after early departure when
compared with traditional students who drop out (Choy, 2001). Researchers analyzing results of
national surveys designed to follow students over time have also found similar results for first
generation students. The National Education Longitudinal Study (2000) is a study that tracks
cohorts of students from the 8th grade through college. The researchers note that these students
are less likely to remain enrolled and graduate compared to traditional students, even when
controlling for family income, academic preparation, and support from family and educators.
Studying this population of students is becoming more important due to increased
numbers of first generation students enrolling in college (Mitchell, 1997; Padron, 1992;
Terenzini et al., 1996). As with any group of students who may be disadvantaged transitioning to
college, Inkelas et al. (2006) suggest that student affairs professionals implement approaches to
improve student involvement. This involvement may help students build a greater connection
with the university they attend, which in return may help retain these students and promote
academic achievement. Unfortunately, this population of students has not been studied to
determine if participation in campus recreation programs is related to their academic success.

11	
	

	

Clearly, it is necessary to determine if many of the reported benefits that have been found due to
campus recreation participation in other populations holds true for first generation students.
FRESHMEN STUDENTS
Freshmen students are another subgroup of concern when focusing on college student
success outcomes. Although student success outcomes may be defined by GPA or amount of
credits completed per academic year, retention is often a concern of university administrators due
to the higher attrition rate of freshmen compared to students in other class standings (Kampf &
Teske, 2013). Considering that 20-30% of freshmen may not be retained from the first to second
year in college, it is imperative to study this subgroup to determine strategies to improve
retention (Kampf & Teske, 2013). Mallinckrodt & Sedlacek (1987) suggest that since freshmen
attrition rates are higher than any other academic class, universities often offer programs
designed to reduce attrition, with particular emphasis on this group. Further, Miller (2011)
suggests that students not involved in campus communities early in their college careers are less
likely to persist. Therefore, it is necessary to determine programs on campuses that encourage
participation and promote retention.
Campus recreation departments offer various programs students may take part in that
promote the continuation of enrollment at universities. Miller (2011) suggests campus recreation
programs are an important entity in improving retention because they may play a role in
improving students’ sense of belonging at the university, which may improve persistence and
increase overall satisfaction with the university. Campus recreation programs are especially
important for freshman retention efforts since the majority of program participants are freshmen
and sophomore students (NIRSA, 2010). Although several cross sectional design based studies
have been successful in identifying a positive relationship between participation in campus

12	
	

	

recreation programs and retention (Danbert et al., 2014; Kampf & Teske, 2013; McElveen &
Rossow, 2014), other researchers have not found the same results with freshmen students
(Lindsey & Sessoms, 2006). Therefore, it is necessary to continue to study this subgroup to
determine the relationship between campus recreation participation in the first year of college
and retention.
Danbert et al. (2014) suggest that even small differences in retention can be substantial to
a university. The researchers obtained membership data from the university’s campus recreation
center and academic data from the university Registrar to determine if campus recreation center
membership was related to academic success measures. The researchers obtained high school
GPA, cumulative college GPA, college credits completed, one year and two year retention, and
class standing. After comparing the academic success outcomes of the members and
nonmembers, Danbert et al. (2014) found that campus recreation members completed more
credits and had higher GPAs than nonmembers. After the first year, members were more likely
to be retained compared to nonmembers. The same was found after the second year. Although
not statistically significant, campus recreation members were 3.5% more likely to be retained by
the second year, compared to nonmembers. While a difference of 3.5% may not seem
substantial, it resulted in almost 1600 students being retained at this large, Midwest University.
The financial benefit to the university for retaining even a small percentage of students can be
great. Kampf & Teske (2013) agree that financial gains for retaining students is important, as is
the goal to justify programs on campuses that support retention. Acknowledging attrition rates
are highest with freshmen students, Kampf & Teske (2013) studied the relationship between club
sports participation, being a campus recreation employee, and student recreation center entry
count data and retention among freshmen at a large, four-year, residential university. The

13	
	

	

researchers obtained campus recreation participation and employment data from the university’s
Division of Student Affairs Department and demographic data, precollege data, latest term oneyear college GPA, and Fall-to-Fall retention status of students from the Office of Institutional
Research. The researchers found that freshmen students participating in club sports were 2.22
times more likely to enroll the following academic year, compared to nonparticipants.
Additionally, students who used the campus recreation center were also more likely to be
retained. Frequency of recreation center use was also related to retention. McElveen & Rossow
(2014) also found encouraging results relating campus recreation participation (intramural
sports) to retention in first time college students. Using a cross sectional design, the researchers
collected intramural participation data from a sports participation database and retention data
from the institutional database through the university’s Department of Institutional Research.
Retention rates included 1) proportion of students who started in the fall 2010 semester and also
enrolled in the spring 2011 semester, 2) proportion of students who enrolled in the spring 2011
semester and also in the fall 2011 semester, 3) proportion of students who enrolled in the fall
2010 semester and also in the fall 2011 semester. After comparing campus recreation participants
to nonparticipants, McElveen & Rossow (2014) reported that the retention rate from fall 2010 to
spring 2011 was 4.7% greater in participants than nonparticipants. Retention was greater in
participants from spring 2011 to fall 2011 at 2.2%. Finally, the retention rate from fall 2010 to
fall 2011 was 5.9% greater in participants compared to nonparticipants.
Lindsey & Sessoms (2006) also studied the relationship between campus recreation
participation and retention for undergraduate students at a small southeastern university.
The researchers distributed the NIRSA Quality and Importance of Recreational Services Survey
(QIRS) to students enrolled in Department of Physical Education and Health classes. The survey

14	
	

	

provides many sections of questions, although the researchers were most interested in questions
focusing on the quality of recreational sports facilities and programs. Furthermore, the
researchers were also interested in the effect students believed campus recreation had on
retention. After collecting the data, the researchers stratified class standing results to determine if
there were any differences in responses about the relationship between campus recreation
participation and retention. When comparing juniors and seniors to freshmen and sophomores,
there was a significant difference in perceived retention due to campus recreation participation in
the junior and senior level students. Although methods of this study were different than those of
studies that showed positive relationships between campus recreation participation and retention,
it is necessary to note the differences in results.
In summary, freshmen students remain a subgroup of interest as attrition rates continue to
be highest in this population as compared with more advanced students (Kampf & Teske, 2013).
While the majority of previous research suggests positive relationships between campus
recreation participation and retention for freshmen students, this has not been found in all
studies. Furthermore, few researchers have studied the relationship between campus recreation
participation and retention past the sophomore year (Danbert et al., 2014). While reasons for
attrition during or after the second year in college are reportedly different (e.g. students’ initial
expectations and the school's ability to deliver and meet those expectations, Boivin, Beuthin, &
Hauger, 1993; Boivin, Fountain, & Baylis, 2000), this group holds the next highest attrition rate
when comparing class-standing attrition (Adelman, 2006) among years in school. Therefore,
positive relationships between campus recreation participation and student success outcomes like
retention justify the importance of such non-academic entities on campuses. Encouraging results

15	
	

	

provide evidence to university administrators for financial support of such departments, although
consistent results are necessary for continued support.
RESEARCH AIMS:
Specific Aim 1: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) and the following academic success measures:
credits completed per academic year and GPA.
H 1.1. Campus recreation program participants will have more credits completed per
academic year and a higher GPA compared to nonparticipants.
Specific Aim 2: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) and the following academic success measures:
credits completed per academic year and GPA in first generation students.
H 2.1. First generation campus recreation program participants will have more credits
completed per academic year and a higher GPA compared to first generation
nonparticipants.
Specific Aim 3: To determine if a relationship exists between amount of campus recreation
participation (club sports, intramural sports, group fitness classes) and the following academic
success measures: credits completed per academic year and GPA.
H 3.1. Students who participate in more than one campus recreation option will have
more credits completed per academic year and higher GPA compared to students
participating in only one, or no campus recreation options.
Specific Aim 4: To determine if a relationship exists between amount of campus recreation
participation (club sports, intramural sports, group fitness classes) and the following academic
success measures: credits completed per academic year and GPA in first generation students.

16	
	

	

H 4.1. First generation students who participate in more than one campus recreation
option will have more credits completed per academic year and higher GPA compared to
students participating in only one, or no campus recreation options.
Specific Aim 5: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) in freshmen students and the following
academic success measures: one-year retention and two-year retention.
H 5.1. Freshmen campus recreation program participants will have more favorable oneyear retention and two-year retention profiles compared to freshmen nonparticipants.
Specific Aim 6: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) in first generation freshmen students and the
following academic success measures: one-year retention and two-year retention.
H 6.1. First generation freshmen campus recreation program participants will have more
favorable one-year retention and two-year retention profiles compared to first generation
freshmen nonparticipants.
ORGANIZATION OF THE DISSERTATION
This dissertation will be organized into seven chapters. Chapter one includes the
introduction and research aims. A review of the literature is covered in chapter two. Chapters
three, four, and five include the study methodology for the different aims. Chapter six will
include results of the data collected. Finally, chapter seven will include a discussion and
summary of results, strengths and weaknesses of the study, and suggestions for future research.

17	
	

	

REFERENCES

18	
	

	

REFERENCES

Adelman, C. (2006). The toolbox revisited. Paths to degree completion from high school through
college. Washington, DC: US Department of Education.
Astin, A. W. (1984). Student involvement: A development theory for higher education. Journal
of College Student Personnel, 25(4), 297-308.
Attinasi, L. C., Jr. (1989). Getting in: Mexican Americans’ perceptions of university attendance
and the implications for freshman year persistence. Journal of Higher Education, 60, 247–
277.
Becker, C., Cooper, N., Atkins, K. & Martin, S. (2009). What helps students thrive? An
investigation of student engagement and performance. Recreational Sports Journal, 33, 139149.
Boivin, M., Fountain, G., & Baylis, B. (2000). Meeting the challenges of the sophomore year.
in L.A. Schreiner & J. Pattengale. (Eds.), Visible solutions for invisible students: Helping
sophomores succeed (Monograph No. 31). (pp. 1-18). Columbia, SC: University of South
Carolina, National Resource Center for the First-year Experience and Students in Transition.
Boivin, M. J., Beuthin, T. M., & Hauger, G. (1993). Why Christian students leave Christian
colleges: Evaluating the dynamics of adjustment in a Christian community. Journal of the
First-Year Experience & Students in Transition, 5(1), 93-125.
Bowen, W. G., & Bok, D. (1998). The shape of the river: Long-term consequences of
considering race in college and university admissions. Princeton, NJ: Princeton University
Press.
Brandon, C. (2010). Five year party. Dallas, TX: Benbella Books Inc.
Choy, S. (2001). Students whose parents did not go to college: Postsecondary access,
persistence, and attainment. US Department of Education, National Center for Education
Statistics, 2001-126.
Danbert, S. J., Pivarnik, J. M., McNeil, R. N. & Washington, I. J. (2014). Academic success and
retention: The role of recreational sports fitness facilities. Recreational Sports Journal, 38,
14-22.
Darling, R. A. & Smith, M. S. (2007). First-generation college students: First-year challenges. In
M.S. Hunter, B. McCalla-Wriggins & E.R White (Eds.), Academic Advising: New insights for
teaching and learning in the first year: pp. 203-211. Manhattan, KS: NACADA Monographs.
Forrester, S. (2015). Benefits of collegiate recreational sports participation: Results from the

19	
	

	

2013 NASPA assessment and knowledge consortium study. Recreational Sports Journal 39,
2-15.
Hall, D. A. (2006). Participation in a campus recreation program and its effect on student
retention. Recreational Sports Journal, 30, 40-45.
Henchy, A. (2011). The influence of campus recreation beyond the gym. Recreational Sports
Journal, 35, 174-181.
Henchy, A. (2013). The perceived benefits of participating in campus recreation programs and
facilities: A comparison between undergraduate and graduate students. Recreational Sports
Journal, 37, 97-105.
Inkelas, K. K., Daver, Z. E., Vogt, K. E. & Leonard, J. B. (2007). Living-learning programs and
first-generation college students’ academic and social transition to college. Research in
Higher Education, 48(4), 403-434.
Inkelas, K. K., Vogt, K., Longerbeam, S., Owen, J. & Johnson, D. (2006). Measuring outcomes
of living-learning programs: Examining college environments and student learning and
development. Journal of General Education, 54(4), 294-328.
Ishitani, T. T. (2003). A longitudinal approach to assessing attrition behavior among firstgeneration students: Time-varying effects of pre-college characteristics. Research in Higher
Education, 44, 433-449.
Kampf, S. & Teske, E. J. (2013). Collegiate recreation participation and retention. Recreational
Sports Journal, 37, 85-96.
King, J. E. (2002). Crucial choices: How students’ financial decisions affect their academic
success. Washington, DC: American Council on Education.
Lindsey, R. R. (2012). The benefits and satisfaction of participating in campus recreational
sports facilities and programs among male and female African American students: A pilot
study. Recreational Sports Journal, 36, 13-24.
Lindsey, R. & Sessoms, E. (2006). Assessment of a campus recreation program on student
recruitment, retention, and frequency of participation across certain demographic variables.
Recreational Sports Journal, 30, 30-39.
Lindsey, R., Sessoms, E. & Willis, G. (2009). Impact of campus recreational sports facilities and
programs on recruitment and retention among African American students: A pilot study.
Recreational Sports Journal, 33, 25-34.
Lohfink, M. M. & Paulsen, M. B. (2005). Comparing the determinants of persistence for firstgeneration and continuing-generation students. Journal of College Student Development,
46(4), 409-428.

20	
	

	

London, H. B. (1989). Breaking away: A study of first-generation college students and their
families. American Journal of Education, 97, 144-170.
London, H. B. (1996). How college affects first-generation college students. About Campus, 23,
9-13.
Lower, L. M., Turner, B. A. & Petersen, J. C. (2013). A comparative analysis of perceived
benefits of participation between recreational sport programs. Recreational Sports Journal,
37, 66-83.
Mallinckrodt, B. & Sedlacek, W.E. (1987). Student recreation and the use of campus facilities by
race. Journal of Student Affairs Research and Practice, 24, 28-32.
McConnell, P. J. (2000). What community colleges should do to assist first-generation students.
Community College Review, 28.3, 75-87.
McElveen, M. & Rossow, A. (2014). Relationship of intramural participation to GPA and
retention in first-time-in-college students. Recreational Sports Journal, 38, 50-54.
McGrath, M. & Braunstein, A. (1997). The prediction of freshmen attrition: An examination of
the importance of certain demographic, academic, financial and social factors. College
Student Journal, 31(3), 396-408.
Miller, J. J. (2011). Impact of a university recreation center on social belonging and student
retention. Recreational Sports Journal, 35, 117-129.
Mitchell, K. (1997). Making the grade: Help and hope for the first-generation college student.
Educational Resources Information Center.
Moffitt, J. (2010). Recreating retention. Recreational Sports Journal, 34, 24-33.
NIRSA/NASPA (2010). NIRSA/NASPA consortium campus recreation impact study-descriptive
summary report. Retrieved from
http://www.nirsa.org/docs/Discover/Research/Campus_Rec_Impact_Study.PDF
Nunez, A. & Cuccaro-Alamin, S. (1998). First-generation students: Undergraduates whose
parents never enrolled in postsecondary education (NCES 98-082). Washington, DC: U.S.
Department of Education, Office of Educational Research and Improvement.
Padron, E. J. (1992). The challenge of first-generation college students: A Miami-Dade
perspective. In: Zwerling, L. S. and London, H. B. (eds.), First-Generation College Students:
Confronting the Cultural Issues (New Directions for Community Colleges, No. 80, JosseyBass, San Francisco, pp. 71-80.
Pascarella, E. T., Pierson, C. T., Wolniak, G. C. & Terenzini, P. T. (2004). First generation

21	
	

	

college students: Additional evidence on college experiences and outcomes [Electronic
version]. The Journal of Higher Education, 75, 249-284.
Pascarella, E. T., Wolniak, G. C., Pierson, C. T. & Terenzini, P. T. (2003). Experiences and
outcomes of first-generation students in community colleges. Journal of College Student
Development, 44(3), 420-429.
Paulsen, M. B. & St. John, E. P. (1997). The financial nexus between college choice and
Persistence. In R. A. Voorhees (Ed.), Researching Student Aid: Creating an action Agenda
(New Directions for Institutional Research No. 95, pp. 65-82). San-Francisco: Jossey-Bass.
Paulsen, M. B. & St. John, E. P. (2002). Social class and college costs. The Journal of Higher
Education, 73(2), 189-236.
Rendon, L. I., Jalomo, R. E., & Nora, A. (2000). Theoretical considerations in the study of
minority student retention in higher education. In J. M. Braxton (Ed.), Reworking the student
departure puzzle (pp. 127–156). Nashville: Vanderbilt University Press.
Richardson, R. C. & Skinner, E. F. (1992). Helping first-generation minority students achieve
degrees. In Zwerling, L. S., and London, H. B. (eds.), First-Generation College Students:
Confronting the Cultural Issues (New Directions for Community Colleges, No. 80, JosseyBass, San Francisco, pp. 29-43.
Stier, W. F., Schneider, R. C., Kampf, S., Haines, S. G. & Wilding, G. E. (2005). A survey of
campus recreation directors at NIRSA institutions: Activities emphasized, student
participation patterns, trends and future offerings contemplated. Kinesiology, Sport Studies
and Physical Education Faculty Publications. Paper 45.
Swail, W. S. (2004). The art of student retention: A handbook for practitioners and
administrators. Austin, TX: Educational Policy Institute. Retrieved from:
http://www.educationalpolicy.org
Terenzini, P. T., Springer, L., Yaeger, P. M., Pascarella, E. T. & Nora, A. (1996). Firstgeneration college students: Characteristics, experiences, and cognitive development.
Research in Higher Education, 37, 1-22.
Thayer, P. B. (2000). Retention of students from first generation and low income backgrounds.
The Journal of the Council for Opportunity in Education, 1-9.
Tierney, W. (1992). An anthropological analysis of student participation in college. Journal of
Higher Education, 63, 603–618.
Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research.
Review of Educational Research, 45(1), 89-125.
Todd, M. K., Czyszczon, G., Carr, J. W. & Pratt, C. (2009). Comparison of health and academic

22	
	

	

indices between campus recreation facility users and nonusers. Recreational Sports Journal,
33, 43-53.
Washke, P. (1940). A study of intramural sports participation and scholastic attainment.
Research Quarterly, 11(2), 22-27.
Zizzi, S., Ayers, S. F., Watson, J. C. & Keeler, L. A. (2004). Assessing the impact of new student
campus recreation centers. NASPA Journal, 41(4), 588-630.

23	
	

	

CHAPTER TWO: LITERATURE REVIEW

INTRODUCTION
The relationship between campus recreation participation and student academic success
has been studied to determine whether this component of campus life might aid students’ efforts
to maximize their college experience (Stier, Schneider, Kampf, Haines & Wilding, 2005). While
some researchers study campus recreation as a whole entity and analyze participation benefits
(Forrester, 2015; Lindsey, Sessoms & Willis, 2009; Hall, 2006; Henchy, 2011; Forrester, 2015;
Todd et al., 2009; Henchy, 2013), others study specific programs within campus recreation to
determine individual program benefits (McElveen & Rossow, 2014; Danbert et al., 2014; Kampf
& Teske, 2013). Furthermore, academic success, may be defined in different ways; therefore, a
variety of success outcomes are measured (Baker, 2008; Danbert et al., 2014; Hall, 2006; Kampf
& Teske, 2013). Regardless, an understanding of previous studies focusing on campus recreation
participation and student academic success is necessary to justify the need for additional research
where past research is lacking. Determining appropriate methods to study the relationship
between campus recreation and academic success outcomes is necessary to provide a strong
argument to university administrators who decide the fate of campus recreation programs
(Danbert et al., 2014; Lindsey & Sessoms, 2006; Kampf & Teske, 2013).
THE RELATIONSHIP BETWEEN CAMPUS RECREATION PARTICIPATION AND
GRADE POINT AVERAGE (GPA)
Intramural sports participation and student academic success have a long research history
dating back to the 1930s and 1940s. An early study by Paul Washke (1940) at the University of
Oregon focused on the effect intramural sports had on GPA in male students. Record keeping of

24	
	

	

participation in intramural sports in 1940 was organized so that each participant had a record that
listed his name, class, organization represented (fraternity), sports participated in and number of
activities he participated in throughout the academic year. These records were kept on file within
the School of Physical Education and were updated by the athletic manager of each organization
whenever a student participated in an intramural sports program. Washke chose to analyze
participation in intramural sports from 1931-1936. To define participation in intramural sports
year round, students who participated in five to eight intramural sports per year were included in
the analysis of the study. Washke (1940) believed that students participating in one or two sports
per year had “spasmodic” participation and therefore would not be considered consistent
participants. After defining participation, the sample size for the five years of intramural sports
participation included 271 students. Variables Washke (1940) controlled for included students’
age, class standing, major, rating on psychological entrance exams, and quartile standing in high
school. Grades earned at the undergraduate level for each participant were pulled from each
record. To compare differences in GPA, a control group (n=271) was developed that included
students who did not participate in intramural sports. A third comparison group was also formed.
This group consisted of all men enrolled at the University of Oregon. To calculate GPA,
participants earned the following points for each credit hour (e.g. A=3 points, B=2 points, C=1
point, D=0 points, F=-1 point). All points per semester were added up and then divided by the
total hours taken to determine the GPA. Furthermore, GPA over all five years was calculated as a
cumulative value. Intramural sports participants scored a GPA of 6.395 for the five-year period.
The non-participant group averaged a 6.273 GPA and the overall male GPA was 6.289.
Although this GPA point scheme is different from that which is currently used at most
universities, participants in the intramural sports group had a higher GPA than nonparticipants

25	
	

	

(+0.122) and the overall male student group (+0.106). When reviewing data per year, the
intramural group showed a higher GPA in all years except 1932-1933 when the nonparticipants
had a higher GPA. Additionally, the overall male group had a higher GPA than the intramural
sports group in three of the five years. Although GPA was higher in the intramural sports group
when analyzing the data over a period of five years, it was not always higher than the nonparticipant and overall male groups for each individual year. These results began the theme of
“mixed results” that have occurred in campus recreation research for the past several decades;
therefore, Washke suggested the necessity to continue to conduct research since few studies have
been conducted prior to 1939.
Baker (2008) found mixed results when studying the relationship between intramural and
club sports participation and GPA. He argued it was beneficial to study the relationship between
campus recreation programs and academic performance because of the gaps in the research that
ignore under-represented college students. Therefore, the purpose of Baker’s study was to use
survey data from the National Longitudinal Study of Freshmen (NLSF) to determine if there
were relationships between extracurricular activities and academic performance; specifically
GPA, in a sample of African American and Latino students. The NLSF was distributed to college
freshmen at 28 universities across the United States. A total of 3,924 students completed the
survey, but since Baker only wanted to focus on under-represented students, he chose to analyze
the data of African American and Latino students only. The analytic sample consisted of 1,907
students from 27 universities. The survey was completed by participants at the beginning of their
freshmen year and at the end of their remaining years of college. Baker considered participation
in recreation sports as involvement in intramural sports and/or club sports. After analysis, the
only significant findings between intramural sports and/or club sports participation and GPA

26	
	

	

were found with the Latina population in the study. Baker determined that there was a negative
effect between participation in intramural or club sports and GPA. Participation in intramural and
club sports was not detrimental to GPA for African American men and women and Latino men.
He suggested that not finding a significant difference in GPA in most study participants was
reassuring in that participation in intramural and club sports is not inimical to GPA for those
populations. Although this study is exceptional because of the longitudinal data analyzed, the
results are strictly representative of African American and Latino/Latina students attending elite
colleges. These results are not generalizable to all students at all universities. Future research
should study why Latina students’ academic success outcomes were negatively affected by
club/intramural sports participation. Overall, this is a unique study focusing on subpopulations of
college students who are usually overlooked.
Zizzi, Ayers, Watson II & Keeler (2004) proposed GPA as a success outcome to promote
the use of campus recreation programs for recruiting purposes at universities. The researchers
believed that if positive relationships could be found between campus recreation participation
and student success measures, campus recreation participation benefits could be displayed during
recruiting efforts to prospective students. Unfortunately, Zizzi et al. (2004) explained that
researchers have not assessed the impact of campus recreation facilities on campus climate and
student activity patterns. Therefore, Zizzi et al. (2004) studied the impact of a new student
recreation center on campus climate measures at a mid-sized, mid-Atlantic United States
university. The researchers distributed surveys to about 23,000 students at the university. Of the
23,000 students, 655 completed the survey. Of this sample, 537 students were campus recreation
users and 118 students were nonusers. Campus recreation users were asked to complete one
version of the survey while the students who identified as nonusers completed another version of

27	
	

	

the survey. For the user’s survey, questions about usage patterns, satisfaction with campus
recreation, facility climate, exercise knowledge, health behaviors, sociocultural factors, and
programming issues were included. In the nonuser’s survey, items that did not pertain to the
nonusers were removed from the survey. To determine academic success, the researchers
evaluated students’ GPA. When reviewing self-reported GPA, it was determined that there were
no significant differences between the users and nonusers (3.11, 3.19; respectively). Although
the researchers found interesting results about motivation patterns, exercise frequency, and
comparing new and habitual exercisers, the GPAs between the users and nonusers did not show a
significant difference, leading to results that differ from those of past research.
Based on the results of early studies presented above, more research is needed to
determine if campus recreation participation and student success are positively related. To this
point, Todd, Czyszczon, Carr & Pratt (2009) acknowledged the need for more research due to
mixed findings of previous studies. Among other non-academic variables, GPA was the
academic outcome of interest Todd et al. (2009) investigated in their study. The purpose of their
study was to compare GPA, physical activity, intake of selected nutrients, BMI, smoking status,
and the use of electronic media between nonusers, low users, moderate users, and high users of
campus recreation facilities. Study participants included undergraduate students enrolled at a
particular institution during a spring semester. After recruiting efforts were completed, 1,034
students comprised the sample. Gender ratio was comparable to the ratio of the total population
of the university (658 women, 376 men). Participants completed an electronic survey that
included the informed consent, demographic data, electronic media use, the International
Physical Activity Questionnaire (IPAQ), the National Cancer Institute diet screener, and
smoking frequency. A subset of the sample was asked to complete the survey again in paper

28	
	

	

format. In addition to these data, GPA was retrieved from the Registrar and campus recreation
visitation data were provided via swipe card system by the campus recreation department. After
analysis, the researchers found that the campus recreation high user group had the highest GPA,
although when grouping the three campus recreation user groups together, GPA was comparable
to the nonuser group GPA. To further examine the relationship between campus recreation
participation and student success measures, Todd et al. (2009) suggested that each campus
recreation program should be evaluated individually to determine its specific benefits. Overall,
Todd et al. (2009) showed positive effects of campus recreation participation for individuals who
would be considered high users, although as a whole when all user groups were combined, there
was not a significant difference in GPA compared with the nonusers. These results add to the
inconsistent findings reported in the literature.
SUMMARY
Mixed results tend to be common when studying the relationship between campus
recreation participation and GPA. While some researchers find a positive relationship between
participating in campus recreation programs and student academic success measures (Paul
Washke, 1940), other researchers do not (Baker, 2008; Zizzi et al., 2004; Todd et al., 2009).
Furthermore, researchers like Baker (2008) suggest that no difference in GPA is still an
important finding suggesting that students who participate in campus recreation programs are
able to do so without it affecting their GPA. Regardless, consistent results between campus
recreation participation and student academic success are needed to solidify the suggestion of a
positive relationship between campus recreation participation and student academic success.
Several suggestions are proposed for future research to achieve more consistent results. First,
additional research is needed at universities across the nation to improve the ability to generalize

29	
	

	

results. Different types of universities and universities with varying student populations are
needed to establish results that may be used at other universities. Furthermore, it is necessary to
study subpopulations of students to determine if they benefit similarly from campus recreation
participation. University undergraduate populations consist of individuals from various
backgrounds. Some of these populations have been shown to have a more difficult time with the
transition from high school to college (e.g. minority students, first generation) (Choy, 2001;
London, 1989; London, 1996; Nunez & Cuccaro-Alamin, 1998). Determining programs that are
related to improving academic success is crucial. Finally, it is suggested that not only are campus
recreation programs as a whole evaluated in their relationship with academic success, but
campus recreation programs should also be analyzed individually to determine the benefits each
program has to offer (Miller, 2011; Moffitt, 2010; Todd, et al., 2009). These recommendations
are reasonable approaches for future research.
THE RELATIONSHIP BETWEEN CAMPUS RECREATION PARTICIPATION AND
RETENTION
Research focusing on campus recreation participation and retention is another essential
exposure and outcome dyad worthy of study to help explain the importance of campus recreation
(Danbert et al., 2014; Lindsey & Sessoms, 2006; Kampf & Teske, 2013). If a positive
relationship between campus recreation and retention is found consistently, it could positively
affect future decisions about funding and support.
Hall (2006) indicated that even though there have been research studies conducted since
the 1970s focusing on campus recreation and its influence on retention, universities need to build
on this research to assess their own campus recreation centers. Since such a large percentage
(78%) of all students at Hall’s university used at least one component of campus recreation, he

30	
	

	

believed it was necessary to interpret the benefits related to campus recreation participation at his
university (2006). Interestingly, his research was qualitative as only four men and four women at
the University of Pacific were interviewed to discuss their involvement in campus recreation
programs. The interviewers asked the participants the following questions: Describe those
activities you participated in growing up. What do you do to stay in shape? Do you consider
recreation a social activity? What role, if any, has recreation played in continuing your education
here at the University of Pacific? Why do you participate in recreational activities? What
recreational facilities do you use on campus? Why? What has your use of recreational facilities
or participation in recreation programs allowed you to do on campus? The goal of the interview
was to determine if participation in campus recreation programs was related to retention. In his
analysis, the researcher searched for keywords and common themes reported by participants. The
most popular themes included physical activeness, diversity, socialization, meeting new people,
friendship, sense of community, faculty relationships, student relationships, small campus,
background/lifestyle, competition, academic/major, fun, and self-esteem. Among many other
benefits of campus recreation participation, students commented on the relationship they
believed occurred between campus recreation participation and retention. When students were
asked how they believed campus recreation influenced continued education at the university,
three participants commented that they continued their education at that particular university due
to their involvement in campus recreation programs. As with every research study, this one had
limitations. Considering the researcher administered the program for this study, there was a
possibility of bias. Also, a small sample size warrants caution to generalizing the results of this
study. Although the participants commented they felt their use of campus recreation facilities
was related to several benefits, only three out of the eight participants believed campus

31	
	

	

recreation influenced their choice to stay at the university. Hall suggested that future research
should use a mixed methodology to obtain information about recruiting, retention, and the
benefits of participating in campus recreation programs. Additional research with a larger sample
size may provide more encouraging results about the relationship between campus recreation
participation and retention.
Miller (2011) studied campus recreation as an entity with which students bonded to
determine if it played a role in their retention. Miller (2011) explained that personal bonds occur
between students and places on campus but unfortunately, campus recreation had not been
studied as a place students had a personal bond with. Evaluating personal bonds between
students and campus recreation provided the possibility that bonding strengthens the chance for
retention. To study campus recreation as an entity on campus with which students had personal
bonds, Miller distributed a 20-question survey about social belonging and retention to
undergraduate students as they walked into the student recreation center. After four weeks of
administering surveys to students, Miller received 453 completed surveys. Students who
completed the survey commented that the campus recreation center provided them with strong
emotional ties due to the ability to interact with new people, an increase in self-confidence,
improved leadership abilities, improved happiness, and an improvement in time management
skills. Study participants also commented that the student recreation center was necessary in
creating a social bonding experience for them. Experiences gained allowed students to meet and
develop friendships with other students, build a sense of community to establish relationships
with other students, helped create a network of friends, and increased students’ trust in their
peers. Finally, Miller found that the campus recreation center not only attracted students to the
university but also was a major reason why they did not leave the university. Retention may have

32	
	

	

been explained by the students’ sense of community at the university due to the relationships
built and social bonds in creating a network of friends. This study also had limitations. These
results are not generalizable and are largely applicable to universities similar to the university
studied. An additional limitation focused on the participant sample. Most participants were
campus recreation users. Therefore, the results of this study may be have been biased. Another
limitation involved the self-report survey that was distributed to participants. Although Miller
administered this survey in a pilot study and α=.82, Miller stated using a self-reported survey
could be a limitation. For future studies, Miller suggested that researchers investigate how each
campus recreation program relates to outcome variables like retention. Additionally, Miller
suggested that conducting similar mixed methodology based studies will assist in building the
knowledge of the influence of participating in campus recreation programs. Overall, Miller
(2011) suggested that the social bonds made at the campus recreation center assisted in
improving the academic persistence for campus recreation participants at the university.
Explaining how social bonds are created at campus recreation centers and how that ultimately
may lead to retention should provide justification for campus administrators to continue funding
for these types of programs. Future research providing similar results will add to the relationship
between campus recreation participation and retention.
Forrester (2015) noted that while previous research has demonstrated the benefits of
participating in campus recreation programs, many studies are lacking in sample size and multiinstitutional designs. Therefore, he investigated the benefits of participating in campus recreation
sports using a large multi-institutional sample. The 2013 NASPA Assessment and Knowledge
Consortium was distributed to students at 38 universities across the United States. Of those who
completed the survey (n=33,522), 29,142 students attended public universities and 4,380

33	
	

	

attended private universities. Approximately 75% of respondents indicated they used campus
recreation sports facilities, programs and services. Participants indicated campus recreation
programs (66.6%) and campus recreation facilities (73.9%) were important in their decision to
stay at their university. Study limitations include data being collected via self-report. In addition,
the large sample size produced a statistically significant finding that might not be “practically
significant”. However, the positive results suggest that campus recreation center directors should
recognize and promote the benefits campus recreation programs and facilities may offer to the
university as a means to meet university retention goals.
Some researchers have studied the relationship of campus recreation with retention and
recruitment of students to determine if participation may assist in both drawing in and retaining
students. Lindsey & Sessoms (2006) recognized the importance of not only attracting
prospective students but also retaining them, so the purpose of their study was to determine the
effect campus recreation use had on recruitment and retention. To examine the perception
students had about participating in campus recreation programs, a modified version of the
National Intramural Recreational Sport Association’s Quality and Importance of Recreational
Services Survey was distributed to undergraduate students at a small university in the southeast
U.S. Students were instructed to complete questions on the QIRS related to demographic data
such as class standing, ethnicity and age in addition to questions about the importance and
quality of campus recreation programs. The results of this survey showed that 31% of survey
respondents reported campus recreation was either important or very important when deciding
the university they wanted to attend. Additionally, 37.3% reported campus recreation was either
important or very important when deciding to continue at their universities. Although these
findings provide an argument for campus recreation to be used as a recruitment tool, Lindsey &

34	
	

	

Sessoms noted that the results of this study were not generalizable to all universities.
Additionally, researchers conducting studies similar to this should use a more random method of
choosing study participants rather than a convenience sample. Furthermore, the QIRS survey has
had issues with reliability and validity in the past so researchers may want to implement other
surveys or use with caution. Nevertheless, results of this study can be used for universities
similar in size and demographics. Moreover, some methods utilized in this study may be used for
future studies at other universities to further study the relationship between campus recreation
participation and recruitment and retention.
Lindsey, Sessoms & Willis (2009) replicated their previous study by implementing it at a
historically black university. Since research lacks using this subpopulation, the purpose of the
Lindsey, Sessoms & Willis (2009) study was to assess the impact of campus recreation programs
on student recruitment and retention. Undergraduate students in the Department of Health and
Human Performance completed the National Intramural Recreational Sports Association’s
Quality and Importance of Recreational Services Survey to indicate the importance of
recreational sports facilities and programs to their college experience. When asked about how
important recreational sports were when choosing a university, 47% of students reported sports
were important/very important. An even greater percentage (50%) reported the availability of
recreational sports to be important/very important in deciding to stay at their university.
Considering the researchers focused their study at a historically black institution, the results from
this study may be most relevant to other historically black universities of a similar size with
similar campus recreation facilities. Furthermore, these researchers also limited their results by
the fact that they chose a convenience sample as opposed to one more representative of the
university overall. The authors suggested to randomly select participants and continue this type

35	
	

	

of research at all types of universities to strengthen the knowledge of the benefits of campus
recreation.
Henchy (2011) acknowledged the academic strategies university administrators utilize to
recruit and retain students but also noted the importance of nonacademic departments on
campuses that may lead to retaining students. Although Henchy (2011) provided several research
findings that supported campus recreation participation, she did not discuss where the research
was lacking. Therefore, in addition to the research that has been previously conducted, the
purpose of Henchy’s study was to determine the influence of campus recreation participation on
recruitment and retention for students who participated in campus recreation programs.
Additional nonacademic outcome variables; such as quality of student life, interest in staying fit,
and feelings of well-being and stress, were also measured. Henchy sent the National Intramural
Recreational Sports Association/Student Voice Campus Recreation Impact Study survey to 2,500
students at a university located in southeastern United States. The survey included questions
about recruitment, retention, satisfaction, involvement, and benefits students received from
participating in campus recreation programs. Of the 2,500 students who were sent the survey,
only 237 completed it in its entirety. Results showed that 28% of respondents believed that
campus recreation programs had a moderate or strong influence on their decision to attend the
southeastern university and 31% of all respondents noted that campus recreation programs had a
moderate or strong influence on them staying at their university. Although these results agreed
with results previous researchers have found, Henchy stated that almost half the respondents of
this study were graduate students, overall response rate was poor, and the study followed a cross
sectional design. Henchy suggested that future researchers should encourage survey completion
to improve response rates. Furthermore, it may be beneficial to analyze undergraduate students

36	
	

	

results separate from graduate students’ results due to the differences in responses about
choosing and staying at a university (Henchy, 2011).
SUMMARY
Numerous researchers have studied the relationship between campus recreation and
retention. Several have also studied the relationship between recruitment and retention in order to
explain if campus recreation has any influence in attracting students to campus and then retaining
them. While researchers have found positive relationships between campus recreation and
retention (Miller, 2011; Forrester, 2015; Lindsey et al., 2009), some of these findings may be
questionable in how significant these results are for justifying campus recreation to
administrators (Hall, 2006; Lindsey & Sessoms, 2006; Henchy, 2011). Furthermore, the
methodology for the campus recreation and retention literature differs from study to study. While
most researchers use surveys to find a relationship between campus recreation participation and
retention, some use previously developed surveys (Forrester, 2015; Lindsey & Sessoms, 2006;
Lindsey, Sessoms & Willis, 2009), some develop their own survey (Miller, 2011), while others
use a scripted interview to collect responses from their study participants (Hall, 2006). While
most researchers suggest positive relationships between campus recreation participation and
retention, limitations such as poor response rates, using convenience samples, using surveys with
questionable reliability, and self-report all explain the necessity for future research.
THE RELATIONSHIP BETWEEN CAMPUS RECREATION PARTICIPATION AND
MULTIPLE ACADEMIC ACHIEVEMENT OUTCOME MEASURES
Many researchers approach campus recreation participation and academic success
research to determine if campus recreation participation has a relationship with several success
measures to justify the argument for universities to continue to support campus recreation

37	
	

	

(Kampf & Teske, 2013; McElveen & Rossow, 2014; Henchy, 2013, Danbert et al., 2014). While
focusing on one outcome measure is simplistic, studying multiple success outcome measures
would be advantageous if researchers can obtain such data easily via survey or from a Registrar’s
Office. Kampf & Teske (2013) ensured multiple success variables would be measured
considering retention and GPA data would be provided by the Office of Institutional Research at
the university where their research was being conducted. The investigators believed that more
campus recreation research was needed due to the fact that students were not enrolling in college
at the rates they previously did and universities were having trouble retaining them. Considering
many campus recreation centers need to justify funding for programming, Kampf & Teske
(2013) believed research showing the success campus recreation has on student recruitment and
retention was necessary. Therefore, the purpose of Kampf & Teske’s (2013) study was to
investigate the engagement-persistence relationship in a campus recreation program.
Demographic data, precollege data, latest term one-year college GPA, and Fall-to-Fall retention
status were collected from the Office of Institutional Research at the university studied. To
analyze campus recreation participation data, the researchers were able to collect data indicating
participation in club sports, campus recreation student employment status, and recreation center
entry counts from the Division of Student Affairs. Results showed there were not significant
differences in academic variables (ACT score, high school GPA, latest term cumulative one year
college GPA) between club sports participants and nonparticipants but there was a significant
relationship in Fall-to-Fall retention with club sports participants. The researchers also analyzed
the data by developing a regression model to determine factors that would explain retention.
After controlling for academic and demographic data, club sports were significant and kept in the
regression model. Students participating in club sports had 2.22 times greater odds of enrolling

38	
	

	

the following year when compared with nonparticipants. The researchers also found that student
entry counts into the campus recreation center were positively correlated with high school GPA
and latest term one-year college GPA. This was true after controlling for academic and
demographic data. Students who used the campus recreation center were also more likely to be
retained. In fact, the more that students used the campus recreation center, the greater the
correlation with retention. The study had a number of limitations. One of the most important
limitations to review is that the results of this study do not prove causality between exposures
and outcomes, but rather examine the relationship between the exposures and outcomes. There
may be several variables that could have played a role in academic success that were not
measured. Additionally, this study lacks direct measures of persistence within the logistic
regression. Therefore, there may be variables that are significant to influence the outcome
measures but not included in the analysis. Like many other campus recreation studies,
generalizability of results is limited. Since this was a single institution study, results gathered
may be most relevant for other universities of the same size, same recreational facilities, and
same student demographics. Overall, this study offers promising results for participating in club
sports and campus recreation center programs (via card swipe). Although the methods of this
study may be improved, positive results like these may still be used to argue resources for
campus recreation programs (Kampf & Teske, 2013).
McElveen & Rossow (2014) agreed with the purposes of conducting campus recreation
participation research to determine if there is a positive influence campus recreation has on
academic success outcomes. The researchers recognized that many campus recreation
professionals attempt to justify funding for their programs to administrators but are starting to
feel the pressure to provide evidence for continued funding (McElveen & Rossow, 2014). In

39	
	

	

addition to a need for more research to convince university administration of the benefits of
campus recreation programs, McElveen & Rossow suggested that it is critical to focus on
students with the lowest rate of retention; freshmen. Therefore, the purpose of their study was to
examine the relationship between intramural sports participation and academic performance and
retention in first time in college (FTIC) students. Students who participated in intramural sports
in the Fall 2010 and/or Spring 2011 semesters were included in this study. Participation data
were cross-referenced with institutional data to determine GPA per semester, retention from
semester to semester, and retention from year one to year two. Researchers then split the sample
by students who participated in 1) zero intramural sports, 2) 1-3 sports per semester (moderate
participation), 3) 4 or more sports per semester (heavy participation). After analyses were
conducted, McElveen & Rossow found there was not a significant difference in GPA across
groups but there was a difference in retention. From Fall 2010 to Spring 2011, FTIC students
who participated in intramural sports were retained at 4.7% higher rate than FTIC students who
did not participate in intramural sports. From Fall 2010 to Fall 2011, FTIC students who
participated in intramural sports were retained at 5.9% higher rate than FTIC students who did
not participate in intramural sports. Although it is encouraging to link intramural sports with
academic success, it is necessary to review the limitations of this study. Unfortunately, the
investigators did not control for confounding variables. Participation in other student services on
campus, and external and individual variables could have affected the results of this study
suggesting that intramural sports may not be named the sole reason for the academic success
outcomes in this study. Nevertheless, the finding of increased retention of FTIC students who
participated in intramurals was encouraging. The researchers suggested that future research

40	
	

	

should control for confounders, include other campus recreation programs, and be studied at
other universities to increase generalizability of results.
Henchy (2013) suggested it is important to understand the influence of campus recreation
on recruitment, retention, and other factors such as social belonging, sense of campus
community, and relationship building. Previous research has reviewed the social benefits,
recruitment, and retention benefits campus recreation has to offer for undergraduate students, but
studies focusing on graduate students are lacking (Dalgarn, 2001; Miller, 2011; Elkins et al.,
2011; Kampf & Teske, 2013). The purpose of Henchy’s study was to compare undergraduate
and graduate perceived benefits of campus recreation participation. Specific areas Henchy
wanted to study included how campus recreation influenced recruitment, retention, social, health,
and other outcome measures. To obtain undergraduate and graduate students perceptions of
campus recreation participation, Henchy requested students to complete the National Intramural
Recreational Sports Association/Student Voice Campus Recreation Impact Study survey. The
survey focused on questions targeting how students believed campus recreation participation
influenced recruitment, retention, satisfaction, and involvement on campus. When reviewing
recruitment and retention results, Henchy (2013) found that 36% of undergraduate and 24% of
graduate students reported that campus recreation had a moderate or strong influence on their
choice of university. Additionally, when asked if campus recreation influenced students’
decision to stay at the university, 38% of undergraduates and 27% of graduate students
responded that campus recreation had a moderate or strong influence on their decision to stay at
their university. Finally, when asked about campus recreation influencing academic
performance, 41% of undergraduate and 30% of graduate students noted that they believed
academic performance moderately or strongly improved based on participation in campus

41	
	

	

recreation programs. These results are exciting because this is the first study to show that
participation in campus recreation benefits both undergraduate and graduate students. On the
contrary, the study design used for this study was cross-sectional which does not allow for
changes in participation over time. Overall, while the percentage of undergraduate and graduate
students who benefit and continue to attend their universities because of campus recreation could
be higher, there are still a substantial number of students who are positively influenced by
campus recreation programs. This information should be conveyed to higher administration to
justify campus recreation programs and the importance they provide to students at both the
undergraduate and graduate levels.
To review campus recreation participation and multiple academic success measures,
Danbert et al. (2014) collected data on fitness center membership, GPA, credits completed, and
retention to determine if there was a relationship between participation in campus recreation and
the outcome measures. Unlike many university campus recreation center payment methods in
which students are charged a fee as part of their tuition to use the campus recreation facilities,
students at the university studied were required to pay a separate fee for their fitness center
membership. Previous researchers have found that students who pay a separate fee to utilize
campus recreation facilities in their first semester in college have had higher GPAs, complete
more credit hours, and are more likely to be retained when compared to nonmembers. Therefore,
the purpose of this study was to determine the relationship between recreational sports
membership and college student academic success and retention in first time degree seeking
freshmen. Data on first time degree seeking freshmen were provided to the researchers from the
Registrar’s database. Academic data including high school GPA, cumulative college GPA,
college credits completed, one-year and two-year retention, and class standing were provided as

42	
	

	

well. Students were classified as members of the campus recreation center if they purchased a
membership in their first semester. Nonmembers included first semester freshmen who did not
purchase a membership. Comparing members to nonmembers showed that members completed
more credits and had higher GPAs than nonmembers. Following these students longitudinally, it
was found that college GPA was higher in members than nonmembers and members completed
more credits than nonmembers. After the first year, members were more likely to be retained
than nonmembers. The same was true after the second year. Although Danbert et al. (2014)
addressed limitations of previous research (e.g. self-report, lack of longitudinal data, and
insufficient information of how students accessed the campus recreation center); there were
limitations to their research as well. The main limitation mentioned was that confounding
variables were not figured into the statistical analysis of this study. Danbert et al. (2014)
suggested that confounders like high school GPA, gender, and socioeconomic status have the
ability to effect college GPA so it is important to account for confounders. In conclusion,
Danbert et al. (2014) found that several factors of student success were related to participation in
campus recreation participation at a campus recreation facility that required a fee to participate.
SUMMARY
Researchers that study the relationship between campus recreation participation and
several success outcomes seek the possibility of finding several positive relationships between
campus recreation and success outcome variables to justify their argument for appropriate
funding for campus recreation programs. On the contrary, these researchers also run the risk of
adding to the mixed results that have been a continuing issue in campus recreation research.
Researchers like Danbert et al. (2014) found a positive relationship between campus recreation
members and GPA, while other researchers have not found this same relationship (Kampf &

43	
	

	

Teske, 2013; McElveen & Rossow, 2014). The success outcome found to be positively related to
campus recreation participation in many studies is retention. Not only have researchers studied
this relationship in a quantitative manner (through retention rates), but they have also determined
this relationship through survey. Several researchers have found the relationship between campus
recreation participation and retention allowing for a strong argument for campus recreation to
higher administration. Future research should focus on controlling for confounding variables,
encourage higher response rates for studies requiring surveys, and include additional campus
recreation programs to study.
It is evident by previous campus recreation and academic success research that mixed
results have been found. While some researchers reported positive relationships between campus
recreation participation and GPA (Danbert et al., 2014; Todd, et al., 2009; Kampf & Teske,
2013), many have not found this same result (Todd, et al., 2009; Washke, 1940; Baker, 2008;
Zizzi et al., 2004; Kampf & Teske, 2013; McElveen & Rossow, 2014). Most researchers have
found positive relationships between campus recreation participation and recruitment and
retention measures, even though research methods have varied (Lindsey & Sessoms, 2006;
Miller, 2011; Lindsey et al., 2009; Forrester, 2015; Henchy, 2011; Kampf & Teske, 2013;
McElveen & Rossow, 2014; Henchy, 2013; Danbert et al., 2014). Furthermore, few researchers
have reviewed success outcomes like credits completed (Danbert et al., 2014); therefore,
additional research focused on multiple variables of student academic success are necessary.
FIRST GENERATION STUDENTS
Over the past 25 years, the number of first generation students attending colleges and
universities in the United States has increased; causing concern about their education. With 4.5
million first generation students enrolled in institutions across the nation, this subpopulation of

44	
	

	

students needs proper attention when it comes to student success, especially since first generation
students tend to show lower academic performance compared to traditional students (Pryor,
Hurtado, DeAngelo, Blake & Tran, 2010). First generation students are one of the most
disadvantaged populations of students on college campuses (Choy, 2001; Horn & Nunez, 2000;
Nunez & Cuccaro-Alamin, 1998; Warburton, Bugarin & Nunez, 2001). First generation students
tend to be older, come from minority backgrounds, have a disability, are immigrants, are single
parents, English is not their primary language, are independent from their parents, are lowincome, live off campus, do not attend college immediately after high school graduation, attend
college closer to home, are not full time students, work full time, are nontraditional students, are
female, and are over the age of 24 years old (Bui, 2004; Engle & Tinto, 2008). First generation
students are the first in their families to experience college, and many struggle academically and
culturally with the transition into college (London, 1989, 1992, 1996; Choy, 2001; Horn &
Nunez, 2000; Nunez & Cuccaro-Alamin, 1998; Warburton, Bugarin & Nunez, 2001). When
focusing on academic success measures, first generation students are about four times more
likely to leave college after their first year when compared with traditional students (Engle &
Tinto, 2008). Clearly, it is necessary to identify strategies that promote academic success in a
population like first generation students as they continue to enroll in universities across the
United States.
Although first generation students have been studied for a few decades, retention research
on this group has not been as plentiful (Duggan, 2001, 2002; Ishitani, 2003; Somers, Woodhouse
& Cofer, 2000). In particular, first to second year retention research with first generation students
is especially needed due to the fact that a great deal of students drop out of college after their
first year (Duggan, 2001, 2002). Lohfink & Paulsen (2005) identified this gap in the research and

45	
	

	

therefore used data from a national sample to analyze the determinants of first to second year
retention for first generation and traditional students at four-year universities. To determine if
there were reasons for persistence between the first year and second year of college in first
generation and traditional students, Lohfink & Paulsen (2005) used data from the Beginning
Postsecondary Students Longitudinal Survey-BPS: 96/01. Students were considered first
generation if their parents did not have any education past high school (Choy, 2001; Nunez &
Cuccaro-Alamin, 1998). In contrast, traditional students were classified as such if at least one
parent had any type or quantity of education past high school (Somers et al., 2000). In this study,
the independent variables consisted of the following: background characteristics, precollege
achievement, academic, social and financial reasons for choosing a university, institutional
variables, and in-college experiences. These independent variables were analyzed to determine
their likelihood of influencing the dependent variable of first to second year retention. After
logistic regressions were run to analyze the relationships between the independent variables and
first to second year retention, significant differences between first generation and traditional
students were found. Specifically, 76.5% of first generation students persisted from their first to
second year. Students possessing the following characteristics or reported expectations were less
likely to persist from year one to year two: married, females, English speaking, Hispanic, lower
family income, desire to complete a Bachelor’s degree or less (versus students who expected to
complete higher than a Bachelor’s degree), choice of university not based on faculty reputation,
choice of university not based on the opportunity to live at home, attending a private institution,
smaller institutions, or institutions with a lower academic integration index, lower GPA, lack of
satisfaction with their social lives, lower grant aid, and lower work-study aid. Traditional
students were more likely to persist from year one to year two when compared with first

46	
	

	

generation students (82.2% vs 76.5%, respectively). Traditional students possessing the
following characteristics or reported expectations were less likely to persist from year one to year
two: expectation to complete a Bachelor’s degree or less, choice of university was not due to the
school’s reputation, choice of university was not due to lower tuition, lower GPA, lack of
satisfaction with their intellectual growth, social lives, campus climate, participation in fewer
school clubs, and less work-study aid. These results offer important information for first
generation retention literature. While Lohfink & Paulsen (2005) found a couple similarities in
reasoning to leave college between traditional and first generation students, there were more
differences in the reasoning between the groups for leaving. Thus, it may be beneficial to focus
on the characteristics and expectations first generation students possess when developing
retention efforts. Choy (2001) explains that first generation students are many times non-White,
from low-income families, and female. Lohfink & Paulsen (2005) found that race, class, and
gender were specific characteristics that relate to first generation students who are not retained.
Specifically, Lohfink & Paulsen (2005) found first generation students who are Hispanic, come
from a low-income family and are female all influenced first to second year retention
substantially. To better assist these students, it is suggested that programs should be offered
based on those specific students’ wants and needs. Developing programs that are beneficial to
the whole population of students may not benefit specific populations similarly. First generation
students from low-income families likely come from different social classes that may view
financial obligations to universities differently (Lohfink & Paulsen, 2005). Educating these
students on the strategies of funding their education should be available so that persistence is not
affected. Finally, Lohfink & Paulsen (2005) suggested that since females made up the majority
of first generation students in their study, universities should be aware of this and further

47	
	

	

investigate the possibilities of why first generation female student retention is substantially
different. Although women are attending universities at rates higher than males, a greater number
of female first generation students are also attending and may need programs that influence
retention. It is important to identify the reasons first generation students persist to create
programs that offer the best benefits for this population of students.
While Lohfink & Paulsen (2005) shed some light on the differences between first
generation students and traditional students, Pascarella, Wolniak, Pierson & Terenzini (2003)
commented that there is still much to research with this population. Pascarella et al. (2003)
acknowledged the shortcomings first generation students typically have; knowledge about
postsecondary education, low income, lack of secondary school preparation (Berkner & Chavez,
1997; Hossler, Schmidt & Vesper, 1999; Warburton, Burgarin & Nunez, 2001), and believe it is
important to understand more about first generation student experiences. To study this
population of college students, the researchers gathered their sample from five different
community colleges from five different states. Pascarella et al. (2003) randomly chose 144
students from the five universities sampled. The researchers categorized students into three
groups to ensure a more sensitive analysis. Many researchers group students as either first
generation (both parents did not complete a Bachelor’s degree) or traditional/other. Defining
students as traditional or other could group students whose parents have varying levels of
education (e.g. both parents have Bachelor’s degrees, one parent has a Bachelor’s degree and one
does not, etc.). Therefore, Pascarella et al. (2003) categorized groups by first generation, high
parental postsecondary education (both parents earned a Bachelor’s degree or higher), and
moderate postsecondary education (one or more parents completed at least some college
education but no more than one parent earned a Bachelor’s degree or higher). A questionnaire

48	
	

	

was distributed to participants and they were asked to answer the 21 measures relating to
academic and nonacademic experiences while in college. The following areas were measured:
credit hours completed, time spent studying, number of courses taken in natural sciences and
engineering, mathematics, social sciences, technical/pre-professional and arts and humanities,
college grades, course-related interaction with peers, academic effort/involvement, computer use,
reading and writing involvement, hours worked per week, athletic participation, Greek
affiliation, non-course related interaction with peers, extracurricular involvement, and volunteer
involvement. After multiple regressions were run and pre-college confounders were controlled,
the greatest differences in experiences occurred between the first generation students and
students whose parents both completed a Bachelor’s degree. First generation students completed
fewer credit hours, studied less, took fewer courses in the natural sciences, mathematics, and the
arts and humanities, had lower grades, were less likely to join a Greek organization, and worked
more hours per week. Additionally, first generation students were more likely to take
technical/pre-professional courses and less likely to use computers than students whose parents
both completed Bachelor’s degrees. Although first generation students did not perform as well in
science reasoning, were not as open to diversity/challenge, and had lower learning for selfunderstanding, they showed better writing skills and greater internal locus of attribution for
academic success, had a preference for higher order cognitive tasks, and showed more resiliency
in community college than previous research predicted. This study was unique in the fact that
first generation students were not compared to the rest of the population of students, yet were
compared to students whose parents both completed a Bachelor’s degree and students who had
only one parent that may have completed a Bachelor’s degree. This difference in categorization
shows the importance of comparing first generation students to the rest of the student population

49	
	

	

because the significant differences occurred between the first generation students and students
whose parents both completed a Bachelor’s degree.
Stebleton, Soria & Huesman Jr. (2011) added to the work of Lohfink & Paulsen’s in
studying the differences in first generation students' college experiences. Stebleton et al. (2011)
recognized the challenges first generation students face as they attempt to continue their
education at four-year institutions. Since first generation students carry several typical
demographic characteristics of disadvantaged populations (e.g., low-income, minority, English
as a second language students), Stebleton et al. (2011) wanted to better understand first
generation students’ college experiences and compare them to traditional students’ experiences
to determine if valuable suggestions could be promoted. The overall purpose of the study was to
determine if first generation students faced different kinds of stress when compared with
traditional students. Specifically, Stebleton et al. (2011) examined differences in sense of
belonging between first generation students and traditional students, if there was a relationship
between sense of belonging and satisfaction and their level of well-being, and if there were
differences between first generation students and traditional students level of mental well-being
and their usage of mental health services. Stebleton et al. (2011) distributed surveys to 145,150
students at six large universities across the United States. Of the 145,150 students, 58,017
completed the study that focused on student life and development, mental health, and the use of
counseling services. Students were classified as first generation students if they responded that
both parents have not earned a Bachelor’s degree. The researchers analyzed student life and
development, mental health, and the use of counseling services. When asked about sense of
belonging on campus, traditional students reported a greater sense of belonging when compared
with first generation students. Traditional students reported lower levels of depression and stress

50	
	

	

when compared with first generation students. Finally, first generation students reported needing
but not using counseling services at a higher rate than traditional students. In addition to the
suggested outreach that college counselors should implement for first generation students,
Stebleton et al. (2011) also recommended that first generation students be encouraged to
participate in learning communities, service learning opportunities, study abroad programs,
writing intensive courses, and first year seminars to improve social engagement opportunities.
These programs may help students meet other students, faculty and staff of the university and
enhance their sense of belonging with the university.
SUMMARY
Considering first generation students have a more difficult time with the transition into
college and being retained, it is necessary to determine factors that may aid in a successful
college experience. For decades, researchers have studied what makes college more challenging
for this population (Pascarella et al., 2003; Lohfink & Paulsen, 2005), but only recently have
researchers started to study programs that may assist first generation students in achieving
academic success (Stebleton et al., 2011). Considering only a few programs have been studied to
determine the relationship between participation and student academic success, it is necessary to
explore other programs that may promote or be related to academic success outcomes.
FRESHMEN STUDENTS
According to Tinto (2006), student retention is one of the most widely studied areas in
higher education. Early work by Spady (1970, 1971), Tinto (1975, 1987), Astin (1975, 1984),
and Pascarella & Terenzini (1980) focused on student retention over periods of time in college to
determine patterns of attrition. Additionally, strategies for retention have been studied to
determine factors leading to persistence. What was determined decades ago was that

51	
	

	

involvement in the university setting is important and it matters most during the first year of
college (Tinto, 1975, 1987; Astin, 1975, 1984; Pascarella & Terenzini, 1980). Haines (2001) and
Moffitt (2010) have found that campus recreation departments may be that source of
involvement for students to choose and stay at a university. Other researchers have also
attempted to find this link for freshmen students, although results are mixed (Danbert et al.,
2014; Kampf & Teske, 2013; McElveen & Rossow, 2014).
Danbert et al. (2014) collected campus recreation participation data to determine if there
was a relationship between participation and several academic success measures in freshmen
students. Data collected included fitness center membership, GPA, credits completed, and
retention. Fitness center membership was unique at the university studied as students paid a
separate fee (not included in tuition) to obtain access to campus recreation programs. Due to
results found from previous research focusing on similar campus recreation membership,
Danbert et al. (2014) speculated that students would have higher GPAs, complete more credit
hours, and would be more likely to be retained when compared to nonmembers. Hence, the
purpose of this study was to determine the relationship between recreational sports membership
and college student academic success and retention in first time degree seeking freshmen.
Academic data for first time degree seeking freshmen including high school GPA, cumulative
college GPA, college credits completed, one-year and two-year retention, and class standing
were provided by the Registrar. Campus recreation center members were included if they
purchased a membership in their first semester. Nonmembers included first semester freshmen
that did not purchase a membership in their first semester. To determine if there was a significant
difference in academic success between members and nonmembers, GPA, credits completed, and
retention were compared between the groups. Results showed that members completed more

52	
	

	

credits and had higher GPAs than nonmembers. When reviewing one-year and two-year
retention, members were more likely to be retained after the first year and the second year. While
limitations of previous research were taken into account (e.g. self-report, lack of longitudinal
data, and insufficient information of how students accessed the campus recreation center); there
were limitations to this study as well. Danbert et al. (2014) explained that confounders like high
school GPA, gender, and socioeconomic status may effect college GPA so it is important to
account for these confounders. The researchers suggested that future research should incorporate
confounding variables in the statistical analysis to determine the effect they have on college
academic success variables. Overall, Danbert et al. (2014) found that several student success
variables for freshmen were related to participation in campus recreation participation.
Since attrition rates are highest in freshmen students, Kampf & Teske (2013) studied this
population to determine if there were programs that may encourage retention in the first year of
college. Reiterating Tinto’s (1993) theory that a predictor of retention is academic and social
integration, Kampf & Teske (2013) suggested out-of-classroom experiences may be linked with
academic success and should be further studied. Unfortunately, not all out-of-classroom
activities are seen by higher education administrators as links to success. For instance, Brandon
(2010) notes that campus recreation programs have been thought to be nonimportant to the
overall collegiate learning experience and viewed as a “perk” to students. Consequently, these
beliefs lead campus recreation professionals to having to justify funding for programming.
Therefore, Kampf & Teske (2013) believed it was necessary to continue to research the
relationship between campus recreation and recruitment and retention. The purpose of Kampf &
Teske’s (2013) study was to investigate the engagement-persistence relationship with club
sports, campus recreation employment, and student recreation center entry counts (via electronic

53	
	

	

card swipe) in first time, full time students. Demographic data, precollege data, latest term oneyear college GPA, and Fall-to-Fall retention status were collected from the Office of Institutional
Research at the university studied. Participation in club sports, campus recreation student
employment status, and recreation center entry counts were collected from the Division of
Student Affairs. Results for club sports participants and nonparticipants showed there were not
significant differences in ACT score, high school GPA, latest term cumulative one-year college
GPA but there was a significant relationship in Fall-to-Fall retention with club sports
participants. The researchers also reviewed retention by building a regression model to determine
factors that explain retention. After controlling for possible confounders (academic and
demographic data), club sports were a significant variable in the regression model. Club sports
participants had 2.22 times greater odds of enrolling the following year when compared with
nonparticipants. Utilizing the campus recreation center (via card swipe) was also positively
correlated with high school GPA and latest term one-year college GPA after controlling for
confounders. Retention was more likely in students who used the campus recreation center.
Additionally, the more that students used the campus recreation center, the greater the correlation
with retention. As with many studies, this study had its limitations. One of the most important
limitations to clarify is that the results of this study do not prove causality between campus
recreation participation and student success outcomes, but do review the relationship that exists
between the variables. Another limitation is that this study lacks direct measures of retention
within the regression model. Therefore, there may be other variables (e.g. other programs
students participate in) that could influence success but are not included in the analysis. As
reported in other campus recreation studies, generalizability of results is another limitation.
Results from this study may be used for other universities of the same size, same recreational

54	
	

	

facilities, and same student demographics, although may not be relatable to other universities.
Overall, this study offers encouraging results for first time, full time students participating in
club sports and campus recreation center programs (via card swipe). Although continued
research is suggested, results from this study may be helpful for campus recreation professionals
to justify needed resources (Kampf & Teske, 2013).
McElveen & Rossow (2014) believed that since freshmen attrition rates are higher than
any other academic class, institutions should study on-campus programs that may encourage
retention. The researchers reiterated the importance of the findings of Miller (2011) that students
who are not involved early in their college careers, tend to stay uninvolved and are less likely to
be retained. Therefore, the purpose of their study was to examine the relationship between
intramural sports participation and academic performance and retention in first time in college
students. Freshmen students who participated in intramural sports in the Fall 2010 and/or Spring
2011 semesters were included in this study. Intramural sports participation data, GPA per
semester, retention from semester to semester, and retention from year one to year two were
collected. Researchers then split the sample by students who participated in 1) zero intramural
sports, 2) 1-3 sports per semester (moderate participation), 3) 4 or more sports per semester
(heavy participation). Comparing participant and nonparticipant results, McElveen & Rossow
found there was not a significant difference in GPA across groups but there was a difference in
retention. From Fall 2010 to Spring 2011, freshmen students who participated in intramural
sports were retained at 4.7% higher rate than freshmen students who did not participate in
intramural sports. From Fall 2010 to Fall 2011, freshmen students who participated in intramural
sports were retained at 5.9% higher rate than freshmen students who did not participate in
intramural sports. Although the link between campus recreation program participation and

55	
	

	

multiple measures of academic success was not found in this study, the relationship between
intramural sports and retention should be encouraging for campus recreation departments and
institutions seeking strategies to improve retention in freshmen students. Additionally, it is
necessary to review the limitations of this study to determine the next step for future research.
Like other campus recreation and student academic success studies, the researchers did not
control for confounding variables. For example, participation in other on-campus and off-campus
activities was not configured in the analysis of this study; therefore, intramural sports may have
not been the only positive influence on retention. Regardless, the relationship between intramural
sports and retention that was found is encouraging and should be included with other programs
to determine its effect on academic success in future studies.
SUMMARY
With attrition rates ranging from 20-30% in freshmen students, it is imperative to
determine strategies that encourage student involvement early on in students’ college careers
(Kampf & Teske, 2013). Considering mixed results have been found when studying campus
recreation participation in freshmen, it is necessary to continue to research in the hope of
producing more consistent results (Danbert et al., 2014; Kampf & Teske, 2013; McElveen &
Rossow, 2014). Continued positive relationships between campus recreation participation in
freshmen and academic success may help institutions change their perception of campus
recreation as a “perk” and justify using it for improving success measures like retention.
SOPHOMORE STUDENTS
Researchers agree that attrition rates from sophomore to junior year are the second
highest (10-13%) compared to freshmen attrition rates (Lipka, 2006; Adelman, 2006), although
sophomore retention tends to be overlooked since most universities devote retention efforts

56	
	

	

towards freshmen students. Gahagan & Hunter (2006) note that even though research on
sophomore students has been conducted for decades, the sophomore experience is “less well
understood.” What is known is that the second year of college tends to present a different and
possibly more challenging time in the college career (Gahagan & Hunter, 2006). Two topics that
researchers have been able to agree upon when studying the difference in the sophomore
experience include 1) developmental changes that take place within students during their
sophomore years and 2) institutional policies and support from administrators that either help or
hinder students’ progress during their second year (Gahagan & Hunter, 2006).
Schaller (2005) suggests that since universities are aware of freshmen retention issues,
they are quick to offer programs that promote retention to the second year. Unfortunately, less is
known about student retention from the second to third year in college and educators are left
wondering what needs may or should be met to ensure attrition rates are at a minimum. To learn
more about these needs, Schaller (2005) asked sophomore students if they would join focus
groups or would be willing to complete interviews about their sophomore experience. Students
(n=19) were selected from a mid-size, private Catholic university. Questions about spirituality,
campus involvement, and home life were asked within the interviews, although many of the
respondents spoke upon how they viewed themselves, their relationships, and their academic
experiences and decisions throughout their sophomore year. Schaller (2005) found through these
interviews and focus groups that students transitioned through four stages during their college
careers including 1) random exploration, 2) focused exploration, 3) tentative choices, and 4)
commitment. The random exploration stage explains a time during the college experience when
students are still learning about themselves and their new community. Many study participants in
this study were aware of what they wanted and voiced their opinions and frustrations for the

57	
	

	

choices they previously made or bypassed as freshmen. Schaller (2005) believed that these
comments reflected the fact that the participants had reached the focused exploration phase
which demonstrates frustrations felt with current relationships, themselves, and/or their academic
experiences. Some participants stressed about feeling as if they were being forced to choose a
major during their sophomore year and experienced uncertain thoughts about their futures. It was
clear to Schaller (2005) that many students already transitioned from random exploration in
freshmen year to focused exploration due to their experience as sophomores in college. Once
students realized they were making more permanent choices, they transitioned from the focused
exploration phase to the tentative choices phase. The tentative choice phase allows students to
make decisions that will set them in the direction they would like to travel through the rest of
their college careers. Students in this phase felt a different sense of responsibility. One student
said “I pretty much know what I want out of my future, so I’m going to figure out what to do to
get there and I don’t mind doing things myself.” This phase provides a more permanent
realization for the future. The final phase; commitment, was a phase only few sophomore
students experienced. Students in this phase do not question whether they made correct decisions
to get to this more permanent point. They trust they have made a decision that will work out for
them and their futures. Overall, Schaller (2005) indicated that each stage commented on by
students had associated stressors. With random exploration, students may not have found their
confidence yet and may be feeling doubtful about what they should do with their lives. In
focused exploration, students are searching for direction in life while also being able to review
previous choices. While tentative choices and commitment tend to feel more permanent than
random and focused exploration, there may still be some anxiety about more permanent choices
made. Overall, Schaller (2005) found significant concerns by the sophomores studied. Concerns,

58	
	

	

anxieties, and stressors pose different reasons as to why sophomore students may not be retained
as compared to reasons freshmen students are retained.
Gohn, Swartz, & Donnelly (2000) used an exploratory case study design to learn more
about developmental changes experienced by students. Using qualitative and quantitative
techniques, the researchers’ goal was to gather information about how beginning freshmen who
returned for their sophomore years envisioned themselves in relation to graduation. Gohn,
Swartz, & Donnelly (2000) interviewed sophomore level students to study the experience in
depth. Quantitative questions focused on persistence rates between freshmen to sophomore year,
and sophomore to junior year while qualitative questions included: 1) What are the academic
factors that influenced student retention? 2) What are the issues surrounding emotional and
personal support? 3) What is the commitment and aspirations of the sophomore students? 4)
What are the financial factors affecting the decision to continue? Twenty sophomore level
students were selected to participate in the interview. Researchers selected students who would
represent a cross section of the second year undergraduate population based on the following
characteristics: they were enrolled at the university, represented a variety of majors, academic
standing, high school GPA, ACT scores, cumulative GPA, living on campus or off campus,
scholarship status, racial and ethnic background, and gender. Nine of those twenty students did
not participate due to failure to enroll in the semester the data collection was occurring, no
response, did not want to participate, and did not show for the interview. The remaining eleven
students represented the overall sophomore population at the institution. Quantitative measures
were obtained from the university student database. Factors such as second year to graduation
persistence rates, comparison of retention and graduation rates among regional universities, net
loss of beginning freshmen, beginning freshmen enrolled, and profiles of second year students

59	
	

	

were collected. Qualitative data focused on the following areas of concern when studying
sophomore retention: commitment and aspirations, pre-college factors, college factors, academic
issues, and financial concerns. With respect to commitment and aspirations, sophomore students
felt that their confidence and maturity increased along with their optimism of finishing their
degrees in a timely fashion. Another common theme appeared to include the students’
involvement with the campus community. “Having fun, enjoying the campus and community,
working, and some involvement” supported the fact that these students felt like they were at
home. In fact, the researchers noted that the idea of a “sophomore slump” was non-existent. The
students in this study mentioned several themes when asked about surrounding emotional and
personal support including “social integration,” “lack of isolation,” “personal relationships,” and
“community.” Many students noted that they made friends and felt comfortable talking to faculty
and staff. Academic factors that affected students were a little more complex. Students
mentioned that they seemed satisfied with instructors, especially instructors teaching their major
courses. They mentioned concern with teaching assistants but believed they were doing a good
job overall. The students mentioned that they felt their classes were not as challenging, although
they recognized the importance of attending classes and how attendance is related to academic
performance. In contrast, students who were challenged and struggled through freshmen year,
worried about how grades would affect them throughout their collegiate careers. Several students
mentioned their anxiety with grades and maintaining scholarships. This financial anxiety of
paying for college was a major concern to some respondents. While many common stresses and
anxieties about grades were evident in the data collected in this study, there were some
noticeable differences discovered by Gohn, Swartz, & Donnelly (2000). Although it is important
to determine the reasons for sophomore attrition, it may also be as important to determine the

60	
	

	

reasons for retention in this population so universities can pinpoint opportunities for effective
retention efforts.
SUMMARY
Although researchers have suggested factors that may play into the “sophomore slump,”
others have not witnessed its occurrence through research (Gohn, Swartz, & Donnelly, 2000).
The idea that sophomore students face different challenges than freshmen students has merit as
noted by Gahagan & Hunter (2006), although it may be worthwhile to determine if there are
similarities for retention. Gohn, Swartz, & Donnelly (2000) indicated that sophomore students
who were retained commented on how they felt at home at their universities and that
involvement on campus allowed them to feel a sense of belonging, make friends, and feel
comfortable approaching faculty and staff on campus. It is especially important to continue to
study sophomore students since attrition rates can be substantial, although it may also be
noteworthy to compare freshmen reasons for retention and sophomore reasons for retention to
determine if retention efforts may be managed similarly.
OVERALL SUMMARY
The relationship between campus recreation participation and student academic success
has been studied for decades. Recently, one of the main reasons to study this relationship is for
justification for higher administrators at universities to continue to fund campus recreation
programs during a time when universities may be cutting budgets (Danbert et al., 2014; Lindsey
& Sessoms, 2006; Kampf & Teske, 2013). Several researchers have studied campus recreation
participation and student academic success using different methodologies and focusing on
several academic success outcomes. Some study results are successful at finding a relationship
between campus recreation participation and student academic success (Danbert et al., 2014;

61	
	

	

Todd, et al., 2009; Kampf & Teske, 2013), while others do not find a significant difference in
academic success outcomes between campus recreation participants and nonparticipants (Todd,
et al., 2009; Washke, 1940; Baker, 2008; Zizzi et al., 2004; Kampf & Teske, 2013; McElveen &
Rossow, 2014). It has been suggested that future campus recreation participation research should
focus on campus recreation programs individually, several academic success outcomes should be
studied at a time, and more research should be conducted across the nation to help generalize
results. Due to the significant percentage of first generation students at the university within this
study, it is necessary to study first generation students as a subpopulation in addition to the
general population of students. It is obvious that first generation students enter and continue
through college with challenges that traditional students may not face. Rendon (2000) has
suggested that this population of students gain a better sense of belonging with their universities
in ways that are beneficial to them. Campus recreation research has been linked with sense of
belonging, retention, GPA, credits completed, and overall academic success in other populations
(Danbert et al., 2014; Todd, et al., 2009; Kampf & Teske, 2013; Lindsey & Sessoms, 2006;
Miller, 2011; Lindsey et al., 2009; Forrester, 2015; Henchy, 2011; McElveen & Rossow, 2014;
Henchy, 2013). Considering campus recreation participation of first generation students has not
specifically been studied, research is needed to determine if first generation students can benefit
in many of the same ways as the general population of students when using campus recreation
programs. Furthermore, it is necessary to determine if the general population and first generation
students benefit from participating in any individual campus recreation program and if that
participation is related to several academic success outcomes. Additionally, freshmen students
deserve focus when studying academic success. Considering freshmen student attrition can range
anywhere from 20-30%, freshmen seem to be an obvious population that needs attention (Kampf

62	
	

	

& Teske, 2013). While some researchers have found positive relationships between campus
recreation participation and academic success outcomes (Danbert et al., 2014; Kampf & Teske,
2013) others have only found positive relationships between campus recreation participation and
retention (McElveen & Rossow, 2014). Consistent results are needed to confirm the influence
campus recreation programs may have on academic success.

63	
	

	

REFERENCES

64	
	

	

REFERENCES

Adelman, C. (2006). The toolbox revisited. Paths to degree completion from high school through
college. Washington, DC: US Department of Education.
Astin, A. (1975). Preventing students from dropping out. San Francisco: Jossey-Bass.
Astin, A. (1984). Student involvement: A developmental theory for higher education. Journal of
College Student Personnel, 25, 297-308.
Baker, C. N. (2008). Under-represented college students and extracurricular involvement: the
effects of various student organizations on academic performance. Social Psychology of
Education, 11, 273-298.
Berkner, L. & Chavez, L. (1997). Access to postsecondary education for the 1992 high school
graduates (NCES 98-105). Washington, DC: National Center for Education Statistics, U. S.
Government Printing Office.
Brandon, C. (2010). Five year party. Dallas, TX: Benbella Books Inc.
Bui, K. V. T. (2002). First-generation college students at a four-year university: Background
characteristics, reasons for pursuing higher education, and first-year experiences. College
Student Journal, 36, 3-11.
Choy, S. (2001). Students whose parents did not go to college: Postsecondary access,
persistence, and attainment. US Department of Education, National Center for Education
Statistics, 2001-126.
Dalgarn, M. K. (2001). The role of the campus recreation center in creating a community. NIRSA
Journal, 25(1), 66-72.
Danbert, S. J., Pivarnik, J. M., McNeil, R. N. & Washington, I. J. (2014). Academic success and
retention: The role of recreational sports fitness facilities. Recreational Sports Journal, 38,
14-22.
Duggan, M. (2001). Factors influencing the first-to-second-year persistence of first-generation
college students. Paper presented at the annual meeting of the North East Association for
Institutional Research, Cambridge, MA.
Duggan, M. (2002). The effects of social capital on the first-to-second-year persistence of firstgeneration college students. Dissertation Abstracts International, 63(03), 188. (UMI No
AAT 3045365).
Elkins, D. J., Forrester, S. A. & Noel-Elkins, A. V. (2011). The contribution of campus

65	
	

	

recreational sports participation to perceived sense of campus community. Recreational
Sports Journal, 35(1), 24-34.
Engle, J. & Tinto, V. (2008). Moving beyond access: College success for low-income, first
generation students. Washington, DC: The Pell Institute.
Forrester, S. (2015). Benefits of collegiate recreational sports participation: Results from the
2013 NASPA assessment and knowledge consortium study. Recreational Sports Journal 39,
2-15.
Gahagan, J. & Hunter, M. S. (2006). The second-year experience: Turning attention to the
academy’s middle children. About Campus, July-August.
Gohn, L., Swartz, J., & Donnelly, S. (2000). A case study of second year student persistence.
Journal of College Student Retention, 2(4), 271-294.
Haines, D. J. (2001). Undergraduate student benefits from university recreation. Recreational
Sports Journal, 25(1), 25-33.
Hall, D. A. (2006). Participation in a campus recreation program and its effect on student
retention. Recreational Sports Journal, 30, 40-45.
Henchy, A. (2011). The influence of campus recreation beyond the gym. Recreational Sports
Journal, 35, 174-181.
Henchy, A. (2013). The perceived benefits of participating in campus recreation programs and
facilities: A comparison between undergraduate and graduate students. Recreational Sports
Journal, 37, 97-105.
Horn, L. & Nunez, A. (2000). Mapping the road to college: First-generation students’ math track,
planning strategies, and context of support. Education Statistics Quarterly, 2(1). Retrieved
from http://nces.ed.gov/programs/quarterly/vol_2/2_1/q5-1.asp.
Hossler, D., Schmidt, J. & Vesper, N. (1999). Going to college: How social, economic, and
educational factors influence the decisions students make. Baltimore: Johns Hopkins
University Press.
Ishitani, T. T. (2003). A longitudinal approach to assessing attrition behavior among firstgeneration students: Time-varying effects of pre-college characteristics. Research in Higher
Education, 44, 433-449.
Kampf, S. & Teske, E. J. (2013). Collegiate recreation participation and retention. Recreational
Sports Journal, 37, 85-96.
Lindsey, R. & Sessoms, E. (2006). Assessment of a campus recreation program on student

66	
	

	

recruitment, retention, and frequency of participation across certain demographic variables.
Recreational Sports Journal, 30, 30-39.
Lindsey, R., Sessoms, E. & Willis, G. (2009). Impact of campus recreational sports facilities and
programs on recruitment and retention among African American students: A pilot study.
Recreational Sports Journal, 33, 25-34.
Lipka, S. (2006). After the freshman bubble pops. Chronicle of Higher Education,
p. A34. Retrieved from http://chronicle.com/article/After-the-Freshman-BubblePops/4556
Lohfink, M. M. & Paulsen, M. B. (2005). Comparing the determinants of persistence for firstgeneration and continuing-generation students. Journal of College Student Development,
46(4), 409-428.
London, H. B. (1989). Breaking away: A study of first-generation college students and their
families. American Journal of Education, 97, 144-170.
London, H. B. (1992). Transformations: Cultural challenges faced by first-generation students. In
Zwerling, L. and London, H. editors. First-Generation Students: Confronting the Cultural
Issues (New Directions for Community Colleges) 80, winter. San Francisco: Jossey-Bass.
London, H. B. (1996). How college affects first-generation college students. About Campus, 23,
9-13.
McElveen, M. & Rossow, A. (2014). Relationship of intramural participation to GPA and
retention in first-time-in-college students. Recreational Sports Journal, 38, 50-54.
Miller, J. J. (2011). Impact of a university recreation center on social belonging and student
retention. Recreational Sports Journal, 35, 117-129.
Moffitt, J. (2010). Recreating retention. Recreational Sports Journal, 34, 24-33.
Nunez, A. & Cuccaro-Alamin, S. (1998). First-generation students: Undergraduates whose
parents never enrolled in postsecondary education (NCES 98-082). Washington, DC: U.S.
Department of Education, Office of Educational Research and Improvement.
Pascarella, E. T. & Terenzini, P. T. (1980). Predicting freshmen persistence and voluntary
dropout decisions from a theoretical model. Journal of Higher Education, 5(1), 60-75.
Pascarella, E. T., Wolniak, G. C., Pierson, C. T. & Terenzini, P. T. (2003). Experiences and
outcomes of first-generation students in community colleges. Journal of College Student
Development, 44(3), 420-429.
Pryor, J. H., Hurtado, S., DeAngelo, L., Blake, P. L. & Tran, S. (2010). The American freshman:
National norms fall 2010. Retrieved from University of California, Los Angeles, Higher
Education Research Institute website: http://www.heri.ucla.edu/PDFs/pubs/TFS/Norms/

67	
	

	

Monographs/TheAmerucanFreshman2010.pdf
Rendon, L. I., Jalomo, R. E., & Nora, A. (2000). Theoretical considerations in the study of
minority student retention in higher education. In J. M. Braxton (Ed.), Reworking the student
departure puzzle (pp. 127–156). Nashville: Vanderbilt University Press.
Schaller, M. A. (2005). Wandering and wondering: Traversing the uneven terrain of the second
college year. About Campus, July-August.
Somers, P., Woodhouse, S. & Cofer, J. (2000). Persistence of first-generation college students.
Paper presented at the annual conference of the Association for the Study of Higher
Education, Sacramento, CA.
Spady, W. (1970). Dropouts from higher education: An interdisciplinary review and synthesis.
Interchange, 1, 64-85.
Spady, W. (1971). Dropouts from higher education: Toward an empirical model. Interchange, 2,
38-62.
Stableton, M. J., Soria, K. M., Huesman Jr., R. L. (2014). First-generation students’ sense of
belonging, mental health, and use of counseling services at public research universities.
Journal of College Counseling, 17, 6-20.
Stier, W. F., Schneider, R. C., Kampf, S., Haines, S. G. & Wilding, G. E. (2005). A survey of
campus recreation directors at NIRSA institutions: Activities emphasized, student
participation patterns, trends and future offerings contemplated. Kinesiology, Sport Studies
and Physical Education Faculty Publications. Paper 45.
Tinto, V. (1975). Dropouts from higher education: A theoretical synthesis of recent research.
Review of Educational Research, 45, 89-125.
Tinto, V. (1987). Leaving college: Rethinking the causes and cures of student attrition (1st Ed.).
Chicago: University of Chicago Press.
Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition (2nd Ed.).
Chicago: University of Chicago Press.
Tinto, V. (2006). Research and practice of student retention: What next? Journal of College
Student Retention, 8(1), 1-19.
Todd, M. K., Czyszczon, G., Carr, J. W. & Pratt, C. (2009). Comparison of health and academic
indices between campus recreation facility users and nonusers. Recreational Sports Journal,
33, 43-53.
Warburton, E., Bugarin, R. & Nunez, A. (2001). Bridging the gap: Academic preparation and

68	
	

	

postsecondary success of first-generation students (NCES 2001-153). Washington, DC:
National Center for Educational Statistics, U.S. Government Printing Office.
Washke, P. (1940). A study of intramural sports participation and scholastic attainment.
Research Quarterly, 11(2), 22-27.
Zizzi, S., Ayers, S. F., Watson, J. C. & Keeler, L. A. (2004). Assessing the impact of new student
campus recreation centers. NASPA Journal, 41(4), 588-630.

69	
	

	

CHAPTERS THREE, FOUR, AND FIVE: METHODS

ORGANIZATION OF THE METHODS
Chapters three, four, and five of the dissertation are Methods Sections that focused on the
Specific Aims of this study. Chapter three focused on the overall sample of campus recreation
participants, chapter four focused on first generation campus recreation participants, and chapter
five focused on freshmen campus recreation participants. Chapter three addressed 1) The
relationship between participating in campus recreation programs (club sports, intramural sports,
group fitness classes) and academic success measures (credits completed per academic year and
GPA), 2) If a relationship existed between amount of campus recreation participation (club
sports, intramural sports, group fitness classes) and academic success measures (credits
completed per academic year and GPA) (Specific Aims 1 and 3). Chapter four addressed 1) The
relationship between participating in campus recreation programs (club sports, intramural sports,
group fitness classes) and academic success measures (credits completed per academic year and
GPA) in first generation students, 2) If a relationship existed between amount of campus
recreation participation (club sports, intramural sports, group fitness classes) and academic
success measures (credits completed per academic year and GPA) in first generation students
(Specific Aims 2 and 4). Chapter five addressed 1) If a relationship existed between campus
recreation participation (club sports, intramural sports, or group fitness classes) in freshmen
students on retention (one-year retention and two-year retention), 2) If a relationship existed
between campus recreation participation (club sports, intramural sports, or group fitness classes)
in first generation freshmen students on retention (one-year retention and two-year retention)
(Specific Aims 5 and 6).

70	
	

	

METHODS: CHAPTER THREE
Chapter three focused on Specific Aims one and three:
Specific Aim 1: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) and the following academic success measures:
credits completed per academic year and GPA.
Specific Aim 3: To determine if a relationship existed between amount of campus recreation
participation (club sports, intramural sports, group fitness classes) and the following academic
success measures: credits completed per academic year and GPA.
The dissertation study was observational that utilized a cross sectional design.
METHODS: CHAPTER FOUR
Chapter four focused on Specific Aims two and four:
Specific Aim 2: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) and the following academic success measures:
credits completed per academic year and GPA in first generation students.
Specific Aim 4: To determine if a relationship existed between amount of campus recreation
participation (club sports, intramural sports, group fitness classes) and the following academic
success measures: credits completed per academic year and GPA in first generation students.
METHODS: CHAPTER FIVE
Chapter five focused on Specific Aims five and six:
Specific Aim 5: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) in freshmen students and the following
academic success measures: one-year retention and two-year retention.

71	
	

	

Specific Aim 6: To determine the relationship between campus recreation participation (club
sports, intramural sports, or group fitness classes) in first generation freshmen students and the
following academic success measures: one-year retention and two-year retention.
STUDY POPULATION AND RECRUITMENT
The population of interest was college students. The sample consisted of male and female
undergraduate students who participated in club sports, intramural sports, and/or group fitness
classes at Saginaw Valley State University in University Center, Michigan during the 2013-2014
and 2014-2015 academic years. Data were collected as part of participation data that the Director
of Campus Recreation used to evaluate each program. The intention for this study was to
determine if there were relationships between participating in these programs and academic
success outcomes. Positive relationships between participation and academic success measures
add to the argument that non-academic university programs may be related to a number of
academic outcomes such as grade point average, successful completion of coursework, and
retention.
Demographic Variables
Study participants were male and female undergraduate students eighteen years old and
older. Gender and age were determined by the gender and date of birth the participants claimed
on their college entrance applications. For class standing, participants fell under one of the
following categories: Freshmen, Sophomore, Junior, or Senior. Class standing was determined
by the number of credits accumulated by participants at the beginning of the 2013-2014 and
2014-2015 academic year. Participants were classified by class standing using the following
credit scheme implemented by the university Registrar: 1) Freshmen earned 30 credits or less, 2)
Sophomores earned 31-61 credits, 3) Juniors earned 62-92 credits, and 4) Seniors earned 93

72	
	

	

credits or more. Race was determined by participants’ response on their college entrance
applications. Available categories for race included: American Indian, Alaska Native, Asian,
African American, Canadian, Hispanic/Latino, Native Hawaiian/Pacific Islander, International,
Multiple Race, Not Reported/Unknown, White. Subgroups of interest were first generation and
freshmen students. The same demographic variables populated for campus recreation
participants were also determined for first generation and freshmen campus recreation
participants.
ACADEMIC SUCCESS OUTCOMES
Academic success outcome variables were populated for each participant. Academic
success variables in this study included credits completed per year and GPA for all campus
recreation participants. Additionally, one-year and two-year retention were determined for
freshmen campus recreation participants. The total number of credits completed during 20132014 and 2014-2015 was calculated. GPA was calculated as a continuous variable and expressed
on a 0.0-4.0 scale based on the credits earned during the 2013-2014 and 2014-2015 academic
years. Age, class standing, race, Pell Grant eligibility, gender, year of freshmen year, ACT score,
high school GPA, number of transfer credits, AP credits transferred, and post-secondary credits
transferred were also populated. These variables were included as covariates in the analysis as
they could have been possible confounding variables.
CAMPUS RECREATION PARTICIPATION
Club Sport Participation
Club sport participation for the 2013-2014 and 2014-2015 academic year was recorded at
the time it occurred. Demographic data and academic success outcome variables were populated
for each participant and then de-identified prior to analysis.

73	
	

	

Intramural Sport Participation
Intramural sport participation for the 2013-2014 and 2014-2015 academic year was
collected at the time it occurred using an online registration program called IMLeagues®
(IMLeagues.com). Participation in intramural sports was reliant on students’ successful
registration through IMLeagues®. Demographic data and academic success outcome variables
were populated for each participant and then de-identified prior to analysis.
Group Fitness Class Participation
Group fitness class participation for the 2013-2014 and 2014-2015 academic year was
recorded at the time it occurred. Demographic data and academic success outcome variables
were populated for each participant and then de-identified prior to analysis.
Non-Participants
After obtaining the participant list, a nonparticipant group was created. Demographic data
and academic success outcome variables for each non-participant were populated and then deidentified prior to analysis. Propensity score matching allowed the principal investigator to
match nonparticipants to participants to reduce the number and effect of possible confounding
variables. The goal for using propensity score matching was to create a non-participant group
that was similar to the participant group in as many ways as possible except that the
nonparticipants did not participate in club sports, intramural sports, and group fitness classes.
The non-participant group was used as the comparison group for the participant group.
Procedure
Student participation in club sports, intramural sports, and group fitness classes was
recorded for the 2013-2014 and 2014-2015 academic years. Age, class standing, race, Pell Grant
eligibility, gender, year of freshmen year, first generation status, ACT score, high school GPA,

74	
	

	

SVSU GPA, credits completed, transfer credits, AP credits transferred, post-secondary credits
transferred and retention were retrieved from the Registrar at Saginaw Valley State University.
Participant and non-participant data were de-identified prior to analysis.
STATISTICAL ANALYSES
Descriptive statistics (means, standard deviations, and proportions) were used to
determine if the sample had similar demographics as the university as a whole (e.g.
males/females, international population, first generation students). Propensity score matching
allowed a non-participant group to be created as a comparison group for the campus recreation
participant group. Greedy propensity score matching using nearest neighbor was implemented
for this study to determine a one-to-one match. According to (Thoemmes & Kim, 2011), greedy
propensity score matching using nearest neighbor allows researchers to find an untreated unique
observation that has the closest propensity score to a treated unique observation. The propensity
score assigned to each unique observation is based on the covariates included in the model. For
this study, the following covariates were included: age, class standing, race, Pell Grant
eligibility, gender, year of freshmen year, ACT score, high school GPA, number of transfer
credits, AP credits transferred, and post-secondary credits transferred.
Several two-sample tests for proportions (z-tests) were used to analyze differences in
credits completed per academic year, GPA, and one-year and two-year retention between the
participants and nonparticipants (Specific Aims 1, 2, 5, 6). Effect sizes were calculated to
determine the magnitude of effect in significant findings between participation in campus
recreation programs and academic outcomes. The cross-sectional data were analyzed with an
ANOVA to evaluate the relationships between amount of campus recreation participation and
credits completed per academic year (Specific Aims 3, 4). Another ANOVA was conducted to

75	
	

	

evaluate the relationships between amount of campus recreation participation and GPA (Specific
Aims 3, 4). Program frequency groups were categorized as 1) nonparticipants, 2) individuals
who participated in one campus recreation program per year 3) individuals who participated in
two or more campus recreation programs per year. Covariates considered within this model
included age, class standing, race, Pell Grant eligibility, gender, year of freshmen year, high
school GPA, ACT score, transfer credits, AP credits transferred, and post-secondary credits
transferred. Additionally, linear trend analyses were conducted to determine significant
differences between the participant and nonparticipant groups.
Alpha was set at 0.05 for statistical significance and results were reported using M ± SD.
Descriptive statistics, ANOVAs, and linear trend analyses were run using SPSS 23 statistical
software (SPSS, Inc., Chicago, IL, USA). Two-sample tests for proportions (z-tests) were
conducted using Microsoft® Excel Version 15.38 (© 2017 Microsoft). Propensity Score
Analysis was added to SPSS by uploading a plug-in for R (R Core Team (2016). R: A language
and environment for statistical computing. R Foundation for Statistical Computing, Vienna,
Austria). Effect sizes were calculated using a Cohen’s D Effect Size Calculator for Z-Tests
(Social Science Statistics, 2017).

76	
	

	

REFERENCES

77	
	

	

REFERENCES

Stangroom, J. (2017). Z Score Calculator for 2 Population Proportions. Retrieved from
http://socscistatistics.com
Thoemmes, F. J. & Kim, E. S. (2011). A systematic review of propensity score methods in the
social sciences. Multivariate Behavioral Research, 46(1), 90-118.

78	
	

	

CHAPTER SIX: RESULTS

SPECIFIC AIM #1: CREDITS COMPLETED (TOTAL SAMPLE)
Prior to propensity score matching (PSM), there were 8613 unique individuals in the
sample. Of the 8613 individuals, 1442 were campus recreation participants (treated) and 7171
were nonparticipants (controls). After PSM, there were 1442 participants matched with 1442
nonparticipants, leaving 5729 unmatched nonparticipants. None of the data were discarded.
Table 1.0 shows the sample as a total, matched, unmatched, and discarded data.
Table: 1.0. Propensity Score Matching Sample Sizes
Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
7171 1442
1442
1442
5729
0
0
0

Figure 1.0 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows zero unmatched participants. The next two
stratifications show the 1442 matched participants and 1442 matched nonparticipants. Finally,
the last stratification shows the 5729 unmatched nonparticipants. There was a great deal of
overlap in propensity scores (0.0-0.6) between the matched treatment group and the matched
control group. This is meaningful because a small range suggests that if two individuals were
chosen, one from each group, the main relevant difference between individuals would be campus
recreation program participation. Since randomization is not possible for observational studies
like this, matching is used as a way to imitate randomization to create comparable groups to
analyze. To review the difference in the means of the data, PSM was used to show the covariate
balance in the dataset. Figure 1.1 shows the improvement in the covariate balance in the dataset
after PSM. The absolute standardized difference in the mean after PSM was less than the

79	
	

	

absolute standardized difference in the mean prior to PSM, which is essential when ensuring data
are matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.
Figure 1.0. Distribution of Propensity Scores

Figure 1.1. Standardized Difference in Means

80	
	

	

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 1.2). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis.
Figure 1.2. Histogram of Propensity Scores Before and After Matching

Additionally, a histogram of standardized differences was created to show the data as overlaid
kernel density estimates of standardized differences before and after matching (Figure 1.3). The
histogram at the top of the figure shows the standardized differences prior to PSM. The
histogram at the bottom of the figure shows the decreased standardized differences after PSM.

81	
	

	

Like the standardized differences in means figure (Figure 1.1), the closer to 0.0 the difference in
the means are, the more exceptional the match.
Figure 1.3. Histogram of Standardized Differences Before and After Matching

Of the 2884 individuals within the sample, 48.2% were male, 51.8% were female. The
average age of individuals in the sample was 19.4 ± 1.6 years. Additionally, 65.1% lived on
campus, 77.8% were white, 59.8% were Freshman status, 27.8% were Sophomore status, 10.1%
were Junior status, 2.3% were Senior status, 92.8% started in 2011 or later, average ACT was
21.4 ± 6.5, average high school GPA was 3.18 ± 0.99, 14.3% were Pell Grant eligible, 27.9%
were first generation status, and on average transferred in 4.6 ± 11.7 credits.
After PSM matched the participants with nonparticipants, a two-sample test for
proportions (z-test) (Table 1.1) was conducted to determine if there was a difference in credits
completed between the participants and matched nonparticipants. Participants in club sports,
intramural sports, and fitness classes completed an average of 26.7 ± 14.9 credits per year.
82	
	

	

Nonparticipants completed an average of 24.0 ± 14.2 credits. The difference in credits was
significant (ρ < 0.001) suggesting participants completed significantly more credits per year
when compared to the nonparticipants. Cohen’s (1988) effect size value (d = 0.19) suggested a
low practical significance.
Table 1.1. Two-Sample Test for Proportions (z-test) for Participants and Nonparticipants:
Credits Completed
z-Test: Two Sample for Means

Mean
Known Variance
Observations
Hypothesized Mean Difference
z
P(Z<=z) one-tail
z Critical one-tail
P(Z<=z) two-tail
z Critical two-tail

Variable 1
0.666666667
0.22302158
279
0
3.740122446
9.19653E-05
1.644853627
0.000183931
1.959963985

Variable 2
0.512544803
0.25074134
279

SPECIFIC AIM #1: GPA (TOTAL SAMPLE)
Prior to PSM, there were 13,205 unique individuals in the sample. Of the 13,205
individuals, 2119 were campus recreation participants (treated) and 11,086 were nonparticipants
(controls). After PSM, there were 2119 participants matched with 2119 nonparticipants, leaving
8967 unmatched nonparticipants. None of the data were discarded. Table 1.2 shows the sample
as a total, matched, unmatched, and discarded data.
Table 1.2. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
11086
2119
2119
2119
8967
0
0
0

	

83	
	

	

Figure 1.4 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows zero unmatched participants. The next two
stratifications show the 2119 matched participants and 2119 matched nonparticipants. Finally,
the last stratification shows the 8967 unmatched nonparticipants. There was a great deal of
overlap in propensity scores (0.0-0.6) between the matched treatment group and the matched
control group. This is meaningful because this small range suggests that if two individuals were
chosen, one from each group, the main relevant difference between individuals would be campus
recreation program participation. Since randomization is not possible for observational studies
like this, matching is used as a way to imitate randomization to create comparable groups to
analyze.
Figure 1.4. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 1.5 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are
matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.

84	
	

	

Figure 1.5. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 1.6). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis.
Figure 1.6. Histogram of Propensity Scores Before and After Matching

85	
	

	

Additionally, a histogram of standardized differences was created to show the data as overlaid
kernel density estimates of standardized differences before and after matching (Figure 1.7). The
histogram at the top of the figure shows the standardized differences prior to PSM. The
histogram at the bottom of the figure shows the decreased standardized differences after PSM.
Like the standardized differences in means figure (Figure 1.5), the closer to 0.0 the difference in
the means are, the more exceptional the match.
Figure 1.7. Histogram of Standardized Differences Before and After Matching

Of the 4238 individuals within the sample, 51.1% were male, 48.9% were female. The
average age of individuals in the sample was 19.9 ± 2.1 years. Additionally, 58.3% lived on
campus, 75.8% were white, 50.5% Freshman status, 25.8% Sophomore status, 16.4% Junior
status, and 7.3% Senior status, 84.7% started in 2011 or later, average ACT score was 21.0 ± 6.7,

86	
	

	

average high school GPA was 3.08 ± 1.05, 15.5% were Pell Grant eligible, 29.0% were first
generation status, and on average transferred in 3.9 ± 11.2 credits.
After PSM matched the participants with nonparticipants, a two-sample test for
proportions (z-test) (Table 1.3) was conducted to determine if there was a difference in GPA
between the participants and matched nonparticipants. Participants in club sports, intramural
sports, and fitness classes had an average GPA of 3.08 ± 0.61. Nonparticipants had an average
GPA of 2.98 ± 0.69. The difference in GPA was significant (ρ < 0.001) suggesting participants
had a significantly higher GPA per year when compared to the nonparticipants. Cohen’s (1988)
effect size value (d = 0.15) suggested a low practical significance.
Table 1.3. Two-Sample Test for Proportions (z-test) for Participants and Nonparticipants:
GPA
z-Test:	Two	Sample	for	Means	
		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

	
Variable	1	
3.082760736	
0.36930064	
2119	
0	
4.932774922	
4.05348E-07	
1.644853627	
8.10696E-07	
1.959963985	 		

Variable	2	
2.984374705	
0.47367542	
2119	
	
	
	
	
	

SPECIFIC AIM #2: CREDITS COMPLETED (FIRST GENERATION)
Prior to PSM, there were 2625 unique individuals in the sample. Of the 2625 individuals,
412 were campus recreation participants (treated) and 2213 were nonparticipants (controls).
After PSM, there were 383 participants matched with 383 nonparticipants. Twenty-nine
participants could not be matched with nonparticipants, leaving 1830 unmatched nonparticipants.

87	
	

	

None of the data were discarded. Table 2.0 shows the sample as a total, matched, unmatched,
and discarded data.
Table 2.0. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
2213
412
383
383
1830
29
0
0

	

Figure 2.0 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows the 29 unmatched participants. The next two
stratifications show the 383 matched participants and 383 matched nonparticipants. Finally, the
last stratification shows the 1830 unmatched nonparticipants. There was a great deal of overlap
in propensity scores (0.0-0.7) between the matched treatment group and the matched control
group. This is meaningful because this small range suggests that if two individuals were chosen,
one from each group, the main difference between individuals would be campus recreation
program participation. Since randomization is not possible for observational studies like this,
matching is used as a way to imitate randomization to create comparable groups to analyze.
Figure 2.0. Distribution of Propensity Scores

88	
	

	

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 2.1 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are
matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.
Figure 2.1. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 2.2). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis. Additionally, a histogram of standardized differences was created to show the

89	
	

	

data as overlaid kernel density estimates of standardized differences before and after matching
(Figure 2.3). The histogram at the top of the figure shows the standardized differences prior to
PSM. The histogram at the bottom of the figure shows the decreased standardized differences
after PSM. Like the standardized differences in means figure (Figure 2.1), the closer to 0.0 the
difference in the means are, the more exceptional the match.
Figure 2.2. Histogram of Propensity Scores Before and After Matching

Figure 2.3. Histogram of Standardized Differences Before and After Matching

90	
	

	

Of the 766 individuals within the sample, 49.3% were male, 50.7% were female. The
average age of individuals in the sample was 19.7 ± 1.9 years. Additionally, 65.3% lived on
campus, 74.8% were white, 54.4% Freshman status, 32.1% Sophomore status, 11.4% Junior
status, and 2.1% Senior status, 93.3% started in 2011 or later, average ACT score was 20.6 ± 6.2,
average high school GPA was 3.12 ± 0.93, 53.3% were Pell Grant eligible, and on average
transferred in 4.2 ± 12.3 credits.
After PSM matched the participants with nonparticipants, a two-sample test for
proportions (z-test) (Table 2.1) was conducted to determine if there was a difference in credits
completed between the participants and matched nonparticipants. Participants in club sports,
intramural sports, and fitness classes completed an average of 24.9 ± 13.9 credits per year.
Nonparticipants completed an average of 22.5 ± 13.3 credits. The difference in credits was
significant (ρ = 0.013) suggesting participants completed significantly more credits per year
when compared to the nonparticipants. Cohen’s (1988) effect size value (d = 0.18) suggested a
low practical significance.
Table 2.1. Two-Sample Test for Proportions (z-test) for First Generation Participants and
First Generation Nonparticipants: Credits Completed
z-Test:	Two	Sample	for	Means	
		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

	
Variable	1	
24.92428198	
192.609435	
383	
0	
2.496120035	
0.006278005	
1.644853627	
0.012556011	
1.959963985	 		

91	
	

	

Variable	2	
22.47519582	
176.09297	
383	
	
	
	
	
	

SPECIFIC AIM #2: GPA (FIRST GENERATION)
Prior to PSM, there were 4042 unique individuals in the sample. Of the 4042 individuals,
614 were campus recreation participants (treated) and 3428 were nonparticipants (controls).
After PSM, there were 588 participants matched with 588 nonparticipants. Twenty-six
participants could not be matched with nonparticipants, leaving 2840 unmatched nonparticipants.
None of the data were discarded. Table 2.2 shows the sample as a total, matched, unmatched,
and discarded data.
Table 2.2. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
3428
614 588
588 2840
26
0
0

	
Figure 2.4 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows 26 unmatched participants. The next two
stratifications show the 588 matched participants and 588 matched nonparticipants. Finally, the
last stratification shows the 2840 unmatched nonparticipants. There was a great deal of overlap
in propensity scores (0.0-0.7) between the matched treatment group and the matched control
group. This is meaningful because this small range suggests that if two individuals were chosen,
one from each group, the main difference between individuals would be campus recreation
program participation. Since randomization is not possible for observational studies like this,
matching is used as a way to imitate randomization to create comparable groups to analyze.

92	
	

	

Figure 2.4. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 2.5 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are
matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.
Figure 2.5. Standardized Difference in Means

93	
	

	

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 2.6). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis. Additionally, a histogram of standardized differences was created to show the
data as overlaid kernel density estimates of standardized differences before and after matching
(Figure 2.7). The histogram at the top of the figure shows the standardized differences prior to
PSM. The histogram at the bottom of the figure shows the decreased standardized differences
after PSM. Like the standardized differences in means figure (Figure 2.5), the closer to 0.0 the
difference in the means are, the more exceptional the match.
Figure 2.6. Histogram of Propensity Scores Before and After Matching

94	
	

	

Figure 2.7. Histogram of Standardized Differences Before and After Matching

Of the 1176 individuals within the sample, 49.6% were male, 50.4% were female. The
average age of individuals in the sample was 19.9 ± 1.8 years. Additionally, 60.0% lived on
campus, 75.1% were white, 45.9% Freshman status, 30.1% Sophomore status, 19.6% Junior
status, 4.4% Senior status, 85.3% started in 2011 or later, average ACT score was 20.4 ± 6.3,
average high school GPA was 3.08 ± 0.98, 54.1% were Pell Grant eligible, and on average
transferred in 3.8 ± 11.9 credits.
After PSM matched the participants with nonparticipants, a two-sample test for
proportions (z-test) (Table 2.3) was conducted to determine if there was a difference in GPA
between the participants and matched nonparticipants. Participants in club sports, intramural
sports, and fitness classes had an average GPA of 3.03 ± 0.59. Nonparticipants had an average
GPA of 2.95 ± 0.67. The difference in GPA was not significant (ρ = 0.050) suggesting there was
not a significant difference in GPA between the participants and nonparticipants.

95	
	

	

Table 2.3. Two-Sample Test for Proportions (z-test) for First Generation Participants and
First Generation Nonparticipants: GPA
z-Test:	Two	Sample	for	Means	
		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

	
Variable	1	
3.025986395	
0.34846938	
588	
0	
1.956699964	
0.025191377	
1.644853627	
0.050382754	
1.959963985	 		

Variable	2	
2.954166667	
0.44369692	
588	
	
	
	
	
	

SPECIFIC AIM #3: CREDITS COMPLETED (TOTAL SAMPLE - AMOUNT OF
PARTICIPATION)
Prior to PSM, there were 8613 unique individuals in the sample. Of the 8613 individuals,
1442 were campus recreation participants (treated) and 7171 were nonparticipants (controls).
After PSM, there were 1442 participants matched with 1442 nonparticipants, leaving 5729
unmatched nonparticipants. None of the data were discarded. Table 3.0 shows the sample as a
total, matched, unmatched, and discarded data.
Table 3.0. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
7171
1442 1442
1442
5729
0
0
0

	

Figure 3.0 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows zero unmatched participants. The next two
stratifications show the 1442 matched participants and 1442 matched nonparticipants. Finally,

96	
	

	

the last stratification shows the 5729 unmatched nonparticipants. There was a great deal of
overlap in propensity scores (0.0-0.6) between the matched treatment group and the matched
control group. This is meaningful because this small range suggests that if two individuals were
chosen, one from each group, the main difference between individuals would be campus
recreation program participation. Since randomization is not possible for observational studies
like this, matching is used as a way to imitate randomization to create comparable groups to
analyze.
Figure 3.0. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 3.1 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are
matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.

97	
	

	

Figure 3.1. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 3.2). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis. Additionally, a histogram of standardized differences was created to show the
data as overlaid kernel density estimates of standardized differences before and after matching
(Figure 3.3). The histogram at the top of the figure shows the standardized differences prior to
PSM. The histogram at the bottom of the figure shows the decreased standardized differences
after PSM. Like the standardized differences in means figure (Figure 3.1), the closer to 0.0 the
difference in the means are, the more exceptional the match.

98	
	

	

Figure 3.2. Histogram of Propensity Scores Before and After Matching

Figure 3.3. Histogram of Standardized Differences Before and After Matching

Of the 2884 individuals within the sample, 47.7% were male, 52.3% were female. The
average age of individuals in the sample was 19.4 ± 1.5 years. Additionally, 64.8% lived on

99	
	

	

campus, 77.5% were white, 59.7% Freshman status, 28.4% Sophomore status, 10.2% Junior
status, 1.7% Senior status, 92.9% started in 2011 or later, average ACT score was 21.4 ± 6.5,
average high school GPA was 3.19 ± 0.97, 15.0% were Pell Grant eligible, 27.9% were first
generation status, and on average transferred in 4.8 ± 12.0 credits.
After PSM matched the participants with nonparticipants, an ANOVA was conducted to
determine if there was a difference in credits completed between participants in one program per
year (n=1278), participants in two or more programs per year (n=164) and nonparticipants
(n=1442). Participants in one program completed an average of 26.8 ± 15.0 credits per year.
Participants in two or more programs completed an average of 26.0 ± 14.6 credits per year.
Nonparticipants completed an average of 24.3 ± 14.3 credits per year (Table 3.1).
Table 3.1. Descriptives for Participants and Nonparticipants: Credits Completed

Since the assumption of homogeneity of variance was violated (as shown in Table 3.2), it was
necessary to review the Welch and Brown-Forsythe F-ratios (Table 3.4) rather than the main
table (Table 3.3: ANOVA Results). A statistically significant finding occurred in Table 3.4
(ρ < 0.001). It can be determined that an F-ratio of this size would occur if there was no effect,
therefore providing an overall difference between the groups.

100	
	

	

Table 3.2. Test of Homogeneity of Variances
TotalCred
Levene Statistic
df1
6.811
2
	

df2

Sig.
2881

.001

Table 3.3. ANOVA Results

TotalCred

Between Groups
Within Groups
Total

Sum of
Squares
4416.708
614787.603
619204.311

df

Mean Square
2
2208.354
2881
213.394
2883

F
Sig.
10.349 < 0.001

Table 3.4. Robust Tests of Equality of Means
TotalCred
Statistic df1
df2
Sig.
Welch
10.306
2 453.375 < 0.001
a

Brown10.300
2 680.796 < 0.001
Forsythe
a. Asymptotically F distributed.

Since there was an overall difference between the groups, yet the homogeneity of variance
assumption was violated, a Games Howell Post Hoc Test was conducted to determine differences
between groups (Table 3.5). When comparing the difference in credits completed between
participants in one program and nonparticipants, a significant difference was found (ρ < 0.001).
Participants in two or more programs did not complete significantly more credits per year when
compared to nonparticipants (ρ = 0.298). Finally, there was not a significant difference in credits
completed between the participants in one program per year and participants in two or more
programs per year (ρ = 0.814). Therefore, significance was found only with credits completed
between the participants in one program per year and the nonparticipants.
101	
	

	

Table 3.5. Post Hoc Tests
Multiple Comparisons
Dependent Variable: TotalCred
Games-Howell
(J)
Mean
95% Confidence Interval
(I) NumberofSports NumberofSports Difference (I-J) Std. Error
Sig.
Lower Bound Upper Bound
.00
1.00
-2.53697
.56269 < 0.001
-3.8565
-1.2174
2.00
-1.79357
1.20352
.298
-4.6355
1.0483
1.00
.00
2.53697
.56269 < 0.001
1.2174
3.8565
2.00
.74340
1.21737
.814
-2.1302
3.6170
2.00
.00
1.79357
1.20352
.298
-1.0483
4.6355
1.00
-.74340
1.21737
.814
-3.6170
2.1302
*. The mean difference is significant at the 0.05 level.
*

*

To determine if there were significant differences between the participation and nonparticipation
groups based on level of participation, a linear trend test was run. Reviewing Table 3.6., a
statistically significant difference between the groups was found (ρ < 0.001).

Table 3.6. ANOVA Table for Linear Trend Analysis
ANOVA
TotalCred

Between
Groups

(Combined)
Linear
Unweighted
Term
Weighted
Deviation
Within Groups
Total

Sum of
Squares
4416.708
473.699
3332.045
1084.662
614787.603
619204.311

df
2
1
1
1
2881
2883

Mean
Square
F
Sig.
2208.354 10.349 < 0.001
473.699 2.220
.136
3332.045 15.615 < 0.001
1084.662 5.083
.024
213.394

Figure 3.4. shows the visual representation of the differences in credits completed between the
groups. While there was a significant difference in credits completed between the participants in

102	
	

	

one program and nonparticipants, participants in two or more programs did not complete
significantly more credits than nonparticipants. There was also not a significant difference in
credits completed between the participants in one program per year and participants in two or
more programs per year.
Figure 3.4. Means Plot for Linear Trend Analysis

SPECIFIC AIM #3: GPA (TOTAL SAMPLE - AMOUNT OF PARTICIPATION)
Prior to PSM, there were 13,207 unique individuals in the sample. Of the 13,207
individuals, 2120 were campus recreation participants (treated) and 11,087 were nonparticipants
(controls). After PSM, there were 2120 participants matched with 2120 nonparticipants, leaving
8967 unmatched nonparticipants. None of the data were discarded. Table 3.7 shows the sample
as a total, matched, unmatched, and discarded data.
Table 3.7. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
11087 2120 2120
2120 8967
0
0
0

	

103	
	

	

Figure 3.5 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows zero unmatched participants. The next two
stratifications show the 2120 matched participants and 2120 matched nonparticipants. Finally,
the last stratification shows the 8967 unmatched nonparticipants. There was a great deal of
overlap in propensity scores (0.0-0.6) between the matched treatment group and the matched
control group. This is meaningful because this small range suggests that if two individuals were
chosen, one from each group, the main difference between individuals would be campus
recreation program participation. Since randomization is not possible for observational studies
like this, matching is used as a way to imitate randomization to create comparable groups to
analyze.
Figure 3.5. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 3.6 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are
104	
	

	

matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.

Figure 3.6. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 3.7). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis.

105	
	

	

Figure 3.7. Histogram of Propensity Scores Before and After Matching

Additionally, a histogram of standardized differences was created to show the data as overlaid
kernel density estimates of standardized differences before and after matching (Figure 3.8). The
histogram at the top of the figure shows the standardized differences prior to PSM. The
histogram at the bottom of the figure shows the decreased standardized differences after PSM.
Like the standardized differences in means figure (Figure 3.6), the closer to 0.0 the difference in
the means are, the more exceptional the match.
Figure 3.8. Histogram of Standardized Differences Before and After Matching

106	
	

	

Of the 4240 individuals within the sample, 50.4% were male, 49.6% were female. The
average age of individuals in the sample was 19.9 ± 1.8 years. Additionally, 58.4% lived on
campus, 76.9% were white, 48.1% Freshman status, 28.5% Sophomore status, 17.7% Junior
status, 5.7% Senior status, 86.1% started in 2011 or later, average ACT score was 21.0 ± 6.9,
average high school GPA was 3.08 ± 1.06, 14.7% were Pell Grant eligible, 28.4% were first
generation status, and on average transferred in 4.0 ± 11.6 credits.
After PSM matched the participants with nonparticipants, an ANOVA was conducted to
determine if there was a difference in GPA between participants in one program per year
(n=1184), participants in two or more programs per year (n=236) and nonparticipants (n=2120).
Participants in one program had an average GPA of 3.08 ± 0.61. Participants in two or more
programs had an average GPA of 3.13 ± 0.59. Nonparticipants had an average GPA of 3.03 ±
0.65 (Table 3.8).
Table 3.8. Descriptives for Participants and Nonparticipants: GPA
GPA

.00
1.00
2.00
Total
Model Fixed Effects
Random Effects

N
2120
1884
236
4240

Mean
3.0267
3.0756
3.1346
3.0544

Std.
Deviation Std. Error
.65403
.01420
.60942
.01404
.59160
.03851
.63178
.00970
.63118
.00969
.02670

95% Confidence Interval for
BetweenMean
Component
Lower Bound Upper Bound Minimum Maximum Variance
2.9988
3.0545
.00
4.00
3.0480
3.1031
.00
4.00
3.0587
3.2105
1.40
4.00
3.0354
3.0734
.00
4.00
3.0354
3.0734
2.9395
3.1693
.00137

	

Reviewing Table 3.9, it was determined that the variances of the three groups were not
significantly different since ρ > 0.05. Therefore, the homogeneity of variance assumption has
been met.

107	
	

	

Table 3.9. Test of Homogeneity of Variances

GPA
Levene Statistic
1.149
	

df1
2

df2
4237

Sig.
.317

Table 3.9.1 shows that there was an overall significant difference between the groups
(ρ = 0.007).

Table 3.9.1. ANOVA Results
GPA

Between Groups
Within Groups
Total

Sum of
Squares
3.996
1687.982
1691.978

df
2
4237
4239

Mean Square
F
1.998 5.015
.398

Sig.
.007

	

Since there was an overall difference between the groups, and the homogeneity of variance
assumption was met, a Tukey Post Hoc Test was run to determine which specific groups
differed. When reviewing GPA of participants in one program and nonparticipants, it was
determined that there was a significant difference in GPA between the groups since ρ = 0.038.
There was also a significant difference in GPA between the participants in two or more programs
and nonparticipants (ρ = 0.034). There was not a significant difference in GPA between
participants in one program and participants in two or more programs (ρ = 0.365). Therefore,
significance was found with GPA between the participants in one program per year and
nonparticipants, and participants in two or more programs per year and nonparticipants.

108	
	

	

Table 3.9.2. Post Hoc Tests
Multiple Comparisons
Dependent Variable: GPA
Tukey HSD
Mean
(J)
Difference
NumberofSports
(I-J)
Std. Error
1.00
-.04894*
.01998
2.00
-.10795*
.04331
*
1.00
.00
.04894
.01998
2.00
-.05902
.04358
*
2.00
.00
.10795
.04331
1.00
.05902
.04358
*. The mean difference is significant at the 0.05 level.
(I)
NumberofSports
.00

Sig.
.038
.034
.038
.365
.034
.365

95% Confidence Interval
Lower
Upper
Bound
Bound
-.0958
-.0021
-.2095
-.0064
.0021
.0958
-.1612
.0432
.0064
.2095
-.0432
.1612

	

To determine if there were significant differences between the participation and nonparticipation
groups based on level of participation, a linear trend test was run. Reviewing Table 3.9.3., a
statistically significant difference between the groups was found (ρ = 0.007).
Table 3.9.3. ANOVA Table for Linear Trend Analysis
ANOVA
GPA

Between
Groups

Within Groups
Total

(Combined)
Linear Term Unweighted
Weighted
Deviation

Sum of
Squares
3.996
2.475
3.981
.015
1687.982
1691.978

Mean
df
Square
2
1.998
1
1
1
4237
4239

2.475
3.981
.015
.398

F
5.015

Sig.
.007

6.212
9.993
.037

.013
.002
.847

Figure 3.9. shows the visual representation of the differences in GPA between the groups. There
was a significant difference in GPA between the participants in one program and nonparticipants,
and between the participants in two or more programs and nonparticipants. There was not a
significant difference in GPA between the participants in one program per year and participants

109	
	

	

in two or more programs per year, although there was a trend of an increase in GPA with an
increase in participation.
Figure 3.9. Means Plot for Linear Trend Analysis

SPECIFIC AIM #4: CREDITS COMPLETED (FIRST GENERATION-AMOUNT OF
PARTICIPATION)
Prior to PSM, there were 2625 unique individuals in the sample. Of the 2625 individuals,
412 were campus recreation participants (treated) and 2213 were nonparticipants (controls).
After PSM, there were 381 participants matched with 381 nonparticipants. Thirty-one
participants could not be matched with nonparticipants, leaving 1832 unmatched nonparticipants.
None of the data were discarded. Table 4.0 shows the sample as a total, matched, unmatched,
and discarded data.
Table 4.0. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
2213 412 381
381 1832
31
0
0

	

110	
	

	

Figure 4.0 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows 31 unmatched participants. The next two
stratifications show the 381 matched participants and 381 matched nonparticipants. Finally, the
last stratification shows the 1832 unmatched nonparticipants. There was a great deal of overlap
in propensity scores (0.0-0.7) between the matched treatment group and the matched control
group. This is meaningful because this small range suggests that if two individuals were chosen,
one from each group, the main difference between individuals would be campus recreation
program participation. Since randomization is not possible for observational studies like this,
matching is used as a way to imitate randomization to create comparable groups to analyze.
Figure 4.0. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 4.1 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are

111	
	

	

matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.
Figure 4.1. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 4.2). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis. Additionally, a histogram of standardized differences was created to show the
data as overlaid kernel density estimates of standardized differences before and after matching
(Figure 4.3). The histogram at the top of the figure shows the standardized differences prior to
PSM. The histogram at the bottom of the figure shows the decreased standardized differences
after PSM. Like the standardized differences in means figure (Figure 4.1), the closer to 0.0 the
difference in the means are, the more exceptional the match.

112	
	

	

Figure 4.2. Histogram of Propensity Scores Before and After Matching

Figure 4.3. Histogram of Standardized Differences Before and After Matching

113	
	

	

Of the 762 individuals within the sample, 48.2% were male, 51.8% were female. The
average age of individuals in the sample was 19.7 ± 1.9 years. Additionally, 64.7% lived on
campus, 74.0% were white, 54.5% Freshman status, 30.8% Sophomore status, 12.2% Junior
status, 2.5% Senior status, 93.1% started in 2011 or later, average ACT score was 20.6 ± 6.1,
average high school GPA was 3.14 ± 0.90, 52.6% were Pell Grant eligible, and on average
transferred in 4.2 ± 12.3 credits.
After PSM matched the participants with nonparticipants, an ANOVA was conducted to
determine if there was a difference in credits completed between participants in one program per
year (n=327), participants in two or more programs per year (n=54) and nonparticipants (n=381).
Participants in one program completed an average of 24.7 ± 13.8 credits per year. Participants in
two or more programs completed an average of 25.8 ± 14.2 credits per year. Nonparticipants
completed an average of 23.2 ± 13.7 credits per year (Table 4.1).
Table 4.1. Descriptives for First Generation Participants and First Generation
Nonparticipants: Credits Completed
TotalCredits

N
.00
1.00
2.00
Total
Model

381
327
54
762
Fixed Effects
Random Effects

Mean
Std. Deviation Std. Error
23.2047
13.69883
.70181
24.7370
13.80977
.76368
25.8148
14.22206
1.93538
24.0472
13.79388
.49970
13.78365
.49933
.68559

95% Confidence Interval for
Mean
Lower Bound Upper Bound Minimum Maximum
21.8248
24.5846
.00
59.00
23.2346
26.2394
.00
55.00
21.9329
29.6967
6.00
52.00
23.0663
25.0282
.00
59.00
23.0670
25.0275
21.0974
26.9971

BetweenComponent
Variance

.50254

	

Reviewing Table 4.2, it was determined that the variances of the three groups were not
significantly different since ρ = 0.555. Therefore, the homogeneity of variance assumption has
been met.

114	
	

	

Table 4.2. Test of Homogeneity of Variances
TotalCredits
Levene
Statistic
.588
	

df1
2

df2
759

Sig.
.555

Table 4.3 shows that there was not an overall significant difference between the groups
(ρ = 0.210), therefore, there was not a significant difference in credits completed between the
groups.
Table 4.3. ANOVA Results
TotalCredits

Between Groups
Within Groups
Total

Sum of
Squares
594.737
144201.562
144796.299

df
2
759
761

Mean Square
297.369
189.989

F
1.565

Sig.
.210

	

To determine if there were significant differences between the participation and nonparticipation
groups based on level of participation, a linear trend test was run. Reviewing Table 4.4., a
statistically significant difference between the groups was not found (ρ = 0.210).
Table 4.4. ANOVA Table for Linear Trend Analysis
ANOVA
TotalCredits

Between Groups

(Combined)
Linear Term Unweighted
Weighted
Deviation

Within Groups
Total

115	
	

	

Sum of
Squares
594.737
322.211
588.549
6.188
144201.562
144796.299

df
2
1
1
1
759
761

Mean
Square
F
Sig.
297.369 1.565 .210
322.211 1.696 .193
588.549 3.098 .079
6.188 .033 .857
189.989

Figure 4.4. shows the visual representation of the differences in credits completed between the
groups. There was not a significant difference in credits completed between the participant and
nonparticipant groups, although there was a trend of an increase in credits completed with an
increase in participation.
Figure 4.4. Means Plot for Linear Trend Analysis

SPECIFIC AIM #4: GPA (FIRST GENERATION-AMOUNT OF PARTICIPATION)
Prior to PSM, there were 4042 unique individuals in the sample. Of the 4042 individuals,
614 were campus recreation participants (treated) and 3428 were nonparticipants (controls).
After PSM, there were 588 participants matched with 588 nonparticipants. Twenty-six
participants could not be matched with nonparticipants, leaving 2840 unmatched nonparticipants.
None of the data were discarded. Table 4.5 shows the sample as a total, matched, unmatched,
and discarded data.
Table 4.5. Propensity Score Matching Sample Sizes

Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
3428
614 588
588 2840
26
0
0

	

116	
	

	

Figure 4.5 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows 26 unmatched participants. The next two
stratifications show the 588 matched participants and 588 matched nonparticipants. Finally, the
last stratification shows the 2840 unmatched nonparticipants. There was a great deal of overlap
in propensity scores (0.0-0.7) between the matched treatment group and the matched control
group. This is meaningful because this small range suggests that if two individuals were chosen,
one from each group, the main difference between individuals would be campus recreation
program participation. Since randomization is not possible for observational studies like this,
matching is used as a way to imitate randomization to create comparable groups to analyze.

Figure 4.5. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 4.6 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are

117	
	

	

matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.
Figure 4.6. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 4.7). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis. Additionally, a histogram of standardized differences was created to show the
data as overlaid kernel density estimates of standardized differences before and after matching
(Figure 4.8). The histogram at the top of the figure shows the standardized differences prior to
PSM. The histogram at the bottom of the figure shows the decreased standardized differences
after PSM. Like the standardized differences in means figure (Figure 4.6), the closer to 0.0 the
difference in the means are, the more exceptional the match.

118	
	

	

Figure 4.7. Histogram of Propensity Scores Before and After Matching

Figure 4.8. Histogram of Standardized Differences Before and After Matching

119	
	

	

Of the 1176 individuals within the sample, 50.2% were male, 49.8% were female. The
average age of individuals in the sample was 19.9 ± 1.8 years. Additionally, 58.7% lived on
campus, 74.3% were white, 45.5% Freshman status, 30.2% Sophomore status, 18.5% Junior
status, 5.9% Senior status, 85.1% started in 2011 or later, average ACT score was 20.7 ± 6.0,
average high school GPA was 3.10 ± 0.94, 52.6% were Pell Grant eligible, and on average
transferred in 3.5 ± 11.7 credits.
After PSM matched the participants with nonparticipants, an ANOVA was conducted to
determine if there was a difference in GPA between participants in one program per year
(n=510), participants in two or more programs per year (n=78) and nonparticipants (n=588).
Participants in one program had an average GPA of 3.03 ± 0.59. Participants in two or more
programs had an average GPA of 3.03 ± 0.62. Nonparticipants had an average GPA of 2.92 ±
0.69 (Table 4.6).
Table 4.6. Descriptives for First Generation Participants and First Generation
Nonparticipants: GPA
GPA

N
.00
1.00
2.00
Total
Model Fixed Effects
Random Effects

588
510
78
1176

Mean Std. Deviation Std. Error
2.9243
.68744
.02835
3.0260
.58554
.02593
3.0250
.62382
.07063
2.9751
.64241
.01873
.64095
.01869
.04290

95% Confidence Interval for
Mean
Lower Bound Upper Bound Minimum Maximum
2.8686
2.9800
.00
4.00
2.9750
3.0769
.00
4.00
2.8844
3.1657
1.40
4.00
2.9383
3.0118
.00
4.00
2.9384
3.0117
2.7905
3.1596

BetweenComponent
Variance

.00337

	

Reviewing Table 4.7, it was determined that the variances of the three groups were not
significantly different since ρ = 0.07. Therefore, the homogeneity of variance assumption has
been met.

120	
	

	

Table 4.7. Test of Homogeneity of Variances
GPA
Levene
Statistic
2.658
	

df1
2

df2
1173

Sig.
.070

Table 4.8 shows that there was an overall significant difference between the groups (ρ = 0.025).
Table 4.8. ANOVA Results
GPA

Between Groups
Within Groups
Total

Sum of
Squares
3.031
481.881
484.912

df

Mean Square
2
1.516
1173
.411
1175

F
3.689

Sig.
.025

	

Since there was an overall difference between the groups, and the homogeneity of variance
assumption was met, a Tukey Post Hoc Test was run to determine which specific groups
differed. When reviewing GPA of participants in one program and nonparticipants, it was
determined that there was a significant difference in GPA between the groups since ρ = 0.039.
There was not a significant difference in GPA between the participants in two or more programs
and nonparticipants (ρ = 0.393). There was not a significant difference in GPA between
participants in one program and participants in two or more programs (ρ = 1.000). Therefore,
significant difference was only found between the participants in one program per year and the
nonparticipants.

121	
	

	

Table 4.9. Post Hoc Tests
Multiple Comparisons
Dependent Variable: GPA
Tukey HSD

	

Mean
Difference (I(I) NumberofSports (J) NumberofSports
J)
Std. Error
.00
1.00
-.10166*
.03878
2.00
-.10074
.07724
1.00
.00
.10166*
.03878
2.00
.00091
.07793
2.00
.00
.10074
.07724
1.00
-.00091
.07793
*. The mean difference is significant at the 0.05 level.

95% Confidence Interval
Sig. Lower Bound Upper Bound
.024
-.1927
-.0106
.393
-.2820
.0805
.024
.0106
.1927
1.000
-.1820
.1838
.393
-.0805
.2820
1.000
-.1838
.1820

To determine if there were significant differences between the participation and nonparticipation
groups based on level of participation, a linear trend test was run. Reviewing Table 4.9.1., a
statistically significant difference between the groups was found (ρ = 0.025).

Table 4.9.1. ANOVA Table for Linear Trend Analysis
ANOVA
GPA
Sum of
Squares
Between
Groups

(Combined)

Mean
Square

df

F

Sig.

3.031

2

1.516

3.689

.025

.699

1

.699

1.701

.192

Weighted

2.561

1

2.561

6.233

.013

Deviation

.470

1

.470

1.145

.285

Within Groups

481.881

1173

.411

Total

484.912

1175

Linear Term

Unweighted

Figure 4.9. shows the visual representation of the differences in GPA between the groups. While
there was a significant difference in GPA between the participants in one program and
nonparticipants, participants in two or more programs did not have a significantly higher GPA

122	
	

	

than nonparticipants. There was also not a significant difference in GPA between the participants
in one program per year and participants in two or more programs per year.
Figure 4.9. Means Plot for Linear Trend Analysis

SPECIFIC AIM #5: ONE-YEAR & TWO-YEAR RETENTION (TOTAL SAMPLE FRESHMEN STATUS)
Prior to PSM there were 4757 unique individuals in the sample. Of the 4757 individuals,
1040 were campus recreation participants (treated) and 3717 were nonparticipants (controls).
After PSM, there were 1038 participants matched with 1038 nonparticipants. Two participants
could not be matched with nonparticipants, leaving 2679 unmatched nonparticipants. None of the
data were discarded. Table 5.0 shows the sample as a total, matched, unmatched, and discarded
data.

123	
	

	

Table 5.0. Propensity Score Matching Sample Sizes

All
Subsamples
(all cases)

Matched

Unmatched

Discarded

Control Treated Control Treated Control Treated Control Treated
3717 1040 1038 1038 2679
2
0
0

	
Figure 5.0 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows two unmatched participants. The next two
stratifications show the 1038 matched participants and 1038 matched nonparticipants. Finally,
the last stratification shows the 2679 unmatched nonparticipants. There was a great deal of
overlap in propensity scores (0.0-0.7) between the matched treatment group and the matched
control group. This is meaningful because this small range suggests that if two individuals were
chosen, one from each group, the main difference between individuals would be campus
recreation program participation. Since randomization is not possible for observational studies
like this, matching is used as a way to imitate randomization to create comparable groups to
analyze.
To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 5.1 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is essential when ensuring data are
matched as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the
more exceptional the match.

124	
	

	

Figure 5.0. Distribution of Propensity Scores

Figure 5.1. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 5.2). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
125	
	

	

of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis.
Figure 5.2. Histogram of Propensity Scores Before and After Matching

Additionally, a histogram of standardized differences was created to show the data as overlaid
kernel density estimates of standardized differences before and after matching (Figure 5.3). The
histogram at the top of the figure shows the standardized differences prior to PSM. The
histogram at the bottom of the figure shows the decreased standardized differences after PSM.
Like the standardized differences in means figure (Figure 5.1), the closer to 0.0 the difference in
the means are, the more exceptional the match.

126	
	

	

Figure 5.3. Histogram of Standardized Differences Before and After Matching

Of the 4757 individuals within the sample, 48.8% were male, 51.3% were female. The
average age of individuals in the sample was 18.7 ± 1.2 years. Additionally, 78.8% lived on
campus, 72.1% were white, 99.8% started in 2011 or later, average ACT score was 21.7 ± 5.0,
average high school GPA was 3.21 ± 0.72, 14.2% were Pell Grant eligible, 26.2% were first
generation status, and on average transferred in 4.2 ± 8.6 credits.
After PSM matched the participants with nonparticipants, two-sample tests for
proportions (z-test) (Table 5.1, Table 5.2) were conducted to determine if there was a difference
in one-year and two-year retention between the participants and matched nonparticipants.
Individuals who were retained were coded with a “1” while individuals who were not retained
127	
	

	

were coded with a “0.” Participants in club sports, intramural sports, and fitness classes had an
average one-year retention of 0.83 ± 0.38. Nonparticipants had an average one-year retention of
0.68 ± 0.46. The difference in one-year retention was significant (ρ < 0.001) suggesting
participants had a higher one-year retention rate when compared to the nonparticipants. Cohen’s
(1988) effect size value (d = 0.36) suggested a low practical significance.
Table 5.1. Two-Sample Test for Proportions (z-test) for Freshmen Participants and
Freshmen Nonparticipants: 1-Year Retention
z-Test:	Two	Sample	for	Means	
		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

	
Variable	1	
0.829479769	
0.14157948	
1038	
0	
7.929874969	
1.11022E-15	
1.644853627	
2.22045E-15	
1.959963985	 		

Variable	2	
0.682080925	
0.21705565	
1038	
	
	
	
	
	

Participants in club sports, intramural sports, and fitness classes had an average two-year
retention of 0.69 ± 0.46. Nonparticipants had an average two-year retention of 0.56 ± 0.50. The
difference in two-year retention was significant (ρ < 0.001) suggesting participants had a higher
two-year retention rate when compared to the nonparticipants. Cohen’s (1988) effect size value
(d = 0.27) suggested a low practical significance.

128	
	

	

Table 5.2. Two-Sample Test for Proportions (z-test) for Freshmen Participants and
Freshmen Nonparticipants 2-Year Retention
z-Test:	Two	Sample	for	Means	
		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

	
Variable	1	
Variable	2	
0.692678227	
0.562620424	
0.21308038	
0.24631598	
1038	
1038	
0	
	
6.182175317	
	
3.16121E-10	
	
1.644853627	
	
6.32242E-10	
	
1.959963985	 		

SPECIFIC AIM #6: ONE-YEAR & TWO-YEAR RETENTION (FIRST GENERATIONFRESHMEN STATUS)
Prior to PSM there were 1396 unique individuals in the sample. Of the 1396 individuals,
282 were campus recreation participants (treated) and 1114 were nonparticipants (controls).
After PSM, there were 279 participants matched with 279 nonparticipants. Three participants
could not be matched with nonparticipants, leaving 835 unmatched nonparticipants. None of the
data were discarded. Table 6.0 shows the sample as a total, matched, unmatched, and discarded
data.
Table 6.0. Propensity Score Matching Sample Sizes
Subsamples
(all cases)

All
Matched
Unmatched
Discarded
Control Treated Control Treated Control Treated Control Treated
1114
282
279
279
835
3
0
0

Figure 6.0 is a jitter plot distribution of propensity scores. Each circle represents each data point
in the total sample. The upper stratification shows three unmatched participants. The next two

129	
	

	

stratifications show the 279 matched participants and 279 matched nonparticipants. Finally, the
last stratification shows the 835 unmatched nonparticipants. There was a great deal of overlap in
propensity scores (0.0-0.7) between the matched treatment group and the matched control group.
This is meaningful because this small range suggests that if two individuals were chosen, one
from each group, the main difference between the individuals would be whether they participated
in campus recreation programs. Since randomization is not possible for observational studies like
this, matching is used as a way to imitate randomization to create comparable groups to analyze.
Figure 6.0. Distribution of Propensity Scores

To review the difference in the means of the data, PSM was used to show the covariate balance
in the dataset. Figure 6.1 shows the improvement in the covariate balance in the dataset after
PSM. The absolute standardized difference in the mean after PSM was less than the absolute
standardized difference in the mean prior to PSM, which is important to ensure data are matched
as precisely as possible. Ideally, the closer to 0.0 the difference in the means are, the more
exceptional the match.

130	
	

	

Figure 6.1. Standardized Difference in Means

To show how PSM matches participants to nonparticipants, a histogram of propensity scores
before and after matching was created (Figure 6.2). On the left hand side, the histograms show
the distribution of the data for the participants and nonparticipants prior to PSM. The distribution
of the data looks different between the participants and the nonparticipants. On the right hand
side, the histograms show the distribution of the data for the participants and nonparticipants
after PSM. The similarity of the distribution of the matched cases is critical prior to further
statistical analysis. Additionally, a histogram of standardized differences was created to show the
data as overlaid kernel density estimates of standardized differences before and after matching
(Figure 6.3). The histogram at the top of the figure shows the standardized differences prior to
PSM. The histogram at the bottom of the figure shows the decreased standardized differences
after PSM. Like the standardized differences in means figure (Figure 6.1), the closer to 0.0 the
difference in the means are, the more exceptional the match.

131	
	

	

Figure 6.2. Histogram of Propensity Scores Before and After Matching

Figure 6.3. Histogram of Standardized Differences Before and After Matching

Of the 1396 individuals within the sample, 48.9% were male, 51.1% were female. The
average age of individuals in the sample was 18.7 ± 0.8 years. Additionally, 80.5% lived on
campus, 69.7% were white, 99.8% started in 2011 or later, average ACT score was 20.9 ± 4.0,

132	
	

	

average high school GPA was 3.18 ± 0.59, 52.0% were Pell Grant eligible, and on average
transferred in 2.9 ± 7.3 credits.
After PSM matched the participants with nonparticipants, two-sample tests for
proportions (z-test) (Table 6.1, Table 6.2) were conducted to determine if there was a difference
in one-year and two-year retention between the participants and matched nonparticipants.
Individuals who were retained were coded with a “1” while individuals who were not retained
were coded with a “0.” Participants in club sports, intramural sports, and fitness classes had an
average one-year retention of 0.82 ± 0.38. Nonparticipants had an average one-year retention of
0.68 ± 0.47. The difference in one-year retention was significant (ρ < 0.001) suggesting
participants had a higher one-year retention rate when compared to the nonparticipants. Cohen’s
(1988) effect size value (d = 0.33) suggested a low practical significance.
Table 6.1. Two-Sample Test for Proportions (z-test) for First Generation Freshmen
Participants and First Generation Freshmen Nonparticipants: 1-Year Retention
z-Test:	Two	Sample	for	Means	
		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

	
Variable	1	
Variable	2	
0.82437276	
0.681003584	
0.14530311	
0.21801913	
279	
279	
0	
	
3.97293901	
	
3.54956E-05	
	
1.644853627	
	
7.09912E-05	
	
1.959963985	 		

Participants in club sports, intramural sports, and fitness classes had an average two-year
retention of 0.67 ± 0.47. Nonparticipants had an average two-year retention of 0.51 ± 0.50. The
difference in two-year retention was significant (ρ < 0.001) suggesting participants had a higher

133	
	

	

two-year retention rate when compared to the nonparticipants. Cohen’s (1988) effect size value
(d = 0.33) suggested a low practical significance.
Table 6.2. Two-Sample Test for Proportions (z-test) for First Generation Freshmen
Participants and First Generation Freshmen Nonparticipants: 2-Year Retention
	
z-Test:	Two	Sample	for	Means	
	

		
Mean	
Known	Variance	
Observations	
Hypothesized	Mean	Difference	
z	
P(Z<=z)	one-tail	
z	Critical	one-tail	
P(Z<=z)	two-tail	
z	Critical	two-tail	

Variable	1	
Variable	2	
0.666666667	
0.512544803	
0.22302158	
0.25074134	
279	
279	
0	
	
3.740122446	
	
9.19653E-05	
	
1.644853627	
	
0.000183931	
	
1.959963985	 		

OVERALL SUMMARY
The main findings of this study are summarized below. When comparing the total sample
of all participants and nonparticipants, participants of campus recreation programs completed
significantly more credits per year and obtained a significantly higher GPA per year. First
generation participants did not have a significantly different GPA than first generation
nonparticipants, but did complete significantly more credits when compared to first generation
nonparticipants. With respect to whether amount of participation in campus recreation programs
had an effect on credits completed and GPA, participants in one campus recreation program per
year completed significantly more credits than nonparticipants, although participants in two or
more programs per year did not complete significantly more credits than nonparticipants or the
participants in one campus recreation program per year. Participants in one campus recreation

134	
	

	

program and participants in two or more campus recreation programs both obtained a
significantly higher GPA when compared to nonparticipants, although there was not a significant
difference in GPA between participants in one campus recreation program per year and
participants in two or more campus recreation programs per year. There was not a significant
difference in credits completed between any of the first generation groups, although first
generation participants in one campus recreation program per year did obtain a significantly
higher GPA when compared to first generation nonparticipants. First generation participants in
two or more campus recreation programs per year did not have a significantly different GPA
when compared to first generation nonparticipants. There also was not a significant difference in
GPA between the first generation participants in one program per year versus first generation
participants in two or more programs per year. In addition to reviewing academic success
outcomes of all campus recreation participants, first generation campus recreation participants,
all nonparticipants, and first generation nonparticipants, freshmen students were also studied.
Retention was studied with this subgroup to determine if participation in campus recreation
programs was beneficial to persistence over time. When reviewing freshmen campus recreation
participants and freshmen nonparticipants, freshmen campus recreation participants had a
significantly higher one-year retention rate and two-year retention rate. The same result was seen
with first generation freshmen participants. First generation freshmen participants had a
significantly higher one-year retention rate and two-year retention rate when compared to first
generation nonparticipants.

135	
	

	

REFERENCES

136	
	

	

REFERENCES

Cohen, J. (1988). Statistical power analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Lawrence Earlbaum Associates.

137	
	

	

CHAPTER SEVEN: DISCUSSION

CAMPUS RECREATION AND ACADEMIC SUCCESS
The impact campus recreation has on student success has been studied for years, but
unfortunately the contribution campus recreation could make has been ignored due to lack of
substance and consistency of findings in the research (Belch et al., 2001). Belch et al. (2001)
suggests that for years recreational sport professionals have acknowledged the potential
relationship between campus recreation participation and learning, development, and persistence
in college, yet more research is needed for higher administration to be convinced of this
relationship. Furthermore, Banta (1991) and Little & Guse (1988) have suggested this lack of
substance and consistency in the literature have led individuals in higher education to
misunderstand and undervalue results of a relationship between campus recreation participation
and academic success. Considering these issues, it is necessary for continued research to be
conducted and consistency in findings to occur for higher administration to be convinced if a
relationship exists. Studying the academic related benefits associated with participation in
campus recreation programs is especially important considering academic success is a typical
area of assessment conducted by higher administration on a regular basis. There are several ways
to define academic success including students achieving a GPA that meets program and
university requirements, credits completed to fulfill program requirements, and retention to
ensure students are returning each year to continue with their education at their respective
universities. It is important to study each academic success factor to draw appropriate
conclusions. Kampf & Teske (2013) studied retention rates of students who participated in club
sports, used the student recreation center, were employed by the campus recreation department,

138	
	

	

and compared them to nonparticipants. In addition to retention rates, the researchers also
collected GPA data for each group. While club sport participants had higher retention rates than
nonparticipants, they did not differ from nonparticipants when it came to GPA. Kampf & Teske
(2013) suggested that while this particular program may contribute to better social integration
and lead to retention, it might not necessarily affect other academic success outcomes. Therefore,
it is important to study several academic outcomes to determine the relationship between
participation and each outcome. In the current study, credits completed and GPA were compared
in a total participant sample, first generation participants, and nonparticipants. One-year and twoyear retention rates were studied in freshmen and first generation freshmen recreational sport
participants and nonparticipants. Considering first generation students are more likely to struggle
academically and freshmen students are the least likely of all class standings to be retained, it is
of particular importance to study these specific populations that are known to earn lower GPAs,
complete fewer credits, and leave college before graduation (Pascarella et al., 2004; Tinto, 1993;
Ferguson, 1990). While it is necessary to continue research of total populations of college
students to add to the existing knowledge in the field, many universities enroll thousands of first
generation and freshmen students each year. Therefore, it is also worth studying whether first
generation status affects the relationship between campus recreation program participation and
academic success. Understanding the factors that could lead to success in populations that are
known to struggle academically could lead to a better approach when working with these
students.
CREDITS COMPLETED
In the current study, credits completed was evaluated in several ways, including in the
total participant sample, as a dose-response relationship with number of programs students

139	
	

	

participated in, and nonparticipants. The total participant sample completed significantly more
credits than the total nonparticipant sample. A positive finding between participation in campus
recreation programs and completing significantly more credits is encouraging for the progression
through academic programs leading to graduation. It is possible that participation in campus
recreation programs promotes the importance to overcome typical challenges associated with
college. For example, students in the total participant sample may have learned time
management skills that allowed them to participate in extracurricular activities while completing
more credits per year. Similar results were found when Danbert et al. (2014) studied the
relationship between campus recreation membership and credits completed. In their study,
campus recreation members had a higher cumulative credits completed when compared with
nonmembers. Consistency in results is encouraging when justifying the importance of campus
recreation programs.
Interesting results were found when determining if a dose-response relationship existed
for credits completed and participation in the current study. Specifically, participating in one
program per year was associated with more academic success than participating in two or more
programs, or no programs. When determining if more participation was related to more credits
completed, it is possible that students who participated in two or more campus recreation
programs did not believe they had the time to enroll in as many credits as students who
participated in only one campus recreation program per year. Students who participated in one
campus recreation program per year may have been able to balance a heavier course load
successfully as compared with students who spent more time participating in two or more
campus recreation programs per year. Therefore, our data suggest that moderate recreational
sports participation may be optimal in helping improve academic success in terms of credits

140	
	

	

completed. Reviewing the dose-response relationship of participation in campus recreation
programs and retention, Kampf & Teske (2013) found that the more a student used the campus
recreation facility, the greater correlation with retention (Kampf & Teske, 2013). However,
Kampf & Teske (2013) also suggested that the influence of participation on retention is a result
of social integration and may not necessarily affect other academic success outcomes (e.g.,
credits completed, GPA). Students who feel integrated within their campus communities tend to
remain in school, but may not always perform strongly in the classroom or choose to take as
many credits per year. The choice to enroll in fewer classes per year may be a result of the desire
to participate in more extracurricular programs. Consequently, the desire to participate in campus
recreation programs may be beneficial to some academic outcomes, but not all. Considering
Kampf & Teske (2013) studied the dose-response relationship between campus recreation
participation and retention rather than campus recreation and credits completed, it is not
appropriate to compare their results to the current study results. Furthermore, Kampf & Teske
(2013) included students as participants if they simply entered the campus recreation facility
whereas the researchers in the current study included participants in specific programs such as
club sports, intramural sports, or fitness classes within the analysis. Although the definition of
participation was different between the two studies, there are very few studies that focus on the
relationship between amount of campus recreation participation and credits completed. The
results of the current study suggest that for the total sample, participating in one campus
recreation program per year is most advantageous to student success as measured by successful
credit completion.

141	
	

	

GPA
GPA was also studied in the total participant sample, as a dose-response relationship with
number of programs students participated in, and nonparticipants. GPA in the total participant
sample was significantly higher than nonparticipant GPA. These results agree with what Danbert
et al. (2014) found when studying academic outcomes of freshmen campus recreation members.
Danbert et al. (2014) studied the freshmen cohort for four semesters to determine cumulative
GPA between the campus recreation members and nonmembers. Analysis of the data showed
that member cumulative GPA was significantly higher after four semesters when compared to
nonmember cumulative GPA. Results of the current study did not support a dose-response
relationship between amount of recreational sports participation and GPA. That is, students who
participated in any amount of campus recreation programs per year attained higher GPAs when
compared with nonparticipants, although there was not a significant difference in GPA between
participants in one campus recreation program per year and two or more campus recreation
programs per year. This finding makes sense when reviewing the results of the amount of
campus recreation participation and credits completed. Participants in one campus recreation
program completed significantly more credits per year than nonparticipants, but participants in
two or more campus recreation programs did not complete significantly more credits per year
when compared to the nonparticipants and other participant group. Therefore, it is possible that
participants in two or more campus recreation programs per year understood how many credits
they should enroll in per year and still be successful with respect to GPA. Our results differ from
those of Todd and associates (2009) who attempted to determine if a dose-response relationship
existed between campus recreation usage and GPA. These researchers noticed that students who
visited the campus recreation center three or more times per week had a significantly higher

142	
	

	

GPA than moderate users, low users, and nonparticipants. Differences in results could be
attributed to data collection methodology. The current study compared amount of campus
recreation programs students participated in per year, whereas the study by Todd et al. (2009)
focused on average number of visits students made to the campus recreation center via swipe
card system. Todd et al. (2009) also noted that intramural field sports and club sports data were
not included in the analysis because only limited data were available. Since the current study
specifically focused on intramural sports, club sports, and fitness class participation, a direct
comparison with the Todd et al. study is not possible. Brock, Carr, & Todd (2015) examined
campus recreation usage as a dose-response relationship with GPA. The researchers reviewed
usage in two different ways. The first approach to determine if more participation was related to
greater results was to review frequency of campus recreation participation as it related to GPA.
The researchers created high, moderate, and nonuser/low user groups to compare GPAs. The
high user group had a 5.0% higher GPA than the moderate user group and a 9.5% higher GPA
than the nonuser/low user group. The second approach to determine if more participation was
related to greater results was to determine if changes in campus recreation participation in
freshmen were related to changes in GPA. The researchers found that an increase in one day per
week of participating in campus recreation programs did not improve GPA. They also found that
there was a small (2.5%) yet not statistically significant increase in GPA with students who
increased their participation in campus recreation programs by two days per week. When
comparing the first approach by Brock et al. (2015) to the current study, some similarities exist.
We found that regardless of amount of participation, GPA was significantly higher in the
participant groups compared to the nonparticipant group. This finding was also seen in the Brock
et al (2015) study. The high user group and moderate user group had higher GPAs when

143	
	

	

compared with the nonuser/low user group. Differing methods in the second approach by Brock
et al. (2015) prove it difficult to make direct comparisons between the two studies, since we did
not compare “change” in the two variables. In contrast with the findings by Todd et al. (2009)
and Brock et al. (2015), a multi-institutional study conducted by NIRSA (2004) showed negative
results. NIRSA researchers found no difference in GPA between heavy recreational sport users,
light users, and nonusers at several universities. Major differences in the NIRSA study and the
current study methodologies exist. NIRSA relied on self-report as they distributed surveys to
students at universities across the United States to determine if there was a relationship between
campus recreation usage and GPA. Students were required to report their campus recreation
usage and academic variables such as GPA. In the current study, we collected campus recreation
participation directly from the campus recreation department and academic outcome data were
obtained from the university registrar. It is possible that students who completed the self-reported
survey by NIRSA may have found it difficult to recall campus recreation center usage or
overestimated their participation. Students were also required to input their academic
information. It is possible that students did not want to include poor GPA values or could have
estimated their GPA if they were unsure of their GPA at the moment they completed the survey.
Another difference between the NIRSA and present study was that NIRSA researchers collected
data from many institutions that differed in size and student demographics, whereas the current
study data were collected at a single institution. While it is important to collect data using a
multi-institutional approach, it is difficult to compare results of a multi-institutional study to a
study which data are only collected at one institution. Results from a single institution may be
best compared with other universities similar in size and student demographics. Finally, the
study by NIRSA (2004) differed in methodology in the fact that they surveyed students about

144	
	

	

their campus recreation center usage rather than participation in specific programs as evaluated
in the current study. Differences in GPA between participants in different campus recreation
programs may exist due to GPA requirements of campus recreation programs like club sports. At
most universities, using the campus recreation center does not require students to maintain a
certain GPA. As long as students are enrolled in classes or purchase a membership, they may use
the fitness center. This may be the reason the results of the study by NIRSA (2004) did not show
a difference in GPA across any participation level. On the contrary, since there are GPA
requirements for students who participate in club sports, their GPA must meet campus recreation
club sport guidelines, should they exist, and therefore may be higher than other participant
groups that are not required to maintain a certain GPA to participate.
ONE-YEAR AND TWO-YEAR RETENTION
In addition to reviewing credits completed and GPA, retention was also studied to
determine if participation in campus recreation was beneficial to retention over time. In the
current study, one-year and two-year retention in freshmen participants and nonparticipants was
analyzed. Freshmen participants in campus recreation programs were retained at a significantly
higher rate compared to their nonparticipant counterparts. These findings are encouraging
considering the highest attrition rates occur from freshmen to sophomore year (Tinto, 1993;
Ferguson, 1990). In the study by Danbert et al. (2014), more freshmen campus recreation
members were retained after one year and two years when compared with freshmen
nonmembers, although the only statistically significant finding was that members were retained
at a higher rate than nonmembers at the two-year retention mark. This finding is especially
important when thinking about the thousands of students who would return to the university the
following year if they were retained. Although the study by Danbert et al. (2014) was conducted

145	
	

	

at a larger institution than ours, similar methodology was used to analyze data and similar
outcomes were measured. In another study that focused on the retention of freshmen students,
Kampf & Teske (2013) collected census data from a large four-year college in the Midwest. In
addition to demographic data, the researchers were also provided GPA and Fall-to-Fall retention
data. Campus recreation participation data available for these analyses included participation in
club sports, campus recreation employment, and campus recreation center entry counts. Unlike
the current study, Kampf & Teske (2013) analyzed the relationship between each individual
program and Fall-to-Fall retention. Results showed that there was a significant relationship
between club sports participation and retention. After controlling for demographic and academic
variables, club sports participation was still significant in their statistical model. When
comparing club sports participants with nonparticipants, the authors determined that club sports
participants were 2.22 times more likely to be retained the following fall. Although the current
study was not designed to evaluate the effects of each individual campus recreation program,
club sports were included in the analysis as part of the participant data. Kampf & Teske (2013)
suggest that participation in campus recreation programs allow students to feel a sense of
belonging at their universities and therefore may add to the reasoning to stay. It is possible that
the students’ campus community involvement played a part in their decisions to continue the
following year at the university. Their enjoyment and/or sense of belonging could have
influenced their choice to stay. In contrast to the current study, Lindsey & Sessoms (2006)
surveyed their study sample and asked specifically whether students believed campus recreation
facilities or programs directly impacted their decision to attend or continue at a university. The
authors found that only junior and senior level students believed recreational sports facilities or
programs were important factors in deciding whether to attend college at all, and/or or continue

146	
	

	

at their universities. The current study did not include questions relating participation with
retention. Many times there are potential issues with the reliability of self-reported data.
Additionally, Lindsey & Sessoms (2006) mentioned the survey they administered has had issues
with reliability and validity since its implementation and future investigators need to develop
ways to increase the reliability and validity measures. Furthermore, just because students may
have agreed with certain survey questions, does not mean they agreed with others. For instance,
students may not have decided to choose a specific university because of their campus recreation
facilities and programs, but that may have played a large role in staying and vice versa. It is also
possible that freshmen and sophomore students who participated in the study by Lindsey &
Sessoms (2006) did not realize the impact campus recreation programs played in their early
careers as students. Furthermore, only one-year and two-year retention data of freshmen students
was analyzed in the current study. It is not known whether junior and senior level students were
retained at similar or higher rates.
A finding agreed upon generally by many higher education researchers is that students
who have a sense of community due to participation in programs on university campuses may be
retained at higher rates when compared to students who are not integrated into their campus
communities (Belch et al., 2001). Campus recreation departments can play an integral role in
offering students programs that might help them feel like they belong, which is likely to have
occurred in the current study.
FIRST GENERATION STUDENTS
According to the U.S. Department of Education and National Postsecondary Student Aid
Study, first generation students have certain characteristics that may disadvantage them for
successfully completing their degrees. For example, first generation students are more likely to

147	
	

	

be older, come from minority backgrounds, have some kind of disability, English may be their
second language, they could be immigrants, are more likely to be single parents, and are
financially independent (Bui, 2002). Pascarella et al. (2004) stated that first generation students
tend to enroll and earn fewer credits, are more likely to live off campus, work more hours,
participate in less out of the classroom activities, have fewer non-academic peer interactions, and
earn lower grades than their peers, therefore, participating in extracurricular activities can be
very beneficial. In the current study, participation in campus recreation by first generation
students proved to be beneficial as they completed more credits than their nonparticipating
counterparts. This is encouraging for the progression through academic programs leading to
graduation. It is possible that the first generation students in the current study developed the time
management skills needed to participate in extracurricular activities while also being enrolled in
several classes. If true, this result is significant considering Pike & Kuh (2005) found that first
generation students were less likely to be engaged in academic and social activities related to
success in college. It is possible that the negative effect first generation status is traditionally
thought to have on academic success was not as strong in the participants of the current study,
considering there were factors (like campus recreation programs) in these students’ lives that
encouraged success.
When determining if there was a dose-response relationship with participation in campus
recreation programs and credits completed, the total sample completed significantly more credits
when participating in one campus recreation program per year compared with nonparticipants.
First generation students in one campus recreation program and first generation students in two
or more campus recreation programs did not complete significantly more credits when compared
to nonparticipants. Separating the first generation students into participant groups based on

148	
	

	

amount of participation resulted in a different finding than that which was found in specific aim
two (e.g. first generation participants completed significantly more credits than first generation
nonparticipants). Although not statistically significant, it is interesting to note that as
participation increased, so did average number of credits completed (nonparticipants=23.2 ± 13.7
credits per year; participation in one campus recreation program per year=24.7 ± 13.8 credits per
year; participation in two or more campus recreation programs per year=25.8 ± 14.2 credits per
year). Although there was not a significant difference in credits completed between the
participant groups, it is possible that first generation participants learned how to balance
extracurricular activities with a similar course load as their nonparticipating counterparts.
Considering students tend to excel in their collegiate experience when they are involved within
the campus community, encouraging students to participate in extracurricular activities like
campus recreation programs may help the students who are known to struggle the most
academically.
GPA in the total participant sample was significantly higher than nonparticipant GPA,
although GPA was not significantly different in the first generation participants and first
generation nonparticipants. Specifically, first generation participants did not have a higher GPA
when compared with first generation nonparticipants. First generation participants in the current
study may not necessarily be completing courses with significantly better grades despite the fact
this subgroup completed significantly more credits per year when compared with
nonparticipants. Therefore, it is possible that the academic challenges the first generation
students faced in the current study did not relate to retention and completion of classes, rather
this population might have had issues with obtaining higher grades. It appears that this subgroup
performed less than average in their classes, but still passed their classes and obtained credit for

149	
	

	

completing them. Depending on program and university requirements, even a below average
grade may be a passing grade and count towards credits completed in students’ degree programs.
Pascarella et al. (2004) suggests that one of the academic outcomes first generation students
typically struggle with is achieving a high GPA; which agrees with our findings. Furthermore, it
is possible that since the first generation participants in this study were enrolled and completed
more credits per year, they struggled with achieving high grades in their courses due to available
study time being stretched too thin. In addition, it is possible that first generation students might
attempt to take more credits per year due to financial reasons. Some universities allow students
to pay a set tuition as long as they are completing 12-18 credit hours per semester. This might
encourage some first generation students to take the maximum amount of credits per semester at
the same cost, regardless of the outcome in grades obtained. However, students are required to
pay per credit hour at the current study institution. Therefore, it is more likely that first
generation students in the current study were involved in extracurricular activities and completed
more credits per year, yet did not achieve higher GPAs when compared with nonparticipants.
When reviewing amount of campus recreation participation and GPA in first generation
students, first generation student results differed from that seen with the total sample results. In
the total sample, participants in one campus recreation program and participants in two or more
campus recreation programs both obtained a significantly higher GPA when compared to
nonparticipants, although there was not a significant difference in GPA between participants in
one campus recreation program per year and participants in two or more campus recreation
programs per year. First generation students in one campus recreation program had a
significantly higher GPA when compared with first generation nonparticipants, although first
generation participants in two or more campus recreation programs per year did not have a

150	
	

	

significantly higher GPA compared to first generation nonparticipants. As previously stated, first
generation participants may be able to balance participation in extracurricular activities and
completing more credits, but the quality of their coursework may not be higher than
nonparticipants. Participating in two or more campus recreation programs per year and obtaining
a high GPA may not be reasonable if first generation students are also spending time on other
obligations. It is possible that if first generation students in this study were living off campus,
had a job, were caring for family members and then were also participating in two or more
campus recreation programs, they may not have devoted as much time to schoolwork, therefore
their GPAs would be affected accordingly. Furthermore, although first generation participants
may have completed more credits than their nonparticipating counterparts, they may not have
earned significantly better grades. Therefore, it is possible that with the challenges and
responsibilities faced by first generation students, there may be a recommended amount of
campus recreation participation for the most beneficial effect.
Participation in recreational sports may also benefit other academic outcomes such as
retention. According to the National Center for Education Statistics’ Beginning Postsecondary
Study, first generation students were four times more likely to leave their universities after the
first year when compared with non-first generation students (Engle & Tinto, 2008). In the current
study, freshmen campus recreation participants were studied to determine if they had a higher
one-year and two-year retention rate when compared with freshmen nonparticipants, and it turns
out that they did. Furthermore, when studying whether first generation freshmen participation
had the same effects on one-year and two-year retention, it was determined that first generation
freshmen were retained at a higher rate when compared to first generation freshmen who did not
participate in campus recreation programs. This finding is encouraging for campus recreation

151	
	

	

departments and higher administration because not only are freshmen the class with the highest
attrition rates, but first generation students have traditionally been a population with low
retention rates in postsecondary education. Early participation in campus recreation programs is
exciting considering Terenzini et al. (1994) found that first generation students tend to delay
participating in extracurricular activities on campus until they have determined how to integrate
these activities into their academic calendars, even though involvement in these types of
activities may prove to be beneficial in the long run. Tinto (1993) stated that college students’
activities and experiences have a more profound effect on academic success and persistence than
precollege experiences. In the case of the first generation participants in this study, it is possible
that experiences gained by students participating in campus recreation programs played a
significant role in their choice to stay. Garcia (2015) agreed with Tinto (1993) and suggested that
in order to succeed, first generation students should consider getting involved in co-curricular
activities on their campuses. The results of the current study provide similar evidence for campus
related activity participation, in the form of recreational sports. For the most beneficial academic
outcomes, it may be necessary for first generation students to find an ideal amount of
participation that allows for socialization benefits to occur while also providing time for
academic and other obligations.
FRESHMEN STUDENTS
According to Tinto (1993), more than half of all attrition at four-year institutions occurs
in the freshmen population. Ferguson (1990) predicts freshmen attrition to be as high as 60% of
the total attrition seen on college campuses. Many researchers have argued that attrition rates in
freshmen students have not changed since the 1970s (Levitz & Noel, 1989; Tinto, 1993).
Considering the financial implications of a substantial number of students leaving universities

152	
	

	

before their second year, university administrators should promote and support options available
on their college campuses to decrease such attrition.
According to a multi-institutional study by NIRSA (2010), the majority of campus
recreation participants are freshmen and sophomore students. The same finding occurred in the
current study with freshmen and sophomore students making up the majority of the samples in
research aims 1-4. The population of interest in research aims five and six were exclusively
freshmen to narrow focus on retention through the critical first two years of college. The current
study found that freshmen students who participated in campus recreation programs had
significantly higher one-year and two-year retention rates when compared to freshmen
nonparticipants. Focusing in on even a smaller subgroup (i.e., first generation freshmen); results
showed that first generation freshmen participants had significantly higher one-year and twoyear retention rates when compared to first generation freshmen nonparticipants. Tinto (1993)
and Garcia (2015) have suggested the importance of involvement in extracurricular activities on
campus. It is possible the freshmen participants and first generation freshmen participants in this
study enjoyed their experiences in campus recreation programs and gained a sense of belonging
due to their participation. Thus, retention could be at least partially affected by involvement in
extracurricular activities like club sports, intramural sports, and/or fitness classes. Support for the
role of campus recreation and freshmen retention was provided by Danbert et al. (2014).
Researchers collected student information from a university database and participation data from
the campus recreation department. Focusing on freshmen students, the researchers evaluated
campus recreation center membership and academic success measures. Danbert et al. (2014) and
researchers of the current study were interested in determining one-year and two-year retention
rates of freshmen campus recreation participants. The two studies found similar results

153	
	

	

suggesting freshmen participants had a higher retention rate than freshmen nonparticipants. The
main difference in results was that the study by Danbert et al. (2014) did not find one-year
retention between the members and nonmembers to be statistically different. The researchers
argued that the practical significance is worth mentioning considering the number of students
who were retained at the university was a high amount and would have impacted the university
financially if they were not retained. The financial issue is important as higher education
administrators review the implications of various retention rates. Other researchers found
different results. Lindsey & Sessoms (2006) found that only junior and senior level students
believed recreational sports facilities or programs were important factors when deciding to attend
college or continue at their universities. The major difference between the study by Lindsey &
Sessoms (2006) and the current study is the data collection method. Lindsey & Sessoms (2006)
surveyed their study participants and asked specifically whether students believed campus
recreation facilities or programs directly impacted their decision to attend or continue at a
university. The current study did not require participants to respond to a question requesting
perception about a direct relationship between participation and retention. Furthermore, the
current study found that freshmen participants and first generation freshmen participants were
retained at a higher rate (both one-year and two-year) when compared with freshmen
nonparticipants and first generation freshmen nonparticipants, respectively. Lindsey & Sessoms
(2006) surveyed students from each class level to determine their opinions about the impact of
campus recreation on retention. It is also important to note that Lindsey & Sessoms (2006)
suggest the reliability and validity of the survey instrument could have been an issue. Although
there are mixed results relating participation in campus recreation and retention for freshmen
students, it is important to acknowledge the programs that may encourage retention for this

154	
	

	

population. Considering this population holds the highest attrition rate at universities across the
nation, it is prudent for higher administration professionals to understand the value of campus
programs, including recreational sports, as avenues to positively impact and effect freshmen
students.
LIMITATIONS OF THE STUDY
Generalizability of the results of this study is limited to universities similar in size and
student population to the university used in this study. A multi-institutional approach is
suggested when attempting to generalize results, although this approach tends to be difficult
when data are collected in methods other than survey. Furthermore, when multi-institutional data
are collected, it is crucial to ensure data are collected properly with a method that can be used
across universities. When multi-institutional research is not possible, continued research
conducted at various universities differing in size and student population help add knowledge to
the field. Another limitation of the current study is the possible influence of additional and/or
unknown confounding factors that may have played a role in the effect of the outcome variables.
Although propensity score matching was used to compare participants to nonparticipants based
on many covariates that were collected (i.e., age, class standing, high school GPA, etc.), there
may have been additional factors related to recreational sports use that had an effect on academic
success. The researchers of the current study did not have access to student participation in other
on-campus or off-campus activities that might have played a role in their academic success (i.e.,
wellness program offerings, fundraising activities on campus, etc.). This study focused on
undergraduate student participation in intramural sports, club sports, and fitness classes.
Although there are other programs offered through campus recreation at the university (i.e.,
weight loss challenges, wilderness experiences, personal training, etc.), the data for all programs

155	
	

	

were not collected consistently over the collection time period. Furthermore, although data via
swipe card reader were available for potential fitness center usage, the investigators could not
specify whether students used the fitness center, weight room, walked to a classroom or event
within the recreation center, received treatment in the athletic training room, or met a faculty or
staff member in the recreation center. Having data for all campus recreation programs would
provide a more robust dataset representative of all programs provided by campus recreation.
Additionally, having data from all campus recreation programs would provide the opportunity
for researchers to compare academic success outcomes of individual programs. Researchers like
Kampf & Teske (2013) found differences in retention rates when comparing effects of individual
campus recreation programs. It may be relevant to study individual campus recreation programs
in the future if funding decided by higher administration limits offer-ability. Lower et al. (2013)
compared the benefits of participating in specific campus recreation programs (i.e., group fitness,
intramural sports, club sports). Instead of comparing the benefits of participating versus not
participating, the researchers compared the results of each campus recreation program along with
results of nonparticipants to determine the impact of specific programs. Lower et al. (2013)
stated that reviewing the effects of each campus recreation program might help administrators
analyze each program’s strengths and weaknesses and determine whether programs work
together to meet university objectives. Since college administrators determine program funding,
it may be beneficial to pinpoint programs that provide the best impact for students. From a
research perspective, analyzing the effects of each campus recreation program may help campus
recreation professionals and administrators understand their benefits, and hopefully, encourage
participation. The current study included data collected at a university that requires a minimum
GPA for participation in club sports. This requirement is typical for club sports offered at

156	
	

	

universities across the United States. Combining club sports participants with participants from
other campus recreation programs as one participant group might increase selection bias since
the user group could potentially have a higher GPA due to the GPA requirement. Analyzing each
campus recreation program and comparing academic success outcomes amongst the groups
would decrease the possibility of including a participant group that is required to maintain a
minimum GPA. Finally, reviewing the effects of each individual campus recreation program may
help understand the benefits each program has to offer. For instance, club sports participants
could potentially meet most days of the week if they are practicing, competing, and traveling to
compete. Therefore, the time commitment for club sports and other high time demand activities
may convince participants to take fewer credits in the semester(s) they participate. If credits
completed is an academic outcome being studied but certain campus recreation programs pose a
higher time commitment, participation could exert a negative influence on the academic
outcome. However, students who participate in club sports may enjoy their experiences so much
that they continue at the university for another year and are retained. Therefore, whenever
possible, it is important to study each individual campus recreation program to determine the
benefits each program may offer. Although some programs may not influence all academic
outcomes in a positive manner, they may still contribute to a successful college experience.
STRENGTHS
Although there were a number of limitations of the current study, there were many
strengths worth mentioning. Using propensity score matching as the first step in the analysis was
a factor that contributed to the strength of this study. The purpose of propensity score matching
is to decrease the selection bias in an observational study (Randolph, Falbe, Manual, Balloun,
2014). Olmos & Govindasamy (2015) suggest that propensity score matching can strengthen

157	
	

	

quasi-experimental and observational studies to decrease the number of potential confounders
that are included in the pathway of a treatment and outcome. Essentially, propensity score
matching decreases the possibility of other factors’ effects on the outcome (Olmos &
Govindasamy, 2015). Furthermore, it has been argued that propensity score matching is
beneficial because if specific cases cannot be matched, they may be dropped (Olmos &
Govindasamy, 2015). Ultimately, matching is an important analysis because it uses every data
point that can be matched so that fair comparisons can be made. A relatively large sample size
was another strength of this study. Considering data were collected over a couple of years,
thousands of data points were available to potentially be included in propensity score matching
and in the final dataset. Very few cases were dropped due to inability to match. Another strength
of the study was the way data were collected. Participation in intramural sports, club sports, and
fitness classes were collected electronically. Participation data were not collected based on
whether a student swiped an ID or completed a paper copy registration form. Rather, data for
intramural sports, club sports, and fitness classes were collected electronically allowing for
confidence in safe storage and easy access for analysis. The researchers did not need to rely on
survey data that would require participants to recall participation and academic outcomes from
previous years since the campus recreation department provided participation data and the
university registrar provided demographic and academic data. The number of academic
outcomes included in this study was another major strength. Since academic success is multifaceted with many possible causal exposures, studying several factors that may define academic
success was helpful. Finally, considering many variables could have played a role in the
academic success of this sample, it is encouraging that many positive results were found between

158	
	

	

the participation in campus recreation programs and academic success outcome variables, even
though individual effect sizes were not high.
RECOMMENDATIONS FOR FUTURE RESEARCH
Future research should include multi-institutional studies, employing similar
methodology, that review the academic outcomes of campus recreation participants at different
size and types of universities. Continued research in the area of campus recreation and student
academic success is needed to provide higher administration with more generalizable results.
Another recommendation for future research includes the possibility of collecting data
documenting other on-campus student participation. The current study included only campus
recreation program participation data, although many student affairs departments have access to
student participation in other on-campus programs. While this would call for a very complex
study, it could provide researchers the opportunity to use more covariates in the propensity score
matching analysis to decrease the possibility of other confounding variables that may have an
effect on academic success.
Another recommendation for future research would be to increase the data collection
period. In the current study, data were collected for only two years; therefore, data like retention
data, could only be analyzed through sophomore year. University administrators regularly review
retention from freshman through senior year, graduation in four years, and graduation in six
years. Engle & Tinto (2008) found that after following a cohort of students for six years, 43% of
the first generation student population left their institutions without earning degrees. Of the first
generation students who left sometime within the six years, 60% left after their first year. The
current study was successful in determining the retention of first generation students after their
first and second year of college, but the value of our study results could have improved if we had

159	
	

	

been able to follow these students throughout their college careers. Studying these students
longitudinally would not only determine retention from year to year, but would also determine if
students graduated within the six-year time frame that many universities encourage for their
undergraduate population. Additionally, data across all campus recreation programs should be
collected to determine the relationship between participation in individual programs and
academic success. Such data could help explain the role of specific programs on student success
in various academic areas. Finally, individual versus group exercise program effects should be
studied to determine the benefits each type of program has to offer. While participation in
general may influence academic success, the sense of belonging students may gain due to
participating in certain programs, such as group exercise or intramural team sports, is worth
studying.
CONCLUSIONS
The purpose of this study was to examine the relationship between campus recreation
participation and undergraduate academic success. We also performed further evaluation of first
generation students to determine if the same effects of participation occurred. It was found that
participants of campus recreation programs (total sample) completed significantly more credits
per year and obtained a significantly higher GPA per year. Amount of participation in campus
recreation programs had a positive effect on credits completed and GPA. Participants in one
campus recreation program per year completed significantly more credits than nonparticipants.
Participants in one campus recreation program and participants in two or more campus recreation
programs both obtained a significantly higher GPA when compared to nonparticipants. First
generation participants completed significantly more credits when compared with first generation
nonparticipants. Additionally, first generation participants in one campus recreation program per

160	
	

	

year had a significantly higher GPA when compared with first generation nonparticipants.
Freshmen campus recreation participants and first generation freshmen participants had a
significantly higher one-year and two-year retention rate when compared with freshmen
nonparticipants and first generation freshmen nonparticipants, respectively. The findings of this
study should be encouraging for students, campus recreation professionals and higher
administration. Students who are academically successful and are retained are needed for a
university to operate. Programs that provide positive relationships with such university academic
goals and objectives justify their need for support with results such as those displayed in this
study. Furthermore, it is suggested that students who feel a sense of social belonging due to
participating in campus recreation programs may be more likely to become involved in other
university activities and hence become more integrated into the university community (Miller,
2011). Students who are involved in several campus activities fit Tinto’s (1975) student
integration model that suggests the overall success of college students is due to their integration
into the campus community. Campus recreation has the possibility to not only influence
academic success, but also encourage a better sense of belonging and integration into the campus
community.

161	
	

	

REFERENCES

162	
	

	

REFERENCES

Banta, T.W., Bradley, J., & Bryant, J. (1991). Quality and importance of recreational services:
Technical manual and survey. Corvallis, OR: NIRSA.
Belch, H.A., Gebel, M., & Maas, G.M. (2001). Relationship between student recreation complex
use, academic performance, and persistence of first-time freshmen. Journal of Student
Affairs Research and Practice, 38(2), 254–268.
Brock, M, Carr, J. W., Todd, M. K. (2015). An examination of campus recreation usage,
academic performance, and selected health indices of college freshmen. Recreation
Sports Journal, 39, 27-36.
Bui, K. V. T. (2002). First-generation college students at a four-year university: Background
characteristics, reasons for pursuing higher education, and first-year experiences. College
Student Journal, 36, 3-11.
Danbert, S. J., Pivarnik, J. M., McNeil, R. N. & Washington, I. J. (2014). Academic success and
retention: The role of recreational sports fitness facilities. Recreational Sports Journal,
38, 14-22.
Engle, J. & Tinto, V. (2008). Moving beyond access: College success for low-income,
first generation students. Washington, DC: The Pell Institute.
Ferguson, M. (1990). The role of faculty in increasing student retention. College and University,
65 (2), 127–135.
Garcia, V. (2015). First generation college students: How co-curricular involvement can assist
with success.
Huang, Y. R. & Chang, S. M. (2004). Academic and co-curricular involvement: Their
relationship and the best combination for student growth. Journal of College Student
Development, 45(4), 391-406.
Inkelas, K. K., Daver, Z. E., Vogt, K. E. & Leonard, J. B. (2007). Living-learning programs and
first-generation college students’ academic and social transition to college. Research in
Higher Education, 48(4), 403-434.
Kampf, S. & Teske, E. J. (2013). Collegiate recreation participation and retention. Recreational
Sports Journal, 37, 85-96.
Levitz, R., & Noel, L. (1989). Connecting students to institutions: Keys to retention and success.
In M. L. Upcraft & J. N. Gardner (Eds.), The freshmen year experience (pp. 65–81). San
Francisco: Jossey-Bass.

163	
	

	

Lindsey, R. & Sessoms, E. (2006). Assessment of a campus recreation program on student
recruitment, retention, and frequency of participation across certain demographic
variables. Recreational Sports Journal, 30, 30-39.
Little, S. L., & Guse, D. (1988). Campus recreation services: An enterprise in higher education.
Journal of Physical Education, Recreation & Dance, 59(8), 62–63.
Lower, L. M., Turner, B. A. & Petersen, J. C. (2013). A comparative analysis of perceived
benefits of participation between recreational sport programs. Recreational Sports
Journal, 37, 66-83.
Miller, J. J. (2011). Impact of a university recreation center on social belonging and student
retention. Recreational Sports Journal, 35, 117-129.
National Center for Education Statistics. (2007). Digest of education statistics 2006.
Washington, DC: U.S. Department of Education.
National Intramural Recreational Sports Association. (2004). Value of Recreational Sports in
Higher Education. Corvallis, Oregon: NIRSA.
NIRSA/NASPA (2010). NIRSA/NASPA consortium campus recreation impact studydescriptive summary report. Retrieved from
http://www.nirsa.org/docs/Discover/Research/Campus_Rec_Impact_Study.PDF
Olmos, A. & Govindasamy, P. (2015). Propensity scores: A practical Introduction using
R. Journal of Multidisciplinary Evaluation, 11(25), 68-88.
Pascarella, E. T., Pierson, C. T., Wolniak, G. C. & Terenzini, P. T. (2004). First generation
college students: Additional evidence on college experiences and outcomes. The Journal
of Higher Education, 75, 249-284.
Pike, G. R., & Kuh, G. D. (2005). First- and second-generation college students: A comparison
of their engagement and intellectual development. The Journal of Higher Education, 76,
276–300.
Randolph, J. J., Falbe, K., Manual, A. K. & Balloun, J. L. (2014). A step-by-step guide to
propensity score matching in R. Practical Assessment Research and Evaluation, 19(18),
1-6.
Terenzini, P. T., Rendon, L. I., Upcraft, M. L., Millar, S. B., Allison, K. W., Gregg, P. L.,
Jalomo, R. (1994). The transition to college: Diverse students, diverse stories. Research
in Higher Education, 35(1), 57-73.
Tinto, V. (1975). Dropout from higher education: A theoretical synthesis of recent research.
Review of Educational Research, 45(1), 89-125.

164	
	

	

Tinto, V. (1993). Leaving college: Rethinking the causes and cures of student attrition
(2 Ed.). Chicago: University of Chicago Press.
nd

Todd, M. K., Czyszczon, G., Carr, J. W. & Pratt, C. (2009). Comparison of health and academic
indices between campus recreation facility users and nonusers. Recreational Sports
Journal, 33, 43-53.

165