A STATISTICAL ANALYSIS OF STUDENT ATHLETES AT STETSON UNIVERSITY
By
APRIL COATES
A SENIOR RESEARCH PAPER PRESENTED TO THE DEPARTMENT OF MATHEMATICS AND COMPUTER SCIENCE OF STETSON UNIVERSITY IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF BACHELOR OF SCIENCE
STETSON UNIVERSITY
2005
ACKNOWLEDGMENTS
I would like to start of first by acknowledging Dr. Erich Friedman. If not for him, I would not have a project to present. Thank you for allowing me to investigate what started as just a casual lunch conversation. I would also like to thank Dr. John Tichenor and Patti Sanders for all of their assistance in the collection of data. I would like to give special acknowledgement to Dr. Will Miles for his assistance throughout the semester. Thank you for your continuous support through computer crashes and frantic mental breakdowns. To my mom and brother, thank you for your constant love and support. Your confidence in me is what got me through this project. Lastly, I would like to say thank you to my second family—all of the professors of the Stetson University Math Department. Thank you all for pushing me beyond what I thought were my limits.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ----------------------------------------------------------------------------
|
2
|
LIST OF TABLES ---------------------------------------------------------------------------------------
|
5
|
LIST OF FIGURES -------------------------------------------------------------------------------------
|
6
|
ABSTRACT ----------------------------------------------------------------------------------------------
|
7
|
CHAPTERS
|
|
BACKGROUND -----------------------------------------------------------------------------------
Stetson University Athletic Department Mission Statement----------------------------
Financial Eligibility---------------------------------------------------------------------------
1.3. Data Collection---------------------------------------------------------------------------------
|
8
8
8
9
|
REGRESSION MODELS-------------------------------------------------------------------------
Two Variable Regression--------------------------------------------------------------------
Principle of Least Squares --------------------------------------------------------
Variable Interaction-----------------------------------------------------------------
Residual Analysis-------------------------------------------------------------------
Multiple Regression--------------------------------------------------------------------------
2.2.1. General Additive Multiple Regression Model ---------------------------------
2.2.2. First-Order Model -----------------------------------------------------------------
2.2.3. Second-Order No-Interaction Model -------------------------------------------
2.2.4. First-Order Predictors and Interaction ------------------------------------------
2.2.5. Complete Second-Order Model -------------------------------------------------
2.3. Categorical Variables----------------------------------------------------------------------
2.3.1. Dichotomous Variables ----------------------------------------------------------
2.3.2. Multi-Category Variables -------------------------------------------------------
|
12
12
13
14
15
1718
18
19
19
20
21
2123
|
DISTRIBUTIONS ----------------------------------------------------------------------------------
Normal Distributions -------------------------------------------------------------------------
Determining Underlying Distributions ----------------------------------------------------
Hypothesis Testing and Significance Level -------------------------------------
Chi-Square Distribution ------------------------------------------------------------
Goodness-of-Fit Test ---------------------------------------------------------------
|
25
25
27
28
29
30
|
PRELIMINARY ANAYSIS ----------------------------------------------------------------------
Linear Regression ----------------------------------------------------------------------------
Residual Plots ---------------------------------------------------------------------------------
Goodness-of-Fit Test-------------------------------------------------------------------------
|
31
31
32
33
|
CONCLUSIONS -----------------------------------------------------------------------------------
|
36
|
REFERENCES ------------------------------------------------------------------------------------------
|
37
|
BIOGRAPHICAL SKETCH --------------------------------------------------------------------------
|
38
|
LIST OF TABLES
TABLE
Cost of Attendance---------------------------------------------------------------------------------
Hypothesis Testing for One Proportion --------------------------------------------------------
Hypothesis Testing for Two Proportions -------------------------------------------------------
Frequency Table -----------------------------------------------------------------------------------
5. Expected Values -----------------------------------------------------------------------------------
|
10
28
29
33
34
|
LIST OF FIGURES
FIGURE
Sample of Linear Model----------------------------------------------------------------------------
Residual Plot-----------------------------------------------------------------------------------------
First-Order Model ----------------------------------------------------------------------------------
Second-Order No-Interaction Model ------------------------------------------------------------
First-Order Predictors and Interaction Model --------------------------------------------------
Complete-Second Order Model ------------------------------------------------------------------
Categorical (no interaction) -----------------------------------------------------------------------
Categorical (interaction) ---------------------------------------------------------------------------
Normal Distributions -------------------------------------------------------------------------------
Standard Normal Distributions -------------------------------------------------------------------
Normal Curve Probabilities -----------------------------------------------------------------------
Chi-Squared Distribution --------------------------------------------------------------------------
13. Baseball Percentage vs. GPA --------------------------------------------------------------------
14. Baseball Regression Models ---------------------------------------------------------------------
15. Baseball Residual Plots ---------------------------------------------------------------------------
|
12
16
18
19
20
21
22
23
26
26
27
29
31
32
32
|
ABSTRACT
A STATISTICAL ANALYSIS OF STUDENT ATHLETES AT STETSON UNIVERSITY
By
April Coates
May 2006
Advisor: Dr. Will Miles
Department: Mathematics and Computer Science
Student-athletes have very different roles in the eyes of society. There are those in society which feel athletics are top priority while academics come second. On the other hand, there are others who believe athletes are students first and athletics are merely an extracurricular activity. The Stetson University Athletic Department strives for excellence in the classroom and on the playing field. However, are these student-athletes succeeding in the classroom? I approached the registrar, Dr. John Tichenor, to see what data would be available for the analysis on student-athletes over the past seven years. Prior to receiving the records, I began research on various methods needed to analyze the data—two variable regression, multivariable regression, residuals, categorical variables, and the goodness-of-fit test.
Does the amount of athletic scholarship granted to an athlete correlate with academic performance? Do athletes tend to shy away from the more demanding majors? Are there sports that place a higher standard on academics? Are males and females really on the same playing field? I hope to answer these and other questions that arise next semester when I can analyze the complete data set.
CHAPTER 1
BACKGROUND
Stetson University has been housing Division One athletes since it joined the Atlantic Sun Conference in 1985 [7]. Currently, Stetson students may participate in nine varsity sports: basketball, volleyball, crew, golf, tennis, soccer, cross country, baseball and softball. The NCAA maintains that every athlete be held to specific standards in order to remain eligible. For example, student-athletes must be full time students (twelve credit hours) and must maintain a minimum of a 2.00 grade point average [7].
STETSON UNIVERSITY ATHLETIC DEPARTMENT MISSION STATEMENT
Stetson University strives for individual student-athletes to achieve excellence in both the classroom and on the playing field. The university places a lot of emphasis on the student-athletes having a well-rounded college experience. Below summarizes Stetson University Athletic Department’s ideal college experience:
The Stetson University Athletic Department strives to provide students with a sound educational experience through a holistic and collaborative athletic program that allows students to develop intellectually, spiritually, socially, and physically. Excellence is pursued through participation in a successful Division I, NCAA program, superior coaching, interaction among coaches, faculty, students, and staff, and a diversity of student-athlete activities based on a liberal-arts education. Students develop leadership through sport participation and community activities. In unison with the University Mission, the Athletic program helps students pursue truth by actively recruiting and providing a diverse and caring environment that values and commits to the rights and fair treatment of all people regardless of race, religion, or gender. The Athletic Program (Department) encourages its student athletes to be morally sensitive and contributing citizens through active forms of social responsibility [7].
1.2. FINANCIAL ELIGILIBILTY
Just like at other colleges and universities, student-athletes at Stetson University may be awarded scholarships for their athletic ability. These scholarships can range from a small stipend to aid covering tuition, room, board and books. However, these scholarships and grants, are only awarded for a one year period. Upon the completion of an academic year, each athlete’s athletic abilities and eligibility will be considered when deciding whether or not the grants will be renewed. Stetson University has the right to withdrawal the student-athlete’s scholarship if he or she has willingly withdrawn from the sport, is no longer eligible to compete, or is guilty of a serious misconduct [7].
1.3. DATA COLLECTION
In order for the results of this project to have a statistically significant result, the collection of data should be as large as possible. Thus, the data collected spans over the past seven years, from 1997 to 2004, and includes over five hundred student-athletes. Included in the data set are variables such as, sport, sex, race, home state, SAT scores, declared major, Stetson University cumulative grade point average as of the end of the fall of his or her sophomore year, Stetson University athletic scholarships and non-athletic Stetson University scholarships granted.
When the data was collected, there were a few concerns that were raised. While we wanted the data set to be as large as possible, we could not include all athletes over the time span available like originally planned. If this was to happen, student-athletes would more than likely be included in the data set more than once. For instance, suppose there is a student who played soccer for each of the four years he attended Stetson University. This particular soccer player would be included in the data set four times. In order to eliminate counting student-athletes more than once, it was determined to only use one class each academic year.
The original thought was to use all freshmen student-athletes; however, variables such as grade point average, would have not been an accurate estimate of academic performance after only one semester. Many freshmen are overwhelmed from the adjustment to college life and academic performance is not usually at its best. Therefore, the next thought was to include sophomores. As a sophomore, students have had time to adjust to their new environments and by this time, many students have declared their major. However, the students are still at a point where their grade point average is not heavily affected by their major.
Another concern that was raised was in regards to the amounts of athletic scholarship granted to the student-athletes. Since the data extended over the past seven years, it was essential to incorporate the changes in tuition, room and board, books, as well as other fees not included in tuition. Looking at the table below, one can see the drastic increases over the seven years this project includes [8].
-
Year
|
Cost of Attending Stetson University
|
1997-1998
|
$21,220
|
1998-1999
|
$23,180
|
1999-2000
|
$24,140
|
2000-2001
|
$25,255
|
2001-2002
|
$26,440
|
2002-2003
|
$27,925
|
2003-2004
|
$29,685
|
Table 1. Total Cost of Attendance
For example, if a woman playing basketball received an athletic scholarship worth $12,000 in 1998 was compared to one of her teammates who also received an athletic scholarship for $12,000 when she played three years later, on the surface, it appears that both are women are receiving the same amount of aid. However, by examining the chart, in 1998 one can see that it costs approximately $23,180 to attend Stetson University for one year, compared to nearly $26,440 in 2001. Thus, in all actuality, the scholarship granted in 1998 was worth more than the one awarded three years later in 2001. Therefore, in order to integrate the increase in cost over the past seven years, rather than simply looking at the amount of athletic scholarship granted, the focus will be on the percentage of total cost—tuition, room and board, books and other fees—covered by the athletic scholarship granted.
CHAPTER 2
REGRESSION MODELS
When examining data, it is often essential to determine the relation or correlation between variables. For instance, in a simple case where two variables are being compared, one of the variables, x, is independent and known in advance, while the other, y, is dependent on x. While it is impossible to predict the actual value of the dependent variable, it is possible to create a model that can estimate the expected value, called E(Y) [3]. In the simple case, often times, a linear regression model can be used. However, in a more complicated data set, the model may be exponential, power, logistic, or reciprocal [2].
TWO VARIABLE REGRESSION
As previously mentioned, the simple case of regression is one comparing two variables. In order to estimate E(Y), the predicted value of y, take a random sample of n points, . So, E(Y) depends on the value of x, the independent variable. Thus, E(Y) =. In other words, the predicted value of y, let’s call it Y, can be written as a function of x. Assuming that the E(Y) is linear yields,, a regression curve. Begin by fitting a straight line through the sample points [3]. An example of a linear regression model is shown in Figure 1.
Share with your friends: |