Date of Award

Spring 2014

First Advisor

Dr. Thomas H. Short


Boys Hope Girls Hope (BHGH) is a nonprofit organization that strives to help students in middle school and high school build character and reach their full potential in the classroom. This organization accepts 18 students into their residential program who are bright, motivated, and who come from challenging household situations and provides them with an education and living accommodations throughout their high school career. BHGH created a new program in 2007 called Hope Prep that allowed more students to participate, and therefore created a need for a more efficient way to sort through the larger number of applicants. Since BHGH prides itself on developing students into young adults with good character, I have used a character development survey to act as the dependent variable that will allow BHGH to determine which students to accept into the program.

I have chosen a Likert-type scale survey to measure character development since I will be able to associate a numerical value with the results, which will represent an overall score, and allow an easy comparison of the Hope Prep program student applicants’ characters. In order to create a model that will provide BHGH a more efficient way to determine which applicants could benefit the most from their program, I have created a hypothetical prediction model based on simulated data. This model can be used to predict how much the students’ characters will develop if accepted into the program. I was able to combine two data sets that were provided to me by BHGH to create a file that contains background information about some Hope Prep program applicants as well as current Hope Prep members. The background information that I have for the students includes gender, school, region, grade, Math OGT scores, Reading OGT scores, and BHGH test scores for various disciplines. I had to impute many of the OGT test scores in order to create a more complete data set for my prediction model. After imputing the test scores, I created random survey test scores to represent the possible character development of each of the students from the beginning of the program to the end. I then used a linear regression model to create an equation that would use the background information and a pre-program survey test score to predict a post-program test score. Although this prediction model cannot be used by BHGH in its current state due to the heavy reliance on simulated data, this model shows BHGH what they could do with an infusion of real data.

Included in

Psychology Commons