Determining final grade predictors

Determining Final Grade Predictors

Executive Summary

The purpose of this analysis is to attempt to predict the student achievement at the end of the semester, specifically, the final grade of a regular undergrad course.

The data in the model includes 202 observations. The initial model includes the following independent variables: influence of the teacher, the ACT score, the SAT score,the gender, the age, the major field of studies, the GPA, and whether the student has transferred from other school or not.

To predict student achievement more accurately, some other variables that could be tested would be: relationship with classmates, parents’ support, house environment, number of courses the student is taking, whether he works part time, affinity to the course subject,parents field of work, student’s field of work, quality and variety of books in which the course is based, among others.

In addition, an analysis of the prediction of the score of the final exam is included. For this analysis, the score of the final exam is predicted based on the previous grades the student has been receiving during the semester. The regression is presented in Annex B.Introduction

Schools have been always interested in determining factors related to student academic success. Examples of factors they use as predictors are recommendation letters, essays, previous GPA, SAT and ACT scores for prospective undergraduate students, GMAT and GRE for prospective graduate students, as well as TOEFL for international students. However, there are other variables which are equallyas important that could have a strong influence on the student’s academic success. Even more, maybe some of the variables that schools use in order to rate students are not really significant.

The purpose of this study is to accurately determine the effect of some of the independent variables that schools use in order to predict the academic success of their prospective students against someother variables that may have influence too. In this case, we only analyze the variables that are tested for prospective undergrad students and their influence on the final grade of the sample.

In addition to the common independent variables like ACT, SAT and GPA, we have identified several other variables that might affect the student achievement. One of the most important is the quality of theprofessor’s teaching ability Substantial knowledge and skills are critical for students achievement. Some studies have been made regarding this issue, and one researcher stated: “What the evidence suggests most strongly is that teacher quality matter and it should be a major focus of effort to upgrade the quality of schooling. Skilled teachers are the most critical of all schooling inputs.”(Ronald Ferguson in “Paying for Public Education: New evidence of how and why money matters.” Harvard Journal of Legislation, 28 (summer 1991), pp 465-498).

Several studies have been done about the influence of age on learning. Unfortunately, most of them are related only to child development, and there is no literature about age influence in scholastic achievement based on people over 16 years old.However, it is logical to assume that age is an important predictor of academic successl. Usually, people who are much younger or much older than the average student age find more difficulties in studying.
The table below shows the age distribution of the students of UCF for the fall 2002 period. This data is provided by the Office of Institutional Research.(http://pegasus.cc.ucf.edu/~irps/character/current.html)

|Age Distribution |
|Freshmen |18 |Graduate |31 |
|Sophomore |21 |Adv. Graduate |34 |
|Junior |23 | |...