Abstract:

A Discriminate Multivariable Statistical Analysis was used to develop a research for studying factors that facilitate or inhibit student’s persistence and, based on these findings, to build a predicted model in a Higher Education Institution at Mexico. A total of 500 students of five different academic programs were included in the study. An eight steps methodology was used to carry out the study: 1) Problem definition, 2) Hypothesis definition, 3) Variable selection 4) Questionnaire design and validation, 5) Survey development, 6) Data processing, 7) Discriminate analysis performing, and 8) Conclusions and recommendations.
The starting diagnostic data for the problem definition shows up the retention rates for the programs under study as follows: 54% for Computer Systems Engineer, 63% for Electromechanical Engineer, 58% for IE, and 83% for Management program. A set of nine factors were included in the study.
A questionnaire with 43 questions was answered by the freshmen students in a Mathematics class. After the first semester we studied the students that were persistent at the next semester and those who drop out college. With these results we developed a discriminant analysis.
The results of the analysis reveals statistical significance for the following factors: GPR from previous level studies (High School), Positive motivation for academic success, Methods and techniques for studying, Qualified tutoring assistance, and Regular and on time attendance to classroom sessions. Finally, a probabilistic matrix was developed to predict the student retention with an 80% of confidence level.
This information is fundamental to improve the efficiency of the University in terms of students’ retention, students’ persistence, and programs for reinforcing skills and competences for better student’s performance.

Keywords:

discriminant multivariable analysis, student persistence, student retention.