Predictive Analytics for Student Retention: Group vs. Individual Behavior

Perhaps the most problematic element associated with individual prediction is that there can be little variation between individual student behaviors at any particular university, with the result that most students have a fairly similar predicted probability of success (return). [...]the problem is...

Full description

Saved in:
Bibliographic Details
Published in:College and university 2013-07, Vol.88 (4), p.63
Main Authors: Davis, Michael, Burgher, Karl E
Format: Article
Language:English
Subjects:
ACT assessment
Behavior
Class size
College admissions
College students
Grade Point Average
Higher education
School Holding Power
Secondary schools
Student Behavior
Studies
Success
Survival analysis
Variables
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Perhaps the most problematic element associated with individual prediction is that there can be little variation between individual student behaviors at any particular university, with the result that most students have a fairly similar predicted probability of success (return). [...]the problem is less that our overall predictions are not accurate than that we are not able to differentiate among individual students given available information. The biggest difference between the two samples was that the non-conditionally admitted students were less subject to selection bias. Because these students had only to attain a minimum high school gpa of 2.5 to be admitted, there was no concern that an admissions counselor chose to admit them on some unknown basis.
ISSN:0010-0889