Related Experiment Video
Updated: Nov 28, 2025
Bridging the Technology Divide in the COVID-19 Era: Using Virtual Outreach to Expose Middle and High School Students to Imaging Technology
Published on: September 28, 2022
Demographic and Admission Predictors of Students with Perceived Difficulty in Entry-Level Doctor of Physical Therapy
Amy E Heath1, Edward Mahoney, Addie Middleton
1Dep. of Physical Therapy, Western Michigan University, 1903 W. Michigan Ave., Kalamazoo, MI 49008, USA. Tel 269-387-8361. amy.heath@wmich.edu.
Background:
Admission committees have the difficult task of selecting candidates with the greatest likelihood of success for their programs and the profession. Because of limitations in defining the successful candidate, we attempted to predict who will become a "student with perceived difficulty" within a doctor of physical therapy (DPT) program using data available during the time of application.
Methods:
A retrospective analysis of 479 students from three entry-level DPT programs. The dependent variable was student with perceived difficulty status. Student characteristics were compared using unpaired t-tests (or non-parametric equivalent) and chi-squared tests. Receiver operating characteristic curves were constructed for variables significantly associated with student status to compare the predictive capabilities of the student characteristics and identify cutpoints that maximized sensitivity and specificity. We examined the predictive capabilities of clusters of characteristics that differed significantly between groups by calculating likelihood ratios and estimating odds ratios from logistic regression.
Results:
The cluster of characteristics that best identified students with perceived difficulty was prerequisite GPA <3.7, Analytical Writing GRE <4, and attended >2 undergraduate institutions. Twenty students met these criteria and 8 (40%) were identified as students with perceived difficulty. The positive likelihood ratio for this cluster of characteristics was 6.9 and the odds ratio was 8.7 (95% CI: 3.2, 23.0).
Conclusions:
These results suggest that this cluster of variables, available at the time of admission, can be used to identify students whose progress in the program may need to be more closely monitored and who may benefit from additional services to minimize difficulties for the student and faculty.
