A Distribution-Free Approach for Comparing Growth of Knowledge

The longitudinal testing of student achievement requires the solution of several new problem areas. In this article, several small groups of medical students at the University of Limburg Medical School in Maastricht, The Netherlands, are compared with respect to their performances. The results indic...

Full description

Saved in:
Bibliographic Details
Published in:Journal of educational measurement 1994-03, Vol.31 (1), p.51-65
Main Authors: Tan, E. S., Imbos, Tj, Does, R. J. M. M.
Format: Article
Language:English
Subjects:
Academic Achievement
Analysis of variance
Change
Comparative Analysis
Educational research
Foreign Countries
Item Response Theory
Knowledge Level
Longitudinal Studies
Medical Education
Medical schools
Medical Students
Modeling
Netherlands
Nursing students
P values
Prediction
Proportions
Random allocation
Regression analysis
Statistical Distributions
Stochastic Analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The longitudinal testing of student achievement requires the solution of several new problem areas. In this article, several small groups of medical students at the University of Limburg Medical School in Maastricht, The Netherlands, are compared with respect to their performances. The results indicate, that, despite the possession of more knowledge at entrance, students with a low rate of growth of knowledge in the first year demonstrate a lower level of knowledge after the second academic year and continue to do so throughout the academic program when compared to students who show a higher rate of growth of knowledge in the first year. The analysis has been carried out using a distribution-free version of a longitudinal IRT-model suggested by Albers, Does, Imbos, and Janssen (1989). Furthermore, growth of knowledge has been described by means of a general regression model. Statistical inferences are possible using a randomization design extended to the situation where the observations are time-dependent proportions of correct answers.
ISSN:0022-0655
1745-3984
DOI:10.1111/j.1745-3984.1994.tb00434.x