Title: Models for Studying Strategic Test-Taking Behavior?

Author: Niels Smits

Abstract:
In the Netherlands new forms of testing such as compensatory and progammatic
testing (in both not all individual exams need be passed to receive a diploma)
have become popular, which may have detrimental consequences such as strategic
learning behavior and, as a result, hiatuses in knowledge and ability. To
detect such effects the standard test-theoretical approaches and linear models
are unfit [1]. Here, I consider several alternative approaches, such as Latent
class/profile analysis [2], Mixture regression analysis [3], Model-based
recursive partitioning [4] and compare them in exam data from courses of my
own department where these effects are expected due to the setup of the rules
for passing/failing.

1. Dorresteijn, C. van, Kan, K. J., & Smits, N. (2023). Absence of evidence is
not evidence of absence: On the limited use of regression discontinuity
analysis in higher education. Assessment & Evaluation in Higher Education,
48(1), 16–26. https://doi.org/10.1080/02602938.2021.2016606

2. Yocarini, I. E., Bouwmeester, S., Smeets, G., & Arends, L. R. (2020).
Allowing course compensation in higher education: A latent class regression
analysis to evaluate performance on a follow-up course. Assessment &
Evaluation in Higher Education, 45(5), 728–740.
https://doi.org/10.1080/02602938.2019.1693494

3. Grün, B., & Leisch, F. (2008). FlexMix version 2: Finite mixtures with
concomitant variables and varying and constant parameters. Journal of
Statistical Software, 28(4), 1–35. https://doi.org/10.18637/jss.v028.i04

4. Zeileis, A., Hothorn, T., & Hornik, K. (2008). Model-based recursive
partitioning. Journal of Computational and Graphical Statistics, 17(2),
492–514. https://doi.org/10.1198/10618600SX319331