Explore Testing Performance and Learning Behaviors

ABSTRACT

This study explores various approaches to investigate participants’ testing performance and learning behaviors in a computer-based spatial rotation learning program. Using multivariate learning and assessment data, including responses, response times, learning times and selected covariates, a comprehensive data analytic framework is developed that not only utilizes the test level information but also the item level information. This top-down and multivariate data analytic framework can shed light on conducting exploratory analysis with high-dimensional and mixed-type multivariate data, especially on how to aggregate information from the test-level and item-level. The findings about participants’ testing performance and learning behaviors are valuable in guiding the design of an adaptive learning platform in the future and can also provide some support in developing confirmatory statistical methods to model testing and learning behaviors.

KEYWORDS:

• Clustering analysis
• multicategory logit models
• mixted-type data
• response times
• learning times
• learning behaviors

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. The normality assumption of the above two paired t-test is violated, and thus, a generalized Yuen s robust test using trimmed mean is used in R package WRS2 (Mair & Wilcox, 2018). Similar for the two paired t-test at TC2.