Resolving mixtures of strategies in spatial visualization tasks.

Area of Science:

• Cognitive psychology
• Educational measurement
• Psychometrics

Background:

• Traditional test theory models, rooted in trait psychology, inadequately represent complex cognitive structures and problem-solving strategies.
• Understanding individual differences in performance and learning necessitates incorporating knowledge structures and strategy use.

Purpose of the Study:

• To propose and illustrate a method for integrating qualitative distinctions in cognitive strategies into quantitative test theory models.
• To demonstrate how mixed test models can capture individual differences in strategy use for spatial visualization tasks.

Main Methods:

• Utilized response latency data from spatial visualization tasks solvable by distinct strategies (mental rotation vs. rule-based).
• Applied a mixture modeling approach within test theory to differentiate individuals based on their likely strategy use.
• Assumed consistent strategy use within individuals across tasks, with discussion on strategy-switching extensions.

Main Results:

• Demonstrated that response latencies can differentiate between mental rotation and rule-based strategies.
• Showcased the feasibility of using mixture models to represent heterogeneous strategy use within a test population.
• Provided empirical illustration of the proposed modeling approach.

Conclusions:

• Mixed test theory models offer a viable framework for incorporating cognitive strategy information into measurement.
• This approach enhances the understanding of individual differences in cognitive performance and learning beyond traditional trait-based models.
• Future research can extend this to model dynamic strategy-switching behavior.