Modelling multiple problem-solving strategies and strategy shift in cognitive diagnosis for growth

Area of Science:

• Educational Psychology
• Cognitive Science
• Psychometrics

Background:

• Problem-solving strategies are intentional actions, distinct from unintentional skills.
• Strategy-oriented instruction benefits low-achieving students more than skill-oriented instruction.
• Existing longitudinal cognitive diagnosis models (CDMs) do not account for strategy shifts.

Purpose of the Study:

• To propose a longitudinal CDM that models both multiple strategies and strategy shifts.
• To differentiate between strategy choice and skill implementation processes.
• To provide diagnostic information on both strategy choice and skill mastery.

Main Methods:

• Developed a novel longitudinal cognitive diagnosis model (CDM).
• Separated the strategy choice process from the skill implementation process.
• Conducted simulation studies for parameter recovery and model validation.
• Performed empirical data analysis to demonstrate model application.

Main Results:

• The proposed model effectively models between-person strategy variation and within-person strategy shifts.
• Simulation results indicate good parameter recovery for the new model.
• Ignoring strategy shifts can lead to inaccurate diagnostic information.
• Empirical analysis demonstrated the model's ability to measure strategy shift and skill growth.

Conclusions:

• The new longitudinal CDM offers a more comprehensive understanding of student learning.
• It enables accurate measurement of strategy shifts and skill development over time.
• This model can inform more personalized and effective educational interventions.