Do Adaptive Representations of the Item-Position Effect in APM Improve Model Fit? A Simulation Study

Area of Science:

• Cognitive Psychology
• Psychometrics
• Educational Measurement

Background:

• The item-position effect, where an item's position influences response, is a known phenomenon in testing.
• Existing models for this effect are often inflexible, limiting their applicability.

Purpose of the Study:

• To evaluate if adapted representations offer a more appropriate model for the item-position effect compared to traditional ones.
• To investigate the item-position effect using simulated data mirroring Raven's Advanced Progressive Matrices (APM).

Main Methods:

• Confirmatory factor analysis models were employed to test various representations of the item-position effect.
• Simulated data, reflecting the covariance patterns of Raven's APM items, were utilized for the analyses.

Main Results:

• No significant improvement was observed when using adapted representations for modeling the item-position effect.
• The study found that existing, simpler models were as effective as adapted ones.

Conclusions:

• Adapted representations did not substantially enhance the analysis of the item-position effect in this context.
• Further research may be needed to explore other potential factors influencing the item-position effect or to refine modeling approaches.