Using Vision + Language Models to Predict Item Difficulty
This work offers a method for automated psychometric analysis and item development, which could benefit educators and test designers by streamlining the creation of data visualization literacy assessments.
This project explored using GPT-4.1-nano to predict the difficulty of data visualization literacy test items for U.S. adults. The multimodal approach, combining visual and text features, achieved the lowest mean absolute error (MAE) of 0.224, outperforming vision-only (0.282) and text-only (0.338) methods.
This project investigates the capabilities of large language models (LLMs) to determine the difficulty of data visualization literacy test items. We explore whether features derived from item text (question and answer options), the visualization image, or a combination of both can predict item difficulty (proportion of correct responses) for U.S. adults. We use GPT-4.1-nano to analyze items and generate predictions based on these distinct feature sets. The multimodal approach, using both visual and text features, yields the lowest mean absolute error (MAE) (0.224), outperforming the unimodal vision-only (0.282) and text-only (0.338) approaches. The best-performing multimodal model was applied to a held-out test set for external evaluation and achieved a mean squared error of 0.10805, demonstrating the potential of LLMs for psychometric analysis and automated item development.