Using Vision + Language Models to Predict Item Difficulty
Fuente:
arXiv
Saved in:
| Main Author: | |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911486234853376 |
|---|---|
| author | Khan, Samin |
| author_facet | Khan, Samin |
| contents | 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. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2603_04670 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Using Vision + Language Models to Predict Item Difficulty Khan, Samin Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition I.2.7; K.3.2 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. |
| title | Using Vision + Language Models to Predict Item Difficulty |
| topic | Artificial Intelligence Computation and Language Computer Vision and Pattern Recognition I.2.7; K.3.2 |
| url | https://arxiv.org/abs/2603.04670 |