Analysis code and scoring data for: Academic Performance of Large Language Models Across an Undergraduate Physiotherapy Curriculum: A Subset Analysis

This repository contains the analysis code and supporting data for the study "Academic Performance of Large Language Models Across an Undergraduate Physiotherapy Curriculum: A Subset Analysis" (Acar, Öztürk, Arslan, & Alaca; Scientific Reports, under revision). Contents:- scoring_data.csv: Course-level scoring data for 34 undergraduate physiotherapy courses, with midterm, final, and total examination scores for GPT-5.0 and Claude 4.5 Sonnet, course domain classification (Basic/Clinical), and examination item-format categorization.- bootstrap_spearman_CI.R: R script implementing percentile bootstrap (5,000 resamples; seed = 2026) for 95% confidence intervals around Spearman rank correlations between midterm and final examination scores. Reproduces Table 4 of the manuscript.- README.md: Full documentation of repository contents, column descriptions, reproducibility instructions, and methodological notes. Live model session outputs were not retained, as both platforms' temporary/incognito session modes preclude persistent storage. The scoring records provided here constitute the primary record of the evaluation and are sufficient to reproduce all inferential analyses reported in the manuscript.