Understanding Training Load Management through Contextual Experience Sampling Method: An Exploratory Study

Effective training load management is essential for optimizing athletic performance and preventing injuries. While experience sampling method (ESM) and large language models (LLMs) offer new opportunities for enhancing the accuracy and depth of subjective training load reporting, the design of LLMs-generated contextual ESM prompts remains underexplored. This study investigates university badminton athletes’ perceptions of different LLMs-generated contextual ESM prompt styles, including simple quantitative, data-guided, and empathetic conversational format, to collect subjective training load data. Through semi-structured interviews conducted immediately after training sessions, participants evaluated the usefulness and acceptability of each type of prompt. Findings indicate a strong preference for data-guided prompts that integrate objective training metrics with supportive questions. Based on the outcomes of the exploratory study, we also propose a set of principles for designing high-quality contextual ESM prompts to support more effective and reliable self-reporting of training load.