Description

This Master Prompt would focus on the responsible integration of computer-assisted qualitative data analysis software (CAQDAS), computational tools, generative artificial intelligence and large language models into thematic-analysis workflows. It would guide researchers in determining which analytical activities may appropriately be supported by tools such as NVivo and emerging LLM systems, including data organization, code retrieval, preliminary code generation, code clustering, theme comparison, documentation and visualization.

A major emphasis would be placed on distinguishing analytical assistance from analytical authority. Researchers would critically evaluate AI-generated codes and themes for contextual accuracy, cultural sensitivity, emotional nuance, semantic fidelity, interpretive depth and consistency with the chosen qualitative methodology. Human researchers would retain responsibility for methodological decisions and final interpretations. This theme is strongly supported by the uploaded literature: one methodological investigation found that ChatGPT produced efficient and structured coding outputs but frequently missed emotional nuance, idiomatic expressions and culturally embedded meanings, requiring bilingual human review for semantic and contextual validity. Another study separated LLM-supported analysis into code extraction, code combination and theme generation and concluded that LLMs are most appropriately used as structured, human-supervised analytical assistants rather than replacements for qualitative researchers.