How to develop an Interview Guide with ChatGPT (3 strategies)

Qualitative interview guides can be upgraded quickly by using ChatGPT in two targeted ways: generating large pools of candidate questions and steering those questions with psychological theory—especially approaches tied to cognitive interviewing and memory retrieval. The payoff is practical: researchers can brainstorm faster, then refine prompts into questions that help participants stay cognitively engaged and recall concrete experiences they might otherwise struggle to remember.

A common starting point is to let ChatGPT draft an entire guide from a broad study description—such as a hypothetical project on nurses’ experiences during the pandemic, focusing on barriers and how they were overcome. That “open” method can produce a usable structure and introduction, but it often lands on generic or mismatched details, and the output may vary from day to day. In the presenter’s experience, the generated guide can include categories that look plausible yet are hard to justify for the specific study focus, meaning researchers still need to supply their own logic for what matters and why.

Instead of relying on a single all-at-once draft, the recommended workflow begins with the study’s key questions—here, “What barriers did nurses face during the pandemic?” and “How did they manage to overcome these barriers?” Those core questions are written down first, then researchers brainstorm their own additional question ideas. ChatGPT is then used to expand that brainstorming: one question at a time, with a detailed prompt asking for many options that account for memory limitations and encourage multiple perspectives. This approach helps overcome the real-world constraint that researchers can only generate so many ideas and theme groupings before they start repeating themselves.

The second strategy uses theory to improve question quality. Rather than treating interview questions as purely creative prompts, ChatGPT can be asked to draw on psychological theories relevant to participant recall and engagement. The transcript highlights cognitive interviewing—an approach rooted in memory retrieval concepts and influenced by techniques used in police interrogations. The goal is to design questions that prompt detailed recollection, reducing the chance that participants omit key events simply because they cannot retrieve them.

Because theory suggestions can be inconsistent across runs, researchers are encouraged to treat ChatGPT as a direction-finder, not an authority. It may propose memory retrieval theory, resilience theory, or other cognitive frameworks; some will fit, others won’t. Researchers then select the most relevant theory and ask ChatGPT to generate question ideas grounded in that framework.

The end result is a more “digging deeper” interview guide: questions become more specific and experience-focused, such as prompts that ask participants to walk through a particular time window of a memorable shift. The transcript repeatedly stresses a boundary: ChatGPT should not replace the researcher’s judgment, dissertation writing, or full analysis. Used carefully, it can still save substantial time and help make interview guides more professional by aligning them with established theory and better recall-oriented questioning.