Explore·Fieldwork·Augmentation·Emerging·EXP-008
Diary Study Adaptive Prompting
Value hypothesis
Fixed prompt schedules treat every participant identically. Adaptive prompting builds longitudinal data from each participant's actual narrative, following threads the common schedule would have missed.
Quality · Insight
Diary studies involve participants logging experiences, behaviours, or feelings over time in response to researcher prompts. In this adaptive version, an LLM analyses each participant's prior entries and proposes personalised prompts tailored to their emerging patterns - identifying threads worth pursuing, following an emerging trend, and adjusting cadence if engagement wanes. The researcher reviews proposed adaptations before deployment, maintaining methodological control while allowing the study to respond dynamically to what participants actually report.
Risks in application
Bias Bleed
AI adaptations inject researcher-like decisions throughout the study at scale; those decisions carry the AI's biases and are harder to audit retrospectively than a static prompt set. The resulting dataset reflects those choices without making them visible.
Shallow Solutions
Adapted prompts may appear contextually sensitive but pursue superficial parts of the narrative while missing deeper, more significant patterns.
Expertise that differentiates
Research and Insight
Distinguishing a thread worth pursuing from a participant idiosyncrasy worth noting, based on what the study needs to answer rather than what the participant finds compelling to report.
Behavioral Reasoning
Reading engagement patterns over time to judge when an adapted prompt risks becoming intrusive or leading, or overwhelming.
AI Fluency that assures
Task Delegation
Delegation requires setting threshold around which prompting will shift: what counts as an interesting thread, what engagement levels warrants cadence change, what may even signal the need for human intervention. Without these parameters, adaptation follows frequency rather than methodological significance, or research ethics.
Creation Diligence
Structural review before adaptations are deployed helps manage quality. Because each adaptation is AI-generated, and can impact study validity, evaluation against objectives is recommended.
Related
↓
Enables
Possible Indicators
Data richness
Depth and specificity of diary entries relative to a static-prompt study baseline
Emergent theme coverage
Proportion of participant- specific threads generating new insights relative to a fixed prompt schedule