Daily Decisions
A web platform for studying how older adults seek and use advice when making repeated financial decisions, and whether that depends on who the advice is attributed to and how much of it there is.
A 2 × 2 mixed design
Participants check in twice a day for seven days and make one decision each time, fourteen in all. Every decision starts with an independent answer and confidence rating. Only then is the advisor named and advice offered, followed by a final answer. Each participant stays in one information condition (the columns); the same advice is attributed to an AI advisor on seven decisions and a human advisor on seven (the rows). Nothing in a demonstration is recorded.
AI recommendation
One recommended option with a stated confidence and an expandable rationale, attributed to an AI advisor.
Open demonstration →AI-selected materials
Three materials about the situation, each opening in place, attributed to an AI advisor.
Open demonstration →Human recommendation
The identical recommendation, confidence, and rationale, attributed to a human advisor.
Open demonstration →Human-selected materials
The identical three materials, in the same order and format, attributed to a human advisor.
Open demonstration →Researcher dashboard
The 2 × 2 with its main effects, participant histories, trial-level interaction replay, and CSV exports for R or Python.
Implemented at submission
- An independent initial answer before the advisor is named, then a separate final answer
- Advice-seeking gate, so exposure is recorded apart from uptake
- Identical advice across AI and human attribution, verified automatically for every matched pair
- Fourteen scenarios with on-screen evidence records; misleading versions contain one checkable factual error
- Twice-daily scheduling with outcome reveal at the next check-in and accumulating points
- Task-bounded interaction telemetry, researcher monitoring, trial replay, and CSV export
Supported by the R21 award
- Expert validation of the scenario bank and norming with older adults
- Reminder, retention, and missed-session workflows
- Security, data-governance, and production deployment work
- Accessibility and usability testing with older adults