Effort: Medium
What to notice
Optionally check experience and uncertainty, then offer basics or a direct task path.
Where it fits
- Fits
- First-time users of specialist services with varied experience.
- Use it when
- When answers can lead to genuinely different support.
- Avoid it when
- When a pseudo-assessment would force everyone through the same path.
How to build it
- Ask separately about current skills and uncertainty.
- Offer brief practice and a route to skip explanation.
- Keep confidence reports separate from observed task results.
Measure
- Main metric
- Share of eligible users independently finishing the first task
- Guardrail
- Confidence–accuracy mismatch and errors after skipping guidance
- How to measure
- Allocate comparison groups before exposure and use equal follow-up and eligibility. Track Share of eligible users independently finishing the first task alongside Confidence–accuracy mismatch and errors after skipping guidance.
When it backfires
Do not fix ability labels or publish confidence answers.
Alternatives
Theory behind it
Self-efficacy
Belief in being able to perform a task relates to starting and persisting.
Practitioner framework
Confidence calibration
Treat reported confidence and actual correctness as different information.
Limited evidence
Evidence limits: An editorial web application; conversion effects on this interface are unmeasured. Confidence messages do not teach skills. Theory from clinical and learning settings does not directly predict sign-up lift.