Data Integrity

How we build
the archive.

The Regret Index is only as valuable as the integrity of its data. This page explains exactly how we collect, structure, verify, and protect every submission in the archive.

Collection Principles

Structured, not freeform

Every submission follows the same schema: the decision itself, the context at time of decision, the outcome after a defined timeframe, and a regret score (0-10). This structure is what makes the data comparable and searchable.

We deliberately do not collect free-form essays. Structured data scales; essays don't.

Outcome-first, not advice-first

Submitters report what actually happened — not what they think others should do. This removes the advice-giving bias that corrupts most crowdsourced wisdom. We ask: "What would you tell yourself?" not "What should others do?"

The regret score is self-reported. We do not infer it from sentiment analysis or external signals.

The Submission Schema
decisionMade
The Decision
A concise statement of what the person chose to do.
category
Category
One of 9 standardised life domains.
ageAtDecision
Age at Decision
The submitter's age when the decision was made.
yearOfDecision
Year
When the decision was made. Enables temporal analysis.
whatHappened
Outcome
What actually happened as a result of the decision.
regretLevel
Regret Score (0–10)
0 = no regret. 10 = would undo if possible.
wishIKnew
What I Wish I Knew
The key piece of information that was missing.
adviceToOthers
Advice
What the submitter would tell someone facing the same choice.
Bias Reduction

Survivorship bias: We actively recruit submissions from people who made decisions that did not work out. Regret scores below 3 and above 8 are equally represented in the moderation queue.

Recency bias: We require a minimum elapsed time between the decision and submission to ensure the submitter has lived with the outcome, not just reacted to it.

Social desirability bias: All submissions are fully anonymous. There is no social identity attached to a submission once it enters the archive.

Moderation

Every submission enters a moderation queue before appearing publicly. Our moderation checks for:

  • Completeness — all required fields populated
  • Coherence — the decision and outcome are logically connected
  • Authenticity — no promotional content, no fictional scenarios
  • Safety — no content that could identify or harm a third party

Rejected submissions receive a reason. Submitters can revise and resubmit once.

Longitudinal Tracking

Decisions look different over time.

A career pivot evaluated at 6 months often feels different at 3 years and dramatically different at 5. We follow up with submitters at 1, 3, and 5-year intervals to update their regret score and outcome narrative. These longitudinal data points are the most valuable in the archive.

1-Year Follow-up
Immediate outcome
3-Year Follow-up
Mid-term perspective
5-Year Follow-up
Long-term truth

Follow-up participation is opt-in. Email reminders are sent only to users who explicitly enrolled.

Ready to contribute?

Add your experience to the archive. Fully anonymous.