Evidence
What we measure, what it can support, and what it cannot.
A pilot's value comes from the honesty of its reporting, not the size of its claims. This page states our method and our limits before any result exists.
Methodology
How the pilot is run
- 01
Design
Curriculum objectives are written as observable capabilities before any teaching happens, and each one is mapped to a rubric criterion.
- 02
Collect
We collect process data (attendance, submission, completion), rubric-scored artefacts, and self-report at three points: entry, midpoint, exit.
- 03
Score
Artefacts are scored by a trained assessor against the published rubric. A ten-percent sample is double-marked to check consistency.
- 04
Report
Findings are published with their classification tier, sample size, and the specific limitations that apply. Null and negative results are published too.
Classification
Every statement we publish carries a tier
Tiers are printed next to each claim in any report. A claim without a tier is not a claim we have made.
| Tier | What it means | Evidential weight |
|---|---|---|
| Design intent | A stated aim of the curriculum. Not a finding. | Lowest |
| Process data | Counted activity: attendance, submissions, completion. | Low |
| Self-report | What learners say about their own experience. | Moderate, biased |
| Facilitator observation | Rubric-scored observation by a trained assessor. | Moderate |
| External review | Assessment by a reviewer outside the Institute. | Highest available |
Limitations
Known and stated up front
- A single cohort of at most twenty-four self-selected volunteers cannot support generalisation.
- There is no control group and no randomisation. Any change observed may be caused by something other than the programme.
- Self-report is subject to demand characteristics: learners who paid and stayed are motivated to report improvement.
- Assessors are not independent of the Institute in the pilot. Double-marking reduces, but does not remove, this bias.
- The rubric itself is unvalidated. It is a working instrument in its first version.
- Effects are measured over twelve weeks. Nothing here says anything about durability beyond that window.
Refusals
Claims we will not make
- That the programme rewires, retrains, or physically changes the brain
- That it heals trauma or resolves a psychological condition
- That it teaches reliable detection of lying or intent
- That it produces a guaranteed outcome in work, income, or relationships
- That any finding here is scientifically proven, peer-reviewed, or replicated
Corrections
How we fix what we get wrong
Published material is versioned. A substantive correction is never applied silently.
Raise
Anyone may submit a correction request by email with the passage and the issue.
Assess
We respond within ten working days stating whether the correction is accepted and why.
Publish
Accepted corrections are applied with a dated note, and the document version increments.
No corrections have been issued because no findings have been published. The corrections log will appear here once the first report exists.
Cohort 001 report
Reserved for real results
No pilot findings yet
Cohort 001 concludes on 7 December 2026. The full report — method, results, limitations, and raw counts — will be published here, including any result that does not support the programme.
Nothing on this site is evidence of effectiveness