Prove
All outcomes
Learning

Run an Original Study (High School)

8 weeks · 0 milestones

Run your planned study, keep an honest log, analyse it with the planned test, and defend the results and their limits.

Milestone map

Milestone map

3 milestones

Collect the data as planned

3–6 weeks

Run the study exactly as your dated plan says. Keep a log of every session and every change, and keep raw data untouched.

Proof required

Submit the raw data file (anonymised), the collection log, and a note of any change from the plan and why.

What gets checked

  • Data matches the planned method
  • Every change logged
  • Raw data kept separate from cleaned data

Common mistakes

  • Changing the method silently
  • Throwing out inconvenient data
  • Names or personal details in the data

Powstik Guide

Raw data and an honest log.

Steps

  1. Follow the plan.
  2. Log date, conditions and issues each session.
  3. Save raw data read-only.
  4. Record any change and why.
  5. Remove names before sharing.

Template

Date | session | conditions | issues

Changes from plan:

What gets sent back

  • No collection log.
  • Silent method change.
  • Identifiable data shared.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

  • Checked by the verifiers you invite to your panel (up to 3). Pick people who can judge the work, not friends.
  • Compare the log to the plan
  • Check data is anonymised

You'll sign in first, then come straight back here.

Analyse it with the planned test

2 weeks

Run the analysis you planned. Report the result, the effect size and the uncertainty — including if the result is not what you hoped.

Proof required

Submit the analysis file or spreadsheet, a results table, and one chart, noting any analysis you added beyond the plan.

What gets checked

  • Planned test run first
  • Effect size and uncertainty reported
  • Extra analyses labelled as extra

Common mistakes

  • Trying tests until one is significant
  • Only reporting p-values
  • Hiding a null result

Powstik Guide

The planned analysis, honestly reported.

Steps

  1. Run the planned test.
  2. Report effect size and a confidence interval.
  3. Make one clear chart.
  4. Label any extra analysis as exploratory.
  5. Write what the result means in one paragraph.

Template

Planned test:
Result:
Effect size:
Uncertainty:
Exploratory (extra):

What gets sent back

  • Switched test to get significance.
  • No effect size.
  • Missing chart.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

  • Checked by the verifiers you invite to your panel (up to 3). Pick people who can judge the work, not friends.
  • Check the test matches the plan
  • Ask what the effect size means in plain words

You'll sign in first, then come straight back here.

Explain results to a reviewer

1 week

Walk a reviewer through your data and results. They should ask about weaknesses, and you should name the limits of your study plainly.

Proof required

Submit the review session log and a limitations section (200+ words) you wrote after the session.

What gets checked

  • Reviewer challenged the method
  • Limits named honestly
  • At least one change made after review

Common mistakes

  • Defending every weakness away
  • Reviewer only praised
  • No change after review

Powstik Guide

Results that survived questions.

Steps

  1. Present data and results.
  2. Invite hard questions.
  3. Log each challenge.
  4. Write limitations.
  5. Change one thing because of the review.

Template

Review session log
Date:            Duration:
Reviewer name:
Reviewer role and experience:

Challenge 1 (method and results):
  What they asked:
  My answer:
Challenge 2:
  What they asked:
  My answer:

Change I made after the session:

What gets sent back

  • No challenges logged.
  • Generic limits.
  • Nothing changed.

Resources

Foundationstart here

Depthgo deeper

What a verifier looks for

  • Checked by the verifiers you invite to your panel (up to 3). Pick people who can judge the work, not friends.
  • Ask what could explain the result other than the hypothesis
  • Check limits are specific

You'll sign in first, then come straight back here.

We use analytics to improve Powstik. No ads, ever.