Audit a URL →

Sign in Sign up free

Loading…

Weekly · Sunday Morning

Five longevity claims.
Worth your time.
Every Sunday.

Last Sunday

Rapamycin extends mouse lifespan 9–14%
NAD+ precursors reverse aging in humans
Blue-light glasses improve sleep quality
Walking 10 min after meals blunts glucose
Verdict change — caffeine half-life downgraded

5 audits·1 verdict change·0 affiliate links·Every Sunday

Joining 1,247 readers — founders, MDs, and longevity-obsessed people who fact-check their own health decisions.

Browse

Decisions Claims Content Creators

Product

Audit a URL Sunday roundup API (soon)Pricing

About

FAQ Methodology Editorial Standards Corrections Press & brand RSS feed

Legal

Terms of Service Privacy Policy DMCA

Finally, something serious in this space.
Because we actually read the papers.

© 2026 Whalespan. Independent, evidence-graded longevity.

Loading…

Power, p-values, and statistical significance — Whalespan

Content›Video›Power, p-values, and statist…

Power, p-values, and statistical significance

Peter Attia11m · 2023-09-09

Support Rate

41%

7 of 17 claims supported

Support rate

7 of 17 claims supported

7 supported10 partial / not

Claims from this content

13

✓Well SupportedUnderpowering an experiment by not including enough subjects can lead to false negatives.✓Well SupportedAn overpowered study may find statistically significant results that are clinically irrelevant.◐Partially SupportedThe 0.05 p-value threshold is a standard for rejecting the null hypothesis.◐Partially SupportedP-values should be as close to zero as possible, with a maximum threshold of five percent (0.05) typically used.◐Partially SupportedWhen evaluating a study, determine if it was correctly powered for its intended effect size and statistical significance level.

More from Peter Attia

Peter Attia on Zone 2 Cardio

Peter Attia on Rapamycin (with Matt Kaeberlein)

Attia walks back the clinic protocol: too aggressive for healthy adults

Discussion

Native comments, Twitter mentions, and Reddit threads about this content — surfaced together so the conversation isn't fragmented across platforms.

M

Share your read of this content…

Posting publicly as you · linked to your social profiles

Sort by:

AV

A. Volkovreader

4h

Bookmarking — the dossier-vs-overview split is the right call. Most of the time I want overview; sometimes I want receipts.

NP

N. Park

2d

Would love a "what would change this verdict" RSS feed. Sign me up if it exists.

◐Partially Supported

A smaller effect size in a clinical trial requires a dramatically increased number of subjects per arm.

◐Partially SupportedPower in statistical analysis is defined as 1 minus beta, where beta is the probability of a false negative.

◐Partially SupportedStatistical significance is achieved when the null hypothesis is rejected because the p-value is less than a pre-stated level.

◐Partially SupportedA false positive is the probability of rejecting the null hypothesis when it is true.

◐Partially SupportedA hypothetical trial with a 25% placebo success rate and 35% treatment success rate requires 459 participants for 90% power and 358 participants for 80% power.

◐Partially SupportedThe false negative rate is typically allowed to be between 10% and 20%, which corresponds to a statistical power of 80% to 90%.

◐Partially SupportedUnderstanding statistical power is crucial for designing any clinical trial.

✕Not SupportedSetting the p-value threshold too low risks discarding potentially relevant information.