B Evidence Grade B Graded accuracy verified by the review judge Safe to Cite Recent

26.1%

26.1% -- inflated false alarm rate for A/B tests with continuous peeking (2025)

2025  · Evan Miller

Continuous peeking in A/B testing inflates the false alarm rate from 5% to 26.1%, according to calculations by Evan Miller.

Cite this stat Share How we verify
Attribution: Evan Miller View source
Lighthouse Research Team
Last verified:  ·  How we verify data →
Source verified Evidence graded
Value
26.1%
Unit
Percentage
Data Vintage
2025
Entity
A/B tests with continuous peeking
B
Strong, reliable evidence

Sourced from a well-regarded research firm or industry leader with transparent methodology and a reasonably large sample.

61/100 confidence

Single source - not yet corroborated

Extraction
75
Source Trust
60
Citation Chain
20
Independence
30
Methodology
45
Freshness
95
Evan Miller Primary Tier B
https://kirro.io/ab-test-sample-size-formula ↗
Confidence: 53/100 Vintage: 2025 Channel: Deep Crawl
Source Article
A/B Test Sample Size Formula, Explained With Examples
The A/B test sample size formula is: n = (Zα/2 + Zβ)² × [p₁(1−p₁) + p₂(1−p₂)] / (p₁−p₂)² If that looks like gibberish, you’re not alone. Most articles either throw this formula at you with zero explanation, or skip it and say “just use a calculator.” This guide does both. We’ll break down every piec…
Original Research
Evan Miller · Original study ↗

Evan Miller's statistical calculation on peeking

Methodology documented
Period Value Change
2025 (current) 26.1%
2026 26.1% 0%
Industrymarketing
Plain Text
A/B tests with continuous peeking: 26.1% inflated false alarm rate. Source: Evan Miller (2025). Via Lighthouse Intelligence -- https://lighthousedata.io/data/peeking-inflates-false-alarm-26-1pct
HTML Embed
<blockquote cite="https://lighthousedata.io/data/peeking-inflates-false-alarm-26-1pct" style="border-left:3px solid #2563eb;padding:12px 16px;margin:16px 0;font-family:system-ui,sans-serif;background:#f0f4ff;"><strong>26.1%</strong> -- inflated false alarm rate for A/B tests with continuous peeking<br><small>Source: Evan Miller (2025) · <a href="https://lighthousedata.io/data/peeking-inflates-false-alarm-26-1pct" target="_blank" rel="noopener">Lighthouse Intelligence</a></small></blockquote>
Share
Share on X Share on LinkedIn

Use this data in your work

Copy the citation, embed the widget, or explore thousands more verified marketing statistics -- all free.

Copy citation Explore similar stats →