The top-5,000 test is not statistically significant | Trakkr Research
Only the top-5,000 subset test is non-significant; the full-sample test is significant.
Methodology: Built from HTTP scans of 37,894 AI-cited domains, linked to 337,362 citations and 882 citation snapshots in the Trakkr corpus.
Claim
The top-5,000 citation comparison returns p=0.85; the full sample returns p<0.001 with reported r=-0.065.
Why it matters
Report both cohorts and avoid interpreting a non-significant test as proof of no effect.
Supporting metrics
| Metric | Value | Context |
|---|---|---|
| Top-5,000 Mann-Whitney p-value | 0.85 | Subset of the top 5,000 cited domains; not the full-sample test. |
| Full-sample Mann-Whitney p-value | <0.001 | Full 37,894-domain sample. The public JSON rounds this value to zero. |
| Full-sample reported effect size r | -0.065 | Small reported magnitude; not an estimate of causal citation lift. |
Related pages
Continue through the same study cluster.
- which industries adopt llms txt most - Related answer page
- does llms txt help you even if it does not raise citations - Related answer page
- llmstxt adoption by sector tracker - Related tracker page
Data & Sources
- The llms.txt Effect - Flagship study behind this page
- Page JSON - Machine-readable companion file