"If your experiment needs statistics, you ought to have done a better experiment"
About this Quote
Rutherford’s jab lands with the clean violence of a lab door slammed shut. “If your experiment needs statistics” is framed as a kind of confession: you built something so noisy, so ambiguous, that you now have to ask mathematics to rescue the story. The line performs a hard-nosed ideal of science where a proper experiment produces an effect so unmistakable you can see it without squinting at p-values. It’s not just anti-statistics; it’s pro-clarity, pro-instrumentation, pro-design. Statistics, in this worldview, is the mop you reach for after you’ve spilled uncertainty everywhere.
The subtext is also a status play. Early 20th-century physics prized dramatic signal-to-noise: cloud chambers, radioactive decay, spectra you could almost point at. Rutherford, an experimentalist’s experimentalist, is policing the boundary between robust causal demonstration and what he’d likely dismiss as “inference work” dressed up as rigor. The insult is that leaning on statistics suggests you never controlled the system in the first place.
Context matters because the quote has aged into a weapon people use against whole fields - especially the messier ones: psychology, medicine, social science - where effects are subtle, humans vary, and “better experiment” often means “impossible experiment.” In those domains, statistics isn’t a crutch so much as the steering wheel.
That’s why the line still works: it’s a perfect expression of a seductive fantasy, that nature will speak in boldface if you ask the question correctly. It flatters the experimenter, and it needles everyone else.
The subtext is also a status play. Early 20th-century physics prized dramatic signal-to-noise: cloud chambers, radioactive decay, spectra you could almost point at. Rutherford, an experimentalist’s experimentalist, is policing the boundary between robust causal demonstration and what he’d likely dismiss as “inference work” dressed up as rigor. The insult is that leaning on statistics suggests you never controlled the system in the first place.
Context matters because the quote has aged into a weapon people use against whole fields - especially the messier ones: psychology, medicine, social science - where effects are subtle, humans vary, and “better experiment” often means “impossible experiment.” In those domains, statistics isn’t a crutch so much as the steering wheel.
That’s why the line still works: it’s a perfect expression of a seductive fantasy, that nature will speak in boldface if you ask the question correctly. It flatters the experimenter, and it needles everyone else.
Quote Details
| Topic | Witty One-Liners |
|---|---|
| Source | Later attribution: 1001 Quotations to inspire you before you die (Robert Arp, 2018) modern compilationISBN: 9781788400510 · ID: FGNADwAAQBAJ
Evidence: ... If your experiment needs statistics, you ought to have done a better experiment.” Ernest Rutherford Attributed c. 1920 British physicist Ernest Rutherford, often described as the “father of nuclear physics,” began his career at ... Other candidates (1) Ernest Rutherford (Ernest Rutherford) compilation33.3% hrough his discovery and interpretation of rutherford scattering during the gold foil experiment |
More Quotes by Ernest
Add to List
