p-hacking

Fresh Trending

2 views · Added 1d ago · 2 definitions

1
When researchers mess with their numbers until they get a p-value under .05, even if they don’t know what they’re doing. It's like cheating at a math test with a calculator full of lies.
A study says eating gummy bears makes you smart, but they just picked the best results out of 20 experiments.
A scientist claims a new coffee will cure baldness, but they only used the data that looked good.
A nutritionist says chocolate is healthy, but they ignored all the times it made people gain weight.
2
Making up stats so the results look good, usually because someone paid for it. They pick the numbers they want and say it's real, even if it's total garbage.
A drug company says their pill works, but they only reported the times people got better, not the times they got worse.
A fitness coach says a new workout burns 1000 calories a day, but that’s only if you dance like a madman and never eat.
A food company says their cereal is healthy, but it's just the best of 100 failed experiments.
xs