null hypothesis
In statistics, the default assumption of no effect or no difference; tests can fail to reject it but cannot 'prove' a positive evidentiary claim of safety.
As defined in Nassim Nicholas Taleb on the Precautionary Principle and Genetically Modified Organisms
Where it comes up · 15
- Nassim Nicholas Taleb on the Precautionary Principle and Genetically Modified Organisms3 claims
- Roland Fryer on Race, Diversity, and Affirmative Action3 claims
- Gary Taubes on Why We Get Fat2 claims
- Ep. 11 - Letting It All Go2 claims
- Machine Intelligence and Risk Management1 claim
- Jim Manzi on the Oregon Medicaid Study, Experimental Evidence, and Causality1 claim
- Noah Smith on Whether Economics is a Science1 claim
- Matt Ridley on the Evolution of Everything1 claim
and 7 more
Most important claims · 10 of 21Search all claims for “null hypothesis” →
- Everyone, including Taubes himself, is susceptible to confirmation bias: a study supporting your view is accepted after reading only the abstract and discussion, while a contradictory study is scrutinized for an hour to find reasons to reject it; framing one's position as the null hypothesis at least forces a demand for remarkable evidence to overturn it.Gary Taubes on Why We Get Fat
- Operating in an underpowered environment creates two problems simultaneously: a very high chance of false negatives (missing real signals), and a high risk that any statistically significant result is a false positive or a grossly inflated effect size—the smaller the power, the greater the average inflation of detected effects.John Ioannidis on Statistical Significance, Economics, and Replication
- Discrimination measurement in observational data is effectively a one-sided test: a residual disparity remaining after controlling for education, experience, and skill could be discrimination, or it could be mismeasured variables (noisy or inadequate proxies for skill), so pinpointing discrimination requires excellent data and techniques.Roland Fryer on Race, Diversity, and Affirmative Action
- Understanding discrimination is equivalent to understanding the causal effect of race on an outcome, which is extraordinarily hard because you cannot randomly assign race; researchers must rely on natural experiments or proxies like randomizing racially-indicative names on resumes.Roland Fryer on Race, Diversity, and Affirmative Action
- Critics of digital media make three mistakes: they overestimate their own agency as designers of attention-manipulation mechanisms, they underestimate the conscious agency users bring to their consumption choices (a given attention allocation may be exactly what the user wants), and they fail to recognize that current behaviors are adaptive experiments responding to a genuinely transformed information environment with orders of magnitude more information than our scarcity-evolved processing methods can handle.Venkatesh Rao on Waldenponding
- Until the 1960s the conventional wisdom held that carbohydrates were inherently fattening, as documented from Brillat-Savarin's 1825 observations to a 1963 British nutrition article opening 'Every woman knows that carbohydrates are fattening'; this historical consensus, plus the supporting biochemistry, means the carbohydrate hypothesis should be the null hypothesis requiring remarkable evidence to reject.Gary Taubes on Why We Get Fat
- Every macroeconomic theory that makes bold strong predictions has cases where those predictions failed to pan out; if subjected to rigorous statistical testing—even generously treating the theory as the null hypothesis—all of them would be rejected, which is why macroeconomists have retreated to using theories as thought experiments that merely 'organize our thinking' by mapping assumptions to conclusions.Noah Smith on Whether Economics is a Science
- The p-value is the probability that an observation as extreme as the data would occur by chance given that the true effect is zero (the null hypothesis is correct); it is not the probability that the result is due to the treatment.Jim Manzi on the Oregon Medicaid Study, Experimental Evidence, and Causality
- Western ways of thinking, embodied in the scientific testing of the null hypothesis, build in a cynicism of 'I don't believe it, show me' that closes the doors to spiritual experience, because westerners fear being conned if they allow themselves irresponsible emotional faith they cannot justify to a hard-headed scientist.Ep. 04 - Guru Found
- The widespread claim that we became moral faster because of supernatural beliefs has not been demonstrated; religions have taken credit for morality for so many centuries that we cannot get behind the veil to test it, and the null hypothesis—that we can be moral through reason without religion—deserves a chance.Matt Ridley on the Evolution of Everything