regression
A statistical technique that estimates the relationship between an outcome variable and one or more explanatory variables.
As defined in How to Avoid Lying With Statistics (with Jeremy Weber)
Other definitions
Statistical technique for modeling relationships between variables; ridge regression and generalized least squares are regularization methods that prevent overfitting
Surviving the Meme Stock Bubble | Cliff Asness
Where it comes up · 10
- Russ Roberts (and Robin Hanson) on Truth and Economics 01/26/20093 claims
- Emily Oster on Pregnancy, Causation, and Expecting Better2 claims
- Joshua Angrist on Econometrics and Causation2 claims
- How to Avoid Lying With Statistics (with Jeremy Weber)2 claims
- Crowdsourcing Predictive Algorithms1 claim
- Machine Learning in Investing1 claim
- Dwarkesh Patel Podcast - Terence Tao – Kepler, Newton, and the true nature of mathematical discovery1 claim
- Jim Manzi on the Oregon Medicaid Study, Experimental Evidence, and Causality1 claim
and 2 more
Most important claims · 10 of 15Search all claims for “regression” →
- Kepler derived his third law (the square-cube relationship between orbital period and distance) from only about six data points by curve-fitting, which is statistically unreliable; he was lucky it generalized, and he likely de-emphasized it relative to his first two laws because he intuitively sensed six points warranted tentative conclusions.Dwarkesh Patel Podcast - Terence Tao – Kepler, Newton, and the true nature of mathematical discovery
- Machine learning models can be organized in a two-by-two grid: classification versus regression on one axis and supervised versus unsupervised on the other, with linear and logistic regression, neural networks, clustering, and trees placed within it.Machine Learning in Investing
- Machine learning is essentially regression and curve fitting, and a fitted model can predict well even when the analyst does not know what the input axes represent, as with Numerai's roughly 50 anonymous dimensions.Crowdsourcing Predictive Algorithms
- A randomized controlled trial is the gold standard because, unlike regression or matching that attempt to hold observable factors equal, random assignment ensures that even unobserved or unmodeled differences between treatment and control groups are approximately equally distributed, subject only to sampling error.Jim Manzi on the Oregon Medicaid Study, Experimental Evidence, and Causality
- Regression analysis cannot credibly isolate the employment effect of a minimum wage that applies to about 3 percent of the population when the outcome measured is the overall unemployment rate affected by many other factors, because controlling for enough variables is an unverifiable assumption.Russ Roberts (and Robin Hanson) on Truth and Economics 01/26/2009
- Friedman and Schwartz's Monetary History of the United States was decisive because it relied on looking at facts such as money supply growth and measured inflation rather than sophisticated regression analysis, and it forced opponents to confront monetary explanations of inflation.Russ Roberts (and Robin Hanson) on Truth and Economics 01/26/2009
- Establishing causation is genuinely hard, and in public policy where causal mechanisms are more complex than in pregnancy, progress is limited: studies rarely get definitively refuted, instead a new study comes along with different methodology yielding slightly different conclusions, so it is hard to ever draw strong causal conclusions despite incremental knowledge gains.Emily Oster on Pregnancy, Causation, and Expecting Better
- One should never run a regression without a clear purpose, because meandering through data invites being led to manufactured stories: a researcher gets a hunch, subsets the data until the story emerges, and ends up selling a coherent but fabricated narrative--a danger amplified by modern abundant data and instant computation.How to Avoid Lying With Statistics (with Jeremy Weber)
- Even with vast data and many variables, social scientists never have all the relevant data (moods, childhoods, genetics), so theory-free exploration is dangerous: practitioners pretend they have enough data, discard what seems irrelevant, and become prone to confirmation bias rather than confronting the inherent incompleteness of social-science data.How to Avoid Lying With Statistics (with Jeremy Weber)
- The apparent earnings advantage of attending selective private colleges over public ones largely disappears once you condition on where students applied and were admitted, implying the raw difference reflects pre-existing ambition and ability rather than a causal college effect.Joshua Angrist on Econometrics and Causation