Browse Teach Curate

Turning an Effect Estimate into a P-Value with Its Standard Error

Effect size alone is uninterpretable; the estimate's standard error is what converts it into a test statistic and then a p-value. This node is why P = 2.6 x 10^-18 for the lead variant is a statement about precision in a sample of 12,901 and not a statement about how large the effect is.

Questions this Concept answers

Why can a very small p-value sit alongside a very small effect estimate?

jeremy Video

Statistical Inference for Regression Coefficients in Linear Regression

This covers the statistical inference theory for ordinary least squares (OLS) regression coefficients: under the standard linear model assumptions (Gaussian errors), the OLS estimator beta-hat follow…