Understanding Statistical Noise in Research: 3. Noise in Regression Analysis.

In regression analysis, the adjusted R 2 value describes the proportion of the variance in the dependent variable that is explained by the independent variables in the equation. What remains is unexplained or residual variance. Residual variance has two components: the contribution of unmeasured variables (residual confounding) and measurement error (noise). These concepts are explained using examples.