Regression slope
= Regression slope
{title2=$\beta$}
In a <simple linear regression> with <conditional mean> $\mathbb E[Y\mid X=x]=\alpha+\beta x$, the slope $\beta$ is the change in that mean per unit increase in the predictor. The <ordinary least squares> estimate is $\widehat\beta=S_{xy}/S_{xx}$ when $S_{xx}>0$. It describes a <statistical association>; a causal interpretation requires further assumptions.