Home >  Tutorials > assumptions of linear regression

Tutorials : assumptions of linear regression

The study of the assumptions are based on the standardized residuals (differences between the observed and the predicted values divided by the standard error).


The non-variability of the residuals' variance is evaluated by a plot of the residuals against the predicted values. No particular form should be observed, i.e. the residuals should homogeneous around zero (in-between -2 and 2).


The residual distribution can be described by an histogram and a quantile-quantile plot (Q-Q plot). This last plot represents the quantiles of the sample against the quantiles of the normal distribution: the point have to be aligned on the first bisector.