Regression Analysis and Experimental Design attempt to create a regression model that allocates the variation in the data to the factors being tested. The difference between the values obtained from the process and the values predicted by the regression model are known as the 'residuals'.
The residuals should conform to a normal distribution. If they do not then the patterns in the residuals can give important insights into how well the regression model represents the process under study and evidence of special causes. Residual analysis is an important part of the analysis in regression analysis and experimental design.