Experimental designs can be single- group, three- group, four- group, parallel -group, counter balanced or Latingroup, rational group. Residuals are estimates of experimental error obtained by subtracting the observed response from the predicted response.