L9 orthogonal array was designed before the experimental run. During the experiments, it was revealed that some parameters were invalid. So, instead of nine experiments, seven experiments were conducted. What effects will have upon the onward analyses?