From the simulation study, we found that the JS-type control charts usually have significant improvement when the true mean is in a known region which should not be too wide, and v in the James–Stein estimator is selected to be a point in this region.