Question

In studying the uniformity of polysilicon thickness on a wafer in semiconductor manufacturing, Lu, Davis, and Gyurcsik (Journal of the American Statistical Association, Vol. 93, 1998) collected data from 22

independent wafers: 494, 853, 1090, 1058, 517, 882, 732, 1143, 608,590, 940, 920, 917, 581, 738, 732, 750, 1205, 1194, 1221, 1209, 708. Is it reasonable to model these data using a normal probability distribution?

Fig: 1