The article “Applying Design of Experiments to Improve a Laser Welding Process” (J. of Engr. Manufacture, 2008:

1035–1042) included the results of a half replicate of a 24 experiment. The four factors were: A. Power (2900 W, 3300 W), B. Current (2400 mV, 3600 mV), C. Laterals cleaning

(No, Yes), and D. Roof cleaning (No, Yes).

a. If the effect ABCD is chosen as the defining effect for the replicate and the group of eight treatments for which data is obtained includes treatment (1), what other treatments are in the observed group, and what are the alias pairs?

b. The cited article presented data on two different response variables, the percentage of defective joints for both the right laser welding cord and the left welding cord. Here we consider just the latter response. Observations are listed here in standard order after deleting the half not observed. Assuming that two- and three-factor interactions are negligible, test at level .05 for the presence of main effects. Also construct a normal probability plot.
8.936 9.130 4.314 7.692 .415 6.061 1.984 3.830