The article ??Case Study Based Instruction of DOE and SPC?? (J. Brady and T. Allen, The American Statistician, 2002:312??315) presents the result of a 24??1 factorial experiment to investigate the effects of four factors on the yield of a process that manufactures printed circuit boards. The factors were A: transistor power output (upper or lower specification limit), B: transistor mounting approach (screwed or soldered), C: transistor heat sink type (current or alternative configuration), and D: screw position on the frequency adjustor (one-half or two turns). The results are presented in the following table. The yield is a percent of a theoretical maximum.

a. Estimate the main effects of each of the four factors.

b. Assuming all interactions to be negligible, pool the sums of squares for interaction to use in place of an error sum of squares.

c. Which of the four factors, if any, can you conclude to affect the yield? What is the P-value of the relevant test?

A B C D Yield -1 -1 -1 79.8 -1 1 69.0 1 72.3 -1 71.2 91.3 95.4 92.7 1 1 91.5