The experimental results revealed that the main effects of factor B
(plating bath temperature), factor G (shared anodes), and factor H (anode bar
connection) were statistically and practically significant. In fact, the
independent contributions of these three main effects accounted for nearly 60%
of the observed variation. In addition, it was apparent that several of the
second order effects were also very important. In Pareto terminology, it was
concluded that the "vital few" effects had been successfully isolated from the
"trivial many" in terms of XYZ copper plating thickness variability.
A graphic presentation of the overall results is shown in Figure B-29.
Notice that the horizontal axis of the bar chart displays a number. In order
of sequence, 1 = factor A, 2 = factor B, 3 = factor C, and so on. The vertical
axis of the bar chart reflects the percent contribution that the related
effect had in the total scheme. For example, it is apparent that bar 8 (factor
H or "anode bar connection") extends to the 20% mark on the vertical scale.
This indicates that the anode bar connection accounted for approximately 20%
of the observed experimental variation.