Once the multi-vari study was completed, the team met again to statistically design an experiment which would identify the leverage-independent variables. The first step in this process was to brainstorm all variables which could potentially impact the variance within the intelligent board area of a PCB and from board-to-board within a rack. (Remember the multi-vari study revealed that the problem was related to the variance and not the mean.)
The brainstorming matrix presented in Figure B-25 was used to help the team
stay focused during brainstorming. Focus was given to those factors which were
not formally controlled on an ongoing basis such as those variables which did
not have formal control mechanisms or knobs.