When the brainstorming process was complete, 36 variables had been
identified as potentially causal in terms of the variance (refer to Figure
B-25). The key to successful experimentation was the increased focus in
brainstorming as a result of knowing the matrix intersect.
Following the brainstorming sessions, the team developed a rigorous method
of variable selection and prioritization. The key activity relating to this
method involved rank ordering all of the variables on several unique
multi-point rating scales. Such scales were inclusive of but not limited to:
By weighing the various scales and making the necessary computations, it
was possible for the team to quickly prioritize the 36 variables.
The experimental variables are listed in Figure B-26. The control variables
are defined in Figure B-27. The experimental levels or test settings are also
defined, as well as level tolerances. It required two days for the team to
arrive at this point. It was necessary to use the full time to maintain the
experiment's integrity and ensure that quality inputs and assumptions would
likewise ensure quality output.