I think Michael Edmiston's Type I/Type II brainstorming is a good start,
but the division is too black and white and too 1 dimensional (and looking
back now that am done, perhaps I'm getting too complex and
multidimensional).
First of all, I'm assuming that the ultimate purpose is to find a
manageable set of plausible answers from which a more detailed, final
solution could be developed. And there are other possible purposes, like
ice-breakers, teaching problem-solving skills, or making administrators
feel important.
I see four dimensions to a brainstorming process, that could be put on a
continuous scale of 0 (low) to 1 (high).
Dimension I: Complexity and/or importance of the problem
Dimension II: Thoughtfulness of the proposals, (from off the cuff,
more-or-less random to fully refined, optimized solutions)
Dimension III: Thoughtfulness of the responses, (from either meaningless
or nonexistent responses to articulate, well reasoned critiques)
Dimension IV: Time spent, (from a few minutes for a few people to a few
hours for many people)
Having any of these factors close to an extreme is inefficient for
brainstorming. For a trivial question there is no point in having a bunch
of people pretend not to know the answer, while a truly complex problem
requires more forethought than brainstorming can provide. Expending too
little thought on proposals or responses leads nowhere, while too much
thought within the session wastes time. And of course, you need enough
time to make progress but not so much time that you are simply going in
circles.
Furthermore, not only do the factors need to be kept away from extremes,
they should all be kept reasonably close to each other. All four should be
reasonably matched. To have your kids decide where to go out to eat,
something near 0.1's for all categories would be reasonable. To find out
why your data suddenly is giving unexpected results, perhaps 0.6's. For
Bush's Security Council to explore if/how to attack Iraq, 0.8's.
The original barometer problems is something like 0.2 - 0.9 - 0.4 - 0.9.
Here, the effort is out of proportion with the original problem.
What Michael seem to be objecting to is a process near X - 0 - 0 - 1, which
violates both of my rules for brainstorming. No matter what the value of
X, it isn't an efficient brainstorming process.