Page 301 - Six Sigma Advanced Tools for Black Belts and Master Black Belts

OTE/SPH
OTE/SPH
August 31, 2006
JWBK119-18
Taguchi Methods
286 3:6 Char Count= 0
foundations and likely pitfalls awaiting their users. Maintaining a studied silence on
4
concerns such as the potential problems with the L 9 (3 ) design (not to mention the
even more theoretical issues), the camp has found (and continues to find) it difficult
to bring itself to openly conceding any shortcomings in the Master’s teaching.
18.6.5 Backlash and alternatives

It may be noted that while there may be excessive passion among Taguchi’s followers,
some criticisms against Taguchi methods have been no less pointed. Thus ‘accumula-
tion analysis is not a useful technique and to teach it to engineers would be a serious
mistake’. 41 And ‘maybe in other parts of the world, you can sell engineers simplistic
solutions as profound enlightenment. But around here, we don’t just buy something
on faith . . . Mr. Taguchi, pack up your “methods’’ and get out of my lab. DO NOT
TALK to any of my engineers or technicians. Do not suggest any other ways to be
more “efficient’’ or “optimized’’. Just get out’. 56 The latter quote did not come from a
statistician, but an engineer disappointed with what Taguchi methods purport to offer
to improve engineering design. More questions have also been raised in engineering
circles lately. 57,58 Actually, in terms of ‘market share’ today, Taguchi methods appear
to have an edge over the other ‘brands’ of experimental design (Box, Shainin, etc.)
outside academia. Although the use of statistical designs in a special way without
acknowledging their sources may not be as serious as industrial espionage, which the
Japanese have from time to time been accused of, American industry is now being
warned against being unwittingly done in by the Japanese (or Japanese methods) on
the engineering front, not just in trade. In fact, a curious love--hate feeling, not unlike
what takes place in a Toyota or Honda showroom, has been apparent during many
conference or workshop discussions on Taguchi methods in the USA.
In any case, as is now generally agreed, Taguchi methods on balance do have the
potential to bring about first-cut improvements in industrial applications. However,
owing to their theoretical imperfections and other limitations, success cannot be as-
sured in every instance, much like an angler who can go to the water conveniently,
but cannot be sure of a catch every time a line is thrown. (This point was amply
demonstrated earlier on through the of experiment A and B in Section 18.5.5.) Ap-
plied statisticians would rather get the basic philosophy laid out and understood, 14
and go for methodologies that are on firm theoretical grounds which can lead users
to the best possible results. Exemplified by response surface methodology, 30 the aim
is the best solution, not marginal improvements, hence their strategy is more akin to
mountaineering. Herein lies the contrast in choices: should one opt for the routines of
an oriental set piece for fishing, hoping for the best, or should one acquire all the nec-
essary tools to meticulously execute an ambitious, high-reward mountain climbing
expedition, exercising individual insights and judgments along the way?

18.7 TRAINING AND LEARNING

The above metaphorical alternatives have emerged from debates between Taguchi
advocates and skeptics, which are never short of colorful claims and counterclaims.
The Taguchi controversy is damaging when it comes to introducing statistical design