There are three approaches to the design of experiments, the classical, Taguchi, and Shainin

Characteristic Classical Taguchi Shainin
Technique Fractional Factorials, EVOP, etc. Orthogonal arrays Multi-vari, variable search, full factorials
Effectiveness Moderate (20% to 200% improvement)
Retrogression possible Low to moderate (20% to 100% improvement) Retrogression likely Extremely powerful (100% to 500% improvement) No retrogression
Cost Moderate
Average of 50 experiments High
Average of 50 to 100 experiments Low
Average of 20 experiments
Complexity Moderate
Full ANOVA required High
Inner and outer array multiplication, S/N. ANOVA Low
Experiments can be under-stood by line operators
Statistical Validity Low
Higher order interaction effects confounded with main effects
To a lesser extent, even 2nd order interaction effects confounded Poor
No randomization
Even 2nd order interaction effects confounded with main effects
S/N concept good High
Every variable tested with all levels of every other variable
Excellent separation and quantification of main and interaction effects
Applicability Requires hardware
Main use in production Primary use as a substitute for Monte Carlo analysis Requires hardware Can be used as early as prototype and engineering run stage
Ease of Implementation Moderate •Engineering and statistical knowledge required Difficult
Engineers not likely to use technique
Even line workers can conduct experiments