Two-Level Factorial Taguchi Designs: Significance Testing for Factor Effects of Signal-to-Noise Ratios

摘要

Signal-to-noise ratios are used in Taguchi designs to determine optimal settings of parameters. This paper analyzes the distribution of signal-to-noise ratios in two-level factorial designs and derives approximate t-tests for the factor effect of signal-to-noise ratios. A simulation study shows that the t-test is very accurate for the "smaller-the-better" and the "nominal-the-better" while it is acceptable for the "larger the better" signal-to-noise ratio. Additional simulations show the power of the t-tests.