Larger is better: Select when the goal is to maximize the response.
Nominal is best: Select when the goal is to target the response and you want to base the signal-to-noise ratio (S/N ratio) on standard deviations only.
Nominal is best: Select when the goal is to target the response and you want to base the signal-to-noise ratio (S/N ratio) on means and standard deviations (default).
Use adjusted formula for nominal is best: Select to use the adjusted formula. To compare the nominal-is-best and the adjusted formulas, go to Methods and formulas for Analyze Taguchi Design and click Signal-to-noise ratio for static designs.
Smaller is better: Select when the goal is to minimize the response.
Use ln(s) for all standard deviation output
Use the natural logarithm transformed standard deviations as the response variable in the response table, regression/ANOVA results, and on the main effects and interactions plots.
Consider using the natural log transformation on the standard deviation for the following reasons:
To stabilize the variability of the standard deviations
To avoid negative values for the fitted values, which can occur when using untransformed standard deviations. With a log transformation, all fitted values are positive, which makes them easier to intuit.
To avoid multiplicative effects on variability. Using this option transforms multiplicative effects into additive effects, which are better suited to linear models.