Analysis of a Robust Design Experiment - Means Tab

Analyzing Taguchi Designs

Select the Means tab of the Analysis of a Robust Design Experiment dialog to access options to view the marginal means.

Display design and S/N ratios

Click the Display design and S/N ratios button to display the results for each run of the experiment in a spreadsheet. The last column will contain the respective Eta values (S/N ratios as chosen on the Design & Analysis of Taguchi Robust Design Experiments dialog).

Descriptives for untransformed vars

Click the Descriptives for untransformed vars button to display a spreadsheet with the means, sums, and standard deviations for the untransformed dependent variables.

Marginal means

Click the Marginal means button to display a spreadsheet with the marginal means, parameter estimates, standard deviations, and standard errors for each level for each factor. These descriptive statistics pertain to the average Eta values (S/N ratios), not to the untransformed variables. The parameter estimates can be interpreted as deviations of the means from the overall mean Eta value (m) for the respective levels of the respective factors. For example, a parameter estimate of lA1 = 1.5 for level 1 of factor A means that, all other things being equal, setting this factor to this level resulted in a mean of m + lA1.

Means plot

Click the Means plot button to display a plot of means for each factor in the design. In this plot, optimum settings for each factor can be easily identified. Note that the 2 * standard error limits in this plot (indicated by the dashed lines) are based on the MS residual computed from the factors that were selected into the model the last time that either the Summary: Analysis of variance option on the ANOVA tab or Eta under optimum conditions option on the Accumulation Analysis Results - Optimum Eta tab were selected. By default (if neither one of these options has previously been used), all main effects are included in the model.