Re: Minitab decimal places - How to change?
Just found this thread and want to add some (hopefully useful) additional informations:
First:
You cannot change the numbers of decimal places in Minitabs session window (unfortunately).
For some methods / features you can get them if you store results in the worksheet, but not for ANOVA-tables. (That works for metrics such as mean Stat > Save descriptive statistics and e. g. coefficients from GLMs Storage > Coefficients).
Second:
You can get more information from your MSA-data if you run Stat > ANOVA > GLM in addition to Stat > Quality Tools > MSA > Gage R&R(crossed). Enter your data as response and write parts testers parts*testers in the model field. If the interaction parts*testers is not significant, you can delete it from the model field. To get the same model Minitab uses in the MSA-menu, you have to put additionally testers parts (only main factors without interaction term parts*testers) in the random field below the model field.
The ANOVA-table for GLMs provides one more decimal place in column SS, MS and F than ANOVA-tables for Gage R&R(crossed). The number of decimal places for p-values remains 3 in both outputs.
Third:
p-values and F-test-values contain different informations. Columns with SS, MS and F are based on the dispersion in your data/response due to different sources (parts, testers, parts*testers). p-values are probabilities which answer the question "does the changes in my response happen by chance (p>alpha) or is there a real effect due to changes in source-settings (p<alpha)?"
alpha is usually set to 5%=0.05 (risk for type I error), sometimes to 1% if the risk has to be smaller. (Values of alpha below 1%=0.01 are very unusual.) So if you take the three decimal places for p-values Minitab and many other statistical programs provide you can make the decision if a specific source has a real effect on your response.
