Producing normal quantile plots in Minitab.

A Probability Plot in Minitab serves the same purpose as a normal quantile plot as described in the text. The default format for probability plots in Minitab differs from what the text uses in three ways:

• the axes are transposed so the data values are on the horizontal axis and the expected value is on the vertical axis in Minitab's default
• the expected value is reported as a percentile rather than a z-score in the Minitab default
• curves corresponding to 95% confidence intervals are included in the Minitab default.

You can use the Minitab default. The central question remains the same: To what degree do the points fall along a straight line? The more the points fall along the line, the more closer the distribution is to being normal. With the Minitab default, granularity will show up as points lined up vertically. With the text's choices, granularity will show up as points lined up horizontally. You don't need to worry about the precise meaning of the 95% confidence interval curves. You can use these in making a judgment about normality. If most of the points fall within the curves with a non-systematic scatter, assuming normality is probably fine.

Here's steps you can take to have Minitab produce a plot in the same style as the text:

• Go to Graph:Probability Plot....
• Hit OK for the default choice of Single.
• In the dialog box, set up your desired graph variable.
• In the dialog box, hit Scale....
• In the Scale dialog box, check the Transpose X and Y option on the front Axes and Ticks tab.
• In the Scale dialog box, go to the Y-Scale Type tab and choose Score.
• Hit the OK button to return to the main dialog and then OK again to get the graph.

If you want to get rid of the confidence interval curves, go to Distribution... on the main dialog box (next to the Scale... button you used above). Go to the Data Display tab and uncheck Show confidence intervals.