Statview

Note: StatView currently has no procedure for automatically generating the Normal Quantile plot. First use Sort under the Manage menu to order the observations from smallest to largest. Then use Formula under the Manage menu to construct a variable p = (i-.5)/n where i is the rank order of the observation (from smallest to largest). Next, use Formula to construct a new variable Quantile = ReturnNormal(p,0,1). Finally, use one of the scattergram options to plot the Data variable against the Quantile variable.

Note: JMP's procedure for plotting the Normal Quantile plot also gives a confidence band around the best fitting line through the points. Regression lines are explained in Chapter 12. The intercept of the line equals the mean (the intersection of the red and greed lines above) and the slope equals the standard deviation. The confidence band around this line is not a useful test of the normality assumption. Also, this plot displays key percentiles across the top of the graph.

Important Note: The Normal Quantile procedure in Minitab reverses the role of the vertical and horizontal axes compared to most programs.