5.3.4. Normal Probability Plot

The normal probability plot generates the Y-Axis values from data and therefore only one data column is plotted at a time. If the data lies on a near-straight line, then it is said to conform to the normal distribution. By default, an Anderson-Darling Test of normality is also performed and reported in the legend. Smaller p-values indicate non-normality.

Edit → Data Series dialogue allows connecting data points with lines or drawing a line of best fit with or without confidence intervals. It is possible to plot probabilities or complementary probabilities.

Data itself is plotted on the X-axis with all scaling options available (see Scale Type) and the corresponding Y-axis (expected normal probability) values are computed from the inverse normal cdf employing a scale transformation and plotted with a probit scale. The following approximations to the normal scores are supported (see Blom, G. 1958 and Conover, W. J. 1980), where Blom transformation is the default:

Blom scores:

Tukey scores:

Van der Waerden scores:

Note that due to these transformations a Normal Probability Plot is different from X-Y Plots with a probit axis.