Published online by Cambridge University Press: 18 January 2010
Graduating error distributions by families of curves has a long and distinguished history. In seeking a class of frequency distributions which graduate navigational data, Anderson and Ellis are motivated by the well-known shortcomings of the normal distribution which so often fails to do justice to the data in the tails of the distribution. They generalize the one parameter (σ) zero-mean gaussian family to a two parameter (α, β) family which is in fact the Pearson Type VII class. They then observe that this class graduates published navigational distributions very wells.