Abstract
Stata tip 141: Adding marginal spike histograms to quantile and cumulative distribution plots
Highlights
Such additions help clarify what a quantile plot means and how it can be explained as stacking values in terms of their associated cumulative probabilities
Marginal histograms add no more information to a quantile plot
They offer a complementary view of each distribution, affording another way to think about distribution level, spread, and shape—and fine structure too
Summary
Some intriguing examples given by Harrell (2015) prompt experiment with adding marginal spike histograms to quantile and (empirical) cumulative distribution (function) plots. This tip explains why this might be helpful and gives sample code using the official command quantile. Such additions help clarify what (for example) a quantile plot means and how it can be explained as stacking values in terms of their associated cumulative probabilities. Marginal histograms add no more information to a quantile plot They offer a complementary view of each distribution, affording another way to think about distribution level, spread, and shape—and fine structure too. See Galton (1889, 38) for (in modern terms) a quantile plot and a histogram sharing the same vertical magnitude axis
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have