A Graphical Technique for Enhancing Scatterplots with Moving Statistics

Abstract
A graphical technique is presented which enhances the information on scatterplots by plotting three smooth cluves of movie statistics which slunmarize the change in the distribution of Y given X. The three curves involve the tue of the midmean ithe average of all observations betweell the qluxtiles), and two new statistics: the lower semi-midmean (the midmenn of all observations below the median), and the upper semi-midmean (the midmean of all observations above the mediate). The technique is used in the analysis of air pollutiou data to show that emissions of primary pollrltnnts in the Camden-Philadelphia urban complex are partly responsible for high ozone concentratiotls 37 km tO the southeast in Ancora, New Jersey, a nonluban area, where concentrations of primary pollutants are low, but where ozone daily maxima freqltently exceed the Federal Standard.