Stereoscopic Models of Multivariate Statistical Data

Abstract
Equations are given for the coordinates required in plotting stereographic projections of 3-dimensional objects. Applications are concerned with statistical data which represent unknown response surfaces or solids and where a visual model would aid analysis or interpretation. Model sizes compatible with low cost stereoscopic viewing devices are recommended and discussed. Illustrations include response surfaces in fertilizer trials, and an example of a multiple regression predicting mechanism. It is assumed that the labor of calculating coordinates and plotting projections have been factors limiting use of stereoscopic models. With computers in general use for statistical purposes, the calculations are no longer a chore. Hand plotting is equivalent to constructing ''scatter'' diagrams, already a vital and acceptable task with some statistical data. Electronic plotters capable of fully automating the task are on the market.