Abstract
Generalized workspace and clothing design problems often arise in which two anthropometric constraints must be considered simultaneously in order to accommodate some specific target percentage of the population. In the theoretical (and unlikely) instance in which the two variables are perfectly positively correlated, the problem is readily solved using univariate percentile information. However, in the more realistic case in which the two variables are less than perfectly positively correlated, bivariate percentile charts are required. These bivariate charts are rarely available but may be computer-generated from the univariate data and the correlation between the variables after assuming some appropriate bivariate distribution (usually Gaussian). However, such computer access is not always available especially when a quick estimate is needed.
This paper presents a simplified approach to bivariate design based on the workable assumption that the bivariate target percentage will be met by using the same (to be determined) univariate cutoff value for each variate. This cutoff value depends on the target percentage value and on the degree of correlation between the variables and the assumption that the data are adequately represented by a Gaussian bivariate distribution.
The method takes advantage of simple charts prepared expressly for this purpose and several of which are presented herein. The method also has utility in a number of practical and common problems as well as being suitable for student use.
Get full access to this article
View all access options for this article.