Estimating the Covariates of Historical Heights

Abstract

Data on human height can provide an index that may measure more accurately changes in the standard of living than the more conventional real wage index. Height data, like those on real wages, are relatively abundant and extend back to the seventeenth century. In a previous paper, we developed and tested procedures for estimating the mean and standard deviation of the distribution of human height when the sample is distorted to an unknown extent by missing observations at lower heights. The purpose of this analysis is to extend our techniques so that the covariates of height can be estimated. Such an extension is necessary when trying to draw inferences about the causes of shifts over time in the height distribution so that changes in sample composition can be controlled.