Sunday, January 03, 2010

Height and Aging, 33 and 1/3

This post is an extension of Kevin’s previous post on heights. I cannot find a longitudinal dataset which contains information on objectively measured height over the life-cycle. Despite this, I present some cross-sectional birth-cohort data which is consistent with the hypothesis that aging effects height.

I use two samples from the Health Survey for England, which are ten years apart. Both surveys contain objectively measured height data for a large, nationally representative, sample of the English population. There are over 13,000 observations in each study. Most importantly, height is objectively measured under the same clinical conditions and using the same equipment in both surveys. Details of the height measurement procedure are contained in the Health Survey for England User Guide (AFAIK).

The figure below shows the quadratic prediction of height as a function of year of birth controlling for gender. This figure is both consistent with Kevin’s previous post, which showed no height loss at age 42 (1963), and the hypothesis that there is a negative relationship between height and age, the divergence of the trends as age increases.

1 comment:

Alan Fernihough said...

The above result is not driven by differential mortality, unless you believe that shorter men and women live longer. In fact, I would argue that differential mortality is underestimating the height-loss effect since the height-health gradient is positively sloping.