Abstract
Mid-parental heights are widely used to help assess an individual child's growth. However, the methods in use vary, and most make no allowance for extremes of parental height. This study aimed to examine the actual distribution of parental heights in a survey population and the relation with their children's heights. The heights of 419 representatively sampled children aged 8-9 years were compared with their reported mid-parental heights, all expressed as standard deviation scores (SDS). These confirmed previous predictions that 90% of the children's heights would fall within 1.5 SDS (approximately two centile spaces) of their mid-parental heights. However, where parents were unusually tall or short, their children were relatively less tall or short, respectively, and the mid-parental height was a poor predictor of attained height. A simple calculator for expected height centile is described that automatically adjusts for this regression to the mean. Of 13 children below the second centile for height, eight were within two centile spaces (90% range) of their mid-parental height SDS. However, when allowance was made for regression to the mean, only three of 13 were within the 90% range. Although mid-parental height provides a useful guide to expected height centile for children and parents of average stature, it can be misleading when used to assess short children.