The secular trend in the height of primary school children in England and Scotland from 1972-1980.

Data from 29 230 children in 22 areas in England and 6 in Scotland were used to estimate the secular trend in height of primary school children over the period 1972-1980. The children were from 15 birth cohorts, born 1961-1975. The method of analysis was selected in order to give estimates from mixed longitudinal data, in which the number of measurements differed from cohort to cohort and children with incomplete data could not be assumed to be comparable to those measured at every age. When the overall goodness-of-fit to height data of polynomials in age was examined, for children measured on 7 or 8 occasions, it was found that cubic polynomials were necessary to describe the growth of boys and girls over the age range 5.0 to 11.0 years, but that cubic and quadratic coefficients could not be interpreted for individual children. No evidence of a mid-growth spurt in height was found. Models fitted to mean heights showed that there was a secular trend of increasing height over the period studied, which was greater when estimated at age eight years than at age five. At age 8 the trend was greatest in Scottish boys, 1.5 cm per decade, and least in English girls, 0.5 cm per decade. The data do not distinguish between a trend due to earlier maturation and a trend resulting in increased adult height; further data will show whether the lack of positive trend at age five estimated from the later cohorts was due to a recent cessation of secular trend.