The term secular changes refers to increases or decreases in populations over time that are associated with dates of birth. An increase in median weight for 10-year-old females in a particular country from 1950 to 1990 is a secular change. Secular increases are said to be positive, whereas secular decreases are said to be negative. The terms increase and decrease describe changes in numerical values and not the possible consequences. A positive secular change in weight may have negative health and social consequences. The data needed to analyze secular changes must come from repeated surveys of a national, ethnic, or regional population spanning at least 10 years. These surveys must use the same sampling procedures and techniques of measurement. Migration to or from the study population between the surveys may invalidate the conclusions. Additionally, differences between surveys may be affected by selective changes in response rates, which could be associated with social pressures on those with unusual body size.

Data from repeated surveys of a defined population can be compared in several ways. Some report secular changes in means for measures of size during childhood (Bodzsár, 1998; Jaeger, 1998). This approach may be dictated by the nature of the earlier reports, but the utility of the information provided is limited. Usually, secular changes are much larger in the upper parts of the distributions than in the means; this may have health consequences in the case of weight, BMI, and skinfold thicknesses.