Methodology makes a difference

In the United States, data from the National Health and Nutritional Examination Survey (NHANES) show sharp rises in overweight/obesity rates among children and adolescents between 1976-1980 and 1988-1994, and again, between 1988-1994 and 1999-2002.²⁰

In Canada, because of variations in the methods used to collect information on height and weight, it is difficult to pinpoint when overweight and obesity rates increased. Rates for 12- to 17-year-olds, based on directly measured height and weight, can be calculated for four reference years: 1978/79, 1981, 1988, and 2004 (Chart 17) (see Data sources and analytical techniques). These data reveal a small decrease in adolescents’ overweight/obesity and obesity rates between 1978/79 and 1981,²¹ and then, substantial increases between 1981 and 1988, and again, between 1988 and 2004. Calculations based on self-reported data show a stabilization of rates from 1994/95 to 2003. But between 2003 and 2004, when the collection method changed from self-reported to measured data, overweight/obesity and obesity rates rose sharply. This is not surprising, as self-reports tend to yield lower rates of overweight and obesity.^{2, 3, 4, 5}

Chart 17. Trends in overweight/obese and obesity rates, household population aged 12 to 17, Canada excluding territories, selected years 1978/79 to 2004

Comparisons of the average height and weight of adolescents (ages 12 to 17) in 2003 (self-reported) and 2004 (measured) illustrate these tendencies. In 2004, the average directly measured height of boys and girls was a third of an inch less than the 2003 averages based on self-reports. The directly measured average weight of boys in 2004 was 3 pounds more than in 2003, and for girls, 6 pounds more. As a result, one-year increases in the prevalence of overweight and obesity among adolescents were substantial.

Another problem with overweight and obesity rates based on self-reports is variation in the mode of collection. Self-reported data from face-to-face interviews result in higher obesity rates than do data collected from telephone interviews.²² In 1994/95, almost all interviews were conducted in person; in 1996/97 and 1998/99, most were by telephone; in 2000/01, it was approximately half and half; and in 2003, about one-quarter were in person.

For children aged 2 to 11, reported and measured data reveal a different bias. The data for 1994/95 to 2002/03 come from the National Longitudinal Survey of Children and Youth (NLSCY), in which parents reported the child’s height and weight (Charts 18 and 19). A comparison with the directly measured 2004 CCHS data suggests that when parents report on behalf of their children, overweight and obesity rates are higher, largely because parents tend to underestimate their children’s height. It is possible that they report the height when the child was last measured, which could be inaccurate given how quickly children of these ages grow. If height is routinely underestimated, the result would be substantial overestimates of overweight and obesity.

	Chart 18. Trends in overweight/obese and obesity rates, household population aged 6 to 11, Canada excluding territories, selected years 1978/79 to 2004
	Chart 19. Trends in overweight/obese and obesity rates, household population aged 2 to 5, Canada excluding territories, selected years 1978/79 to 2004