Background: The causes of the "dual burden" of stunting and obesity remain unclear, and its existence at the individual level varies between populations. We investigate whether the individual dual burden differentially affects low socioeconomic status Peruvian children from contrasting environments (urban lowlands and rural highlands), and whether tibia length can discount the possible autocorrelation between adiposity proxies and height due to height measurement error. Methods: Stature, tibia length, weight, and waist circumference were measured in children aged 3-8.5 years (n=201). Height and body mass index (BMI) z scores were calculated using international reference data. Age-sex-specific centile curves were also calculated for height, BMI, and tibia length. Adiposity proxies (BMI z score, waist circumference-height ratio (WCHtR)) were regressed on height and also on tibia length z scores. Results: Regression model interaction terms between site (highland vs. lowland) and height indicate that relationships between adiposity and linear growth measures differed significantly between samples (P<0.001). Height was positively associated with BMI among urban lowland children, and more weakly with WCHtR. Among rural highland children, height was negatively associated with WCHtR but unrelated to BMI. Similar results using tibia length rather than stature indicate that stature measurement error was not a major concern. Conclusions: Lowland and rural highland children differ in their patterns of stunting, BMI, and WCHtR. These contrasts likely reflect environmental differences and overall environmental stress exposure. Tibia length or knee height can be used to assess the influence of measurement error in height on the relationship between stature and BMI or WCHtR. Am. J. Hum. Biol. 26:481-490, 2014. © 2014 Wiley Periodicals, Inc.