Scholarlink

<p>First proposed in 1961, Taylor’s Power Law relates variance (S<sup>2</sup>) and mean (m), expressed as S<sup>2</sup> = am<sup>b</sup>. While this relationship holds in ecology and other fields, it has seen less application in population health. Here, using Taylor’s Power Law, we assess the relationship between variance and mean body mass index across populations in low- and middle-income countries. We extracted adult body mass index data from 236 nationally representative, population-based samples of women aged 20–49&#xa0;years and men aged&#xa0;18&#xa0;years and older, collected between 1994 and 2019 (n = 2,594,211 women, n = 455,479 men). We used log–log linear regression to estimate the scaling exponent (b) and intercept parameters. In women, Taylor’s Power Law showed a strong fit (b = 4.0, R<sup>2</sup> = 0.72, n = 202 surveys), compared to b = 3.9 and R<sup>2</sup> = 0.46 in men (n = 34 surveys). Across survey periods, intercepts increased, while scaling exponents remained stable or declined in some subpopulations, suggesting rising baseline variability in body mass index alongside a slowing of variance growth with continued mean increases. The strength of Taylor’s Power Law relationship varied by social and spatial context, including age, residence, socioeconomic status, and survey period. This approach may offer a useful framework for describing and monitoring population-level variability in body mass index.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Body mass index variance scales with the population mean according to Taylor’s Power Law in 236 survey samples from 68 countries

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年8月5日

查看全文
  • Goutham Swaminathan,
  • Fahad Razak,
  • Daniel J. Corsi

摘要

First proposed in 1961, Taylor’s Power Law relates variance (S2) and mean (m), expressed as S2 = amb. While this relationship holds in ecology and other fields, it has seen less application in population health. Here, using Taylor’s Power Law, we assess the relationship between variance and mean body mass index across populations in low- and middle-income countries. We extracted adult body mass index data from 236 nationally representative, population-based samples of women aged 20–49 years and men aged 18 years and older, collected between 1994 and 2019 (n = 2,594,211 women, n = 455,479 men). We used log–log linear regression to estimate the scaling exponent (b) and intercept parameters. In women, Taylor’s Power Law showed a strong fit (b = 4.0, R2 = 0.72, n = 202 surveys), compared to b = 3.9 and R2 = 0.46 in men (n = 34 surveys). Across survey periods, intercepts increased, while scaling exponents remained stable or declined in some subpopulations, suggesting rising baseline variability in body mass index alongside a slowing of variance growth with continued mean increases. The strength of Taylor’s Power Law relationship varied by social and spatial context, including age, residence, socioeconomic status, and survey period. This approach may offer a useful framework for describing and monitoring population-level variability in body mass index.