Christine L Gray1, Whitney R Robinson. 1. From the aDepartment of Epidemiology, University of North Carolina at Chapel Hill, Gillings School of Global Public Health, Chapel Hill, NC; and bUniversity of North Carolina at Chapel Hill, Carolina Population Center, Chapel Hill, NC.W.R.R. and C.L.G. were supported by the 2012 Junior Faculty Development Award, University of North Carolina at Chapel Hill.
Abstract
BACKGROUND: In childhood obesity research, the appearance of height loss, or "shrinkage," indicates measurement error. It is unclear whether a common response--excluding "shrinkers" from analysis--reduces bias. METHODS: Using data from the National Longitudinal Study of Adolescent Health, we sampled 816 female adolescents (≥17 years) who had attained adult height by 1996 and for whom adult height was consistently measured in 2001 and 2008 ("gold-standard" height). We estimated adolescent obesity prevalence and the association of maternal education with adolescent obesity under 3 conditions: excluding shrinkers (for whom gold-standard height was less than recorded height in 1996), retaining shrinkers, and retaining shrinkers but substituting their gold-standard height. RESULTS: When we estimated obesity prevalence, excluding shrinkers decreased precision without improving validity. When we regressed obesity on maternal education, excluding shrinkers produced less valid and less precise estimates. CONCLUSION: In some circumstances, ignoring shrinkage is a better strategy than excluding shrinkers.
BACKGROUND: In childhood obesity research, the appearance of height loss, or "shrinkage," indicates measurement error. It is unclear whether a common response--excluding "shrinkers" from analysis--reduces bias. METHODS: Using data from the National Longitudinal Study of Adolescent Health, we sampled 816 female adolescents (≥17 years) who had attained adult height by 1996 and for whom adult height was consistently measured in 2001 and 2008 ("gold-standard" height). We estimated adolescent obesity prevalence and the association of maternal education with adolescent obesity under 3 conditions: excluding shrinkers (for whom gold-standard height was less than recorded height in 1996), retaining shrinkers, and retaining shrinkers but substituting their gold-standard height. RESULTS: When we estimated obesity prevalence, excluding shrinkers decreased precision without improving validity. When we regressed obesity on maternal education, excluding shrinkers produced less valid and less precise estimates. CONCLUSION: In some circumstances, ignoring shrinkage is a better strategy than excluding shrinkers.
Authors: Sarah E Barlow; Sonal R Bobra; Michael B Elliott; Ross C Brownson; Debra Haire-Joshu Journal: Obesity (Silver Spring) Date: 2007-01 Impact factor: 5.002
Authors: Michael Spratt; James Carpenter; Jonathan A C Sterne; John B Carlin; Jon Heron; John Henderson; Kate Tilling Journal: Am J Epidemiol Date: 2010-07-08 Impact factor: 4.897
Authors: Robert J Kuczmarski; Cynthia L Ogden; Shumei S Guo; Laurence M Grummer-Strawn; Katherine M Flegal; Zuguo Mei; Rong Wei; Lester R Curtin; Alex F Roche; Clifford L Johnson Journal: Vital Health Stat 11 Date: 2002-05
Authors: Cynthia L Ogden; Robert J Kuczmarski; Katherine M Flegal; Zuguo Mei; Shumei Guo; Rong Wei; Laurence M Grummer-Strawn; Lester R Curtin; Alex F Roche; Clifford L Johnson Journal: Pediatrics Date: 2002-01 Impact factor: 7.124
Authors: David S Freedman; Hannah G Lawman; Asheley C Skinner; Lisa C McGuire; David B Allison; Cynthia L Ogden Journal: Am J Clin Nutr Date: 2015-09-16 Impact factor: 7.045
Authors: David S Freedman; Hannah G Lawman; Liping Pan; Asheley C Skinner; David B Allison; Lisa C McGuire; Heidi M Blanck Journal: Obesity (Silver Spring) Date: 2016-03-17 Impact factor: 5.002