Mis-placed inference: Circular logic in spatial prediction of obesity

Mark Daniel

    Research output: Contribution to journalEditorialpeer-review


    It has been established for several decades that the spatial distribution of chronic disease, including obesity, co-varies with spatial variation in the characteristics of populations and population context [1], [2], [3]. The article by Fitzpatrick et al. [4] asserts extending this substantial literature by showing such relations for 500 cities in the U.S. The letter by Feldman, in this issue, commendable for logic and conciseness, critiques this claim on methodological grounds. Feldman’s point can be restated thus: Fitzpatrick et al.’s conclusion, “… a clear connection between obesity prevalence, income inequality, and racial and ethnic population composition across census tracts in the 500 largest U.S. cities” is inconsequential as an artefact reflecting the use of these variables in the estimation of the obesity prevalence rates with which they were found to be associated.
    Original languageEnglish
    Pages (from-to)403-404
    Number of pages2
    JournalObesity Research and Clinical Practice
    Issue number5
    Publication statusPublished - Sept 2018


    Dive into the research topics of 'Mis-placed inference: Circular logic in spatial prediction of obesity'. Together they form a unique fingerprint.

    Cite this