Advertisement
Research Article

The Relationship of Sugar to Population-Level Diabetes Prevalence: An Econometric Analysis of Repeated Cross-Sectional Data

  • Sanjay Basu mail,

    basus@stanford.edu

    Affiliation: Stanford Prevention Research Center, Department of Medicine, Stanford University, Palo Alto, California, United States of America

    X
  • Paula Yoffe,

    Affiliation: Department of Integrative Biology, University of California, Berkeley, California, United States of America

    X
  • Nancy Hills,

    Affiliation: Department of Epidemiology & Biostatistics, University of California San Francisco, San Francisco, California, United States of America

    X
  • Robert H. Lustig

    Affiliations: Department of Pediatrics, University of California San Francisco, San Francisco, California, United States of America, Philip R. Lee Institute for Health Policy Studies, University of California San Francisco, San Francisco, California, United States of America

    X
  • Published: February 27, 2013
  • DOI: 10.1371/journal.pone.0057873

Reader Comments (8)

Post a new comment on this article

Are the regressions population weighted?

Posted by jaylusk22 on 07 Mar 2013 at 04:33 GMT

I am curious to know if the regressions are population weighted as indicated in the text? When I download portions of the data (admittedly not the exact same as the authors are using), I get significant differences in the weighted and unweighted models with the unweighted results being more similar to those reported in the text. I report the differences here:
http://jaysonlusk.com/blo...

No competing interests declared.

RE: Are the regressions population weighted?

jaylusk22 replied to jaylusk22 on 07 Mar 2013 at 10:52 GMT

I believe that I answered my own question. The above comment was in reference to regressions run using "national prevalence" as the dependent variable. When I switch and use "IGT comparative prevalence", which is apparently some sort of age-adjusted measure, then i find more similar results to that reported in the paper when the weights are applied. Interestingly, the sugar result is not significant in the unweighted model.

No competing interests declared.