2 thoughts on “Coding error sinks cancer study”

  1. I would suggest that the retraction in question was due to a programming error rather than to a coding error. Coding error implies that an incorrect variable value i.e code was assigned to the raw data during data management which is not case here.
    A simple coding error example would be reversing the variable values/codes where 1=Female and 2=Male

  2. Not germane to this paper, but I can’t help telling the worst ‘coding error’ (in the sense of Stratis Ioannou’s post above) story I’ve personally experienced:

    I was collecting data from about 15 labs for a big meta-analysis of diabetes pedigrees. Only after much head-scratching about discrepancies in the results did I discover (I don’t recall how anymore) that 14 labs used “M” for “male” and “F” for “female” and 1 lab used “M” for “mother” and “F” for “father”!

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.