On Wed, Jan 27, 2016 at 10:12:22AM -0500, Erik Bray wrote: > On Tue, Jan 26, 2016 at 9:47 AM, Joshua Ryan Smith Ph.D. > <[email protected]> wrote: > > Titus and Greg, > > Do you have a list of these cases on the web, or is this an informal thing > > between you two? I'd be interested in seeing the list. > > I used to use the example of the Reinhardt & Rogoff bungled Excel > spreadsheet in classes, until at one SWC workshop a learning in the > class--an economics professor--took me aside and told me that he knows > Carmen Reinhardt and Ken Rogoff personally. He asserted that the > overall impact of that error was overblown (I don't have enough > information personally to evaluate that). He also said that they > personally would not be happy about their research being misused to > justify austerity measures. Though I think their personal opinion is > beside the point--all that matters is how right-wing politicians and > pundits chose to interpret their flawed data. Though maybe in the > context of the original goals of their research, and ignoring the > context in which it was interpreted, the flaws may be less impactful.
The Reinhart & Rogoff spreadsheet error is a good object lesson for open data and reproducibility and automation, for sure -- http://www.bloomberg.com/bw/articles/2013-04-18/faq-reinhart-rogoff-and-the-excel-error-that-changed-history -- but I like the protein structure sign error as an example: http://www.nature.com/news/2010/101013/full/467775a.html One important point to make is this: we can make a good guess that there are many errors that are not caught, because nobody cares enough to find them and/or the original authors don't publicize them if they do find them. So we only hear about the big and contentious ones! (One of my papers from 2004 is missing the first gene in a list, due to an off-by-one error. Didn't change the conclusions, wasn't worth a retraction or a comment, but I did fix the source code.) best, --titus _______________________________________________ Discuss mailing list [email protected] http://lists.software-carpentry.org/mailman/listinfo/discuss_lists.software-carpentry.org
