Oh well done. I did look for something like this, but didn't find it. Absolutely perplexed that a journal would not just give the authors the right to correct their own stuff regardless of how 'not serious' it might seem.
Nina Mazar posted some additional info on her OSF page about this paper.
From the comments on the pdf she uploaded:
"In this and the following paragraph, we mistakenly qualified pooled ANOVA SEs as SDs and the other way around. The corrections are noted below."
That might account for some or all of the discrepancies you found?
And from a separate summary file you can also see there:
"In December 2012 we notified Psychological Science about some errors in the reporting of the results, and requested guidance on how to submit a corrigendum. In response to our email, the editorial office informed us that our corrections were not serious enough to warrant publishing a correction. We therefore made publicly available for download an annotated version of the published paper, with our corrections in the text and in comments. This pdf used to be available since 2013 on Nina Mazar's website, while at the University of Toronto. That same file is now available on OSF."
I checked the wayback machine and that claim seems valid.
"It’s more that I find it vanishingly unlikely that there are two researchers who deserve to be in the line of fire while everyone else is a paragon of virtue."
Of course and your quoting of Stapel is bang on. But psychology, particularly of the variety of social psychology that has appeared in Psychological Science, is not serious about pursing best scientific or best publishing practices. SIPS is about protecting reputations from accountability for dubious publication practices. not improving social psychology.
For those percentages where they used SDs instead of SEs, using empirical SDs/SEs isn't ideal to compare percentages, because SD is determined by the percentage (binomial distribution). So it would be nice to do the comparison of the two percentages using Fisher's exact test, chi-squared test, or Barnard's test, and check that a similar (and statistically significant) p-value is found.
Here are some links:
https://osf.io/ryusz/
https://web.archive.org/web/20150915154355/http://www-2.rotman.utoronto.ca/facbios/file/PS_2010_MazarZhong-withCorrectionsComments.pdf
Oh well done. I did look for something like this, but didn't find it. Absolutely perplexed that a journal would not just give the authors the right to correct their own stuff regardless of how 'not serious' it might seem.
Nina Mazar posted some additional info on her OSF page about this paper.
From the comments on the pdf she uploaded:
"In this and the following paragraph, we mistakenly qualified pooled ANOVA SEs as SDs and the other way around. The corrections are noted below."
That might account for some or all of the discrepancies you found?
And from a separate summary file you can also see there:
"In December 2012 we notified Psychological Science about some errors in the reporting of the results, and requested guidance on how to submit a corrigendum. In response to our email, the editorial office informed us that our corrections were not serious enough to warrant publishing a correction. We therefore made publicly available for download an annotated version of the published paper, with our corrections in the text and in comments. This pdf used to be available since 2013 on Nina Mazar's website, while at the University of Toronto. That same file is now available on OSF."
I checked the wayback machine and that claim seems valid.
"It’s more that I find it vanishingly unlikely that there are two researchers who deserve to be in the line of fire while everyone else is a paragon of virtue."
Of course and your quoting of Stapel is bang on. But psychology, particularly of the variety of social psychology that has appeared in Psychological Science, is not serious about pursing best scientific or best publishing practices. SIPS is about protecting reputations from accountability for dubious publication practices. not improving social psychology.
For those percentages where they used SDs instead of SEs, using empirical SDs/SEs isn't ideal to compare percentages, because SD is determined by the percentage (binomial distribution). So it would be nice to do the comparison of the two percentages using Fisher's exact test, chi-squared test, or Barnard's test, and check that a similar (and statistically significant) p-value is found.