Based on data from Altmetric.com. Altmetric is supported by Macmillan Science and Education, which owns Nature Publishing Group.

Every once in a while, an academic argument turns into a social-media spectacle. In 2013, a paper concluded that a person’s approach to happiness can shape gene expression (B. L. Fredrickson et al. Proc. Natl Acad. Sci. USA 110, 13684–13689; 2013). The study and its first author, psychologist Barbara Fredrickson of the University of North Carolina at Chapel Hill, received much media attention at the time, but a new report1 by US and UK researchers says that the findings were “artifacts of dubious analyses and erroneous methodology”.

Stuart Ritchie, a human-intelligence researcher at the University of Edinburgh, UK, evidently enjoyed the show, as revealed by his tweet:

“Ouch” tweeted John Foxe, a neurologist at Albert Einstein College of Medicine in New York City, in reply.

The 2013 paper by Fredrickson et al. looked at two different routes to happiness: hedonism, or the pursuit of pleasure, and eudaimonism, which the authors defined as “striving toward meaning and a noble purpose beyond simple self-gratification.” The researchers used personality questionnaires to gauge levels of hedonism and eudaimonism in 80 healthy volunteers. They also collected blood samples to measure the expression of an array of immune-system genes. The researchers concluded that the most hedonistic subjects — although they may have felt happy — showed physiological signs of stress and adversity. Compared with the eudaimonists, hedonists overexpressed genes that promote inflammation and underexpressed genes for antibody production. The latest paper reanalysed the data in several ways. Among other things, the researchers “replaced Fredrickson et al.’s psychometric data with random numbers and continued to find very large numbers of apparently statistically significant effects”. The authors concluded that the original statistical analyses were “fatally flawed, to the point that their claimed results are in fact essentially meaningless”.James Coyne, a health psychologist at University Medical Center in Groningen, the Netherlands — and a co-author on the latest study — carried the argument further in a PLoS blog post. He suggested that the Fredrickson paper was an “exceptional example of the kind of nonsense, pure bunk, you can find in a prestigious journal”. He specifically queried the way that the questionnaires were scored, and offers to donate US$100 to the American Cancer Society if the original authors “produce the unpublished analyses that justified this idiosyncratic scoring”. PNAS declined to comment.Fredrickson and her co-author Steve Cole, a psychologist and bioinformatician at the University of California, Los Angeles, did not respond to Nature’s request for comment. However, they did send a letter to PNAS saying that the rebuttal by Coyne and colleagues was itself full of statistical missteps. They write that they have already been able to replicate the study results and that they used an “extensively validated” questionnaire and “established” scoring techniques to measure levels of hedonism and eudaimonism.Coyne wrote in his blog post that the letter didn’t clear up his doubts. The $100 offer, he said, still stands.