Skip to main content
Premium Trial:

Request an Annual Quote

Flawed Analysis Feud

Premium

Some academic sleuths say a research paper that claimed to find correlations between the expression of a set of 53 genes and happiness levels (as measured by a questionnaire) is flawed and based on a false positive finding.

Specifically, the investigators claim that the statistical method, called RR53, used in the 2013 paper, A Functional Genomic Perspective on Well Being, is flawed, according to Discover Magazine's Neuroskeptic blog.

"Even when fed entirely random psychometric data, the 'RR53' regression procedure generates large numbers of results that appear, according to these authors' interpretation, to establish a statistically significant relationship between well-being and gene expression," the authors state in their paper. Both of the papers were published in The Proceedings of the National Academy of Sciences.

"We believe that this procedure is, simply put, totally lacking in validity," they write.

The lead author of the reanalysis study and the aggressive words above is Nicholas Brown. He is a grad student at the University of East London who last year also debunked a psychology paper about a notion loosely termed the 'positivity ratio.'

The authors of the genomic perspective on well being paper, including lead author and University of North Carolina at Chapel Hill Professor Barbara Frederickson, have responded, also in PNAS. They say that the conclusions the critics have drawn contain serious factual errors. Their response also points out that the paper's analysis has been replicated, and that the technique Brown and colleagues used to critique the paper is invalid.

The Neuroskeptic author notes that he/she is cited in the acknowledgments on the Brown paper and served in an advisory capacity. The Neuroskeptic author also decided to look at the data again. Running 10,000 RR53 simulations using genetic data from the original paper but using random numbers in place of the happiness questionnaire scores returned a false positive rate of 55 percent, "far higher than it should have been," according to the Neuroskeptic blogger.

"In my opinion, whatever else may be right or wrong with Fredrickson et al.’s paper, their central analysis was flawed and their headline results are probably false positives."