The study, published in the American Journal of Preventive Medicine, suggests that around one in 11 publications on nutrition and obesity in leading journals may overstate the results of their research findings - such as inappropriately describing a correlation as a cause-and-effect relationship and generalising a study’s claims to large groups of people even when the study population was quite different.
"This is troubling because such statements likely influence policymakers, clinicians, other researchers, and the public into making decisions without an accurate understanding of the supporting science, which may have unjustified costs," said the research team - led by Professor Nir Menachemi from the University of Alabama at Birmingham in the US.
"They also have the potential to be amplified and disseminated to a larger audience when they are reported by journalists, who are a key source for public information about scientific discoveries," said the team. "Given that, by and large, journalists dutifully convey the claims made in scientific papers, overstatement of results poses a concern."
According to the results of the study, public health journals had a 'significantly higher' prevalence of publishing overreaching statements when compared to medical, nutrition, and obesity journals, "especially with respect to reporting associative relationships as causal or making policy recommendations based on observational data."
Menachemi and colleagues tracked how often authors overreached in the summary of their findings by searching research articles published in eight leading journals in either 2001 or 2011, in order to examine changes in reporting over time. The team found 937 papers—377 from 2001 and 560 from 2011.
In 8.9% (around one in 11) of the studies, the team reported that findings had been overstated in the abstract - with studies from 2011 more likely to overreach than 2001 papers.
Overreaching statements were also found to be more common in unfunded studies compared to funded studies, regardless of what type of group paid for the study; while a higher number of co-authors was also associated with a reduced likelihood of presenting overreaching statements.
"This trend may be because funded researchers are selected for superior knowledge or skills; have greater resources (as a result of their funding); or are subject to oversight from the funding agency, all of which may translate into a more straightforward presentation of their scientific work," wrote Menachemi and colleagues.
Although those overstatements may be unintentional, they can distort what doctors, policymakers, and the general public know about nutrition, the researchers said.
The team noted that their work is an extension of a project originally funded by The Coca Cola Company, but noted that the Company had no role in the design, execution, or reporting of the current study.
Source: American Journal of Preventive Medicine
Volume 45, Issue 5, November 2013, Pages 615–621, doi: 10.1016/j.amepre.2013.06.019
" Overstatement of Results in the Nutrition and Obesity Peer-Reviewed Literature"
Authors: Nir Menachemi, Gabriel Tajeu, Bisakha Sen, et al