5 Comments »Posted on Thursday 14 May 2009 at 6:43 pm by Jacob Aron
In Getting It Wrong, Health & Medicine

ResearchBlogging.org

We complain about it all the time. Journalists take a small study with some preliminary findings and write it up as the story of a century. The sensationalising of science news is certainly a problem in the media, but a new study suggests perhaps we are too quick to blame the journos.

A paper published in the Annals of Internal Medicine examines the content of 200 randomly selected press releases from 20 academic medical centres in the US. The analysis by lead authors Drs Steven Woloshin and Lisa Schwartz shows that press officers are just as bad when it comes to exaggeration.

The press releases split in to 113 that focused on human research with the remaining 87 covering animal or laboratory research. On the human side, 40% reported on studies limited by factors like small sample sizes. Of the same group, 42% failed to provide caveats explaining the limits of the research.

Things get worse for the animal and laboratory studies press releases. Despite the majority claiming the relevance of the research to human health, 90% failed to mention potential difficulties in extrapolating the results to people.

In total, 29% of releases were rated by the authors as exaggerating the importance of research. Animal research was more likely to be exaggerated than human. It’s not just the press officers grandstanding however. Most press releases contain quotes from the scientists involved, and 26% of these were found to overstate research importance.

The authors admit that their findings would be stronger if backed up by an analysis of the press coverage resulting from these releases, but say the study is still important because press releases are known to be influential. A previous study suggests that as many as one third of news stories rely mostly or completely on a press release.

S. Woloshin, L. M. Schwartz, S. L. Casella, A. T. Kennedy, & R. J. Larson (2009). Press Releases by Academic Medical Centers: Not So Academic? Annals of Internal Medicine, 613-618


  1. 5 Comments

  2. Many years ago for a Rhetoric of Science and Technology course, I wrote a paper that showed press releases were primarily a form of epideictic rhetoric – used to praise scientific discoveries, not just inform. I still think it’s a pretty sound theory.

    By Alison on Thursday 14 May, 2009 at 8:36 pm

  3. You might be interested in the commentary on this paper that appeared in the Columbia Journalism Review’s “The Observatory” blog at http://www.cjr.org/the_observatory/to_report_or_repeat.php?page=all.
    There was an earlier piece that also addressed news media response to preliminary research and their inability to distinguish between coorelation and causation. Find it at http://www.cjr.org/the_observatory/facebook_and_procrastination.php?page=all.

    By Earle Holland on Thursday 14 May, 2009 at 8:59 pm

  4. Of course press releases are going to talk up their institution’s research. That’s their job. Journalists have a responsibility to critically examine the claims made in the press release rather than just reproducing them.

    By Sam Wong on Friday 15 May, 2009 at 12:04 am

  5. Thanks for highlighting this original analysis from Woloshan et al., I’ll be looking forward to read it up soon. It is interesting to me that academic press releases – by their vociferous promotion of published and unpublished institution-based research – are habitually misrepresenting data in different ways (i.e.: omiting methodology flaws or limits in human trials). This adds to the fact that good portions of health professionals AND science journalists are poor “health educators”. While physicians may have trouble explaining coronary heart disease in simple terms to 9th grade-educated eldery patients, science journalist may lose subtle (yet important) insights in reporting a RCT with multiple outcomes to his-her readers.

    Earl, I do believe Ben Goldacre, at badscience.net has more to say on the press’ (and even scientist’s) trouble in distinguising correlation and causation.

    By Nicolas (BScPharm) on Friday 15 May, 2009 at 5:40 am

  6. Sam Wong is quite right: a Press Release can’t be expected to be objective, but we expect a journalist to be able to cut through the rhetoric and get to the truth.

    The problem is that, especially in science and also in other spheres, press releases routinely just end up being re-worded and published under journalist’s names.

    Nick Davie’s book Flat Earth News documents this in great detail (not specifically about science) and attributes it mainly to economics, it’s a lot cheaper to pay a journalist for 10 minutes to rewrite a PR than for 6 hours to really research something.

    By Neuroskeptic on Friday 15 May, 2009 at 10:36 am

Sorry, comments for this entry are closed at this time.