Scientists tend to think that the media coverage of science is terrible, and they blame it on the journalists. They routinely back up their opinion by telling stories of failed encounters with reporters. One of my colleagues once spent more than 30 minutes on the phone answering a journalist's questions. When the story came out, the only trace of their conversation was one unrepresentative and largely meaningless quote. My colleague felt like he'd wasted his time and been made to look foolish in public. He's not alone; many scientists have stories like this. They're not surprised that science is often misrepresented and over-hyped in the news, because they expect that journalists are more likely than not to bungle the story.
But a recent study suggests that journalists aren't the weakest link. The source of misrepresentations and exaggerations in science news stories is often much closer to the scientists themselves: press releases put out by researchers' own institutions. Surveying hundreds of news stories and press releases about medical research, a group of scientists at Cardiff University found that most exaggerations and misrepresentations of science in print news "did not occur de novo in the media but was already present in the text of the press releases produced by academics and their establishments."
A group of scientists at Cardiff University found that most exaggerations and misrepresentations of science in print news "did not occur de novo in the media but was already present in the text of the press releases produced by academics and their establishments."
The findings are striking. When a press release had no exaggerations or misleading claims, relatively few—less than 20 percent—of the related news stories carried misleading claims. But when a press release did include an exaggerated or misleading claim, the majority of the associated news stories also featured exaggerations and misleading claims. The study wasn't designed to prove that press releases are the actual source of exaggerations in the news, but the evidence strongly suggests that they are. The authors found that, of the news stories that contained quoted sources, 72 percent of those stories had quotes taken from the press release. When press releases included specific numbers from a research paper, most news stories included those numbers as well. More than a third of the press releases examined contained misleading statements or exaggerations, so the bad influence of academic institutions on science reporting is very likely substantial. (It should be noted here that the best journalists will actually read the papers they cover and, if time allows, communicate with the scientists directly, but even then, bad press releases can still have an effect.)
Scientists might be tempted to lump university press officers in with journalists and dismiss them both as unreliable, but there is a problem with that: Research press releases are typically written with major input from the scientists themselves, and not issued without the scientists' approval. What we have here is a failure to communicate. Scientists are having trouble getting their story across to their own in-house press officers.
In a sense, this could be viewed as good news, because the problem is something that scientists and their institutions could easily fix. As the study authors wrote, "if the majority of exaggeration occurs within academic establishments, then the academic community has the opportunity to make an important difference to the quality of biomedical and health related news." The trick is to figure out how to improve the quality of university press releases.
Perhaps the biggest problem is getting scientists to care. Researchers generally don't read press releases, nor do they think of them as reliable sources of information. University public relations offices may issue them every day, but these cover only a fraction of all papers published by the university’s researchers. This means that most scientists will only deal with their institution's press office a handful of times over the course of a decades-long career. With researchers' interest and expectations so low, it's a challenge to get them invested in the process of putting out press releases. Ben Goldacre, a physician and the author of a Guardian column devoted to skewering bad science, suggested in a paper for the BMJ that this can be done by targeting scientists' most sensitive pressure point: their reputations. Goldacre proposed that the authors of press releases should be named, and that researchers should be included among them. "This would create professional reputational consequences for misrepresenting scientific findings in a press release," he wrote, "which would parallel the risks around misrepresenting science in an academic paper."
Even if scientists do become more invested in the press releases that discuss their research, there is still the challenge of communicating their message effectively while keeping it accurate. To be effective, writers need to understand their audience. When scientists write for their colleagues, they expect a hefty amount of skepticism. The result is somewhat paradoxical: Scientists will carefully temper their claims, to show their peers that they understand the limits of their data. At the same time, they'll make bold statements about the overall significance of their research program, knowing that their colleagues will put such statements in context and not be misled.
When scientists use the same rhetorical strategy with the public, they run the risk of misleading their audience. This seems to be especially true when it comes to relating non-human research studies to human health. Biomedical scientists who work with non-human animals or cultured cells can make a strong case that their work is relevant to human disease. But that work is unlikely to have a direct and immediate impact on actual clinical practice. In the press release study, statements that overemphasized the immediate human relevance of non-human studies were a particularly potent form of exaggeration: When a press release made such an exaggeration, nearly 90 percent of the associated news stories contained a similar exaggeration. When the press release didn't exaggerate the human relevance of the study, fewer than 10 percent of news stories exaggerated it.
It's no secret that communicating complicated research findings in a way that is accurate, compelling, and readable to the general public is difficult. Press officers are likely to have more experience with this than scientists, and so there is an opportunity for them to help scientists understand how their statements will be received by readers. But that requires scientists and press officers to work well together, which in turn depends on scientists taking the work of their university's PR office seriously. They should do so. The results of the press release study show that scientists have a chance to improve science coverage in the media, rather than just complain about it.