clock menu more-arrow no yes mobile

Filed under:

"Unprecedented!" "Amazing!" "Novel!": the rise of hype in scientific journals


In the brutally competitive world of academic publishing, it can be hard to get noticed. So in recent decades, many researchers have turned to overhyping their studies.

A team of researchers from the Netherlands tried to quantify the rise in hype by studying the titles and abstracts of scientific papers published in the PubMed database between 1974 and 2014. They wanted to see how often adjectives such as "unprecedented," "amazing," "groundbreaking," and "promising" were used.

They found a ninefold increase in frequency during the period — and published their results in BMJ. This graphic from Nature tells the tale:


More specifically, the frequency of positive words increased from 2 percent in 1974 to 17.5 percent in 2014. By contrast, the frequency of negative words increased by a much smaller margin, from 1.3 percent to 3.2 percent.

As a control, the researchers randomly selected value-neutral words — such as "blood," "condition," "disease," and "experiment" — to see if they found a similar rise. But there, they found no such change. They also found no change in positive and negative language in published books. So the use of hyperbole seemed to be exclusive to the research papers.

"Apparently scientists look on the bright side of research results," the study authors concluded. "But whether this perception fits reality should be questioned." After all, too much hype can actually be misleading — or even harmful.

Journalists speak in superlatives, too

Journalists also seem to be employing hyperbole more frequently. Earlier this year, a group of American cancer researchers, writing in this week's issue of the journal JAMA Oncology, measured how often 10 superlatives were used to describe new cancer drugs in the press. (The list of terms included "breakthrough," "game changer," "miracle," "cure," "home run," "revolutionary," "transformative," "life saver," "groundbreaking," and "marvel.")

Those researchers also searched Google News to see when these phrases were used in connection with cancer drugs, and then read the articles to see if there was any evidence to back up the wild claims.

In half the cases, terms like "breakthrough," "miracle," and "cure" were used to describe medicines that hadn't even been approved yet by regulators. So not only were these drugs not yet available to patients, they also hadn't even passed scrutiny by the Food and Drug Administration.

Even worse, about 14 percent of the time these terms were used, the drugs actually had no human data behind them. This means journalists were calling a drug a "miracle" or "cure" when it had only been tested on mice or cell cultures.

So who's to blame for all this hype?

According to one interpretation of the new BMJ study, the pressure on scientists to find amazing discoveries in order to secure funding might be to blame.

As the study's author, Christiaan Vinkers, told Nature, their findings "fit our own observations that in order to get published, you need to emphasize what is special and unique about your study." So researchers may feel the need to pepper their work with more attention-grabbing language in order to stand out.

On the journalism side, a good chunk of the hype was coming from doctors who were quoted using language like "miracle" or "game changer" to describe new cancer drugs, often without good justification. This jibes with other studies finding that a great deal of medical hype in the press often comes from universities and hospitals themselves.

Whatever the reason, scientific hype in medicine has become a real problem. Overselling can not only mislead patients but can also help inform misguided policies, pressure regulators to speed through substandard drugs, and, worst of all, create false hope in people who may be desperate for cures that never come.

Sign up for the newsletter Sign up for Vox Recommends

Get curated picks of the best Vox journalism to read, watch, and listen to every week, from our editors.