James Heathers is a postdoctoral fellow at Northeastern University. He looks for mistakes to have fun. He talks to Lulu Garcia-Navarro, of NPR, about errors published in scientific articles.
LULU GARCIA-NAVARRO, HOST:
We will now examine the persistent strength of a bad science and a group of volunteers who are trying to fight it. Take the famous study done in 1998 by a British scientist who suggested a link between childhood vaccines and autism. It was removed more than a decade later. And its author, Andrew Wakefield, has had his medical license revoked. But the myth that vaccines cause autism persists today with serious consequences for public health. It's that kind of bad science that James Heathers is constantly researching. He is a postdoctoral researcher in behavioral science at Northeastern University. And he had a decisive influence on the publication errors and the retraction of these documents. Heathers says that there are two ways to tell a paper that it might have a problem.
JAMES HEATHERS: The first is your sense of spider as a researcher, which is very difficult to explain. I guess it's a series of heuristics, a kind of mental shortcuts on what might not be right with something when you read it. And often it directs you to the second thing you do. And the second thing you do is try to mathematically determine if there is something inconsistent in the document, if there is a test that does not work or numbers that can not exist or something similar. Between the two and, of course, the fact that researchers will sometimes tell you if they have papers in their field that they do not trust – they think – have problems with – sometimes they'll tell someone else Like me.
GARCIA-NAVARRO: But is there not already a system in place for this? I mean, if you are published in a peer-reviewed journal, it's literally that you publish your article in the scientific community, and that people then have the opportunity to comment.
HEATHERS: Well, that's absolutely what it is. And this is not the end of the story. Peer review is often necessary and very good. The problem with really bad research is that they are not really designed to detect problems. This is done on the premise that everything underpinning the research is going well. This is happening under some kind of trust umbrella. There is not really a culture of scientific criticism of bad science in peer reviews.
The other thing that happens, of course, is that if you write an article that is deeply problematic and that is rejected by a newspaper and the journal writes to the author and says: no, look. We do not believe it. It's wrong. This was done incorrectly. These have errors that we can not explain – all that the authors will do, is to send them elsewhere. This is sometimes called the newspaper shopping. This happens, of course, with perfectly correct and accurate papers. But finally, there is one – there is a kind of saying in science that says that almost anything can be published somewhere.
GARCIA-NAVARRO: James, Wakefield's study was finally withdrawn. But science is a discipline that builds on itself, is not it? – with new research citing old research. Have you ever found something like that, something so serious that could have an effect on public health?
HEATHERS: There are things that look a lot like this.
GARCIA-NAVARRO: It's very disturbing.
HEATHERS: Well, if that sounds scary, it probably should be. Bad medical studies can yield greater results in one way or another. So imagine that we have six studies and one of them is a problem for some reason. And when we add the results of the six studies and we do a meta-analysis, one and the same bad study changes the result of the meta-analysis. This can affect how governments buy drugs, how front-line hospital care is delivered, and what is considered standard care based on evidence or not. At worst, there are problems of this type that currently exist. I'm afraid I can not speculate more than that.
GARCIA-NAVARRO: At this moment.
HEATHERS: Right now, watch this space.
GARCIA-NAVARRO: James Heathers is a member of Northeastern University. And he is the host of a podcast "Everything Hertz". Everyone has a podcast. Thank you so much.
HEATHERS: (laughs) You're welcome. Thank you.
Copyright © 2019 NPR. All rights reserved. Visit the Terms and Conditions and Permissions pages of our website at www.npr.org for more information.
NPR transcripts are created on a deadline by Verb8tm, Inc., an NPR contractor, and are produced using an exclusive transcription process developed with NPR. This text may not be in its final form and may be updated or revised in the future. Accuracy and availability may vary. The authoritative recording of NPR programming is audio recording.