The trouble with science journalism

The trouble with science journalism is that for various reasons—including, but not restricted to laziness—it tends to uncritically accept the claims made in a paper. A case in point was an article that I picked up on recently because it's about work in my field, and I remembered the research in question from a conference last year.

The work in question is a pretty cool project at Cornell. Now as a piece of swarm robotics work this stuff is pretty much cutting edge - they've managed to get simple individual units to perform an organised group behaviour with very limited processing capability on each unit, and only local control. I have as yet to see what's so exciting about the state of the art in swarm robotics, but this work is state of the art.

However, New Scientist, BBC News and Ars Technica have all uncritically repeated, this work is being billed as Self Replication.

As I (and I suspect most people) understand it self replication means what the vast majority of biological organisms do: re-organising the materials in their environment to creat copies of themselves. In a sense, the Cornell modular robots are doing this, but the catch is in what materials they take from the environment. The environment in which this robot exists is pre-seeded with fully functioning robot modules, whereas clearly a baby is not assembled from pre-made parts. This system is equivalent to if humans wandered round in a world full of working limbs and torsos and heads, and we simply picked the right set of parts from the environment and assembled them to make a child.

This is not a damning criticism of the work itself—although I do believe they are overselling their work by playing semantic games—because in their Nature paper (free to download and only a page long, so read it yourself) they make a subtle argument about a continuum of self-replication, on which this work represents a point close to the simplest extreme. In their terms it is at least arguable that these robots are self-replicating. However, I do think it's getting reported in a highly misleading way, from the Nature paper being title Self-reproducing machines, through the New Scientist article describing these machines as self-cloning, and the Beeb parroting that [t]he experiment shows reproduction is not unique to biology. There is a pretty huge gap between what the experiment really involved and what these reports make it sound like it does.

I don't intend this as a tirade against either the Cornell group (who have put what I think is a better characterisation of their work online in a FAQ) or science journalists. Unfortunately scientists do have to hype their work in the never-ending competition for funding, and non-specialist publications have no choice but to get stories written by people who are not experts in that particular field. I think that to a small extent this does represent a failure of the peer-review process, in that the reviewers of that Nature paper should have insisted that the article's title and subheader be changed to something less dramatic, but again this is relatively small issue considering that the article's content is good. My main point is that we as scientists have to present the counter-point when we do see this kind of misrepresentation happening in a field we really know about, and as lay readers we have to take everything outside our area of expertise with a large grain of salt. It's not even safe to assume that peer-reviewed articles are free of errors and over-grandiose claims, let alone articles in the daily papers.


Trackback URL for this entry is: How science reporters should do their job
Excerpt: About a year ago, Eldan Goldenberg had a post complaining about the lousy job that reporters do when covering science....
Weblog: Mano Singham's Web Journal
Tracked: August 22, 2006 04:52 AM


Even when they are critical of the research, though, it's usually not in a very intelligent way.

Posted: May 18, 2005 10:21 PM

The press will also jump on any words they think will grab readers' attention. Especially words like "cold fusion" as allegedly related to sonoluminescence around 2002.

Posted: May 19, 2005 06:48 AM


I agree with the criticism of science reporting that you raise but the causes may be deeper than just cowardice or lack of understanding of science by journalists and their editors. The analyses of the media by people like Ben Bagdikian (The New Media Monopoly) and others suggest that the reasons for this form of "balanced reporting" are structural, not personal.

On another note, I noticed on Planet Case that there appears just an excerpt to the post and people have to click the link for the full article. This is a nice feature. How did you do it?

Also, the header on your blog with your name etc. is messed up when I view it, with some words overflowing into other areas. I am using a Mac with Safari browser, and that may be the cause of the problem but I thought I'd let you know in case other browsers also have problems.

Posted: May 19, 2005 11:36 AM

Tom: thanks for the link. I was aware of the issue, but that's by far the most detailed critique I have seen of it, so it will be useful.

Sherri: absolutely. I see it particularly clearly in my own field - when I give a talk to non-specialists one thing I stress a lot is how far short the real state of the art is from realising the works of Isaac Asimov.

Mano: do you have either links or book references for the sources you mention? It sounds like I should read some Bagdikian.

Posted: May 19, 2005 12:57 PM

regarding the technical issues with the blog:

There are two text boxes on the "New Entry" form in Movable Type (or at least in the blog@case install of MT, which is the only one I've used). Whatever you put in "Entry Body" becomes the part that shows up as an excerpt, and whatever you put in "Extended Entry" becomes the rest of the body that shows up when people click on the link. The only problem with this is that my RSS feed just shows the excerpted paragraph without a clear indication that there is more to be read - if anyone knows how to change this I'd be grateful if you could tell me.

Thanks for pointing that out about the header. It happens in Firefox too, and I know that a fair proportion of my readers do use one of those two browsers (given that I know most of you personally...) I'm in two minds about fixing it though because whenever I start tinkering with a blog template I find it a gigantic timesink, and I'm planning on completely replacing this template with something homebrewed sooner or later, so I'm inclined to not tinker with this one until I have a whole day to spare making the new one. Maybe I should just shorten the message as a temporary fix.

Posted: May 19, 2005 01:04 PM

Ben Bagdikian is the former Dean of the Graduate School of Journalism at the University of California at Berkeley. His book The Media Monopoly (now updated and released as The New Media Monopoly) is a classic.

You can link to articles and interviews with him at

Robert McChesney is professor of communication at the University of Illinois at Urbana-Champaign and he has written a lot about the media too. Here is a link to his website

where you can find links to his articles and a list of his books.

An old but good book is Manufacturing Consent by Noam Chomsky and Edward Herman.

The header problem is not bad and not worth a huge expenditure of time to correct. I was just alerting you in case you were unaware.

Thanks for the information about the excerpting.

Posted: May 19, 2005 01:49 PM

Post a comment