pageok
pageok
pageok
Media miscoverage of the role of man-made chemicals in disrpupting human or animal reproduction:

That's the topic of my media column in today's Rocky Mountain News. The column also expresses skepticism about the benefits of Gannett buying Colorado's leading college newspaper, about media coverage of Obama and Clinton, and about Maureen Dowd.

runape (mail):
"Sens. John McCain and Clinton have, at least, been the subjects of biographical books, but the only in-depth biography of Obama is his autobiography."

This is not correct; David Mendell published a biography last year.

Incidentally, while Dowd was technically inaccurate (and should issue a correction), Clinton certainly did indirectly aim the hat/cattle remark at Obama (e.g., her reference to Kirk Watson's appearance on MSNBC).
2.23.2008 8:13pm
Dave N (mail):
The Gannett newspaper chain might buy the CSU student newspaper, The Rocky Mountain Collegian. Although Gannett has not interfered with the two Florida college newspapers it currently owns, the Columbia Journalism Review reports that after Gannett took over the Des Moines Register and The Courier-Journal in Louisville, Ky., it dumbed them both down, mandating short, local fluff pieces at the expense of serious reporting. It's hard to see a benefit from a Gannett takeover.
Amen. That is exactly Gannett's M.O. They take a perfectly good newspaper and dumb it down beyond recognition. It wants fluff and dreck that it can then feed into USA Today. Can I think of anything good about Gannett? Honestly, no.
2.24.2008 2:02am
Brian Mac:
Unfortunately, such distortions aren't restricted to mass media outlets. Ever compared the way that research findings are written up in the "news" sections of scientific journals, to the research articles themselves?
2.24.2008 8:15am
Anonperson (mail):
I think part of this problem is due simply to the fact that news, fundamentally, relies mostly on its entertainment value. A further problem is that scientific papers are intentionally (and rightly so) written to be suggestive, but precise.

This is because the nature of science is incremental. So, a paper often will not be conclusive. But, to be interesting, it must explain and suggest why it is relevant.

So, consider the case of chemicals and reproduction. If there is no causation, and the correlation is simply a coincidence, then there is nothing interesting here. The paper is a non-paper. So the paper must make a case for how its results are evidence of causation, or at least suggestive of causation. At the same time, it can't overstep its claims, or else the reviewers will shoot it down.

This nuance is easy for scientists to navigate, because we do it all the time, but seems to often confuse the media and the public.

You might ask, why not write conclusive papers only? That's due partly to the pressure to publish. If you only write papers that are conclusive, you will have a lot fewer papers. Probably more importantly, though, is that you will never get it funded. No one will fund you for a massive, conclusive research project on a particular question unless you have preliminary results of some kind, have established a track record, etc.

For example, suppose I think that cell phone towers cause autism. Would you, as a taxpayer, give me $N million to study this, with absolutely no preliminary results to suggest that there might really be causation? Of course not. So how do I get the preliminary results? Well, I publish small, cheap studies that suggest there is something more to it. That piques your interest, and hopefully you will give me more money for a bigger study, etc.

The media gets a hold of it, they miss all the qualifiers in the text, they don't understand the statistics or the math, etc. Then, they themselves are under pressure to simply, to reduce, to condense. Their readers don't want to actually have to think hard.
2.24.2008 10:04am
BT:
Anonperson thank you for that explanation. I never knew how that system worked. Some things make a whole lot more sense now.
2.24.2008 10:34am
Pyrrhus (mail) (www):
Parts of this article are very opaque. Where exactly did the 1% a year claim come from?
2.24.2008 11:20am
Fub:
Anonperson wrote at 2.24.2008 10:04am:
This is because the nature of science is incremental. So, a paper often will not be conclusive. But, to be interesting, it must explain and suggest why it is relevant.

So, consider the case of chemicals and reproduction. If there is no causation, and the correlation is simply a coincidence, then there is nothing interesting here. The paper is a non-paper.
Only if the researcher, or the research community, or the popular journalists, has some vested interest in proving causation over non-causation.

A well designed and executed experiment which shows anything, correllation or non-correllation, causation or non-causation, reasonably definitively is generally far more useful than an experimental result which is at best a Rorschak ink blot.
2.24.2008 11:27am
Brad D. Bailey (mail):
David,

I am delighted to know that you believe my alma mater is Colorado's leading college. Or at least it has the best college newspaper in the state.

Class of 1980
2.25.2008 6:37pm