Back to index of science pages by Donald Sauter.

Questionable science in the media

While I have tremendous respect and admiration, and even amazement, for what scientists have figured out over the centuries, I also have a touch of ambivalence towards "science", as you may have noted in such pages as "The the human race is special"; "The Big Bang, explosion or expansion?"; and my "Evolution FAQ".

Maybe it's something to do with my own biases, or some selection effect at work, or my resentment of "scientific thought" always being held up as somehow different from and much loftier than the stuff an ordinary person uses to get himself safely across the street, but it's sure seemed to me that all the science-related articles I've ever read or heard in the media contain a dose of nonsense - often a big dose.

This has led me to the notion that scientists are not externally policed and that there is minimal internal policing, and this gives scientists carte blanche to take advantage of the public's awe of science.

As "unscientific" as these impressions are, here are just a few examples that come to mind. In my experience, they represent the norm, not aberrations.

A few years ago [writing this in the late 1990s] a study came out that said that left-handed people die 10 years earlier than right-handed people. Now, we're not talking 10 days, or 10 weeks, even - but 10 years! We're not dealing with barely measurable differences in beak sizes - we're talking 15 percent differences in life expectancy. That's a healthy chunk of a person's life!

If such a claim is correct, shouldn't it be a trivial matter to present it in a clear and irrefutable way? On the other hand, if the researchers blundered, shouldn't it be a simple matter to present hard numbers showing that they were wrong?

The public was flummoxed for months before the follow-up article appeared saying, well, no, lefties don't really die 10 years earlier. But even the follow-up didn't make it clear how much earlier lefties do die, if at all earlier. Shouldn't this be about as simple a question for science to answer as they come, maybe suitable for a junior high school science project? Would it be so hard for someone to contact survivors of the deceased and ask whether he was right- or left-handed?

In May, 1990, a report came out that "Mother Eve" - the female ancestor of all of us - had been tracked down. She lived in Africa 207,000 years ago. The logic had something to do with mitochondrial genes, which are passed only by mothers. The things mutate like clockwork (never mind the inherent contradiction) and by examining different hair root genes in 199 modern women, researchers conjured up the time and place (!) of the original from which all these genes descended.

Make sense? I promise, you can read every account of this finding and it doesn't get any more swallowable than that. Still, the media nod knowingly, and the public nods knowingly.

(There is a web page, www.cs.wisc.edu/~krisna/misc/eve.html , which lifts the clouds of confusion ever so slightly. It does explain what is meant by Mitochondrial Eve, but gives impossible-to-swallow scenarios whereby she may be some - or many - generations later than the "actual" Eve, meaning the first human woman. It doesn't pretend to attempt to explain how "clockwork mutations" allow us to pinpoint her dates and location.)

Another group eventually did make it to press with the obvious - that the original Eve study was just dumb - but that wasn't until September, 1993! (To my dismay, but not surprise, their refutations were not presented clearly, either.)

If that isn't bad enough, there was an equally incredible study nailing the father of us all - by examining identical genes in 38 modern people. I'll bite: how do you trace something back 270,000 years when all you have to go on is the endpoint? - not even a single preceding data point which would possibly allow you one fantastic mother-of-all-extrapolations! This was the subject of a sarcastic letter, reproduced further down, which I wrote to National Public Radio after listening to their embarrassing presentation.

In December 1997, a study said that boys start out as girls, but nothing in the article supported the claim. It said that if a male embryo doesn't produce male hormones, "the female body blueprint will prevail". So??? The male embryo does produce male hormone - probably because it's male, I guess. Anyhow, I suppose this was good enough to get a few scientists a bit of much-needed attention.

In August 1997, scientists discovered a 117,000-year-old footprint in South Africa. Not only did they not give the dating method, but whatever the method was, it would apply to the material the footprint was made in, not the foot-shaped void itself.

The big point here is not lefties or Eves or footprints; it's scientific reporting. When something is true, why should there be any difficulty presenting it plausibly, if not compellingly? My advice is, when you read a science article and it doesn't make sense, don't blame yourself. It's not your fault. If you're intelligent enough to read, the writer should be able to make you understand. If he doesn't, the problem is on his end.

Call him on it. Tell your newspaper or magazine that what they printed didn't make sense. Lean on them to explain it to you - or get them to admit that they didn't really understand what they were printing, in which case tell them to go back to the scientists for a clarification. I'd like to think that scientists would want their findings presented accurately and clearly.

Is the situation any better in the scientific literature? I wonder about this, even. Consider Scientific American, which lies somewhere on the border between scientific journals and the popular press. I used to struggle to make sense of those articles from beginning to end - always without success. Then I found others who admitted they could never follow them through, either. A physicist I knew once explained that that was inherent in the articles themselves; they were all ghostwritten by writers who did not know the subject matter. Whether or not ghostwriters are used, that is no excuse for a science article, or any kind of article, not making connected sense from beginning to end.

Consider the fairly recent case in which Alan Sokal, a physicist at New York University, wrote an article "close to pure gibberish" which got published in a "respected social-science journal", Social Text. (AP, Washington Times, 1996 May 18, page A5).

Actually, when you think about it, it shouldn't be surprising to find articles in scientific journals full of errors and fallacies. A scientific project must break new ground and, in this day and age, is typically so esoteric that it is unreasonable to think there are others who are expert enough to capably judge it. If there were, then the work must have already been done by someone else. See the paradox?

A first-hand experience involving errors in a scientific journal came at the National Radio Astronomy Observatory. My task was to program a model of the density of molecular hydrogen in the region surrounding a star, based on a paper by Hollenbach, Werner and Salpeter in the Astrophysical Journal, 163:165-180, 1971 January 1. Almost all of the equations in that paper needed for the computer model had typos and major blunders which needed to be ironed out in order to make the program work. I'm sure those equations looked mighty impressive to the referees of the article!


Letter to: All Things Considered/National Public Radio. May, 1995.

I'd like to share with you the results of a fascinating study I just completed.

We know that editorial changes can find their way into literary works that are republished over the years. I set out to determine when Edgar Allen Poe's story "The Tell-Tale Heart" was written. In the sentence which I scrutinized, there was absolutely no difference among 38 modern editions. This implies that all 38 are so close to the original that they haven't had a chance to change at all. Therefore, all editions evolved from a manuscript written about 270,000 years ago. Well, not necessarily the original manuscript, but a small group of original manuscripts. And, to be honest, this date is so uncertain, it could be as much as 800,000 years ago. This is not a long time by literary standards.

Pretty dumb, eh? Well, the logic is identical to that used in your report on the Yale study which examined genes on the Y-chromosome to date the common ancestor of all men.

Satire aside, there is a very serious - and common - problem here. Can you honestly claim that you yourselves understood what you were saying in that report? My suspicion is that you are guilty of the line of reasoning that goes, "Well, I don't really understand this, but scientists are smart, so it must make sense." You allow yourself to be snowed by the scientists; the public, in turn, is snowed by the media.

But the lion's share of the blame goes to the scientists, themselves - in particular, to those whose work may be tainted by a desire for the publicity that goes with sensational findings; to those who don't police their colleagues energetically enough; and to those who do not take the care to see that the media get it right when the findings are valid. Note that objective, scientific matters should be among the easiest for the media to present simply, clearly and compellingly.

In the case at hand it is almost humorous to note that our 270,000 year old father was tracked down using identical DNA in various subjects, while a previous study used differences in DNA to pinpoint our 207,000 year old mother. The media presentation of neither study made clear how one can, by knowing the endpoint of a sequence, determine its starting point. The Eve study - even though you hauled it out again - has been debunked, so we needn't worry about her apparent penchant for older men (if only by 63,000 years.)

My closing request is, that you never present anything you don't understand. Adopt the position, "Make us understand what you are saying - or we will simply ignore it." This would force the scientific community to tighten up its act.

P.S. Encouraged by the success of my first study, I am now examining small, identical areas on 38 modern Coke bottles.

P.P.S The sentence in Poe's story was, "Madmen know nothing."

 


Contact Donald Sauter: send an email; view guestbook; sign guestbook.
Back to Donald Sauter's main page.
Rather shop than think? Please visit My Little Shop of Rare and Precious Commodities.

Parents, if you're considering tutoring or supplemental education for your child, you may be interested in my observations on Kumon.