NetNews Usenet Archive 1992 #20/NN_1992_20.iso/spool/sci/research/1092

Internet Message Format | 1992-09-14 | 8.2 KB

Path: sparky!uunet!mcsun!uknet!root44!hrc63!mrcu!paj
From: [e-mail]-mrc (Paul Johnson)
Newsgroups: sci.research
Subject: Re: Newsweek Article: is Science Censored?
Message-ID: <1955@snap>
Date: 14 Sep 92 09:54:17 GMT
References: <[e-mail]>
Reply-To: [e-mail]-mrc (Paul Johnson)
Organization: GEC-Marconi Research Centre, Great Baddow, UK
Lines: 142

In article <[e-mail]> [e-mail] (Danny Keren) writes:
>'"Publicity... would certainly follow," fretted the editor of one top
>journal. "A possible general panic," predicted a researcher. Both were
>explaining why a a study linking childhood leukemia to flourescent
>lights should not be published. That fear trumped the conclusion of
>other reviewers - scientists who evaluate whether a manuscript should
>be published in a journal who called the article "intriguing" and an
>"extraordinary piece of deductive reasoning". The paper was rejected.'
>
>This is the beginning on an article in the Sep. 14 of Newsweek (p. 63)
>which concerns how scientific research is sometimes barred from
>publication, for fear of panic and hysteria which might occur
>if the conclusions would become well known. Other examples cited
>include studies relating bladder and rectal cancer to chlorine in
>drinking water, and bone cancer to fluoride.

On the other hand, they were not kidding about the panic.  A recent
BBC program in the "Antenna" series investigated this from the other
side.  One of their examples was a piece of research which seemed to
link injections of Vitamin B12 shortly after birth (done for some good
medical reason, I can't recall exactly what) with childhood leukemia a
few years after.  The findings were very tentative.  The study was
basically a trawl through epidemiology statistics looking for any
correlations.  This is always dangerous.  These findings were
presented at an obscure conference.  A few days later they were
splashed over the front pages of tabloid newspapers with no indication
that there was any doubt that the effect was real or any indication of
the degree of risk.  The scientists concerned were very embarassed,
but there was nothing they could do.  A reporter did phone one of them
up, but refused to take "no" for an answer.  The scientist was
presented with a choice: either you talk to me, or we publish our own
version anyway.

The problem is that "dog bites man" is not news.  An editor looking
for a story is not going to run a headline saying "There might just be
a slightly increased risk of leukemia from B12 injections, but even if
there is its probably less than the risk of not having the injection".
No, he is going to say "B12 INJECTIONS CAUSE LEUKEMIA: Scientists
working at XYZ University have found that .....".  The latter version
sells newspapers.  It also terrifies couples with young children
because they immediately conclude that their child is going to get
leukemia.  Remember that not only did the papers not indicate how
tentative this finding is, but also the vast majority of the
population are statistical illiterates who would not recognise a
correlation coefficient if it bit them!  This is a bad situation, but
it is reality.  And this kind of thing also gets science a bad name
(yes, it *should* get the tabloids a bad name, but people see the
message, not the media).  I can hear the anti-science lobby now: "How
can we trust scientific findings?  Remember that leukemia scare last
year, turned out to be a load of nonsense".

By the way, the reason that trawling through statistics looking for
correlations is dangerous is as follows:

A significant correlation is one greater than 95%.  But if you look at
random data in 20 different ways, you are going to find a significant
correlation 1 - 0.95^20 = 0.64 of the time.  In other words a study
looking at random data 20 different ways has a greater than evens
chance of finding something "significant".

Disclaimer: I am not a statistician.  I hope I got that right.

Paul.
--
Paul Johnson ([e-mail]).        | Tel: [phone] ext 3245
--------------------------------------------+----------------------------------
These ideas and others like them can be had | GEC-Marconi Research is not
for $0.02 each from any reputable idealist. | responsible for my opinions
Newsgroups: alt.conspiracy,alt.censorship,sci.research
Subject: Re: Newsweek Article: is Science Censored?
References: <[e-mail]>
Reply-To: [e-mail]-mrc (Paul Johnson)

In article <[e-mail]> [e-mail] (Danny Keren) writes:
>'"Publicity... would certainly follow," fretted the editor of one top
>journal. "A possible general panic," predicted a researcher. Both were
>explaining why a a study linking childhood leukemia to flourescent
>lights should not be published. That fear trumped the conclusion of
>other reviewers - scientists who evaluate whether a manuscript should
>be published in a journal who called the article "intriguing" and an
>"extraordinary piece of deductive reasoning". The paper was rejected.'
>
>This is the beginning on an article in the Sep. 14 of Newsweek (p. 63)
>which concerns how scientific research is sometimes barred from
>publication, for fear of panic and hysteria which might occur
>if the conclusions would become well known. Other examples cited
>include studies relating bladder and rectal cancer to chlorine in
>drinking water, and bone cancer to fluoride.

On the other hand, they were not kidding about the panic.  A recent
BBC program in the "Antenna" series investigated this from the other
side.  One of their examples was a piece of research which seemed to
link injections of Vitamin B12 shortly after birth (done for some good
medical reason, I can't recall exactly what) with childhood leukemia a
few years after.  The findings were very tentative.  The study was
basically a trawl through epidemiology statistics looking for any
correlations.  This is always dangerous.  These findings were
presented at an obscure conference.  A few days later they were
splashed over the front pages of tabloid newspapers with no indication
that there was any doubt that the effect was real or any indication of
the degree of risk.  The scientists concerned were very embarassed,
but there was nothing they could do.  A reporter did phone one of them
up, but refused to take "no" for an answer.  The scientist was
presented with a choice: either you talk to me, or we publish our own
version anyway.

The problem is that "dog bites man" is not news.  An editor looking
for a story is not going to run a headline saying "There might just be
a slightly increased risk of leukemia from B12 injections, but even if
there is its probably less than the risk of not having the injection".
No, he is going to say "B12 INJECTIONS CAUSE LEUKEMIA: Scientists
working at XYZ University have found that .....".  The latter version
sells newspapers.  It also terrifies couples with young children
because they immediately conclude that their child is going to get
leukemia.  Remember that not only did the papers not indicate how
tentative this finding is, but also the vast majority of the
population are statistical illiterates who would not recognise a
correlation coefficient if it bit them!  This is a bad situation, but
it is reality.  And this kind of thing also gets science a bad name
(yes, it *should* get the tabloids a bad name, but people see the
message, not the media).  I can hear the anti-science lobby now: "How
can we trust scientific findings?  Remember that leukemia scare last
year, turned out to be a load of nonsense".

By the way, the reason that trawling through statistics looking for
correlations is dangerous is as follows:

A significant correlation is one greater than 95%.  But if you look at
random data in 20 different ways, you are going to find a significant
correlation 1 - 0.95^20 = 0.64 of the time.  In other words a study
looking at random data 20 different ways has a greater than evens
chance of finding something "significant".

Disclaimer: I am not a statistician.  I hope I got that right.

Paul.

--
Paul Johnson ([e-mail]).        | Tel: [phone] ext 3245
--------------------------------------------+----------------------------------
These ideas and others like them can be had | GEC-Marconi Research is not
for $0.02 each from any reputable idealist. | responsible for my opinions