From: Matt Jensen (mattj@newsblip.com)
Date: Mon Dec 18 2000 - 15:57:09 PST
On Mon, 18 Dec 2000, Bill Stoddard wrote:
> > (BTW, I'm surprised when FoRKers use anecdotal evidence to counter
> > statistical evidence
> On a more personal note, would you like it if decisions were made for you by
> your government based on 'studies' of the 'average' person?
If the studies describe threats that the typical person poses to me, then
yes, I would consider supporting such decisions, even if I feel I pose no
threat to others. If the studies describe threats the typical person poses
only to himself, like eating too much red meat, then it's probably nobody
else's business, and no laws should be passed.
For example, take the use of cell phones while driving. I feel I know
when I can handle that and when I can't. If I'm on a busy road, I tend to
pull over to talk; if I'm on a busy highway, I call them back later. But
I was almost hit *twice* last week, while on my bicycle, by drivers using
cell phones. I don't know if the percentage of "dangerous" cell phone
drivers is 2% or 80%, but while I feel confident I'm safer than most of
them, I'm also quite willing to support a well-written law to control
that activity.[1]
(I think the interesting question is, if I hadn't been nearly hit by these
drivers, if I had no personal experience like that and only read the
results of a study, would I still be willing to support a law?
Practically, that depends on the perceived quality of the study, and on
the person.)
> > The fact that Uncle Charlie smokes two packs a day
> > and is 80 years old does not dispute the statistical proof that smoking is
> > likely to kill people.)
>
> Bad analogy. There is a good cause and effect relationship between tobacco use
> and disease, regardless of where the user lives, culture, education level,
> professional training, training on 'how to use tobacco properly :-)',
> political beliefs, or whatever. Comparing this to 'gun control' is, ummm, not
> rigorous :-). Guns are not a disease, despite what the CDC would have you
> believe.
I didn't say guns are a disease, I cited a study with a statistically
signficant result. You can argue against absolutes with anecdotes ("No
smoker has ever lived to 80." "Oh yeah? My Uncle Charlie did."). But you
can't argue against statistics with anecdotes.[2]
You can only argue against statistics with other statistics, and by
challenging the methodology and applicability of the original statistics.
You mention some factors that might not have been controlled for in the
study (education level, etc.), which is the right way to attack it. Citing
anecdotes is the wrong way, in my view, which is why I think the smoking
analogy is good.
Or, to choose an example where the evidence isn't so lopsided, you don't
argue against a study which says cell phones cause brain tumors by telling
us that you've used cell phones for years and your MRI is clean. You
argue the study itself, or you try to duplicate it to show different
results, or you show better studies.
> Have you seen Gary Kleck's reports?
> http://www.guncite.com/gcwhoGK.html
I'll put it on my reading list, thanks!
-Matt Jensen
NewsBlip.com
Seattle
[1] I have no specifics to offer for such a law, but I think it's
reasonable for me to have my phone on so I can know who's calling me, like
a beeper, so long as I don't actually answer the phone. Just
brainstorming here.
[2] Of course, anecdotes work emotionally, and politicians often find they
work better than statistics at persuading people, but I'm talking about
what's effective at making a logical argument.
This archive was generated by hypermail 2b29 : Mon Dec 18 2000 - 16:11:56 PST