Harvey Newstrom wrote:
>
> Mike Lorrey wrote,
> > Harvey Newstrom wrote:
> > > Exit polls done by the media seem to match the actual computer counts in
> > > most cases.
> >
> > And who does the exit polling? The media, which has a proven bias. The
> > fact is that exit pollsters have to correct exit poll data because
> > pollsters ALWAYS project more votes for democrats than what actually
> > occurs.
>
> You are absolutely right. But the margin is small. The fact that they
> ALWAYS do it makes it totally consistent and repeatable, which is why they
> are capable of correcting it as you noted above. Their may be bias in their
> methods, but they do fairly correct the data to present more accurate
> results.
Yes, but what about the impact? Most people know that their one vote
will have little impact on an election either way. When the media report
the democrats are winning by even a small margin, this discourages
republicans from going and voting, expecting the election to be lost.
This is why final margins in 'called' elections are almost always larger
by a greater degree than the margin at the point the election is
'called' by the media's exit polls than when the media doesn't publicly
call it: Enough people on the side that is behind figure it is a lost
cause, and decide not to waste their time voting. Unless the party has
some means of shipping in voters on a moments notice (or in getting
people to vote twice, or vote for dead people: All Democrat tactics),
the act of calling an election before polls close creates a bias in the
final results. The fact that the media has a bias for Democrats means
that this will always benefit Democrats more than Republicans.
>
> This is common in statistics. A lot of sampling methods turn out to be
> biased on way or the other. One theory for this particular bias is that
> liberals are more likely to talk to liberal media, whereas conservatives are
> less likely to trust or cooperate with the media.
That's a nice neat way to let the media off the hook. I don't buy it. I
have personally seen exit pollers ignore people with conservative
leanings and select people who look 'liberal' to participate in exit
polls, whether its done by hair length, clothing quality, or skin color
is irrelevant, it is profile that is at least as wrong as the police
profiling of blacks that liberals complain of.
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:40:06 MDT