"Grey Satterfield" wrote , "Jack Linthicum"

wrote:


John P. Mullen wrote:

The first line of the lead paragraph is, "Joe Lieberman and Ned Lamont
are in a statistical tie in the race for United States Senate in
Connecticut."

Because of the margin of error, they are, in fact, tied. The poll
estimates the true proportion by use of a sample and if a difference is
within the margin of error, it is considered a tie, not a statistical
tie, because there is no statistical evidence that it is not a tie.

Putting it another way, any result of statistical sampling is
statistical in nature. Even a lead of ten points is "statistical,"
because we do not know the true population and it could be very
different. Millions of young soybean plants would give their all to
print the redundant word "statistical" in every published report of any
statistical result were to be added.


Statistically yours,

John Mullen
No statistics were harmed in the generation of this email.

And another example of how a certain political bent keeps people with
that philosophy from actually reading 'anything' for content.
Comprehending and understanding is, of course, never a possibility.

Now who's being "emotional"? "Give us the gift to see ourselves as others
see us," indeed.

Nice riposte, Grey....

Now, with all of us reduced to the rational and clear as to the statistical
nature of statistics, we'd best collectively admit that since none of us
vote in Connecticutt, we'll have little to say as to who the winner in
November is likely to be.

We should also agree that polls and surveys, whatever their nature and
intent, fall into four categories...the first and largest those which almost
any fool could have predicted the outcome going in and which are taken for
purposes of confirmation or to support the high incomes of pollsters.
Second comes the polls which turn out, after the fact, to have been correct.
These are the ones pollsters and political operatives post on their walls
to celebrate their wisdom and foresight. Position number three is held by
all those many polls and surveys which emerge as dead solid wrong. All
involved work hard to forget these as soon as possible. Bad exit polls?
Don't try to fix'em. Banning them is preferable. Pollsters who deliver the
wrong results are equally consigned to unmarked graves as rewards for their
inaccuracies. But even the best of pols and coprporations may fall victim
to inaccurate polling/testing/surveying. Raised to imagine CocaCola's
marketing department and strategizing to be first rate and cutting edge,
would any of us have been willing to say early on that "New Coke" was worse
than a bad idea.

In last place fall the polls which reflect statistical or augury by internal
animal organs (what actually happens in most polling shops) ties. They are
carefully preserved along with the quaintly crafted legends of "Margin of
Error" to allow those who created and took them to preserve some
credibility. I once had a very stute political operative tell me that if
"our side" commissioned a poll and it came out to be tied or nearly so, that
meant we were way behind, for every poll, even those carried out be
media-supported and non-partisan bodies, will be at some point, either in
creation or in application, biased towrd a desired conclusion, and
pollsters ineviotably create polls to reinforce the conceptions or
misconceptions of those paying the bill.

The state in question does have some tradition of electing both
"Independents" (nominal, admittedly) and Party members (with distinctly
model adherence to the public principles and platform of their
affiliations). Our ancestors apparently distrusted the residents of the
state enough to accord them the notorious "Wooden Nutmeg", awarded in a
commercial vein, but likely a statement on loyalty and fidelity as well.

But all the fooferaw of primary victories and party participation aside, the
decision in November will be made by two groups, the first and vastly the
larger, those who didn't vote at all in the Primary, the uncaring and the
unwashed, but bestirring themselves on some bright Autumn morn to trundle
off to the polls. Name ID gives Lieberman some edge in this bunch. The
second group are those Republicans who view their own candidate as
unelectable or unpalatable and cast ballots for another. Even in
Connecticutt, Lieberman should far outpoll Lamont among this group. After
all, Lowell Weiker (sp?) used to get elected based on substantial votes from
members of the other party.

At this point in time and probably until late the night of Election Day,
none of us really know who'll win. With only two viable candidates, any
guess we might individually make will be only half wrong, but in politics,
half wrong is all wrong. I'd tend to give the edge to the smarter
campaigner, the one able to uncover and exploit "pockets" of voters who need
to be cajoled to the polls. One guess....If the traditionally Democrat
minority voters turn out for Lamont, he'll win. If a modest portion of them
support Lieberman out of loyalty/name ID, he'll win, as will he if
minorities, seeing no real "stake" in the outcome, sit home or turn out in
modest numbers. I've no "feel" at all for the smaller number of Hispanic
voters in the state, but unlike African American voting blocks in many US
states, the Black voters in Connecticutt seem far less "pulpit controlled"
than in many locales, far less so than Chicago or NYC. I would expect to
see more Black elected officials (especially the Senator from Illinois) on
the Lamont campaign trail than the traditional faces such as the Reverend
Jackson.

TMO