home *** CD-ROM | disk | FTP | other *** search
- Comments: Gated by NETNEWS@AUVM.AMERICAN.EDU
- Path: sparky!uunet!paladin.american.edu!auvm!ATSDHS2.EM.CDC.GOV!LXP6
- X-Delivery-Notice: SMTP MAIL FROM does not correspond to sender.
- Return-Path: LXP6@ATSDHS2.EM.CDC.GOV
- Encoding: 39 TEXT
- X-Mailer: Microsoft Mail V3.0 (beta-2)
- Message-ID: <2AA43821@router.em.cdc.gov>
- Newsgroups: bit.listserv.stat-l
- Date: Wed, 2 Sep 1992 07:37:00 EST
- Sender: "STATISTICAL CONSULTING" <STAT-L@MCGILL1.BITNET>
- From: LXP6@ATSDHS2.EM.CDC.GOV
- Subject: "polls" does not equate to Statistics
- Lines: 38
-
- I thought that this was supposed to be a statistical bulletin board.
-
- The 'gibberish' in the original message is NOT the mathematical abstraction
- of repeated sampling, but the lack of distinction between an expected outcome
- and actual outcomes. It may be easier to visualize someone flipping a coin
- six times. The expected number of heads in six flips of a fair coin is
- three. But the actual observed number may be as small as 0 or as great as 6
- (each could occur with a probability of _________).
-
- Clearly, statistics must be applied not in a vacuum but in the milieu of a
- real world. BUT...
-
- ------------------------------------------------------------------------------
- REPLY FROM: Pallos, L. Laszlo
- Date: Wed, 2 Sep 1992 01:30:56 EDT
- Subject: RE: Election polls
- In-Reply-To: Message of Tue, 1 Sep 1992 16:01:00 EDT from <CHENX@CUA.BITNET>
-
- >%Each poll has a level of reliability (something like +-4% 95 times out of
- >%100). What this means is that if we were to repeat the poll 100 times, the
- ^^^^^^^^^^^^^^^^^^^^^^^^^
- >%actual (population) percentage of voters supporting Clinton would lie in
- >%the range 56% +- 4% (or 52% to 60%) in the first poll and 48% +- 4% (or 44%
- >%to 52%) in 95 out of the 100 polls taken.
-
- I wonder if Clinton and Bush are frequentists...? They'd love to be able
- to repeat the poll 100 times. If they had that kind of money I suspect
- they'd have pulled out long ago like the other billionaire who decided
- that politics wasn't as much fun as it seemed at first.
-
- Ahhh, the states of the world required to have this repeated sampling
- gibberish be true.... :-)
- ^^^^^^^^^
- Woodrow Wilson School of Public and International Affairs,
- Princeton University, Princeton, NJ 08544
- and
- Department of Political Science, University of Rochester,
- Rochester, NY
-