You are viewing an obsolete version of the DU website which is no longer supported by the Administrators. Visit The New DU.
Democratic Underground Latest Greatest Lobby Journals Search Options Help Login
Google

Reply #88: Lets get this in context: [View All]

Printer-friendly format Printer-friendly format
Printer-friendly format Email this thread to a friend
Printer-friendly format Bookmark this thread
This topic is archived.
Home » Discuss » Topic Forums » Election Reform Donate to DU
Febble Donating Member (1000+ posts) Send PM | Profile | Ignore Thu Aug-18-05 04:39 AM
Response to Reply #82
88. Lets get this in context:
Here's the link to the document containing the table:

http://www.exit-poll.net/election-night/MethodsStatemen...

It's from the Methods Statement for the E-M National Exit Poll. A similar table for the State Exit Poll can be found here:

http://www.exit-poll.net/election-night/MethodsStatemen...

Both are preceded by this explanatory passage:

All samples are approximations. A measure of the approximation is called the sampling error. Sampling error is affected by the design of the sample, the characteristic being measured and the number of people who have the characteristic. If a characteristic is found in roughly the same proportions in all precincts the sampling error will be lower. If the characteristic is concentrated in a few precincts the sampling error will be larger. Gender would be a good example of a characteristic with a lower sampling error. Characteristics for minority racial groups will have larger sampling errors.

The table below lists typical sampling errors for given size subgroups for a 95% confidence interval. The values in the table should be added and subtracted from the characteristics percentage in order to construct an interval. 95% of the intervals created this way will contain the value that would be obtained if all voters were interviewed using the same procedures. Other non-sampling factors, including nonresponse, are likely to increase the total error.


{My bold}. Note that the table gives typical figures for an unspecified "characteristic" and that figures are rounded to the nearest whole number.

The context in which your image was taken was a piece by MP here:

http://www.mysterypollster.com/main/2004/12/what_is_the...

which contains links to further sources that discuss the design effect.

Having said all that, the MoE is largely irrelevant to the present discussion. Nobody disputes that the discrepancy between the poll and the count was massively significant, and that something other than sampling error, clustered or not, must account for it. One question is - what? Another question is - was Ohio special?

The E-M report

http://www.exit-poll.net/election-night/EvaluationJan19...

(p 28) tells us that the problem did not lie in precinct selection p . The discrepancy was at precinct level. So to check out Ohio, the sensible question is not so much: was Ohio's estimate outside the MoE, but was the WPE in Ohio an outlier?

Well, it was higher (-10.9) than the average of -6.5 for all 1460 precincts in the state samples (pp31-33), but Connecticut, Delaware and Vermont all had mean WPEs greater than -15, and New Hampshire, which, at least according to two apparently disinterested sources

http://www.invisibleida.com /

http://www.votenader.org/media_press/index.php?cid=413

had a "clean" recount, also had a higher mean WPE of -13.6. Alabama, Missouri, New York and North Carolina were also all more "red-shifted", according to the mean WPE figures, than Ohio. So Ohio doesn't particularly stand out; moreover, the mean WPE in Ohio was less than in one state (NH) where an apparently properly conducted recount confirmed the count rather than the poll.

Look, there is plenty of evidence for fraudulent activity of all kinds regarding the election. I take Land Shark's point that all a court needs is a preponderance of evidence. But it is my considered view that the exit polls do not constitute evidence that meets anything like the 51% confidence mark, never mind a 95% confidence mark. Whereas the evidence that Kenneth Blackwell obstructed a fair recount in Ohio is damn near 100%, as is the evidence that more Democratic votes than Republican votes were lost in Ohio due to voter suppression, whether that suppression arose from systemic negligence or malice. It wouldn't have taken much for this to have swung Ohio, and thus the presidency, for Bush. And it probably wouldn't have made much impact on the exit poll.

I don't see the point in making a vulnerable case when you have a watertight one. It simply saves your opponents the trouble of making their own straw men.




Printer Friendly | Permalink |  | Top
 

Home » Discuss » Topic Forums » Election Reform Donate to DU

Powered by DCForum+ Version 1.1 Copyright 1997-2002 DCScripts.com
Software has been extensively modified by the DU administrators


Important Notices: By participating on this discussion board, visitors agree to abide by the rules outlined on our Rules page. Messages posted on the Democratic Underground Discussion Forums are the opinions of the individuals who post them, and do not necessarily represent the opinions of Democratic Underground, LLC.

Home  |  Discussion Forums  |  Journals |  Store  |  Donate

About DU  |  Contact Us  |  Privacy Policy

Got a message for Democratic Underground? Click here to send us a message.

© 2001 - 2011 Democratic Underground, LLC