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 #39: Well, there are two issues here [View All]

Printer-friendly format Printer-friendly format
Printer-friendly format Email this thread to a friend
Printer-friendly format Bookmark this thread
Home » Discuss » Topic Forums » Election Reform Donate to DU
Febble Donating Member (1000+ posts) Send PM | Profile | Ignore Fri Oct-10-08 05:18 AM
Response to Reply #28
39. Well, there are two issues here
One concerns specificity and sensitivity.

A fraudulent election is unlikely to be accompanied by an exit poll that matches the count.
But a non-fraudulent election is quite likely to be accompanied by an exit poll that matches the count.

In other words, exit polls are sensitive to fraud (will give a "true positive) but are not specific to fraud (will also give a "false positive").

The second is that if you get a "positive" (a discrepant exit poll) one way of determining whether or not it is a "true or false positive" (i.e. whether it really indicates a fraudulent election, or whether it is a "false positive" - i.e. it has raised a flag when the count is not fraudulent) is to see whether the degree of discrepancy is correlated with advantage to the party that appears to be benefiting by the alleged fraud.

Which is what I did, and it wasn't.

Now "proving a negative" is impossible in statistics, so what I did was to estimate how small the effect would have to be to still have a high probability of showing up in my analysis. And the answer was: "small".

So I can't tell you that no fraud affected the exit poll discrepancy - it may have done. What I can tell you is that if it did, it was a sufficiently small effect that the discrepancy itself must have been largely caused by something else. And if it was largely caused by something else, there is no reason to trust it at all as an indicator of fraud.

It also means that electronic fraud is extremely unlikely to have been on a scale of millions, which is what it would have to be to have robbed Kerry of the popular vote.

But it tells us nothing about Ohio, because the number of precincts in Ohio was too small for the exit poll analysis to rule out large scale vote theft (it certainly didn't indicate it, though). So even if the exit poll discrepancy in Ohio WAS due to fraud, the exit poll sample was not large enough to give the statistical power to demonstrate it.

Short version: exit polls only have enough statistical power to give a positive indicator of fraud at national level; they will also give false positives; in 2004 a check on whether the positive was false yielded the result that it almost certainly was.

And that e-voting theft was not a major factor in Bush's popular vote win.
Printer Friendly | Permalink | Reply | 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