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 #51: STATISTICS LESSON: DO NOT TRUST THESE NUMBERS!! [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
c-macdonald Donating Member (61 posts) Send PM | Profile | Ignore Sat Nov-27-04 02:23 AM
Response to Original message
51. STATISTICS LESSON: DO NOT TRUST THESE NUMBERS!!
As a statistician (and a hardcore Kerry supporter) I really have to step in and provide some analysis of these numbers and hopefully teach everyone a little about statistics along the way. But first, I want to make an important point:

STATISTICS DO NOT PROVE ANYTHING - they can only be used as a guide or a tool to inspire further investigation. Statistics are not "evidence" or "proof" of vote fraud (no matter how much we all want it to be true) but rather a method of uncovering irregularities that should be investigated further.

Now on to these numbers TIA has posted.

His math is essentially correct BUT HIS CONCLUSIONS ARE NOT. The odds of that many exit polls being that far outside their MOE are very small indeed, but there are several problems with his analysis.

1. As noted in a post above, TIA assumes a normal, random distribution in the exit polls and that is not the case. The problem with statistics is that they rely extremely heavily upon the notion of "randomness" of the samples taken - if the samples are not truly random, the results are skewed and are basically useless.

2. The MOE and Standard Deviation calculations don't seem to be accurate because, as I said before, they assumed a simple random sample of voters. The exit pollsters were not placed randomly around the country but rather strategically grouped so as to estimate "clusters" or "groups" of similar precincts and counties together. (See www.mysterypollster.com for more information on this)

3. Another wrench in the problem is the actual randomness of the voters being polled at each precinct. The pollsters are not professional pollsters but mostly college students looking for extra money and they were trained over the phone and given an instruction packet on how to do the exit polling. They also had to take breaks from polling throughout the day to call in their data every few hours. There is also the theory of oversampling Kerry voters which, while not a prevaling theory, probably does have at lest a slight effect given that Republicans are likely to ignore a pollster than a Democrat (since the pollsters are decked out in ABC-CBS-NBC-AP-etc. apparrell and over 50% of repubs don't trust the media, compared to just 38% of Dems).

All these factors create some more uncertainty as to the "randomness" of the sampling and will create a significantly larger MOE than the one TIA has calculated. The actual calculations are much more complex (Freeman does an excellent job with that aspect in his newest version of his paper: http://www.appliedresearch.us/sf/epdiscrep.htm ).

I do want to commend TIA on his efforts because I know it took him a long time to put it all together. I wish I could agree with your conclusions but sadly I cannot.

I know we all want to do our part to help uncover the fraud in the election, but I urge you not to jump to any conclusions about any of the statistical analyses that come out. So far, every paper has been disproven or cast into doubt (including the papers refuting the pro-fraud papers) so just remember that statistics are fun because they can be easily manipulated and your inherent bias will almost always show up in your analysis unless you make a concerted effort to overcome it.

So, to conclude, don't believe everything you see and hear when it comes to statistics. Especially the ones that are being rushed out in the wake of the election because they aren't submitted to rigorous expert analysis and criticism before they're released and most likely have errors in them. TAKE EVERYTHING WITH A GRAIN OF SALT and just hope that somehow, in some way, the truth will come out.

In the meantime, start changing the terms of political debate by using progressive frames to argue our values and our beliefs.
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