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 #86: Well, the problem is [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 locked.
Home » Discuss » Topic Forums » Election Reform Donate to DU
Febble Donating Member (1000+ posts) Send PM | Profile | Ignore Sun Jun-17-07 02:50 AM
Response to Reply #83
86. Well, the problem is
you can't tell whether it accounts for the difference or not unless you know the actual variance, which you don't. All the computation would tell you is what it would have to be if the 16% value were within the MoE of the observed value.

And it may well be true that "suburbs are by far the most likely to have significant boundary changes". But the point is that in a small sample of big city precincts, it would only take a few to have significant boundary changes to push the variance in apparent turnout through the roof. Or indeed, a small number of precincts with extremely large actual turnout increases, or decreases. Small samples are vulnerable to large leverage by individual data points, and if the datapoints are clustered (as they will tend to be in this instance) then this will compound the problem.

The NEP conducts state polls in each state, with, typically, 30 to 50 precincts in each state. In 2004 it was between 1460 precincts. These are used for predicting the state races. However, to get a picture of the national electorate, a sample of these precincts forms the "National" poll, which does not predict the outcome (because it couldn't, given your Electoral College vote system) but is simply used for cross-tabulations to show who voted for whom, and why. It is selected to be as representative as possible of the country as a whole - if you simply threw all the precincts in the entire poll into the calculation, you'd have too many from small states and not enough from large states.

But this means that if you are dealing with turnout, which is a precinct level value (one turnout value per precinct) the Ns are very small, even though they are large at voter level.
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