PDA

View Full Version : Very Telling Data From Ia Poll - Is Ron @ 21%?!?!




bc2208
12-08-2007, 03:13 AM
Finally, some usable data. Follow me through on this one.

All data taken from Newsweek poll pdf:
http://www.newsweek.com/media/75/0714_newsweek_poll.pdf

What's great about this poll is they give the data for all polled people, with subsections. This allows us to eliminate biases.

1408 people were polled, of which 5% supported Ron Paul. This is 70 voters.
Of the 275 likely Republican caucus-goers, 8%, or 22, support Paul.

What I've done is thrown away their idea of likely caucus goers, because we don't fit that definition. Every Paul supporter will vote!

NOW, we can fit all 70 Paul supporters into the likely Republicans group, and instead of 22 of 275, we have 70 of 323.

70/323 = 21.7%!

bc2208
12-08-2007, 03:14 AM
If you remove the "likely goers" from the whole thing, you have 70 of 588, which is 12%, still an improvement. This assumes our supporters are registered. And they better be!

ashlux
12-08-2007, 03:16 AM
100% of Ron Paul supporters voting is a bit too optimistic.

If only... :D

But you're right, I suspect that'll help us quite a bit.

_rp2008_
12-08-2007, 03:53 AM
1408 people were polled, of which 5% supported Ron Paul. This is 70 voters.


This is incorrect. The 1408 figure is the total number of people---Republicans, Democrats, others---polled. The 5% figure is the percentage of Republicans polled who support Paul. It is inappropriate to multiply these two figures. The number of Paul supporters is correctly determined by multiplying the total number of Republicans by the percentage of Republicans who support Paul: 540*5%=27.

You correctly determine the number of Paul supporters likely to vote: 22.

Thus, according to the poll, 22/27=81% of Paul supporters are likely to vote.

Contrast this with Romney: 107/184=58% of Romney supporters are likely to vote.

Thus, the poll seems to reflect your instinct that Paul supporters are more likely to vote than other supporters.