You are viewing m_francis

Previous Entry | Next Entry

Ye Olde Statistician

Razor's Edge
Today's Lesson in Sampling Theory


Tags:

Comments

( 2 comments — Leave a comment )
jjbrannon
Oct. 15th, 2010 01:13 am (UTC)
I'm a mere pawn in the game
I'm a lifelong registered independent who votes in every election for which I have been eligible.

I live in a tiny of polity of 610,000 voters.

I eschew cellphones and Caller ID -- Life's a Banquet! Try the Chef's Surprise! -- so I always answer the phone.

Ergo, I have been part of three polls thus far this political season.

The last time was Monday when the Monmouth University Polling Institute polled my opinion for the Delaware US Senate race.

Thus, Tuesday I -- anonymously -- made the AP and MSNBC via the poll results.

http://www.monmouth.edu/polling/admin/polls/MUP37_DE.pdf

The one flaw in the methodology that leaped to the fore of my mind was the polling of **certain** or **likely** voters who had voted in at least 2 of the last 4 general elections.

As a rule of thumb, regardless of major party, in off year elections as this one, approximately 15% of registered voters turn out for their respective party's Statewide primary and 25% for the general.

This year 32% of registered Republicans -- many of whom are newly added to the rolls -- voted in the primary, offering a narrow lead to O'Donnell.

This poll cannot include this cohort.

However, because I am familiar with the voter demographics, the day after the primary results, I predicted that O'Donnell would not garner even 40% in November.


JJB
princejvstin
Oct. 18th, 2010 12:40 pm (UTC)
"How to Lie with Statistics"

It is still a classic.
( 2 comments — Leave a comment )

Profile

Razor's Edge
m_francis
m_francis

Latest Month

April 2014
S M T W T F S
  12345
6789101112
13141516171819
20212223242526
27282930   
Powered by LiveJournal.com
Designed by Taylor Savvy