From A Poster At Princeton Consortium Re: The Cell Phone Effect

Peter D // Oct 24, 2012 at 1:45 pm

OK. Quiet down. Original research here.

Pollster has 23 polls entirely conducted after the first debate on October 3. 2 of them are internet polls, which I threw out.

8 of them polled cell phones. 11 of them did not. Of the ones that did not poll cell phones, O led by an average of 0.4%.

Of the ones that did poll cell phones, O led by an average of 3%.

Using the reported % of cell respondents suggests a beta of 0.1. So for each 10% of the electorate that are cell onlies, the O margin increases by 1%.

Based on the latest CDC estimates, 33% of Ohioans 18 and over are cell-only and 18% are cell-mostly.
(http://www.cdc.gov/nchs/data/nhsr/nhsr061.pdf)

The tricky part is translating this into turnout. However, the modal (non-zero) industry weighting of 22% seems fair, if not low. My parents, for example, have stopped answering their landlines due to the deluge of political calls.

Conclusion: the Ohio lead is at least 2.5%, possibly, larger. I remain unconcerned. I may repeat this analysis for Virginia at a later time.

http://election.princeton.edu/2012/10/23/debate-3-obama-wins/#comments