[EL] Corrected subject line: Question about polling

[David Redlawsk]

This is done to improve response rates of young people. Several 
survey organizations use a version like this quite regularly that 
asks for the youngest male "now at home" adding "who is registered to 
vote" for RV efforts. It looks like this one randomly chooses "male" 
or "female".

Even with this, the proportion of younger voters reached is lower 
than their share of the population, but if something like this isn't 
done, the proportion drops precipitously.

It works reasonably well because "the youngest" is distributed 
broadly across households. If you ask for the youngest male in my 
household, it is me, at 54 years old.  It is also really important to 
recognize that whatever the "skew" is of the sample, the reported 
results are weighted to known demographic factors like age, 
race/ethnicity, gender. So if this process pushes the average age of 
the sample down, it is probably just lessening the weight that has to 
be applied to younger people who are much harder to include in the 
sample in the first place.

There needs to be some method to select the respondent in a household 
- some use birthdays, or other random selection. And of course, if 
the poll is using a listed registered voter sample, then we ask for 
the person on the list by name.

Dave Redlawsk

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
David P. Redlawsk
Professor of Political Science
Co-Editor, Political Psychology

Director, Eagleton Center for Public Interest Polling
Eagleton Institute of Politics
Rutgers, The State University of New Jersey
Wood Lawn, 191 Ryders Ln
New Brunswick, NJ 08901-8557
(732) 932-9384 X285
redlawsk at rutgers.edu
http://www.rci.rutgers.edu/~redlawsk


At 11:15 AM 9/24/2012, Scarberry, Mark wrote:

>
>From: Scarberry, Mark
>Sent: Monday, September 24, 2012 7:30 AM
>To: law-election at uci.edu
>Subject: RE: [EL] US Dist Ct ruling on Colorado bar codes on ballots
>
>This question is not about the law but rather about political 
>polling. Perhaps our moderator will let me know if this is not a 
>proper question for the list.
>
>There is a surprising feature (at least to me) of the 
>Politico/GWU/Battleground poll released this morning (which shows 
>the President with a small lead). At the very beginning, for calls 
>made to land-lines, the poll-taker asks to speak to the youngest 
>member of the household who is registered to vote. See 
><http://images.politico.com/global/2012/09/battlegroundpoll.html>http://images.politico.com/global/2012/09/battlegroundpoll.html. 
>Am I misreading the poll questions? If I'm not misreading them, then 
>how can a poll reach a random sample of likely voters (as this one 
>was apparently designed to do) if it skews toward the younger voters?
>
>Mark S. Scarberry
>Professor of Law
>Pepperdine Univ. School of Law
>_______________________________________________
>Law-election mailing list
>Law-election at department-lists.uci.edu
>http://department-lists.uci.edu/mailman/listinfo/law-election
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://webshare.law.ucla.edu/Listservs/law-election/attachments/20120924/04a50418/attachment.html>