2016 Postmortem

regnaD kciN

(26,045 posts) Wed Oct 17, 2012, 06:19 PM Oct 2012

An interesting observation from 538...

Even the best surveys these days only manage to get about 10 percent of people on the phone, while the shoddy ones might struggle to get 3 or 5 percent of voters to return their calls. These percentages have fallen precipitously over the past two decades.

Polling firms are hoping that the 10 percent of people that they do reach are representative of the 90 percent that they don’t, but who will nevertheless vote. But there are no guarantees of this, and it is really something of a leap of faith.

http://fivethirtyeight.blogs.nytimes.com/2012/10/17/oct-16-can-polls-exaggerate-bounces/#more-36238

So, irrespective of everything else in this article, every prediction we've bee seeing involving polls, whether pro-Obama or pro-Romney, is based essentially, not just on a cross-section of the populace that gets called, but on 10% (or 5%, or 3%) of that cross-section that responds? Given that, it's amazing that polls come anywhere near predicting the final result.

3 replies

= new reply since forum marked as read

Highlight:

An interesting observation from 538... (Original Post) regnaD kciN Oct 2012 OP

There was a great segment about this on NPR last Friday. DollarBillHines Oct 2012 #1

GOOD ONE! elleng Oct 2012 #2

Gallup is now claiming that their land/cell list is 50-50 DollarBillHines Oct 2012 #3

DollarBillHines

(1,922 posts)

1. There was a great segment about this on NPR last Friday.

Reply to regnaD kciN (Original post)

Wed Oct 17, 2012, 06:32 PM

Oct 2012

About the difference between the "polls" and the odds houses.

Heck the 7-11 soda-cup survey has never been off by as much as 1% and has Obama up by 57-43.

http://lifeinc.today.com/_news/2012/10/10/14340658-obama-wins-by-landslide-in-7-eleven-coffee-cup-survey?lite

elleng

(131,140 posts)

2. GOOD ONE!

Reply to DollarBillHines (Reply #1)

Wed Oct 17, 2012, 06:49 PM

Oct 2012

Also, poilsters recognized recently problem w land/cell phone calls ;pollsters didn't, and probably still don't know how to deal with distinction, and knowing who/which they're calling, whether they should set up rules, like asking: Is thus a land or cell phone. There's a difference in the demographic, which resullts in different results.

DollarBillHines

(1,922 posts)

3. Gallup is now claiming that their land/cell list is 50-50

Reply to elleng (Reply #2)

Wed Oct 17, 2012, 06:58 PM

Oct 2012

How many times/how often do you answer an Unknown Caller on your cell?

Do you think Gallup actually leaves a call-back number?

Heck no.

Also, the odds-makers have real skin in the game (dollars).

I pay no attention to any poll.

Well, unless it is one that calls me The Coolest Guy in Town.

Reply to this discussion