Hey Political Science geeks & majors: I have a question about surveys

OmahaBlueDog

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 02:58 PM
Original message

Hey Political Science geeks & majors: I have a question about surveys

Edited on Sun Aug-31-08 02:59 PM by OmahaBlueDog

I spoke the other day to someone in the survey business related to market research (not ploitics). They do random outcall phone surveys to gauge public perception of products, services, industries, etc.

One think they've noticed. All of their stats are skewing to an older age demographic. Two reasons. First is that increasingly, people under 30 don't have land lines, and most phone lists don't capture wireless numbers. The second is that younger demographics are far more likely to use caller ID, and don't pick up unknown numbers.

My question: Do these factors tend to skew the polls for Obama or against Obama? Or do Gallup/Zogby/Rassmussen have some way around this?

Printer Friendly | Permalink | | Top

Captain Hilts

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 03:03 PM
Response to Original message

1. Actually, a much broader demographic now has cellulars and no land lines.

It still skews young, but it's changing.

Answering machines are a problem, but, again, you can compensate for that but it costs you time and money.

Traditionally, younger people have lousy turnout records when it came to voting. Something like 17% of eligible voters under 30 showed up in 2004. I think Obama will change that somewhat, but I don't know by how much. So, yes, based on this anticipated increase in youth voting, he might have some numbers slightly depressed by survey techniques.

The problem is, McCain's demographic - older, wealthier, gun-owning, etc. have VERY high voter turnout rates.

Three poli sci degrees and 8 years in political polling here.

Printer Friendly | Permalink | | Top

ddeclue

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 03:06 PM
Response to Original message

2. Against Obama...

The phone #'s for cells are hard to come by so it is hard to get polls using them. Therefore the samples tend to skew older and this will be a more McCain friend demographic.

Doug D.

Printer Friendly | Permalink | | Top

Vincardog

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 03:08 PM
Response to Original message

3. Yes it skews the polls against Obama and yes Gallup/Zogby/Rassmussen count on it

When 80% of the country believe we are on the wrong track; and McSame promises more of the McSame.
The race should be a LANDSLIDE BLOWOUT.

They have to "Qualify" the sample (IE Likely voters) to get any kind of race at all.

They also manufacture crap (like PUMAs) trying to make it close enough to steal.

Printer Friendly | Permalink | | Top

OmahaBlueDog

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 03:49 PM
Response to Reply #3

4. The "wrong track" question can be deceptive

IIRC, a majority said we were on the wrong track in 2004, but 25% of those who responded in the negative thought we were on the wrong track since the nation wasn't moving to the right fast enough.

Printer Friendly | Permalink | | Top

Vincardog

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 06:57 PM
Response to Reply #4

6. I will go with the millions new uninsured and underemployed 'wrong track' myself

Printer Friendly | Permalink | | Top

Captain Hilts

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 04:35 PM
Response to Reply #3

5. "Likely voters" has traditionally been the most accurate population to use for surveys...

FAR better than "registered" voters.

This population has always been preferred for political survey research. It's not someting new and sparkly to 'get' Obama.

Printer Friendly | Permalink | | Top

eshfemme

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 07:00 PM
Response to Reply #5

7. The only problem with that is filtering for likely voters

usually means that they discount the younger demographics that traditionally don't vote reliably as a bloc as the older voters. With the surging turnout amongst the younger voters and the lack of coverage of the impact of the groundwork, there's no way to know if the filtering of the polls will be accurate for the current election cycle. That's the only problem I have with that thinking that "likely" voters are a better group to poll for accuracy.

Printer Friendly | Permalink | | Top

Gormy Cuss

(1000+ posts) Send PM | Profile | Ignore

Sun Aug-31-08 07:05 PM
Response to Original message

8. If the poll is done correctly, the bias can be mitigated with weights

IF they are confident that they're getting representative respondents among the age group and they have enough responses to make the adjustment. Survey research people monitor methods as communications and attitudes about participation change. Mail surveys are hardly ever done anymore because the response rate is so abysmal. I would imagine that the traditional random digit dial telephone surveys are on their way out too in favor of other methods like large panels responding by internet, to cite one example.

Printer Friendly | Permalink | | Top

DU AdBot (1000+ posts)

Mon May 06th 2024, 03:00 PM
Response to Original message

Advertisements [?]

Top

Powered by DCForum+ Version 1.1 Copyright 1997-2002 DCScripts.com
Software has been extensively modified by the DU administrators

Important Notices: By participating on this discussion board, visitors agree to abide by the rules outlined on our Rules page. Messages posted on the Democratic Underground Discussion Forums are the opinions of the individuals who post them, and do not necessarily represent the opinions of Democratic Underground, LLC.