TPM's got exit polls, but they're useless

Bucky

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:41 PM
Original message

TPM's got exit polls, but they're useless

TPM's exit poll are from unweighted data. For instance the Georgia results are certainly off by double digit margins of error, even if Obama sweeps it. But with that caveat stated, here's the exits that TPM is posting

http://www.talkingpointsmemo.com/

GA: C- 25.5, O - 75

CT: C - 45, O - 52.2

IL: C - 29.1, O - 69.6

AL: C - 37, O - 59.6

DE: C - 41.9, O - 55.6

MA: C - 47.3, O - 49.8

MO: C - 45.1, O - 49.8

TN: C - 51.6, O - 41.1

NY: C - 55.6, O - 42.2

NJ: C - 47, O - 52.2

AR: C - 71.2, O - 25.5

OK: C - 60.5, O - 30.4

AZ: C - 44.8, O - 50.5

Printer Friendly | Permalink | | Top

Sulawesi

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:42 PM
Response to Original message

1. I dont understand what unweighted means in this context

Printer Friendly | Permalink | | Top

slick8790

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:43 PM
Response to Reply #1

2. Maybe no delegate totals?

I was confused too.

Printer Friendly | Permalink | | Top

Bucky

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:56 PM
Response to Reply #1

6. Unweighted means they didn't adjust for overrepresented demographics

Like say they interviewed 70% male, but men only represent 35% of those voting Democratic. They'd have to adjust those numbers down. Exit polling is not a very scientific method of surveying because the interviewer has less control over who they select and the scientist supervising the data collection has a harder time ensuring quality control of the field researchers.

Printer Friendly | Permalink | | Top

jefferson_dem

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:57 PM
Response to Reply #1

7. It means...

the sample used to collect the exit poll data is not necessarily reflective of the actual population of voters. Traditionally, polls make sure the sample characteristics are representive, which might mean they must weight certain respondents more than others on the basis of age, race, gender, etc.

Printer Friendly | Permalink | | Top

Bucky

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 07:00 PM
Response to Reply #7

9. But even within a demographic group, it's still not random sampling

Say extroverts liked Obama & introverts like Clinton--you might oversample Obama. Or say uggos like McCain and Hooter's Hotties like Romney and the polling organization over hired college age boys. You'd get a pro-Romney skew in any sampling there.

Field science is not the same as lab science.

Printer Friendly | Permalink | | Top

AJH032

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:48 PM
Response to Original message

3. these seem really off...

Clinton winning Arkansas by 45%? Obama winning Georgia by 50?

huh?

Printer Friendly | Permalink | | Top

slick8790

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:49 PM
Response to Reply #3

4. They match up with the polls in this thread:

http://www.democraticunderground.com/discuss/duboard.php?az=show_mesg&forum=132&topic_id=4414111&mesg_id=4414111

But who knows, could be from the same (flawed) source.

Printer Friendly | Permalink | | Top

Bucky

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:53 PM
Response to Reply #4

5. Must be. They're "unweighted" returns

meaning that they don't reflect the demographics of the electorate.

Printer Friendly | Permalink | | Top

usregimechange

(1000+ posts) Send PM | Profile | Ignore

Tue Feb-05-08 06:58 PM
Response to Original message

8. That would be great for Obama!

Printer Friendly | Permalink | | Top

DU AdBot (1000+ posts)

Thu May 02nd 2024, 03:08 AM
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.