ruggerson
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:48 PM
Original message |
Is the exit polling data raw or has it been weighted? |
jezebel
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:49 PM
Response to Original message |
1. What does weighted mean? TIA. nt |
ruggerson
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:50 PM
Response to Reply #1 |
2. Meaning adjusted for what groups voted in what percentage |
|
if it's just raw numbers, it can show a trend, but it makes the state where there are closer numbers much harder to predict
|
jezebel
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:51 PM
Response to Reply #2 |
3. thanks, I think I get it now. NT |
Muttocracy
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:55 PM
Response to Reply #1 |
9. in a weighted average you take into account the % of each group |
|
in the overall total. I think the OP is asking whether exit poll results where race or sex or location is identified are just totaled, or if they're weighted to estimate the real results.
If you interviewed two men and they both voted for Obama, and two women who voted for Clinton, and just said you expect 50-50 in the results, that's unweighted.
If you took into account the fact that women vote twice as often as men (making up these numbers), or there are twice as many women as men or whatever weighting factor, it would be (for men) 0% for Clinton *1/3 of voters + (for women) 100% for Clinton * 2/3 of voters and you'd expect 33% Obama to 67% Clinton.
Silly numbers I'm using but hope that helps. If you really sampled representatively across the state, weighting wouldn't change things much.
|
MannyGoldstein
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:52 PM
Response to Original message |
4. Don't Worry. It's All Been Taken Care Of. |
helderheid
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:52 PM
Response to Original message |
5. why do they do weighted data? Isn't that dishonest? |
K Gardner
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:53 PM
Original message |
The ones we're looking at from TPM, Drudge and MyDD are raw. |
Coexist
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:53 PM
Response to Original message |
6. the stuff posted in TPM is unweighted - he believes |
|
if you are talking about the stuff on the networks, I don't know.
|
ruggerson
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:54 PM
Response to Reply #6 |
8. So then the TPM/Drudge data |
|
doesn't really tell us much about the close races, just the blowouts.
|
Caution
(1000+ posts)
Send PM |
Profile |
Ignore
|
Tue Feb-05-08 07:53 PM
Response to Original message |
7. Exit polling data is raw. Weighting occurs when one demographic in a generic |
|
poll is over-represented.
Exit polling by it's very nature is raw.
|
DU
AdBot (1000+ posts) |
Wed Apr 24th 2024, 08:53 PM
Response to Original message |