Why Did CBS News/New York Times only poll 283 Democratic Primary Voters?

[PBS Poll-435]

Why in the poll that says Obama has rebounded and that Obama & Clinton are both overwhelmingly preferred against McCain?

They don't have an agenda, do they?

;-)

http://www.realclearpolitics.com/articles/docs/MAY08A-Politics.pdf

[PBS Poll-435]

1. Anyone?

Such a groundbreaking conclusion and they only bothered to ask 283 people?

[jordi_fanclub]

2. Read is a GIFT!... Be BIASED is a CON "characteristic"...

Exactly from your link:

This poll was conducted among a random sample of 671 adults nationwide, including 283 Democratic primary voters, interviewed by telephone May 1-3, 2008.

[PBS Poll-435]

4. LOL!!! ABC Hates Obama, right? Better do Potato-head's show instead. ROFL

And why draw such a broad conclusion from only 283 Democratic Primary voters?

[jordi_fanclub]

It's... guess... the MATH!

In any (and I mean ANY) poll the entired sample only is relevant to the error margin.
This poll have a 4% error margin; if you double the sample you ONLY reduce the error margin NOT the poll result!

[PBS Poll-435]

12. If you double the sample, you might get a different result.

[saltpoint]

16. Is that a mathematical / sociological fact, reliable and valid?

Again, you have improved numbers in both the smaller and larger samples.

[jordi_fanclub]

17. Let's say...

You NEVER get a different result in the sense of a "completely different result".
By doubling the sample you only get a REDUCED error margin... say to 3%.
But the RESULT (as winner-loser) is exactly the SAME.

[hokies4ever]

19. Exactly!

Sure they could have polled a lot more people in order to decrease the margin of error from 4% to 3%, but that wouldn't even come close to substantially changing the results that they got. If it makes you feel better, just add one more point to Clinton then interpret the same results.

Also, this poll is much better than most other polls. They actually asked a very lengthy set of questions instead of just asking who they support.

[Forkboy]

3. I'm upset that it wasn't an AOL poll.

[anigbrowl]

5. Because they're polling the electorate, not just the party base

Also, recall that Indiana is an open primary.

[PBS Poll-435]

7. Adults, nationwide is not the electorate.

Registered, and more accurately "likely," voters are a better gauge.

[anigbrowl]

20. Not all of whom are Democrats. Nice goalpost shift though...

make a general complaint, and then shift to being pedantic when it's addressed.

[FrenchieCat]

6. Do you have time to change your stupid uninformed headline....

that doesn't reflect on the poll you are inquiring about.

[PBS Poll-435]

10. Do you have time to put me on ignore.

Please.

Thanks!

[PseudoIntellect]

8. Lol, those +12/+10 numbers were clearly off.

Didn't realize they polled that few people though.

[saltpoint]

9. Zogby & Gallup also have Obama with improved numbers over the

weekend.

I'm not a huge CBS fan per se, but I don't see evidence that their methodology reflects a bias, particularly.

[PBS Poll-435]

11. Biased? Maybe/Maybe Not.

But sloppy? Yeah.

[saltpoint]

13. Well, what is the threshold for margin-of-error reflection, then?

By what differential is 375 likey voters a closer read than 289 likely voters?

I'm not arguing it is or isn't; only asking into the differential threshold.

And how is it, then, that Gallup would show the same improved numbers albeit with the larger sampling?

And in any case, the New York TIMES endorsed Sen. Clinton for the nomination.

[StevieM]

14. And they have since changed their minds. That's fine, but I disagree with their opinion and I don't

trust their polling.

Steve

[saltpoint]

15. The TIMES' Op-Ed page has offered criticism of campaign tactics but has not

withdrawn its endorsement.

In a larger sample, Gallup reflects the same improved numbers.

[Warren Stupidity]

23. Got link for that?

As far as I know, the NYT has not changed is editorial opinion. If it has, please provide the link to the NYT editorial, not some op-ed piece, that endorse Obama over Clinton. Otherwise I just have to conclude that you invented that 'fact'.

[RichardRay]

21. Not sloppy unless you can tell me how they picked their sample.

If they made a sampling error they can poll 1,000,000 people (think of any online poll where the respondents are self-selected) and end up with absolutely nothing but noise. A really well selected sample gives an improved result with a larger sample size only in the reduction of th MoE.

So, if they're sloppy please trot out exactly how they were sloppy, but don't try to pin it on sample size.

[Hippo_Tron]

18. Instead of asking stupid questions, why don't you read a statistics book?

[Warren Stupidity]

22. It is called a statistical sample

Do you have any idea how polling works? It seems not. CBS might have an agenda, or not, but we know where the NYT stands: they endorsed Clinton in the New York Primary. The poll data indicates a fairly typical statistical sample of the voting population, broken down by various relevant categories.