Wow-only 763 LV were polled for the new Gallup poll

[59millionmorons]

Are you kidding me, this sample size is laughable.

[Feanorcurufinwe]

1. How many votes were discarded as 'unlikely voters'?

:shrug:

[Mr_Spock]

2. Only votes that count are in November

n/m

[Name removed]

3. Deleted message

Message removed by moderator. Click here to review the message board rules.

[59millionmorons]

4. I would reply to you

If I knew what the hell you were talking about. Speak english. There is a reason for a 4% MOE.

[Name removed]

8. Deleted message

Message removed by moderator. Click here to review the message board rules.

[BloodyWilliam]

9. Any MOE +- 5% is acceptable.

And ~800 is extremely standard for national polls.

[Name removed]

5. Deleted message

Message removed by moderator. Click here to review the message board rules.

[Name removed]

6. Deleted message

Message removed by moderator. Click here to review the message board rules.

[Name removed]

12. Deleted message

Message removed by moderator. Click here to review the message board rules.

[BloodyWilliam]

7. Polls can be accurate with samples as few as 99 individuals.

~1000, anywhere around 800 respondents and up, are pretty standard for any poll. As long as the respondents are chosen randomly, any sample over 100 is accurate (even if it'll have a bigass MOE).

[depakid]

11. I'm not sure what, if anything, you know about polling

so it's actually hard for me to reply-

N= 709 may or may not be an adequate sample size, depending on how the multistage clustering was done. In addition, if there are other forms of selection bias, such as non-response bias, it doesn't matter how large the sample is- the results won't be valid.

I'm not going to go over all of the various reasons (including the nefarious ones) that "nationwide" poll results aren't very meaningful or trustworthy- Gallop polls in particualar have a bad reputation, in part due to their shoddy methodology and in part due to their historical affiliation with GOP causes. Suffice it to say that poll results are easily manipulated and the trend in this country over at least the past 10 years is for polls to be used more as persuasive devices than as descriptive tools. As often as not, they attempt to influence public opinion, rather than enumerate it.

[Reciprocity]

10. I was in the Zogby poll.

This poll was conducted Thursday before Kerry's speech. I think they wanted a base to compare to after his speech. I'm in the anyone but Bush camp. My feelings run deep so of course nothing has change on my survey. So I was unable to give Kerry a bump.


Zogby Poll Shows Democratic Ticket Up 5 Points

The poll by Zogby America showed that though the Democratic ticket didn't gain points during the four-day convention, the Republicans lost three percentage points to the undecided category.

The poll of 1,001 likely voters conducted from Monday through Thursday found Kerry-Edwards leading Bush-Cheney 48 percent to 43 percent, with 8 percent undecided. A similar sampling taken July 6-7 had the Democrats up 48 percent to 46 percent, with 5 percent undecided.

The poll had a margin of error of plus or minus 3.2 percentage points.
http://www.reuters.com/newsArticle.jhtml?type=politicsNews&storyID=5822627

[Eric J in MN]

13. In a nation of almost 300 million, would be nice if they could ask 1,000

nt