CNN poll flawed (margin of error +/- 5, small poll)

[MalloyLiberal]

The details of the CNN/WMUR poll are out, after the jump. One thing that surprised me slightly: It has a margin for error of +/-5 points, and was a pretty small poll.

http://www.politico.com/blogs/bensmith/0108/CNNWMUR_No_bounce.html#comments

[Capn Sunshine]

1. New Hampshire is a pretty small state

hey , knock yourself out rationalizing.

It's not gonna help, your girl's campaign advisors are the worst ever.
Like I've been saying for a year now.

[MadBadger]

2. I dont think Malloy supports Hill.

[Teaser]

3. state polls have shitty MOE frequently

They take a smaller sample, and you're more likely to screw up.

That said, a look at a cross section of similar polls (meta-analysis) firms up the MOE a bit.

[Breeze54]

4. All polls have a margin of error +/- 5

:P

[Colobo]

5. All of them have Obama ahead comfortably.

[Skip Intro]

6. All of them were taken all or mostly before last night's debate. nt

[MadBadger]

7. You are being factually inaccurate. The MOE for the NEW CNN poll is +/- 4

The quote you have and link refers to the CNN poll from yesterday that has Obama and Clinton tied. The Actual MOE is 4 percent which is right in line with most other polls. This is the poll linked below that has Obama up 10.

http://i.a.cnn.net/cnn/2008/images/06/topnh11.pdf

[Occam Bandage]

8. That's pretty standard for a state poll. Your OP exposes your ignorance.

[chascarrillo]

9. A +/- 5% MoE does not mean it's a flawed poll.

Really, now.