Comments on: WSJ Poll: Obama Moves Past Wright, Hillary In Trouble

By: Rich Horton

Rich Horton — Fri, 28 Mar 2008 17:42:16 +0000

I also love how he “confirms” he story by taking the word of someone at NBC, and not getting the actual sample information. (I forget why dont we let people be the judge in their own case?? Oh yeah….)

This isn’t some grand mystery. If you have the sample information you release it, pure and simple.

By: Rich Horton

Rich Horton — Fri, 28 Mar 2008 17:38:02 +0000

BTW the TalkingpointMemo piece is simply wrong, although he gives himself ample wiggle room by saying things like:

But I’m pretty sure that’s not what Todd is saying.
What I think he means is
In any case, I don’t know that. But from my experience I’m pretty sure that’s what it means.

Why doesn’t he know for suer?? Because they DIDN’T GIVE THE INFORMATION NEEDED TO BE SURE. Giving that info is pretty standard, so why isn’t it here?

All yoiu have to do is look at the numbers to see they didn’t re-weight the results to account for the oversample.

By: Rich Horton

Rich Horton — Fri, 28 Mar 2008 17:32:04 +0000

It certainly duplicitious when it is being used to say it gives a meaningful measure of the choice of Clinton v. Obama. That isn’t an opinion. That a fact of polling requiring statistical random sampling.

Now, there are legitimate reasons for oversampling, like wanting to have a large enough “n” for meaningful crosstabs (the rationale used here), but when you do that the only meaningful measures exist WITHIN each candidates support. Thus this particular poll can tell you something about what Obama supporters think. It is not a measure of election preference.

And yes, my local paper (The Pioneer Press) reported this as a straight ahead scientific poll, as did MSM outlets everywhere.

I teach this stuff for a living and I often see it misunderstood, but even if you KNOW WHAT TO LOOK FOR in the sample that information was not given, and could only be inferred by doing a little stats on the numbers. How is that normal, fair and above board?

By: Donklephant » Blog Archive » Polls Show Obama Moving Past Wright Controversy

Thu, 27 Mar 2008 21:16:14 +0000

[...] all of you who thought yesterday’s WSJ poll was wrong, here are two more polls to support the [...]

By: Justin Gardner

Justin Gardner — Thu, 27 Mar 2008 16:10:34 +0000

Duplicitous? Kind of a harsh word for a poll that’s weighted…like every other poll in existence.

But you’re most likely not going to accept my explanation so go here:
http://talkingpointsmemo.com/archives/185845.php

By: Rich Horton

Rich Horton — Thu, 27 Mar 2008 15:15:02 +0000

This “poll” is duplicitous as they admit they deliberately oversampled black voters. Since that group breaks 90-10 for one of the candidates the results are meaningless.

By: Dos

Dos — Thu, 27 Mar 2008 14:36:10 +0000

His Wright cancer has gone into remission. But as we speak, my 527 loving named, “Stop the Crazy America-Hating Muslim from Killing White Children” (SCAMKWC) is receiving thousands of dollars and have contracted with a production company to make a 5-7 minute video that…..well, I don’t want to give it away, but if the Swift Boaters were an uncomfortable rectal exam for Kerry, this is going to be a full blow prison gang rape with shanks included. Enjoy, because it ain’t going away.

By: Bubbles

Bubbles — Thu, 27 Mar 2008 05:44:14 +0000

Obama’s not back. Hillary will fail too.

Dude, I don’t know about you, but I’m for Gravel.

What’s better than a committed socialist running under a party that supports anarcho-capitalism?

By: mw

mw — Thu, 27 Mar 2008 05:18:37 +0000

I agree that the Wright issue has been completely marginalized. He has dealt with it well. His only issue now, is that more Democrats want Clinton as the nominee.

By: Avinash_Tyagi

Avinash_Tyagi — Thu, 27 Mar 2008 02:45:20 +0000

He’s even 2 points ahead of McCain note its RV not LV, because Obama gets alot of support from people who aren’t considered Likely voters, I for one would probably not be considered a likely voter, but I will vote this year