Luke Russert must have flunked his college statistics class.

[ShortnFiery]

He's now talking with Andrea Mitchel showing a poll asking the question: "Are you confident of X candidate as commander and chief of the USA?"
50% said "yes" for McCain and 48% said "yes for Obama with a 2% Margin of Error.

Luke Russert is drooling all over himself that Obama has not really made up the gap because McCain's percentage of slightly (2 points) higher.

HELLO?!? College boy? When the difference between results are within the MOE then it is a statistical tie. :eyes:

I KNOW that the M$M is bias for the GOP ticket, however, it seriously pisses me off when they assume that I'm STUPID. :grr:

[atreides1]

1. Consider the Source

One legacy son trying to boost another legacy son!

[HereSince1628]

2. Legacy newscasters...look where that got us with the Wallaces. n/t

[CatBO]

3. Nepotism... It gave us Bush...

It gives us this loser too.

[Lex]

4. he didn't get the job on merit and he's not too bright

nt

[Ichingcarpenter]

5. He never had statistics

I think you can op out of that with a history and communication degree.

But not with a poly/sci degree.

[Overseas]

6. But he passed Kiss Up to the Station Owners 101. //nt

[woolldog]

7. He's a tool

[Critters2]

8. Imagine if they let me design farm machinery just because my dad did.

I'd be just as skilled at it as Luke is at this.

[Heather MC]

35. My father was a Mechanic, so clearly I am qualified to rebuild an engine and send out on the road!

[Secret_Society]

9. I went to school with him and that's very possible

[rch35]

12. I post things twice. nt

[Bake]

10. But math is HARD!

Kinda like presidentin'.

:rofl:

Bake

[Secret_Society]

11. dupe

[rch35]

13. I post things twice. nt

[ShortnFiery]

16. Play nice.

Don't make me tweak you. ;) :hi:

[rch35]

37. Haha, I am just bored and having fun, cant help myself sometimes.

:) :hi:

[ShortnFiery]

15. Wow! He was seriously lacking in SOMETHING. I've seen news models preform more genuinely

than this guy. He just STINKS in the smarmiest, most partisan way. :(

[Viking12]

14. A difference within the margin of error IS NOT a statistical tie

Seems that Luke Russert is not the only one to flunk basics stats.

[ShortnFiery]

17. Bullshit! You have a "statistical dead heat" if the percentages are within the MOE.

But hell, I only tested into and achieved an A- in a PhD. guided psychological statistical design course at my low life "state" university ...

so you just may be correct. :eyes:

[Viking12]

19. Swearing doesn't make it so.

BTW I am PhD that teaches research methods at a state university.

Here's a brief tutorial:

http://www.washingtonmonthly.com/archives/individual/2008_08/014294.php

[ShortnFiery]

20. Check my source? Gee, I'm a Ninja ... one upsmanship.

BTW, I accept your apology. ;)

[ShortnFiery]

21. I guess someone in your life didn't show you adequate attention today?

If you want to have a pissing contest with me, let's take i PM.

If you are out to just "trash" another person, well that is less than classy teach. :(

p.s. I'm so glad I did not continue to Pile it Higher and Deeper. Academics are too often "lost in thought" and caught up with splitting hairs.

Statistics is difficult enough for the general public to understand without such qualifiers.

I stand by my original statement albeit you make some valid points about reqirements for closer analysis of EACH poll's methodology.

Have a nice day. :-)

[Viking12]

22. Let me get this straight.

You're wrong, but stats are hard so we should allow misperceptions to flourish. Is that what you're saying?

[ShortnFiery]

23. I may be TECHNICALLY WRONG, or NOT (we don't know the true methodology) ...

But you are just behaving in an ultra-negative way.

Now, if you genuinely BELIEVE that people are going to look at all the specifics of a polling sample and analyze it as "a whole", then yes, go ahead and say that using the MOE is "full of shit."

I submit that, without the specifics of the sampling methodology, using the MOE as an indicator, is basically sound.

I know that you are going to go ON and ON and ON about how smart you are and how wrong I am, but that's what PhD's do ... get lost in the forest.

Knock yourself out Professor. :eyes:

[Viking12]

24. You're a walking self-contradiction

First, you claim how smart YOU are because you took a course that was guided by someone with a PhD -- YOU made the claim to the authority of the degree, not me. Since you'd established that baseline, I merely pointed out that I actually have a PhD and that you were sorely mistaken. Now, in a startling divergence from your earlier position, people with PhDs are worthless. Your anti-intellectual streak would serve you well at a Palin rally. If your feelings were hurt because someone on a message board pointed out the you are wrong, then you need a different hobby.

[ShortnFiery]

25. Come on, you were rejected by someone today and piling on me helps you to make up for it?

I never said that I was "an authority" ... I typed that I tested into a course reserved for PhD students.

However, this is where you and I disagree. Because people will NOT delve deep into the methodology of Polls, specifically I believe the TAINTED (who knows how they are asking the questions) Polls of this Presidential Election, it is NOT WRONG to use the MOE as a guide.

Most lay people use MOE as a guide because we don't have ANYTHING better if you don't wish to "dive into number crunching" (p values anyone?).

How would you suggest that we interpret THIS poll when they give only the percentages and the MOE?

No, my feelings are not hurt because you are the one who seemingly has something to prove to whomever rejected you earlier. :shrug:

[Viking12]

26. I've had a very pleasant day, thank you.

However, I don't suffer fools well. Your OP was foolish. Your subsequent posts, equally foolish.

[ShortnFiery]

29. I have to admit Professor, I worked my a** off to genuinely earn my Masters, but I didn't like many

of the PhDs whom I encountered. Yes, I do believe that "a significant" percentage of Academia LOOK DOWN upon those of us with average intelligence. Most professors appreciate those of us sporting average intellects but there are some who feel a need to behave like snobs.

No, all I can surmise is that someone hurt them early in life and the way they make up for not feeling like a complete human being is by intellectually putting people in their place ... in a public forum at every damn given opportunity.

You must be "a laugh riot" at parties? ;)

[Viking12]

32. Isn't putting Luke Russert "in his place" the reason you started this thread?

Project much? You've got issues, I'll let you work them out on your own.

I've got a canvassing shift to attend. Have a nice life, I hope things get better for you.

[ShortnFiery]

34. Luke Russert was gushing over the SIGNIFICANCE of this sad poll. He is a public persona, I'm not.

I extend my empathy to any freshman student who dare challenge your intellectual superiority.

You see, unless you are a friend or relative of Luke Russert, your behavior here is akin to "a hyena."

I'm sorry that you must behave in such a manner just to "prop yourself up."

Have a nice life, I hope things get better for you.

[dbmk]

40. You could interpret it as close.

And that there was an outside chance that more actually saw them as equals or an even slimmer chance that Obama actually was leading.

[ShortnFiery]

46. I'll give you that. I withdraw my contention of "a tie" but I place little veracity on one poll.

:-) :hi:

[BlooInBloo]

31. BWAHAHAH!!!!

[Usrename]

36. Let me see if I get what you are saying here.

If there were six polls taken, we would expect one of the six polls to yeild this exact result, even if the two candidates are actually tied.

Is that it?

[dbmk]

41. Not at all what he was saying.

Can't quite figure out where you dug that conclusion out from.

What he was saying was that at 2% MOE and 2% behind(all other things being correct and equal), then there is a 95% chance that McCain is actually doing better.

That is not a tie. Its a bogus term that misrepresents the theory behind polls and it degrades the common knowledgebase. It teaches you to believe stuff that is not true.

[Usrename]

42. From the link he gave, there is only an 83% confidence factor.

Did you look at the table?

http://www.washingtonmonthly.com/archives/individual/2008_08/014294.php

2% lead with 2% MOE shows an 85% chance that the candidate is really in the lead.

There seems to be a one in six chance that the poll is wrong. Where do you get the 95% confidence factor from?

[dbmk]

44. Sorry, I might have messed up when I looked it up

85% seems to be the number, indeed. Does not detract from my point, though, as far as I can tell.

If there is an 85% chance that you are really behind, calling it a tie is not correct.

And that is not the same as saying that there is a 15% chance it would be like that if they were really tied.
There is a 15% chance that Obama is really ahead. By any given number.

[Usrename]

47. Still, there is a one in six chance that the poll is wrong.

I think that falls below the threshold for having ANY statistical significance, which supports the OP's point, doesn't it?

[ShortnFiery]

18. Here's a little reading for you. You're welcome. :-)

http://www.oswego.edu/~srp/stats/stat_tie_2.htm

A poll is said to show a statistical tie when two candidates’ numbers fall within the sampling margin of error, a measure of how confident pollsters are in their results.

[dbmk]

39. Correct

Only if the numbers are the same is it a tie. Statistical or not.

"Statistical tie" is a bogus term. It implies equality, where there usually is none.

If you poll lower you are polling lower. No ifs or buts. Then you can argue whether the difference is significant. But thats another discussion.

[bluedawg12]

48. For the sake of discussion: a statistical tie would be if

two candidates had 50% vote the vote, and the numbers had a say 95% confidence level with-in the first SD?

In another words, the poll numbers are tied and the confidence in the poll makes it statistically likely?

A 50/50 poll with a 50% confidence level would be a sloppy poll, tied, but low confidence level...and.. say..that would be a mclame pollster! :rofl:

[Medusa]

27. The Kid didn't get his job based upon what he knows

they did not hire him for his smarts.

[bushisanidiot]

28. I saw that this afternoon on my lunch hour. He was saying shit like

"The Obama campaign up in Chicago can't be too happy about that." He's clearly shilling for the repukes and I'm surprised Andrea Mitchell didn't question him on his theories because they made no sense. I guess she thought she'd be disrespecting her late friend Tim if she did that.. so Luke get's a free pass to spew whatever talking points he wants.

[On the Road]

30. If the Results are Right At the MOE

there is a 95% chance they are due to a genuine difference rather than sampling error. (There are other sources of error like response error which are not factored into the MOE.) So I would agree with Russert here.

What I disagree with is that this is good news for McCain. Military leadership was one of his chief selling points and supposedly one of Obama's weaknesses. If McCain only has a two-point lead in this department, he's in trouble.

[BlueIdaho]

33. So much for the whole nature/nurture argument...

[David__77]

38. Technically, it's not a tie.

It could be a bigger gap in McCain's favor, or a small one in Obama's favor, at a certain level of confidence. But to speak of a tie is inaccurate. That said, his spin is biased for other reasons.

[ShortnFiery]

45. You and the professor are "technically" correct. However, when you don't know the methodology

I get "less than enthused" by the significance when the numbers are within the MOE.

Admittedly - by strict definition - (oh how this stings) I've been dressed down.

There are NO other polls re: CIC other than the earlier ones that show vast improvement of Obama's percentage to the point where it does fall within the MOE.

No, NOT "hard and fast" a tie .... But let's just say WITH CONFIDENCE, I'd NOT *bet the farm* on this polling data, i.e., one poll of small sampling does not result in SCIENTIFIC conclusions. :evilgrin:

[Cha]

43. Rec'd~ They always act like

we're stupid.