Memo to Mark Penn: Zogby Has a Real Dataset out on HRC's Electability

[CorpGovActivist]

http://news.yahoo.com/s/nm/20071126/ts_nm/usa_politics_poll_dc_2;_ylt=AjLFloLV4eQimowxAfKw30cE1vAI">What a real dataset looks like, Mr. Penn.

http://news.google.com/news?hl=en&ned=&ie=UTF-8&scoring=n&q=zogby+clinton">A round-up of stories about this real dataset.

But the voices in Mark Penn's head tell him that http://www.google.com/search?tab=nw&hl=en&ie=UTF-8&ned=&q=mark%20penn%20360%20electoral%20votes">Senator Clinton has 360 Electoral Votes in the bag.

- Dave

P.S. For poll junkies, http://www.zogby.com/news/ReadNews.dbm?ID=1393">here's the Zogby link, with methodology described.

[dmallind]

1. Actually that's what a self-selected sample looks like. NT

hk

[Captain Hilts]

2. Yes, a textbook meaningless survey. nt

[CorpGovActivist]

3. Isn't Election Day Precisely That? n/t

[dmallind]

5. Not in the same sense no

Sure people can choose whether to vote or not, but since we don't know ahead of time who will decide to vote, we must use probability samples (random samples) not non-probability samples (self-selected samples) to draw valid conclusions about their likely behavior. When you're trying to predict whole-population behavior, you have to use samples where each memebr of that whole population has as close to an equal chance of being selected as any other. No method is perfect of course, but restricting thsi to web surfers who choose to take part in surveys is WAY down the list.

[CorpGovActivist]

9. Replicating the Incremental Added Hassle...

... of throwing on sweat pants and a ballcap, and traipsing down to one's local polling station, is about the only thing missing.

; )

- Dave

[dmallind]

11. Not even close

First off you have to be on the internet, or have access to it. Know who the demographic least likely to be on the internet is? Older women. Know who the highest turnout demographic is? Older women. Big problem for non-probability sample right there.

Then you have to care about online surveys enough to sign up, and have an email account to get notice of them most likely. Who cares enough? The partisans. Why do you see so many "Du thgis poll" posts? because we are likely to be partisans who care enough about meaningless online polls to sway them our way. Same reasons the other side Freeps the same polls. If you think DU feelings towards Hillary are representative of the larger population then....well it's not polite to say really, so I'll assume you get the point and don't think we are representative. The people who fill out these opt-in online polls are the people who care a lot and get involved. The whole swathe of barely-engaged middle of teh roaders who wake up one Tuesday in Novemebr and saunter down to the polls every four years is likely massively underrepresented in these polls.

[CorpGovActivist]

16. What's Your Favorite, Most Puritan Form of Polling?

I agree that all polling types have their own built-in flaws.

I just sort of prefer that there be at least two actual data points before someone like Mark Penn pronounces a trendline.

- Dave

[dmallind]

18. Well really I'd go for a mix

Phone lines are still the best SINGLE method as not enough people are cell only to dramatically affect the outcome, and phones are as close to universal as you can get, but if I really really wanted to get the best sample I'd do a mix of phone, in person and direct mail. I'd maybe add online too, but only as a response option to a contact via one of the other methods or you have the same sample error issues when polls get "freeped" or "DU'd". Each will reach a slightly different subset of the population, and each certainly has their limitations, but the same survey presented in a mix of these different polling techniques would get as good a sample as you can. There are other ways to introduce error of course - but it all starts with the sample.

[CorpGovActivist]

25. I'd Like to See Data Broken out...

... by the number of verified times a voter has cast a ballot; I think that provides a stronger inference to the likelihood of future voting.

- Dave

[Recursion]

45. But you can't "DU" or "freep" this poll; it's not an "online" poll in this sense

The initial population came from people willing to sign up for an online poll, but it was further selected for demographic, etc., criteria. It's not like everybody who goes there and clicks on some radio buttons gets counted.

[w8liftinglady]

4. what does that make me? I took that survey...

[dmallind]

6. a self selected sample respondent

It's not an insult is it? After all I filled it out too. But all it tells us is what a subset of people who are a)online and b) want to take part in surveys thinks. It is less likely to be representative than a random sample would be.

[CorpGovActivist]

7. My Guess? "A Likely General Election Voter" n/t

[book_worm]

8. shades of 2004 all over again. Dean wasn't electable but Kerry was

It's too early to say anybody is not electable.

[CorpGovActivist]

10. I'm Curious to See the Post-Mortem Data Showing Dean...

... would've unseated Bush.

- Dave

[Renew Deal]

13. Dean could have won on politics.

The Dean scream was inevitable. I told everyone that would listen that it was coming and it came. Deans problem was personal, not politics. Kerry won on "electability", but this was a mistake. Edwards and Clark were probably the most electable.

[Renew Deal]

12. Hillary is not electable

Write it down. She will not be president.

Obama has problems too once the pukes start playing the tape of him saying we should give licenses to illegals.

[dmallind]

21. Are you a betting person?

I'll take even odds on Clinton against any candidate you care to name. In fact heck at this point I'd take even money against the field.

[Renew Deal]

22. I don't want to take your money.

Or anyone elses. I don't bet on sports, so there's no way I bet on politics. Just note that I said it. I could be wrong. If being wrong means we have a Democratic president in 2008, I can live with that.

[CorpGovActivist]

26. Which State(s) Do You See Her Flipping? n/t

[dmallind]

29. On a scale of likelihood?

OH - 20 ECV
AR - 6 ECV
IA - 7 ECV
MO - 11 ECV

The above I consider stong or moderate probables

FL - 27 ECV
NM - 5 ECV
NV - 5 ECV
CO - 9 ECV

The above I consider very possible tossups, the remainder unlikely but feasible

VA - 13 ECV
IN -11 ECV

Couple of other longshots where I could see the scenario being doable. NC, KY maybe even TX if the gods are aligned.

The point of this discussion however is all she needs is 18 ECVs. How many states are going to flip blue to red - seriously and without all this bromide about energizing the base (trust me the wingnuts will energize their base against ANY candidate - she's just pre-smeared), and with back up if possible? I can't think of any really. I suppose we could go total long shot and imagine MN continues its rightward drift in spite of all data to the contrary and say she needs 27 maybe. So OH and one other and we're done, even at that stretch. If the blue holds as I think is far more likely, OH is all she needs.

[CorpGovActivist]

33. Two + Two Strategy

If the nominee builds a double-concentric ring strategy, based on holding the line on all blue states, plus:

1. Concentrating 70% of spare resources in two single-flip-win states (e.g., Ohio and Florida); and

2. Concentrating 30% of spare resources in two combination-flip-win states (e.g., Virginia + Kentucky OR Louisiana + Missouri), that gets to the magic 270.

It takes a net pick-up of 19 Electoral College votes.

Close states in 2004:

http://en.wikipedia.org/wiki/United_States_presidential_election%2C_2004#Close_states

- Dave

[dmallind]

34. I'm sure the campaign folks know this very well

Just one of OH and FL would take you well over the bar if you hold the blue states. I would not want to be a TV viewer in either of them coming up to election time.

[CorpGovActivist]

50. Which Campaign Folks?

The DNC?

Or a particular candidate's?

- Dave

[dmallind]

53. All of em or they don't deserve a job

Heck a campaign coordinator for a municipal dogcatcher could look at an electoral map and say OH, FL and maybe - slim maybe - one or two others will determine who wins. Unless they can come up with likely blue-red shifts though, the advantage definitely lies on our side this time round.

[CorpGovActivist]

54. Insert Any Democratic Candidate into Tab A...

... and 270 Electoral College votes pop out of Slot B?

I don't think so.

Moreover, to tag up with the OP, I really don't think 360 Electoral College votes come out of the electoral slot machine just because the voices in Mark Penn's head tell him they do.

- Dave

[dmallind]

55. Whh? Yes of course it does

Insert ANY cabdidate of ANY party and 270 comes up as what you need to win for one thing - that's simple math nobody can argue with. As far as what they have "in the bag" we can only look at present ECV and direction of partisan support. Since Kerry got 252 ECV and the direction overall remains much more towards Democrats than against them, we have a smoewhat stable base for a comparable Dem candidate. I don't think any of the remotely viable Dem candidates are likely to lose many if any blue states barring wild events, so certainly for Clinton/Obama/Edwards/Dodd/Biden/Richardson we can maybe be very generous to Republicans and assume they MAY be able to swing maybe one of the smaller NE states and MN as long shots. I very much doubt again we'll see ANY blue-red switches, at least based on current standings. So essentially it comes down to which candidate can take EITHER OH or FL and we're done OR none of those and a couple of smaller states - say MO and VA for example - and we're done.

Strangely, and rarely, enough, the electoral map actually works in our favor for once. I don't care how energized anti-Hollary voters are. If they can't switch a blue state to red - and which one would it be? - then they need to hold pretty much ALL the red states - and I just don;t think they can do it in today's zeitgeist against any of the major Dem candidates.

[Leopolds Ghost]

47. You think Hillary could flip "NC, KY or maybe even TX?"

[dmallind]

56. long shots but could. However doesn't need to. OH or FL is all

or even failing that a smorgasbord of much more swing states such as a selection of MO IA, AR, VA, NM etc.

The whole point of my post was she (or any Dem nominee) doesn't NEED to get anywhere near swinging the south or taking deep red states. Demographics are powerful things though. I'd bet any money TX goes blue anywehere between now and 2020 for one.

[CorpGovActivist]

57. Demographics Are Indeed Powerful...

... and the map by 2012 is looking increasingly favorable (especially with the Census in 2010).

I see a retrenchment in the bluish trend in 2008, though, absent a candidate who can speak to Reagan Democrats, independents, and moderates.

I wouldn't bet the farm on Ohio or Florida flipping in 2008. I'd concentrate at least 30% of excess resources (those above the minimum to hold the blue line from 2004) on less expensive media markets in purplish states, to cobble together that needed 19.

- Dave

[Tesha]

14. Clinton isn't electable. It's not too early to say that. (NT)

[Blackhatjack]

15. Any way you cut it, Hillary is not a 'Repub Killer' in the General Election...

IMHO she would have a hard time nosing out a win in the General Election if she is the Democratic Nominee.

[CorpGovActivist]

24. It's the Electoral College Math, Stupid...

... to borrow a line from the rapid response team of yore.

- Dave

[Blackhatjack]

35. ... and in the end, that is the ultimate trump card. n/t

[CorpGovActivist]

36. See #33 Up-Post n/t

[suston96]

17. Who ya gonna believe.....?

Clinton wins....... Gallup.

www.gallup.com/poll/102862/Democratic-Candidates-Look-G...


Clinton loses.......Zogby

http://news.yahoo.com/s/nm/20071126/pl_nm/usa_politics_...


Who ya gonna believe.....?

[CorpGovActivist]

19. ... Not Mark Penn, That's For Sure

http://www.google.com/search?hl=en&q=mark+penn+360+electoral+votes

- Dave

[dmallind]

20. Sad but true answer?

The one you want to believe.

Between Obama and Clinton I vacillate daily, but anyone who thinks the aggregate result of all reliable polling does not suggest Clinton is a) in the lead amongst Dems and b) has at least an equal and possibly slightly better chance of winning any GE matchup is being wilfully selective in what they take from the data we have.

I'd be happy with either of em so hopefully I have little bias at work. I suppose I MIGHT have a bit of a bias against the irrational dismissal of evidence I see more strongly in Anti-Clintonites than in her supporters, but no real pull either way candidate wise for me. I'll work the streets and phones for either of em myself.

[CorpGovActivist]

23. Polls That Aren't Normalized against the Electoral College...

... are weaker, in my view.

"National polls" - of whatever ilk - are much less meaningful when they're not normalized against the Electoral College.

- Dave

[dmallind]

27. Oh absolutely agreed there

, and was answeruing the question perhaps more generically than you intended it (not even thinking only of election polls). But again, Hillary is doing better than anyone else in state level polls too.

[CorpGovActivist]

28. *Shrug* Mark Penn Says She Has 360 Electoral College Votes in the Bag...

... and I believe him, because he's Dick Morris's protege.

:sarcasm:

- Dave

[dmallind]

30. I think THAT's a bit optimistic

But then he's only got to be 75% right and we're fine. Don't need 360 after all. You're still every bit as much president with 270.

[CorpGovActivist]

32. Lie BIG...

... or not at all seems to be his motto.

I would've found plausible that he saw a net gain of 30 Electoral College votes, and would've been scouring the data.

- Dave

[Spike89]

31. Big problem with conclusions drawn now

Although it is NOT possible to scientifically prove what will happen next year based on even a PERFECT poll/survey taken now--people insist on doing it anyway. I guess poll junkies can never be cured, but for cryin' out loud, at least recognize that the "facts" you're throwing around and arguing over are trivial at best. The perfect poll, obviously impossible, would still show what people think now, not what they'll think in 12 months.
In case you haven't noticed, the GOP field has been running as much (if not more) against Hillary as they are against each other. For the most part, the Dems are not currently running against a specific GOP candidate. This is a huge factor that will obviously change in the general election. Another huge factor is the candidate fervor within each party at this stage. For example, if an ardent Edwards supporter is asked if they'd vote for Kucinich if he were the Dem candidate, a "thinking" supporter of Edwards would say no, reinforcing the perception of Kucinich as unelectable and helping Edwards somewhat. It DOESN'T mean that next year, the Edwards supporter won't vote for Kucinich, just that they won't say so NOW.
Aside from those two huge factors, the environment is likely to change in 100s of small, and maybe even some large ways. A terrorist attack (or a real "foiled" attack), a war with Canada, invasion by Venusian slug-people, or a spike in the housing index can have a major effect on the mood and likely votes of the majority of the US population that hasn't firmly locked onto a candidate right now. Remember, the majority of the vote is "locked" for one side or another (most DUers will never vote for a GOP candidate, Freepers won't vote Democratic) and at this stage of the elections cycle, it's more like 99% of the people interested in presidential politics are in that "locked" category.
The simple truth is that the swing voters can't even be identified now and even if they were, they don't have hardened or set choices--they probably don't even know who's in contention.

[ThomWV]

37. Its OK, once votes start being cast she will rapidly fade away a la Dean.

[CorpGovActivist]

38. Mark "Potemkin" Penn Tells Her Mountaineers Love Her...

... and want to name all those "Country Roads" after her.

; )

- Dave

[suston96]

39. .....she will rapidly fade away a la Dean.

I know Howard Dean. Howard Dean is a friend of mine.

Hillary Clinton is no Howard Dean.

[CorpGovActivist]

40. Or Bill Clinton n/t

[flyarm]

41. Thank you!!..sincerely ..thank you! ..I wasn't a Dean Voter.. ..

But i would sure vote for him ...alot faster..today ..than the alternative....if you get my drift.

fly

[CorpGovActivist]

46. He'd Make One Helluva...

... White House Chief of Staff.

- Dave

[flyarm]

48. :)..yes he would! eom

[CorpGovActivist]

49. His 50-State Strategy Was Brilliant, and Brilliantly Excecuted (n/t)

[DearAbby]

42. Well...my 2 cents worth

The GOP is praying for a close election, so stealing it wont be obvious. Nothing has changed. those vote flipping machines are still out there.

Hillary nomination could give that close election results they are praying for...jes saying

[The River]

43. Exactly What I've Posted Here Before

Both Hillary and Obama are non-traditional (ie white male) candidates.
For the "enlightened" souls on DU, race and gender do not matter but
in the rest of America it will be a factor.
After the GOP operatives flip just enough votes to steal the election
the official MSM cover story will be all the closet racists and
the closet misogynists, good Democrats in all other aspects, but
who just couldn't vote for a black or a woman.
They will interview hundreds of "Democrats" of both genders
who will own up to the fact they just couldn't vote for a woman
or a black. They will say America just isn't "ready".
How do you disprove this meme when the Media is saying it's true?

[CorpGovActivist]

44. Iowa and New Hampshire...

... could give an early foreshadowing of that meme.

- Dave

[BenDavid]

51. 53 friends across this nation belong to the Zogby International/

interative polling. Out of the 53 friends across this country only 5 received the last zogby polling questionaire, in which the polling results are splashed all over the du. I did not receive mine either. Now what kind of polling results would you expect if you left out 48? That is why this poll is bogus. It is a shame that some polling organization try to fix the game from the start.

so if the front-runner splits the party, we should nominate someone with less support?


Ben David

[CorpGovActivist]

52. Do You Have a State-by-State Breakout...

... of where the 5 and the 48 live?

Thanks.

- Dave