The urban legend of "Dukakis winning" in exit polls (1988) put to rest

[RaulVB]

Or "Lies and the lying liars who...", you get the rest.

ftp://ropercenter.uconn.edu/United_States%5CNBCWSJ/USNBCWSJ1988-NATELEC/version2/usnbcwsj1988-natelec.pdf

Bush Sr.: 54%

Dukakis : 46%

Next topic and try harder this time!

[RaulVB]

1. Kick (n/t)

[RaulVB]

2. Oh well, "we" can't use Dukakis anymore...(n/t)

[RaulVB]

3. What happened guys?

Not playing anymore? :think:

[ClintCooper2003]

4. And here's a link to a bbc article to provide a timeline:

The bbc shows the date as Nov. 9 because they are many hours ahead of us. But the election date itself was Nov. 8, 1988.

http://news.bbc.co.uk/onthisday/hi/dates/stories/november/9/newsid_3655000/3655368.stm

[Bill Bored]

11. How come they say Bush was the FIRST VP in a zillion years to become Pres?

LBJ and Ford don't count?

[Wabbajack]

13. They succeeded to the office

via death and resignation. bush the first was the first sitting VP elected since Van Buren I think.

[Bill Bored]

17. LBJ was elected in '64 though.

But I guess they have some trivial point.

[Wabbajack]

19. When LBJ

was elected he was the already the President, same as Coolidge, Truman ect. pappy bush was Vice President.

[MadisonProgressive]

21. And Gore was the second!

[Bill Bored]

5. Can I play?

Raul,

I can't even get them to write to their reps, what makes you think they'll argue about a 1988 exit poll?! :)

But since you brought it up, I believe the myth was that only the early results favored Dukakis, not the final ones. You haven't disproved that here have you?

[RaulVB]

7. A myth, indeed

Because those "early" exit poll results are NOWHERE TO BE FOUND or seen.

[ClintCooper2003]

6. Hold on a minute. If you look at the "frequency" section, it doesn't...

match the percentages. Bush 5,871 Dukakis 5,546 No response 286
Percentages should be Bush 50.17%, Dukakis 47.39, No response 2.44%.

Of course, if the non-responders were all for Bush it would be Bush 52.61% to Dukakis 47.39%

The actual results were Bush 53.37% to Dukakis 45.65%. (uselectionatlas.org)

[RaulVB]

8. Well, well,

Good catch!

Don't tell me that we have found MORE COOKED NUMBERS!

[ClintCooper2003]

9. I think the 54 - 46 part on the left was simply showing the...

approximate tallied results.

[RaulVB]

10. Yes, that's it (n/t)

[Bill Bored]

12. Well doesn't this prove that there was a Red Shift then? n/t

[RaulVB]

15. What "Red Shift"? (n/t)

[Bill Bored]

18. see post #6

[ClintCooper2003]

28. 50 to 47 does not equal 54 to 46. That's a definite shift.

[Bill Bored]

43. Right, and it's a Red shift, from 50% Repub to 54% Repub.

Not unlike 2004 which shifed from 48% to 51% Repub.

The difference is that in '04, the shift changed the outcome, because it was a close election. But it does not appear to be true that the shift wasn't there in prior years, at least in '88.

Is this right, or did I miss something?

[RaulVB]

44. Since "*" started a political "life" anomalies concerning poll results...

Became common in the US politics.

This data just aimed to debunk the claim that Dukakis "won" the exit polls in 1988.

That's false and I think the point is clear now.

[Bill Bored]

50. Clear but there was a Red Shift in '88 similar to the one in 2004, yes? nt

[m.standridge]

16. weird coincidence here, though

--about the Electoral College.
Some people said Bentsen conceded too soon--remember that. I dimly do. They think he influenced the votes in Illinois, western Michigan, Ohio (uggh) and maybe people still in line in PA and MO.
Back then, that was a bunch of Electoral votes.
And, that's not to mention NM, which was close, but which Dukakis didn't carry.

In other words, according to that theory, Bentsen's concession was premature and self-fulfilling prophecy. Back then, those states carried a lot more Electoral votes than they do now.
Maryland was also very close, probably, as I recall, recountable. And there was data about New Jersey in later years--Ed Rollins admitting to paying "walking around money" and Ed Baumeister writing of off-duty cops hired to intimidate Af-Amer precincts.

When you add all those up, even if Dukakis wasn't leading in the Popular vote, he'd have been doing a lot better in the Electoral College.
Anyone seeing a pattern here?
Massachusetts with a Southerner:
JFK and Johnson, 1960: squeaker in the Popular vote, not so much so in the Electoral College
Dukakis/Bentsen--same pattern? Seemed to have pretty hefty potential in the Electoral College versus the Popular.
Kerry/Edwards--same pattern?

If that's the case, the exit polls aren't as essential in determining who actually won, due to the Electoral College math. In other words, a relatively small number of votes, could have carried it, anyway, and that might not reflect all that much in the exit polls, or might not have a direct connection with exit poll accuracy, nationwide.

How much variation on a theme with each?

[Zynx]

23. Dukakis wasn't going to win.

No pre-election poll showed it even close. I don't want to hear how Bentsen's early concession cost them the race. That's bullshit. The Duke cost them the race and that's a fact.

[m.standridge]

45. Oh, you're on--boy, am I ready for you!

You want pre-election polls for 1988?

5/17/88: Arkansas Gazette--"Harris Survey: Dukakis Widens survey Lead over Bush: Michael Dukakis has widened his survey lead over Vice-President George Bush to 50-43 percent lead. Last month, the Massachusetts governor had a 50-45 percent lead. In fact, Dukakis has gained steadily on Bush since the Harris survey first compared the two just over a year ago..."
5/22/88: Arkansas Gazette--"Dukakis Takes lead In All Major voter Groups: . . . Dukakis now leads Bush by a 54 to 38 percent margin. . .Bush's close association with Reagan has probably hurt him more than it has helped. Reagan's job performance rating from 52 percent in early March to 46 percent..."
8/2/88: Arkansas Gazette--"Atlanta Increased Dukakis lead 16 points"
(c) 8/20/88: Arkansas Gazette--"Dukakis lead falls to pre-convention level . . .Dukakis holds a 49 to 42 percent advantage over Bush..."
8/17/88: Arkansas Gazettte--"And now, A Standoff"
"...Bush is holding a lead in most--though not all--of the national polls. But even Republican pollsters said about 40 percent of the voters--an unusually high proportion this late in the game--have not firmly made up their minds, an indication that the race could shift dramatically in either direction...".

In short, you're on. The polls were suggesting at the very least a very close race.
But you've missed the point, anyway: my whole point is that the Electoral College could make nationwide exit poll results misleading as to who wins a US Presidential race anyway. You have to break them down on a state-by-state basis, to see who is winning in states essential to an Electoral win. Otherwise, you don't know much.

[m.standridge]

46. correction: that last poll was 9/17/88, not 8/17/88

Bush started the "Willie Horton" ad, which was really helping him to kick more butt in states he was ALREADY carrying before he started running it, greatly increasing his likelihood of a Popular vote lead, but not by any means ensuring he'd carry the Electoral College.

Insofar as the Electoral College:
New Mexico was hair-close as it was.
Maryland was so close as to be recountable
Michigan probably fell to Bush because of Bentsen's early concession--the polls hadn't closed there, and there were still people in line to vote even in eastern MI.
Same with Illinois, Missouri, and Pennsylvania--people were in line around the block. And, do I have to mention Ohio? An article by a New York Times columnist suggested Dukakis probably would have won had he run with John Glenn, then a Sen. from OH, as his VP instead of Bentsen.
God, you're on.
Even California was close in the 1988 election. Bush wasn't Reagan out there. I've often wondered what a recount there would have shown.

There was a Florida US Senate race that time, too, that was down to the wire--a harbinger of the 2000 Presidential election, between Connie Mack and Buddy McKay. I've wondered, since then, if we had Ohio-like indicators there of vote-transferance phenomena. No one thought about that kind of thing back then.

The Willie Horton thing was producing a similar effect for Bush, sr., to what incumbency may have produced for W. this time: a big increase in popularity in the states he was already carrying, not any guarantee of a victory in the Electoral College (did I just say that?).

[TankLV]

42. The election was called before the polls in Hawaii OPENED!

That's the main thing I recalled, living in Honolulu at the time.

It was disgusting.

[euler]

14. You really should stop posting documents...

...you don't understand.

The very first paragraph says:

The data were weighted to correctly reflect the outcome of the election.

Whereas the exit poll numbers you think you are debunking are "raw exit poll results"

See here: http://www.washingtonmonthly.com/archives/individual/2004_11/005178.php

The claim that so called 'naysayers' are making is that the raw numbers at precinct closing for previous presidential elections have overstated democratic support before, sometimes by more than the overstatement of democratic support we now see in the 2004 presidential election exit poll.

You can't debunk exit poll raw numbers taken at precinct closing time by pointing to exit poll numbers that have been re weighted to match the actual vote count.

You were more effective when you limited yourself to calling people stupid.

[flintdem]

20. Thanks Euler

I saw the same thing but didn't post it. This Election forum is going nowhere. I'm off to find a progressive site where people live in reality.

Having eyes they do not see...

[Chi]

32. Nice meeting you flintdem

I hope you find what your looking for.

[RaulVB]

41. "Don't let the door..." (n/t)

[MadisonProgressive]

22. Not so fast

"NOTE: Frequency values are unweighted; percentages are WEIGHTED
only for valid responses."

The frequency values are correct and there is a 1% MOE.

[Chi]

31. So which exit poll numbers...

Heya euler
So which exit poll numbers do you believe euler? (50.3 vs 49.7 Dukakis, or 54 vs 43 Bush)

And more importantly...why?
(since you didn't respond to MadisonProgressive I'm assuming he's right)

(don't forget..don't think about an elephant!)

[flintdem]

33. You might want to check post 24... N/T

[flintdem]

24. The Error was in the ABC New Poll not NBC!!!

There were four 1988 Exit Polls! The networks didn't work together in 1988. Here are the poll sizes but the codebooks don't have the marginals.

NBC/Wall Street Journal: 11,703
ABC New: 95,167 (22,785 nationally)
CBS/New York Times: 11,645
LA Times: 6,043

According to the washington post it was the ABC poll that was bogus- the largest sample with the smallest margin of error was off the farthest!!!

http://www.washingtonpost.com/ac2/wp-dyn/A64906-2004Nov20?language=printer

[Bill Bored]

25. Everyone knows the stock market does better under Dem Presidents! n/t

[genieroze]

26. I'm sure * will change that when he MAKES people put their SS into

the market against their will and allow the Enron types to steal it.

[Bill Bored]

27. I was thinking that this could be their way to save the market from

the escalating cost of energy, since they haven't been able to do anything about it, although they could have. Bunch of morons!

BTW, if they wanted to take some of this money and use it for a noble purpose, such as investments in renewable energy, would we still be against it? I know it's really an insurance policy and it's supposed to be in Govt. Bonds, but there could be a way to shift some of it into some needed R&D or production that would certainly pay off in the long run. Just thinking out loud of course. I don't expect anything like this from Bush, but it might be an idea for someone with a brain.

[genieroze]

30. It should be a choice and if a person invests their money and they lose it

that should be their problem not mine. Just as if they invest their money and they get a huge profit then their retirement should reflect that as well.

[RaulVB]

29. We are talking about 1988 here...

Read before posting.

[m.standridge]

48. well, we still had the Electoral College in 1988

and several states then, which were in Dukakis "possible" field, had a lot more Electoral votes than they have today.

[dbonds]

34. Notice that the times the exit polls were 'off' is

when a Bushie was running for office. <Things that make you go hmmm> <No, things that make you go WTF>

[flintdem]

35. Not really significant...

Considering that we have only used exit polls since 1980 and a bush has been a presidential or vice-presidential candidate for six of those seven elections.

[dbonds]

36. Not sure about that

I seem to remember having them in the 70's, I was too young to remember the 60s though. I think they have been around for a while.

As an aside, looking at sample size, our ratings for our favorite tv shows is produced by a national sample of 5000 and that is across hundreds of channels and shows. A sample of 96k is far above that. If we through out that level of sample size then lots of science has just come under question.

[flintdem]

37. Yes and No

Exits polls were used in 1964, but were only first used for Presidential elections beginning in 1980.

http://slate.msn.com/id/1006435/

http://www.americanvoice2004.org/askdave/25askdave.html

[m.standridge]

49. I almost got "off topic" but no harm intended

to your main argument:
exit polls could have been in error before.
My point, is just that, exit polls like that, are only telling you--at best--who's getting the national Popular vote. You have to break them down, state by state, to see who's got a shot at the Electoral College.

That doesn't mean you don't have an important point on this. We should have accurate exit polls, and if they are wildly off, especially when broken down state by state, it's alarming, and an indication of election fraud.
I was referring to pre-election polls because I'd been "challenged" on this by an uninformed individual on this point.

[Mistwell]

47. Those are weighted

That's all dealing with weighted polls...not the unweighted polls cited by Washington Monthly in their Century Foundation data.

[m.standridge]

52. This isn't the first thing I've seen about ABC

and politics. ABC has been a front for CIA people before--remember the guy who was kidnapped in ...China or the Mideast, allegedly an ABC News reporter, who turned out to be a CIA agent?
I've heard a lot of scuttlebutt about ABC people's activities in election offices, too.

rememeber the George Will "debategate" incident in 1980?
In '84, ABC helped distort election results out in Nevada--its stats did. Not all of it is deliberate. But ABC seems to get "infiltrated" by intelligence types from time to time--and, perhaps, even by "plumber" types.

[Mistwell]

38. Nice Try

That's the weighted poll. The unweighted one, which is the one we are using here to claim Kerry won, also showed a massive deviation from the MOE to favor Dukakis.

Compare apples to apples, and you get the same sort of deviations.

[bemis12]

39. It's PROOF I tell ya

Guess he's "not playing anymore".

[RaulVB]

40. "Tag teams" working around here. What is this, WWF? (n/t)

[m.standridge]

51. well, there's something to the exit poll discrepancy thing

there really is. It's most impressive when broken down on a state-by-state level.
I just want to be sure we are clear: it's not the national Popular vote lead that matters, its how they're doing in the states needed for the Electoral College.
That's where W. was in trouble this time.
That's where Sr. was in trouble in 1988.

Dukakis was hanging on in the states that weren't gun states.
And Bentsen was helping him tremendously--no doubt about it--in getting it close in the South and in the border states.


I don't know whether Benten's concession speech was deliberately premature--he said later he thought he'd been misinformed.

But it did have a dramatic impact on the election.
I think it's possible Bush sr did some shenanigans,too, in the states he needed in the Electoral College, that Dukakis had been having leads in: Illinois, Pennsylvania, Michigan, Ohio...They were states that weren't respondinga as "well" to the Willie Horton/gun control thing.