Susq. PA Poll: Close Race

By Big Tent Democrat, Section Elections 2008
Posted on Mon Apr 14, 2008 at 11:57:19 AM EST
Tags: (all tags)

Share This: Digg!

By Big Tent Democrat

In a poll conducted entirely before BitterGate (the poll was conducted from April 6-10), the Susquehanna Univ. PA poll shows a close race with a very high number of undecideds - Clinton 40, Obama 37, Undecided 23. This is a large drop for Clinton from a poll from a month earlier when she led by 14. That said, even without taking into account BitterGate, this poll seems incredibly suspect to me. I'll explain on the flip:

This poll has Obama winning African Americans by only 50-30 but still within 3. That is simply impossible. Let's do our SUSA analysis again.

Let's assume, for the sake of argument, 3 out of 4 A-A undecideds break for Obama, giving him a 65-35 win among A-As, a preposterous result. Obama will win AT LEAST 85% of A-As. This translates, assuming the 14% A-A percentage of the vote that SUSA finds, that A-As will deliver 9% of the TOTAL vote to Obama 9% and 4.5% to Hillary.

Extrapolating from there, assuming an overall 4 point Clinton lead, we can then assume that this polls has Clinton winning whites relatively narrowly - by less than 10. Thus delivering Obama 39% of the total vote. 39% plus 9% equals 48% for Obama and 52% for Clinton.

I believe NEITHER finding. Obama will win much more than 65% of the African American vote (he certainly must be winning by more than 3-2 now.) Clinton will win whites by more than 55-45 (this poll has her with virtually NO lead, at most a couple of points). The numbers simply do not add up, even if one assumes (the poll mentions that past primary voters break for Clinton by 7, also a ridiculously low number) high A-A voter and young voter turnout.

Bottom line, Obama may be down 3, but this poll does not convince me that that is true. I simply do not believe its internals. They strike me as completely not credible. Obama winning A-As 90-10, a 20% A-A vote and Obama capturing 35% of the white vote is the formula for Obama being within 4. Not the results of this poll.

< Right Wing Talking Points | Anybody Worried About Alienating Clinton Supporters? >

Legal

Nothing on this site should be construed as legal advice. TalkLeft does not give legal advice.

TalkLeft is not responsible for and often disagrees with material posted in the comments section. Read at your own risk.

Susq. PA Poll: Close Race | 20 comments (20 topical)

Readers Must Login or Create an Account to Comment

If the poll was accurate (5.00 / 1) (#4)
by ding7777 on Mon Apr 14, 2008 at 12:29:52 PM EST

Obama would not have made the "clinging" comment; he would have said he was within striking distance.

How about "Cling-gate" instead of "Bittergate" since it the clinging part that's elitist.

Bittergate has more of a ring to it (none / 0) (#14)
by ruffian on Mon Apr 14, 2008 at 12:40:30 PM EST

: Bittergate is best, (5.00 / 1) (#18)
by zyx on Mon Apr 14, 2008 at 12:54:16 PM EST

but "Gaffe-gate" has a nice alliterative quality.

I don't believe "Undecideds" (5.00 / 1) (#5)
by cmugirl on Mon Apr 14, 2008 at 12:30:19 PM EST

In any election, when I see people being interviewed a week before the election and listed as "undecided" I'm always suspicious. In this day and age of 24 hour+ media attention, I don't know how anyone is undecided - even if you don't pay attention, people around you talk about it, it's on the front page of newspapers when you pass a newsstand, you'd have to run across something while flipping the channels, etc.

I'm really suspicious of 23% being undecided as of last week.

I think there is some lying being done (5.00 / 1) (#12)
by BarnBabe on Mon Apr 14, 2008 at 12:39:05 PM EST

I do not know when they contact a person if they already know their race or religion. I know the name is 99% giveaway as to gender. Do you know?But I could screw up a poll by lying. I don't but I could. Maybe some people are just having some fun because they are annoyed at being called.

Just one week from tomorrow and we get to vote. Thank goodness. It was one thing to think your vote would not mean anything but once we did, we wanted to get it over with.

Any Poll with 23% undecided is suspect (5.00 / 1) (#13)
by ruffian on Mon Apr 14, 2008 at 12:39:21 PM EST

IMHO

I think any PA poll taken (none / 0) (#1)
by Jeralyn on Mon Apr 14, 2008 at 12:14:48 PM EST

before Bittergate is irrelevant now.

Where did you get the 20% African American figure from? According to QuckFacts Census reports, African Americans are only 10.7% of Pennsylvania's population. I'm not suggesting you aren't correct, just asking.

Gallup and Rasmussen (5.00 / 1) (#10)
by magster on Mon Apr 14, 2008 at 12:38:23 PM EST

Daily tracking polls show no (Gallup) or little (Ras.) downward movement for Obama. While the actual percentages are suspect, I do think these tracking polls ebb and flow with the scandal du jour. If there's no movement by tomorrow, I would guess that Obama weathered the storm.


: will the O blogs stop "explaining" it ? (none / 0) (#15)
by Big Tent Democrat on Mon Apr 14, 2008 at 12:45:23 PM EST

Let the damn thing go away please.
I promise to stop writing about it if they stop writing foolish things about it.
Personally, I think you worry too much about national tracking polls, positive or negative, but let that be the standard please.

The A-A projected turnout is 14% (none / 0) (#2)
by Big Tent Democrat on Mon Apr 14, 2008 at 12:18:00 PM EST

as per SUSA.

I think 20% would be high, but it is what Obama would need to have a chance to win PA.


I predict they will be 17-19% (5.00 / 1) (#3)
by andgarden on Mon Apr 14, 2008 at 12:26:29 PM EST

given that this is a high profile Dem primary.


Every point is worth a point for Obama (none / 0) (#7)
by Big Tent Democrat on Mon Apr 14, 2008 at 12:33:16 PM EST

If it is 20 instead of 14, that cuts the margin by 12.


: Yup (none / 0) (#8)
by andgarden on Mon Apr 14, 2008 at 12:34:50 PM EST

It's why I predict a 10-15 point win for Hillary. The exit poll will still tell a dismal story, though.

A good analysis of PA polls (none / 0) (#19)
by Cream City on Mon Apr 14, 2008 at 04:42:27 PM EST

and their "sensitivity," how to look at outliers and the like there, is at pollster.com now, Jeralyn. Might be worth your time, tight as it must be, to be ready for the next polls we see, dissect, debate, deplore, celebrate, etc. :-)


: Not to ignore you, BTD! (none / 0) (#20)
by Cream City on Mon Apr 14, 2008 at 04:43:52 PM EST

I replied to Jeralyn with the above, then realized it's your diary . . . so you can arm-wrestle over it.:-)

High undecideds (none / 0) (#6)
by cannondaddy on Mon Apr 14, 2008 at 12:32:46 PM EST

Take the number of undecideds, give 60% to Clinton and you have 54%-46%. I think that's pretty consistant with other polls. I think that's my new best guess: Clinton by 8%.

The undecided is very high (none / 0) (#9)
by Prabhata on Mon Apr 14, 2008 at 12:37:50 PM EST

which means they did not press respondents to give their leaning.

: It's a poll that tells us nothing. (none / 0) (#11)
by andgarden on Mon Apr 14, 2008 at 12:38:37 PM EST

well... (none / 0) (#16)
by myed2x on Mon Apr 14, 2008 at 12:49:19 PM EST

obviously any poll with Hillary being challenged is suspect!

Regardless though, This is excellent news!! For Hillary!!

: Huh? (none / 0) (#17)
by Big Tent Democrat on Mon Apr 14, 2008 at 12:50:05 PM EST

WTF?