How Did Pollsters Do in Pennsylvania?
By: Lowell
Published On: 4/23/2008 6:23:45 AM
It looks like the final results in Pennsylvania will give Hillary Clinton a 10-point margin of victory, 55%-45%. How did the pollsters do in predicting this result? Not too badly, overall. Here are the last poll results and the difference from the final results:
Zogby: +10 Clinton (nailed it!)
Suffolk: +10 Clinton (nailed it!)
Insider Advantage: +7 Clinton (off by 3 points)
Strategic Vision: +7 Clinton (off by 3 points)
Quinnipiac: +7 Clinton (off by 3 points)
Zogby tracking: +6 Clinton (off by 4 points)
SurveyUSA: +6 Clinton (off by 4 points)
Mason-Dixon: +5 Clinton (off by 5 points)
LA Times-Bloomberg: +5 Clinton (off by 5 points)
ARG: +16 Clinton (off by 6 points)
As I said, not bad...except for PPP.
PPP: +3 Obama (off by a whopping 13 points!)
What happened there? They completely missed this one, not even close. Ee gads, how embarrassing.
Comments
And then there's the RK poll (Lowell - 4/23/2008 8:37:08 AM)
Looks like RK readers pretty much nailed it, with 85% saying Clinton would win (correct) and an average victory margin among those people of about 6 or 7 points. Not bad for an "unscientific" poll with 93 votes on a blog!
10 points? (Randy Klear - 4/23/2008 9:28:11 AM)
This morning, the
Pennsylvania Department of State is showing it at 54.3-45.7 with 99% of the precincts in. Zogby and Suffolk didn't quite nail it after all.
Statistics! (code - 4/23/2008 2:21:55 PM)
Most of those polls have a MoE of +/- 3 or 4 points. That means +10 to Clinton is actually +7-+13 range, to one standard dev (68% probability of accuracy), not precisely +10. So in reality, Zogby, Suffolk, Insider Advantage, Strategic Vision, Quinnipiac, Zogby tracking, and SurveyUSA all "nailed it" as closely as possible statistically.
Good point. (Lowell - 4/23/2008 3:43:05 PM)
n/t
only margin is not 10, but currently 9 (teacherken - 4/23/2008 10:15:16 AM)
most recent figures I saw an hour ago were 54.69 to 45.31 which is a difference f 9.38 which rounds ot 9
CNN and others are rounding each of the number, which is how the 55 go 45
Right now it's (Lowell - 4/23/2008 10:32:11 AM)
54.3%-45.7%, which if rounded is 54%-46% or an 8-point Clinton win. If not rounded, it's an 8.6-point Clinton win. I'll probably update the diary soon.
Slight warming (Randy Klear - 4/23/2008 10:43:14 AM)
The state numbers appear to be inflating Obama's total in Lancaster County by about 12 thousand votes, so it may well end up closer to 9.
One odd note: Philadelphia has their precinct totals hidden in a password-only section of their web site. Big city business as usual?
That should be warning, of course. (Randy Klear - 4/23/2008 10:44:01 AM)
The warming is global, as we all know.
No matter if its not 10 points (Rebecca - 4/23/2008 10:56:03 AM)
The point is that everything I have read shows that the media inflated Clinton's lead to two digits for several hours and it was not really confirmed by the results. You may say that inflating by one point is no big deal, but since everyone said she had to win by double digits, it is significant in the PR sense. It gave the impression that Clinton is surging. It seems that in the last 3 weeks the media is making a lot of "gaffs" that seem to favor Clinton. We have the ABC debate, and in addition even members of the Democratic Congress (people who know better) can be seen on TV saying Clinton won Texas. This is not a gaff, but a deliberate effort to create a positive view of the Clinton campaign at the expense of the truth.
actually clinton did win the texas primary. (notwaltertejada - 4/23/2008 3:56:43 PM)
and she won by over 100,000 votes. obama of course got more delegates because he won the caucuses with about 5,000 votes. it's really a shame that our nomination system is flawed in this way and if that isn't elitist i don't know what is. you're right in that i don't know why anyone tried to make this look positive when it is such a travesty.
I guess it depends (Rebecca - 4/23/2008 5:31:45 PM)
I guess it depends on what you mean by win. In your terms Al Gore should have been president for the last 7 years. Unfortunately for some we do have rules and a system,and we are also a nation of laws. As flawed as that can be at times its what makes us a civilization.
The Democratic Party in each state has a system too. These are the rules. As someone on another blog said, if you don't like the rules of your own party get them changed or start your own party.
Hillary chose to play by the Democratic rules, .... or did she? Or does she really want to be a monarch? Seems like Hillary would like the Republicans' primary rules better.
Hillary plays by her own rules ... (j_wyatt - 4/23/2008 5:44:25 PM)
That's her sociopathy.
She will say or do anything, because, in her mind, the end justifies the means.
yes...she did go along with the rules (notwaltertejada - 4/23/2008 6:07:54 PM)
however crackpot they may be. seems like a lot of people would like the republican primary rules of winner take all. it would certainly be more decisive and we probably wouldn't be where we are today with them.
acually there aren't many people in the U.S. advocating to establish a monarchy despite your conspiracy theories.
Would it be elitist if Hillary were winning? (Tom Joad (Kevin) - 4/23/2008 7:47:50 PM)
Just wondering...
yes it absolutely would be (notwaltertejada - 4/23/2008 9:09:31 PM)
the fact is you had a primary with over 2.8 million people participating followed by a caucus with about 75,000 participating. clinton's margin of victory in the primary was larger than the total number of participants in the caucus!! but hey if it's good for obama it's good enough for people on here.
Polls...Polls...Polls (TheToe - 4/23/2008 11:00:51 AM)
Also, Raz polled on 4/20 and ended up being off by 5 points (they had Clinton +5)
This is just further proof that polls need to be taken with a grain of salt. Zogby was blasted after Super Tuesday for botching so many of their polls (highlighted by their election eve prediction of Obama +13 in California prediction). Today they are one of two pollsters to nail the Pennsylvania Primary. We are in uncharted waters right now in the race for the Democratic nomination, yet everyday I feel like I am resuscitating one friend after he or she sees a poll that isn't good for Obama and trying to calm down another after he or she sees a poll with Obama up big. PA is a perfect example that its a complete crap shoot as to what will happen on election day.
I hope we can be comfortable with the things we do know for certain. People will vote, someone will win, and someone will lose.
Media saturation (Teddy - 4/23/2008 11:52:13 AM)
bombing of Obama was extremely intense (with only obligatory token attacks on Clinton), and had to influence how the pollsters were conducting their polls, and the answers they received--- no matter how professional polling claims it designs its polling. And, I must say the media blitz worked for Fox, didn't it? Comments from voters actually parroted Fox's drip-drip-drip sound bites, and many did break for Clinton. (Looks to me like a degree of cause and effect).
Remember that Hillary had a lunch meeting with Rupert Murdoch before the Pennsylvania primary. I don't know what they discussed, of course, but I have a pretty good idea after watching how Murdoch's minions "carpet bombed" Obama for four straight weeks, and the rest of the media and pundits quickly got the message and repeated the attacks everywhere in every guise (consider the first 45 minutes of ABC's so-called debate--- a perfect example).
It is true that Obama will face even heavier artillery if he becomes the Democratic nominee, so he must figure out how to respond effectively. It is also true that Hillary need not bask in the warm, helpful Murdoch (Fox) sunshine because, if she becomes the Democratic nominee, the formerly helpful media will begin treating her exactly as they have treated her in the past, only more so.
That's the trouble with "going negative." We can now kiss goodbye real discussion of real issues from now through November.
9.2 vs. 10 (j_wyatt - 4/23/2008 3:50:45 PM)
Presumably, news organizations are in the factual news business. Daily, they report an accurate tally of the war dead, the price of a barrel of gas, what Microsoft is offering for Yahoo ($ 43.6 billion -- not $ 43 or 44 billion).
So why is it being reported that Senator Clinton's margin of victory in PA was in the quote unquote double digits?
Case in point, CNN's Election '08 website is still showing it as 10%.
Could it be they're trying to save virtual newsprint fabricated from digital rainforest pulp by eliding that one space difference between 9.2 and 10?