Public Policy Polling (D)
Clinton 46, Obama 43
So now we have 6 recent PA polls, five of which are consistent to a remarkable degree, showing Clinton between 45 and 50 and Obama between 42 and 45. Plus SUSA with the 56-38 margin. I guess you can decide if SUSA is an outlier, but I will suggest that its outpants are on fire.
Posted on April 9, 2008 12:13 PM
guess so. but again, in the PPP you have faulty demos. @17% for aa's when census puts it at 8-9.
i'd caution you not to get too optimistic. plus the numbers as uniform as they are ex the SUSA poll, indicate recent obama slippage. somewhat of the same trend seen in other states where clinton brings it in the closing days. :>)
Posted on April 9, 2008 12:34 PM
8%? Last time I looked up censis data, African Americans were more like 10% of the population, which would make 17% for a Dem primary rather modest. But anyway, what are we talking about, 17 vs. 14? That might make it a lead of 5-6 instead of 3, still in line with other recent polls.
As to your second point, it is not true that recent PA polls show Obama slippage. Ras is steady since their last poll, Q-poll, SV, and ARG all show Obama gaining slightly. And these are not the closing days--it's 2 weeks till PA.
Posted on April 9, 2008 12:57 PM
"guess so. but again, in the PPP you have faulty demos. @17% for aa's when census puts it at 8-9."
Of course, African Amercians are a much larger share of Democratic primary voters than the general population. Blacks made up about 21% of Kerry's vote in 2004. So, if anything, 17% seems to understate their likely share of the vote.
Posted on April 9, 2008 12:59 PM
Every poll we have this discussion of WHY African-American % representation is so high compared to "census data" --
Do people not understand that this is a poll of LIKELY DEMOCRATIC VOTERS and not the entire state of (fill in the blank) ??
Why must we have this same discussion every time?!
Posted on April 9, 2008 2:34 PM
Democrats are about 48% of the Pennsylvanian registered voters and the population of African-Americans in Pennsylvania is about 10%.
I suspect few are Republicans.
The 17% figure used likely assumes that AA's aren't typically registered voters or as likely to go to the polls as other Demes within the Democratic party...so it has reduced a 20% figure to 17%. But if Obamas registration drive has indeed brought in 200,000 new voters into the party over the last two weeks that imbalance may not only have been eliminated, it may have made other demographics fall behind AA or "youth" registrants.
This is where polling becomes an "art" as much as a science. The use of prior indicators as to whom may be a "Likely Voter" cannot deal with a phenomenon like Obamas campaign. If a demographic like AA's previously would turn out 25% of eligible voters for a primary vs. 60% for whites over 55 years of age...that can be utterly confounded if the AA's (or youth cohort) turns out in significantly higher numbers (and have increased registration).
Survey USA regular throws out about half of their sample as being "not likely" voters. But what if that sample is actually the ones they should be evaluating.
Wouldn't it be nice if SUSA and other survey groups gave us full access to their data (even the throw away data)? It's a pity that we only end up with data that's been massaged, and have no reference of a raw set of data to evaluate that from. Did the massaged data mirror the population at large, or was the extracted data set at high variance from the excluded sample?
Posted on April 9, 2008 3:40 PM
Two things you have to factor in.
One, changes from the 2000 Census on projections of race.
Two, changes from likely african-american voters from the last election.
I think the number should be higher in the 15% range frankly.
Posted on April 9, 2008 4:45 PM
Wasn't there a Pollster report card somewhere on the web (other than Survey USA's? I think SUSA might have a bias in their reporting of other pollsters' success/failure scores.)
Posted on April 9, 2008 6:59 PM
You can always check USAElectionPolls.com. They rate the various pollsters comparing their results to the Election Results. Covers all states and nationally and has a regular rss feed.
Posted on April 9, 2008 11:14 PM
I actually did an assessment of SUSA's "last poll" success to actual election results and they are pretty good when compared to competing polls. But in several States all the polls went far afield, particularly in the South and Southeast. In addition SUSA was way off in Missouri. They do tend to have a more accurate prediction record than other competing polls.
At the same time, if one looks at their long range predictions they are often at variance with the final results. Not unexpected given the fluidity of the electorates during a campaign.
Posted on April 10, 2008 1:57 AM
Comments: (you may use HTML tags for style)
Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.
Please email us to report offensive comments.
See our comment policy here. Note that we require commenters to share their email address via Typekey. We will never share your email address with anyone without your explicit permission.
MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR