Yesterday, the Tribune Review of Pittsburgh released a poll conducted by Susquehanna Polling showing the Pennsylvania presidential race tied at 47-47. This apparently lit a fuse under Karl Rove's ample bum, flooding the PA airwaves with the foulest emanations, and has the R&R boys scurrying around the state like Energizer rats in a maze. Good. Waste of energy, $$ and focus. (Plus, either they missed the October 18 poll by the same outfit which had Romney ahead by 4 points - thus proving Obama has some real momentum here -- or they are really, really desperate now.) In any case, both Susquehanna polls are demographic garbage.
Whenever we see an "outlier" poll (and these two run way against the grain of other PA polls), we try to get behind the headlines to look at the underlying questionnaire and, more importantly, the demographics of the respondents. These are not always easy to get to (see Note 1 below), but in this case, they were available. We then check the demographics against two gold standards: (a) the 2008 CNN Exit Polls (b) the 2010 U.S. Census Bureau data. If there are significant discrepancies between the poll demographics and the standards, we use various methods to adjust and "normalize" the results for a more accurate projection.
The Susquehanna Poll
The three categories in this poll which varied most egregiously from the corresponding categories for Pennsylvania in the CNN 2008 Presidential Election Exit Poll and the 2010 US Census are shown below, with the biggest offenders highlighted.
(Data in cells are percentages. Row total in category may not equal 100 because of rounding)
Source: Susquehanna Poll, CNN 2008 President Exit Poll,US Census 2010
Errata: Second Age column should obviously read 30/44, not 30/34
Analysis
The "thumb on the scale" here could not be cruder or more obvious. Bias your sample population against the young, increase the number of seniors, load it up with white folk and wingnuts, and voilà - Romney ties Obama!
Well, all that can be adjusted fairly easily, by "unskewing" or unbiasing the sample to resemble more normal voting demographics -- we used the 2008 CNN numbers for our model -- and recalculating. We also correct for how Obama/Romney are doing compared to Obama/McCain in 2008. The results are hardly surprising.
Remove Age Bias: Obama +1.15 Romney -1.15
Remove Race Bias: Obama +4.40 Romney -4.40
Compared to 2008: Obama -1.00 Romney +0.80
which all adds up to:
Barack Obama: 51.5% Mitt Romney: 42.3% in Pennsylvania 2012
This represents a 9 point lead for Obama, unassailable even in the unlikely event that the entire 6% undecided broke the other way. The likeliest outcome:
Obama: 54% Romney: 45% not that different from Obama: 55% McCain 44% last time around.
We didn't even bother with the "Ideology" skew, it's so out of bounds and self-selected categories like that are fairly indeterminate, except to point out that for the given sample, Romney is doing 6 points worse than McCain would have done!
Conclusion
The Susquehanna poll is not just an "outlier" but what in the trade is known as an OUT AND OUT LIAR.
Note 1: Most reputable polling organizations will provide the backup questionnaires, demographics and cross-tabs for published polls. Most University affiliates (Suffolk, Marist, Franklin & Marshall etc.) are especially good in this regard, the commercial pollsters less so. Most are signatories to the American Association for Public Opinion Research Transparency Initiative, but significantly
NOT Susquehanna or Rasmussen (or its parent: Pulse Opinion), notorious for its Republican tilt. Be warned.