Election 2012 Siena Poll Updates

Posted at: 10/10/2012 6:01 AM
By: Dr. Timothy Kneeland -- Political Blogger

Since the blog on the State Senate and Assembly races was written a new Siena Poll has come out with a look at the Ted O’Brien and Sean Hanna Senate race.    The poll shows Hanna up by 8 points over O’Brien, 47% to 39%, but a whopping 14% have yet to make up their minds.   This is good news for both candidates.  On the one hand, an aggressive campaign by O’Brien or some help from more well-known and popular Democrats such as Bob Duffy and Andrew Cuomo might swing some of those votes to the Democrat.  On the other hand, Hanna’s ability to run an aggressive campaign through a plethora of ads defining O’Brien through his connection to wayward judges has been very effective.   At this point it is Hanna’s race to lose.

Over in 27th Congressional district House incumbent Kathy Hochul is polling even with challenger Chris Collins the former Erie County Executive with both holding 47% of likely voters.   The six percent of undecided voters, many of whom will likely break for the challenger as well as the fact that Mitt Romney is leading Barack Obama in the presidential polls in that district gives Collins the distinctive advantage.     

Speaking of polls there seems to be a tightening of the presidential race in several national poll of late (post Debate) and in the swing states.   There are two things happening in these polls one is a bounce that Mitt Romney is receiving from his impressive victory of the President in the October 3rd debate.  The other is that pollsters are changing the number of Democrat and Republican voters as the race draws closer.   Frankly, I have suspected all along that many of the earlier polls were over sampling Democrat voters using the models from the 2008 race when 39% of voters were Democrats, 29% Republican, and   30% were Independents.    My feeling all along has been that 2008 was a historic race that will not be duplicated in 2012.   Recent polls by Pew are sampling 31% Democrat, 36% Republican, and 30% independents which are based not only on registered voters but likely voters.   This may be an over corrective but provides some insights into the problems of sampling which must reflect not only voter registration but turnout models and demographic information from cell phone use to age and race.   The best predictor of the outcome of the presidential election since 1952 has been the mid October Gallup Poll.  We will have more to say about this poll next week.  In the meantime the next blog will be about the vice presidential debate scheduled for Thursday.