About the Murphy Trend in NY-20
Mark Blumenthal | March 27, 2009
Topics: NY-20 , Siena Institute
The final poll (release, crosstabs) by the non-partisan Siena Institute certainly gives Democrats reasons to be optimistic: The trend is all Murphy, with the Democrat improving his performance on each successive poll and now holding a slight (though not yet statistically significant) four-point lead (47% to 43%). The trend is consistent with earlier internal polls conducted by the two campaigns. If we focus only on the trend, it is hard not to consider Murphy a slight favorite, especially since the final Siena poll came out of the field with four days left until voting begins on Tuesday.
However, we also see three reasons for caution about these results:
- The margin on the final Siena survey is still within the margin of error, which is to say we cannot be at least 95% confident that Murphy is leading. That does NOT mean the survey amounts to a "statistical tie." If this final snapshot were a perfect random sample of truly likely voters (a very big and essentially impossible "if" for any pre-election poll), we could be roughly 80% confident that Murphy really leads.
- Predicting turnout in this sort of special election-- and thus selecting "likely voters -- is far more difficult than in a November general election. So there is more room than usual for the possibility that all of these surveys are measuring the wrong likely electorate.
- Finally, as of this final Siena poll, Murphy has closed virtually all of Tedisco's early advantage in name recognition. Eight-six (86%) can now rate Murphy compared to 89% that can rate Tedisco, the the Democrat may have little room to grow over the last few days of the campaign.
By Mark Blumenthal | March 27, 2009 11:54 AM | Permalink | TrackBacks (0)
