It seemed like all the major cable news outlets could talk about on Wednesday was the results of a new survey by the Quinnipiac University Poll. That poll (top chart) showed Hillary Clinton would lose to all three of the most likely GOP candidates in some swing states -- including Virginia. In Virginia, she trails Bush and Walker by 3 points, and Rubio by 2 points. Is this true?
Maybe not. A different survey in Virginia, done by Public Policy Polling, had a quite different result (bottom chart). It showed Clinton leading all three of those candidates --Bush by 8 points, Walker by 5 points, and Rubio by 4 points.
Both of these surveys were done in generally the same time period (Quinnipiac between July 9th and 20th, and PPP between July 13th and 15th), had similar survey sizes (Quinnipiac questioned 1,209 people, and PPP questioned 1,170 people), and had similar margins of error (Quinnipiac's was 2.8 points, and PPP's was 2.9 points). How did the two surveys arrive at such different results? Which is correct?
One of these polls is an outlier (a poll that disagrees with all other polls). In my opinion it is the Quinnipiac University Poll. Their results disagree with all of the other polls we have seen, while Public Policy Polling's survey is in line with other polls.
It is possible that Quinnipiac is showing a change happening, but I don't believe it -- at least not yet. It will be interesting to see what future polls show.
Not that polls mean much at this time, but single polls never mean much. The main thing I notice is that undecideds are much larger in the Quinnipiac poll. If you add the undecided difference, it pretty much gives the PPP result. As I learned in analytic chemistry, when you only have two data points, you are far better off averaging them rather than trying to justify one being better than the other. The average puts Clinton ahead in each race.
ReplyDelete