This being an election year, it seems that not a day goes by without a few polls being reported by the media. But what do these polls mean? Do they really tell us what the public is thinking? Clearly, the answer is no. Poll after poll gets released and often, the results vary so much from one to the next that there has to be a problem. Here is an example: in Ohio, Senator Sherrod Brown, the Democrat, is running for re-election against State Treasurer Josh Mandel, the Republican. Two recent polls have been released. One by Quinnipiac was taken up to and including March 26, and it found Brown with a ten point lead (46-36). A second poll by Rasmussen was taken up to an including March 25 (the same date) and it found the racea tie (43-43). The difference is so large that it is outside the margin of error announced for these two polls.
In the presidential contest, similar anomalies are appearing. The last three head to head polls in a Romney-Obama race show Obama up by 1% (Rasmussen), 2% (McClatchy/Marist) and 11% (CNN). All were taken within the last week. One or two is clearly wrong.
The question, of course, arises as to whether or not polling organizations push their results towards their desired outcomes or if all the polling is accurate. Clearly, as the calendar gets close to election day, the polling organizations have to try to be as accurate as possible; actual results that are far off from the polling data from a firm can be the death knell for that firm. But, six months before the elections get held, the pressure to be correct is just not there, and the temptation to push the narrative may be proving overwhelming.
The message from all this is a simple one: don't believe everything you see in a poll.
No comments:
Post a Comment