One Of These Polls Is Not Like The Others

OK, no sooner do I post about how one poll can be misleading than we see an example of it.

Four polls have come out in the last few days: American Research Group, Harris, Pew and CNN/USA Today/Gallup.

Three of the polls have the national horse race in a statistical tie (neither candidate ahead of the other by more than the margin of error). ARG and Harris have Kerry slightly ahead (ARG has Kerry with a statistically significant lead if you take Nader out of the race). Pew has Bush slightly ahead. Zogby polls from the last week or so also show a tied race.

But Gallup has Bush with a 55-42 lead! This is not completely out of the blue. Gallup has been consistently showing more support for Bush all along. If you go back and look at Gallup for months you'll see they showed Bush with leads back when every other poll showed him significantly behind.

The problem with this is that CNN ONLY REPORTS THE GALLUP POLL (guess why). So CNN is not only making the poll results the story rather than jobs, Iraq, Enron, Halliburton etc but they are making what looks like an outlier poll THE story. The result is that millions of people are being told that Bush has a prohibitive lead when in fact it that is almost certainly not true. It is more likely that he is tied with Kerry and could actually be behind given the poll results that CNN doesn't report.

Which brings me to another caveat, don't rely on only one news source!