Surfing the Net today, I came across an opinion poll report at the CNN site which reminded me of one of my pet peeves with regard to the way most news organizations report on and describe such surveys. Here are the figures for this poll:
Before I continue, let me add that the graphic is in error -- the margin of error of this poll, as is made clear in the accompanying text, as well as the sample size, is in fact 3.5%, NOT 3%. Bear that in mind as we continue this discussion.
CNN, in its article, comes to the conclusion that Clinton has moved down and Dole has moved up. WRONG!!!
This error stems from the fact that CNN has incorrectly interpreted the information which the poll has collected. You see, the margin of error is not an extra added attribute to the statistics gleaned from the survey. IT IS AN INTEGRAL PART OF THE INFORMATION WHICH HAS BEEN OBTAINED! It would be a lot less misleading if, instead of including the margin of error as a not-so-important afterthought, the news media reported the figures as follows, taking full account of that margin of error:
Now December
Clinton 43.5-50.5 48.5-55.5
Dole 37.5-
44.5 33.5-40.5
Now, when we look at the comparative figures of the December poll versus the current poll it is NOT apparent that they differ. Quite to the contrary, it becomes apparent that they may very well AGREE with each other. Obviously, it's possible that Dole may have gone all the way from 33.5 to 44.5, an 11 point gain. But it is equally likely that he moved from 40.5 to 37.5, -- in other words, lost ground. In other words, the safe conservative journalistic conclusion from the results of these two polls is that, first, there's no statistically convincing evidence of any movement on the part of EITHER of these candidates, and, second, that since these two polls AGREE WITH EACH OTHER once the margin of error is taken into account, their actual similarity serves to not only bolster the confidence in the accuracy of both surveys but also reinforce the information gleaned from both of them, namely that Clinton and Dole probably stand as follows:
Clinton 48.5-50.5
Dole
37.5-
40.5
Conclusion: Clinton continues to lead Dole narrowly but persistently, with no statistically convincing evidence of movement towards, or away from, EITHER candidate.
If anything I've said is not clear, don't hesitate to e-mail me for elaboration at