Oregon’s Democratic Gov. Kate Brown leads her Republican challenger Rep. Knute Buehler by 3 percentage points in a new poll released Wednesday.
No, the two are in an indistinguishable statistical tie. The Oregonian’s fiction writer even says as much later on:
Brown’s lead is within the margin of error of plus or minus 3.7 percentage points, meaning it’s possible she could pull off a more decisive win or lose to Buehler.
The only legitimate finding is that the two are statistically tied and we do not know which is ahead of the other. Creative news writers make this mistake every week, however. We’ve written about it before.
The other legitimate finding is that reporters should never, ever be left in the room, alone, with a number.
Polls largely measure the effectiveness of prior propaganda and publishing their result reinforces the desired message. False reporting on the outcome of a poll, as the Oregonian has done here, is also propaganda messaging.
Another trick is to forecast the election outcome with, say, 75% likelihood of candidate X beating candidate Y. This is just speculation. Pundits can make these forecasts to generate a “Get on the bandwagon” effect – many will want to associate with the expected winner. Yet the pundit has also given themselves a large “out” by saying “I could be wrong” (25%). Forecasts like these are meaningless but they do sound persuasive, don’t they?