I was just called by a Gallup pollster with a survey about drinking and driving habits from the DOT. Overall it seemed like a pretty well-designed survey, but one formulation struck me as odd. I was asked about several measures that could be taken to reduce DUI, and asked to rate them on a scale of effectiveness:

1. Very effective
2. Somewhat effective
3. Neither effective nor ineffective
4. Not very effective
5. Not at all effective

I don't know what it's called as a concept, but it seems logically impossible for something to be neither effective nor ineffective. If it has no effect, then it is ineffective. The only thing that might fit the definition is something that is not intended to have any effect, like say, the color blue, but that was not the case here.

If I had to resolve this contradiction, I would say that 3 seems the same in meaning as 5, which was clearly not the intent. Better perhaps would be to make the scale from very effective to very counterproductive, with ineffective in the middle. Maybe I think some of the measures would increase DUI!