Just a super-technical warning: If you're working with percentage margins you need to double the MoE since it applies to both findings (thus a +/-5 on a 50-50 poll could be 55-45, or a ten-point swing on a 5% MoE)
I don't think anyone in the thread forgot that, but it appears approximations are in use so I'm not 100%
I think it's actually closer to 1.66%, as the two numbers are technically related.
Sorry, I don't understand what you mean.
Each individual number can range Plus or minus 5, but they can't both be fully stacked in opposite directions to reduce or expand the margin by double the MOE.
That's basically what i remember from stats class, so if anyone else thinks they understand it, feel free to explain.
That doesn't make sense to me at all. I mean, what if there were only two options? I see no conceptual reason this would be true.