Thus, a 4 point lead is not outside the margin of error, because Trump could actually be up to 3.5 points lower, and Hillary could be up to 3.5 points higher.
That's the way I've always understood the MOE to function on such polls.
In any event, if Trump is close, or even up, in states like Colorado, that's a very good sign for him moving into October...
Vote Trump!
It says it’s 41-37. That’s 4 points.
Welcome to the club. You and I appear to be the only two who understand this fundamental attribute of statistical analysis. Good job!
Also, consider that no point within that range is any more likely than any other - the midpoint is reported for ease of understanding. Where this usually matters is when a talking head says, "Candidate A gained 2 points this week." The fact is, even if the new reported number is 2 points higher than last week's, Candidate A could actually have dropped from last week.
It apparently never occurs to those misapplying MOE that, if, as they would have it, MOE had something to do with the margin between two candidates, then what happens when there are more than two candidates?
What this means is that no matter how hard you squint at a yardstick, you can't use it as a micrometer.
Thus, a 4 point lead is not outside the margin of error, because Trump could actually be up to 3.5 points lower, and Hillary could be up to 3.5 points higher.
***************************
I prefer to think of it as, Trump could go up 3.5 points and that vile woman go down 3.5 points.
Anybody heard recent poll results in Mo.?