From a reader re: Zogby vs. Gallup:
The difference in the Zogby (B-7) and Gallup (B+8) Wisconsin polls amounts to about 8 sigma! To put this in perspective, there’s an old statistics joke (perhaps the only one) about a student who asks how rare a 6 sigma event is and is told, “Well, the second coming occurs at 4.” At least one, and probably both of these polls are seriously, seriously flawed. Remember, “margin of error” only makes sense for an unbiased sample