Reply #14: This is great news, but the fact is that it is not outside the confidence interval

[karynnj]

I wish that journalism degrees would require a class in statistics. Here, both the Warren estimate and the Brown estimate have confidence intervals around them - and those confidence intervals overlap. You can not take the difference and then use the MOE for the point estimates and test whether the interval is all above zero. The problem here is that the small sample size leads to a very large MOE.

However, this does not mean that they are equal. The probability that Warren is genuinely ahead is quite high - and can be computed. (It might be good if the polling companies included the calculation so reports could quote it to accurately make the statement that this article tries to make in the wrong way.

I wonder if Brown will be further hurt by his hiding from the MA residents who spent hours camped outside his office. Brown voted, so he was there. It is ridiculous that he avoided his office for hours rather than speak to them. (In fact, imagine a very smart politician, who would send staff with sandwiches (or cookies) and explain Brown's schedule "working for MA" - and finally getting a 5 or 10 minute time with Brown "squeezed between commitments" How would that have coopted the entire effort!