[TomGuy]

the Post would talking about how close it was.

In Arkansas, the 2 Senate candidates were almost tied, one would get a slight lead, then the other -- according to the polls and press.

When the votes were counted, Cotton (R) claimed nearly 57% and Pryor (D) got only 39%.

No poll showed them that close. A UofA poll the week before the election did show Cotton up by 14%, and most who saw that poll, discounted it as wishful thinking. Ironically, it was closer to the final tally, even though it, too, was well under the final total.

Many of the polling organizations need to retool their approaches and methodology, because they have been mostly wrong since the 2010 election cycle.

[MUDDOG]

Many of the polling organizations need to retool their approaches and methodology, because they have been mostly wrong since the 2010 election cycle.

Nate Silver wrote about that in his post-election analysis, and said he was going to try to find out what the statistical/sampling reason for it was.

Of course, polls have become instruments of propaganda as well, so there's that too.