There were two polls out Sunday night - one shows Clinton 4 points up and the other shows her 11 points up. With Rasmussen showing Trump two points ahead that 11 point poll from the Wall Street Journal/NBC News looks very odd. Well here's the rub - its data is obviously flawed. Why the hell was it used at all and why has the media just not picked up on this?
The poll also asked respondents how they voted in 2012. Excluding the did not vote, don't remember respondents the numbers are Obama 57%, Romney 40% with 3% voting for others. The actual result was Obama 51.1%, Romney 47.2%. In other words this poll massively over-samples Dems and under-samples Republicans. And as such its conclusion is obviously wrong - it has Crooked Hillary on 48% and Trump on 37% with the rest split among the other candidates.
If one adjusts back for the Dem oversampling you actually get a result of crooked Hillary 43.18% and Donald Trump 38.23%. That is a lead of less than five percent. I put it to you that using a data set that was obviously flawed was going to give you a flawed result. A half decent pollster would have acknowledged this and adjusted.
This poll was conducted 10-13 October by phone. We know that Trump was gaining ground in other polls conducted on a daily basis last week - you may not like Rasmussen but it showed a clear trend as "pussygate" faded. And I also suspect that phone polls underestimate Trump support because of "shy Trumpsters."
Thus before the liberal media establishment orders the champagne for a Clinton victory party it might like to consider that a sub five point real lead early last week is almost certainly a lead of 2-3 points by now. When one factors in "shy Trumpsters," terrified of admitting to anyone that they are voting for a man demonised by the mainstream press, what the WSJ/NBC poll really shows is that the contest is just too close to call.
Not that the liberal media cares about that. Its narrative will be an 11 point lead for their candidate and Trump being toast. The days when the media offered objective analysis of crooked polls are long gone.