If you are worried about Hillary losing read this
The State of the Race – 11/6/2016
I don’t claim to be a political expert. However, given the weird environment surrounding this Presidential election cycle, not even the experts appear to have a firm enough handle on the race to be able to make confident predictions about the outcome. The best political prediction experts, such as Nate Silver, hedge their bets by not making absolute predictions; they instead provide candidates’ chances of winning. Thus when Silver gives Donald Trump a 36% chance of winning the Presidency, if Trump were to win Silver can say, “See I never told you that his chances of winning was zero”. On the other hand I have followed this election battle about as closely as any layman could and I have noticed factors which makes me confident that Hillary Clinton will ultimately emerge as the victor.
I don’t mean to disparage Nate Silver and his FiveThirtyEight website’s model. On the contrary, based on my knowledge of statistics, I believe that Silver’s complicated model is the best method of making sense of the amazing amount of poll data which is made available daily this close to the election; in the three day period ending today (Saturday, 11/5/16) the results of 331 legitimate state and national polls were reported. I believe that no other political data analyst condenses this massive of poll data into better predictions that Nate Silver.
However, as with everything else in life the FiveThirtyEight methodology has its advantages and its disadvantages. One of its disadvantages this close to the election is inability to pick up on what I will call micro trends, for instance the current tendency the polls over the last few days in some the battlegrounds states to mostly favor one candidate over the other. For instance, Silver rates currently rates Florida as a virtual toss up giving Trump a slim 52.6% chance of winning the state. However, this small Trump advantage is based on the weighted average of 90 Florida polls, including some of the very lightly rated polls which were conducted in early September. I fail to understand how polls conducted in September, or even the many polls in the mix which conducted in early October, have any relevance to the state of the race in Florida today.
On the other hand, I have noted the 7 latest Florida polls which completed their canvassing 11/1 or later favor Hillary Clinton. The results of those 7 Florida polls (with the most reliable polls shown first) were: Clinton +2, Clinton +2, Clinton +1, Trump +1, Clinton +4 and two polls with ties – advantage clearly to Hillary. The same micro trend can be seen in a second toss up state, North Carolina, where Silver give Trump a very slight edge with 51.1% chance of winning. The 4 latest NC polls which completed their canvassing 11/1 or later had the following results: Clinton +2, Clinton +3, Trump +5 and Clinton +3. (Again the most reliable polls appear first.) Again advantage Clinton. The micro trend in the third and final tossup state, Nevada, is not as clear cut. Silver currently gives Trump a slight edge in Nevada with a 50.3% chance of winning the state. The latest polls concluding in November show mixed results: Trump +6, Clinton +3, Clinton +7, Trump +1 and a tie.
In addition, statistical analytical methodologies are only as good the polls that they analyze. When all or most of the polls for a state or states are not corrected by their pollsters for factors which are not obvious, the accuracy of the polls suffer as so will the accuracy the predictions of the analysts who depend on those polls. Two of those factors give Hillary the edge in this election.
Continue reading here:
The State of the Race – 11/6/16
= new reply since forum marked as read
