Saturday, March 3, 2012

Looking for Good Polling Data in the Upcoming Republican Primaries – Don’t

Five Thirty Eight Confirms Lack of Accuracy

The ability to forecast what will happen in the Republican primary races is weak.  It just cannot be done.  There are several reasons for this.  One is that primary polls are often taken by folks who are really not good a polling.  A second is that the race for the Republican nomination is very dynamic, it changes with each new result.  And finally, it seems like a lot of voters have just not made up their minds, and do not do so until they are just about ready to vote.

All of this is confirmed by an analysis of polling errors by Five Thirty Eight’s Nate Silver. 

The short version: the polls have been reasonably good in the last few days before the election. Not perfect by any means — worse than general election polling typically is, for example. But no worse, and probably somewhat better, than in past primaries. . . .

On the other hand, the polls have been pretty awful at most points prior to about three days before the election, seeing surges and momentum shifts that often dissipated.

Here is where Mr. Silver thinks the all important election in Ohio stands before the weekend.  Note:  This data is largely meaningless

Mar. 6
Rick Santorum38.3%81%
Mitt Romney30.419
Newt Gingrich18.00

The results are based on old polling which has large errors.  If the race actually turns out this way it is pure coincidence.

So at this point the polling and the projections for Super Tuesday (which is not all that Super, Ohio being the main contest after Virginia would not let Mr. Santorum and Mr. Gingrich on the ballot, thus rigging the election for Mr. Romney) are not any good.  In fact, everyone is going to have to do this the old fashioned way, that is, actually wait for the results to be in, for the votes to be counted before we will all know who won.

Gosh, that’s a terrible way to have elections.

No comments:

Post a Comment