Common Sense: By the numbers

The real consequence of the back and forth between Hillary Clinton and Donald Trump will be fewer people casting their vote for president. How this will affect the final outcome I do not know. But it does affect the reliability of the polling numbers. The pool of likely voters is probably changing faster than is reflected in the polling data.

Of course, anyone running down ballot for Congress or the state legislature wonders how this will affect them. The old rule of thumb was that a state legislative candidate could expect to pick up one point (which translates into a two-point shift) on his or her total percentage for every five points the candidate for president won his or her district, Thus, if you were going to lose 52 to 48 and your candidate for president wins your district by 10 points, you would probably end up in a 50/50 outcome. The same is not true during a gubernatorial race since turnout is usually not so heavily elevated.

That is the way it had been historically. I do not think that is what is going to happen this year because of the unusually vexatious nature of the attacks by each candidate.

There is also normally a large fall-off from those voting for the top of the ticket to the bottom. In presidential years, it can be as high as 20 percent. This year I think it will be less, as a result of many campaigns pushing to make sure that voters turned off by Clinton and Trump still come out and vote for them, kind of a bottom-up strategy.

* * *

Speaking of polling, the Los Angeles Times is conducting daily tracking polls using a new methodology. It will be interesting to see how its numbers compare to the final outcome inasmuch as, if they are accurate, I suspect national polling will change forever.

Normally a scientifically drawn sample of say 4,000 or 5,000 voters is drawn nationally, and from that larger sample, 1,000 or so interviews are conducted, which then go through a slight adjustment to conform with national demographics in order to increase accuracy.

Since the larger pool from which the smaller sample was drawn has been scientifically selected, you can then come up with the margin of error and an overall reliability percentage. All polling only marks a moment in time so, a week later, you can more or less throw the data out although many polls using the same methodology give you a trend which can be very important.

The LA Times is doing it differently. It chose a national scientifically drawn sample of 2,000 Americans in August and has been polling a drawn sample of that group every few days ever since. It is always a sample drawn from within the same 2,000.

This poll has consistently shown the national race to be within two or three points for weeks, mostly with Donald Trump slightly ahead.

The averages of all standard polling show Hillary with around a five-point lead and have shown her consistently winning since August. It should be noted that neither the LA Times polling nor the national averages give you any sense on how the candidates are doing in individual states, which is all important.

It will be very interesting to see how accurate the LA Times has been.

* * *

The Brooklyn Conservative Party will be holding a reception at Hunter’s Steakhouse on the evening of Wednesday, November 2 at 7 p.m. Radio personality Jim Kerr will be receiving an award named in honor of longtime party activist Jim Gay. David Keene, the founder and longtime chaiperson of the American Conservative Union, as well as a former national chairperson of the NRA, will be the guest speaker. Tickets are $90 per person and there is a journal. For ticket information, please call 718-921-2158.

* * *

The Brooklyn South Conservative Club chaired by Liam McCabe will be hosting radio personality Frankie Russo at an event at 7 p.m. on Wednesday, October 26 at the Leif Bar, Senator Street and Fifth Avenue. All are invited.

Leave a Reply

You must be logged in to post a comment.