Wednesday, January 18, 2017

The Myth Of The Bogus National Polls

I am repeating this post I did near the end of 2016 because with polls now showing Donald Trump as the least popular president going into the inauguration, the president-elect tweeted the following [this from Politico]: "The same people who did the phony election polls, and were so wrong, are now doing approval rating polls. They are rigged just like before." 

But what the PEOTUS Trump said is simply not true.  The pollsters that did the national election polls were not phony or rigged.  As you will see from this previous Tales post, it will show the Real Clear Politics average of the national polls [the same organizations that now show Trump with low approval ratings] turned out to be highly accurate.  It's final poll was within one point of showing Hillary Clinton's popular vote lead [3.1% v. 2.1%].  With the millions and millions of votes cast, it is always a wonder to me how they get so close.  

Yes, the state polls were way off and missed the electoral vote badly [and therefore missed badly the winner of the presidential election], just as they are usually more way off than the national polls, as shown in the midterm election of 2014 that they missed badly.  But that is not what we are talking about here.  Here we are talking about the national polls from national polling organizations that are doing these approval ratings.  As national polls can differ, as they did in the presidential election, always turn to Real Clear Politics that takes an average from the most credible polling organizations with credible scientific methodologies.  

So, check out the following previous Tales post first published in November of 2016.

The following post was published near the end of November, 2016 with an update on Dec. 1, 2016 and another update on Dec. 20, 2016: The RCP Average Of National Polls Was Not That Far Off:

There has been a lot of condemnation of the inaccurate polls in this presidential election, and that is true in the various state polls [that predicted the electoral count].  These state polls seem to always be less accurate than the national polls as the last couple of midterm elections have shown.  In 2014 the polls missed badly the GOP landslide wave that occurred.

But while the battleground state polls were very inaccurate this year, the final results of the popular vote in the 2016 election show that, as usual in presidential elections, the average of the national polls as shown by Real Clear Politics was not that far off.  With most of the vote now in, Real Clear Politics has Hillary Clinton with 48.2% of the vote, Donald Trump with 46.5% and all other candidates with 5.3% of the vote.  

As usual, the RCP average of national polls was not that far off in this 2016 presidential election

While meaningless as far as who the winner was [i.e., the president elect], the national polls "RCP average" of about a 3.2% Hillary Clinton win of the popular vote was pretty close to the actual 1.7% win of the popular vote by Hillary Clinton.  

It is true that some of the individual polls were not close, but a couple were pretty close to the final result.  Here are the best and the worst of the national polls as recorded by Real Clear Politics.

The Top 3 Most Accurate Polls:

1. IBD/Tipp tracking poll had Clinton winning popular vote by 1% [off by only 0.7%]

2. The Bloomberg poll and the 3. Washington Post poll both had Clinton winning the popular vote by 3% [off by 1.3%]

The Most Inaccurate Polls:

1. LA Times/USC Tracking poll predicted a Trump win of the national vote by 3% [off by 4.7%]

2. the Monmouth poll had Clinton winning the popular vote by 6% [off by 4.3%]

3. NBC News and 4. Reuters/Ipsos both had Clinton winning by 5% in the popular vote [off by 3.3%]
UPDATE:  Dec. 1, 2016: The latest numbers from the 2016 election just came out from Real Clear Politics and they show even more, that while some of the individual polls taken by themselves were way off, the RCP average of polls of the national vote was very accurate.  

So, basically on election day, Real Clear Politics showed Hillary Clinton with a 3% edge and with the results almost all in, Hillary Clinton ended with a 2% lead in the national vote.

Pretty darn close if you ask me, and it lets me know in future elections, Real Clear Politics is the place to turn [and not the individual pollsters] for the best accuracy. 
UPDATE: December 20, 2016:  Now with almost all of the final numbers in from the 2016 presidential election it shows how close the Real Clear Politics average was in their final national vote [popular vote] prediction to the actual result on election day.  The RCP final poll had Hillary Clinton winning the popular vote by 3.1%, and they were only 1% off as Hillary Clinton won the actual popular vote, 48.2% - 46.1% = +2.1%. 

Bret Baier pointed this out on his Special Report show on Monday night Dec. 18, 2016, that while the polling in the individual states missed the winner of this election, RCP actually was very close in the popular vote difference and winner of that.  

The final result also means that the Bloomberg Poll and the Washington Post Poll were the closest to the final result as they had Hillary Clinton winning the popular vote by 3% and missed it by only 0.9%

The LA Times/USC tracking poll ended up in last place as they were way off in the national popular number final results. They had Trump winning the popular vote by 3%, while actually Clinton won it by 2.1% -- they were -5.1% off. 

So, while individual national polls missed it, and many state polls missed the election, I say Bravo once again to Real Clear Politics for pretty much nailing the final national presidential poll result. 

No comments: