07-25-2016, 05:26 PM
This article should be interesting. Here is the URL: http://www.nytimes.com/2016/07/23/upshot....html?_r=0
Here is the article:
Here is the article:
Quote:Donald Trump officially became the Republican party’s nominee Thursday night, and on Monday, the Democratic convention begins in Philadelphia. In the coming weeks, you can expect lots of polls — and headlines — suggesting new insight into the state of the presidential race.
With some caveats, our advice is: Don’t pay too much attention to them.
You can see what we mean in the chart above. It shows how much the polling average at each point of a presidential election cycle has differed from the final result. Each gray line represents a presidential election since 1980; the bright green line represents the average difference. In general, as the election nears, the polling average comes closer and closer to the election’s final result — but not for the next few weeks.
History suggests that in the short periods after the conventions, the polling average can often move away from the final result, not toward it. That’s because polls taken in the middle of the convention are often unreliable: Gains made by the party’s nominee can often be short-lived.
One approach is to ignore the polls during this tricky period. Our presidential forecast, which currently gives Hillary Clinton a 74 percent chance to win, does something different: It imposes a small penalty on polls taken when a candidate might be receiving a convention bounce. This penalty, based on conventions since 1980, fades out over the next few weeks.
At the same time, and perhaps more important, the model increases its variance estimate — how uncertain it is about its assessment of the race — during the convention period. Because of this volatility, it makes sense to judge polling over the next few weeks with more skepticism than usual. As a general rule, when variance goes up, it helps the underdog, so it’s possible that the model’s estimate of Mrs. Clinton’s overall chance of winning may decline slightly in the weeks ahead.
Of course, it’s also possible that Mr. Trump will not get a convention bounce at all, reflecting perhaps the most disorganized, unpredictable and bizarre convention in decades.
It’s unclear whether conventions still hold the power they once did. The bounces candidates received in 2008 and 2012 were more modest than some in years past, but television ratings for the conventions remain high. Ratings for the 2012 Democratic convention were the highest since 1992, and ratings for the Republicans in 2008 were the highest since 1976.
Bounce or no bounce, the history lesson remains: Polling averages tend to be volatile in the weeks after national conventions. As the election nears and the short-term effects of the conventions fade, the polling averages will rapidly become more precise.
In other words, if you’d like to take a break from political polls for a few weeks, you won’t be any worse for it. We’ll be here when you get back.
Until then, take a look at this series of charts showing how conventions have affected polling averages in every election since 1980.
What polls say about the 2016 election
Mrs. Clinton held a steady lead going into the conventions.
Trump becomes presumptive nominee
What polls said about the 2012 election
President Obama seemed to receive a bump in polls from the Democratic convention and the “47 percent” video of Mitt Romney released by Mother Jones. But polls tightened again in the month before the election. In 2012, Mr. Obama outperformed his polling averages on Election Day by about three percentage points.
“47 percent” video published
What polls said about the 2008 election
In some ways, the 2008 election mirrors this race, with two Democratic candidates fighting over the party nomination after the emergence of a presumptive Republican nominee. At this point in that election, Mr. Obama and John McCain were nearly tied, but that was before Mr. Obama was the clear Democratic nominee. Once he was, he became the favorite. The most significant movement came in the closing months of the contest, amid the 2008 financial crisis.
McCain: "The fundamentals of the economy are strong."
Obama clinches nomination
What polls said about the 2004 election
John Kerry held a slight lead for a couple of months in 2004, but George W. Bush hung on to his five-point Republican convention bounce for a narrow victory.
What polls said about the 2000 election
With all of the controversy surrounding the 2000 election, and the Florida recount in particular, it is easy to forget that the national polling average in November showed George W. Bush with a comfortable three-to-four-point lead over Al Gore.
But on Election Day, it was Mr. Gore who won the popular vote while losing the Electoral College.
What polls said about the 1996 election
Third-party candidates add an extra complication to presidential polling. Frequently, they perform well in polls early on, but their support tends to fall off.
In 1996, early polls showed the Reform Party candidate Ross Perot at 15 to 20 percent, a number that dwindled to 8 percent by Election Day.
You can see the well-defined convention bounces in 1996, as first the Republican and then the Democratic convention seemed to swing the polling average toward each party temporarily.
In terms of raw numbers, the 1996 election was the worst polling miss since 1980, with the Election Day average missing the final vote by over four points. But the ultimate result was never in doubt; Bill Clinton won by almost twice that margin.
What polls said about the 1992 election
The combination of the Democratic convention in July and Ross Perot’s abrupt departure from the race seemed to be responsible for a swing in polling average toward Mr. Clinton of over 20 points. Mr. Perot re-entered the race in October, but it did not substantially alter Mr. Clinton’s standing in the polls.
Perot drops out
At this stage of the 1988 election, polls were very far from the final result, swinging a full 20 percentage points over the next five months, generally favoring George H.W. Bush after the Republican convention. Mr. Bush did win handily, as expected, though by a slightly smaller margin than polls predicted.
President Reagan led by double digits in most polls at every stage of his re-election campaign, but even these numbers swung wildly after Walter Mondale emerged as the Democratic nominee, and after both conventions.
Mondale clinches nomination
The 1980 contest is a good example of how a candidate’s standing in the polls can collapse.
Jimmy Carter led by around 10 points in early April.
A spate of polls in the week leading up to Election Day suggested that Mr. Carter might have narrowed Mr. Reagan’s lead – three of the nine polls released in the final 10 days of the campaign even showed Mr. Carter ahead – but ultimately, Mr. Reagan won handily.