Sigue el blog por EMAIL. Seguir por EMAIL

viernes, 6 de mayo de 2016

Why Hillary Clinton has a 70% chance of beating Donald Trump

Why Hillary Clinton has a 70% chance of beating Donald Trump: An elections data wonk explains what polls reveal about the general election matchup

Clinton blasts Trump at California rally
In this crazy Presidential primary season, it has been said that data journalism has failed. That is not exactly true — though I do think data journalists themselves have some lessons to learn.
To those of us who look closely at public opinion polls, the near-inevitable rise of this year’s likely nominees Hillary Clinton and Donald Trump has been visible for months. In early January, I arrived at this conclusion based purely on polling data and comparing that to the path of past nominees. Trump's polling numbers never faltered from the day he entered the race last July.
The main lesson, I think, is to keep one's pundit-y judgments separate from direct measurements of public opinion. Looking honestly at the data, what can we now learn about what will happen on Election Day, 186 days from now? Even this far out, polls can tell us something. A purely poll-based view of the race is like a thermometer for the weather — it is an objective measure of conditions now.
And like the temperature on a given day, voter preferences usually change slowly enough that past history can give us an idea of where things are likely to head within an election season. We can get a deeper sense of the interaction between polls and final results by looking at how the last 16 presidential elections unfolded. From early May to Election Day, the typical amount of net movement in polls has been about 12 points, as measured by the margin between the two major candidates. Currently, Hillary Clinton leads Donald Trump by 7 percentage points.

The real odds

Given Clinton’s lead, two out of three times we would expect a likely range of November outcomes of anywhere between a Trump win of 5 percentage points to a Clinton blowout of 19 percentage points. Based on this information, I place the probability of a Clinton win at 70%, or 7-3 odds in favor. That is not a terribly definitive statement, but it has the advantage of being relatively uncontaminated by pundits’ judgements.
And it allows for the fact that once in a while, a surprise can happen. The most extreme such case happened in 1980, the year of the Iran hostage crisis and a national energy crisis. Jimmy Carter led Ronald Reagan in May surveys by about 10 points, but ended up losing by almost as much. However, statistically speaking, Trump supporters should not count on beating the odds in such a spectacular manner because such a comeback has succeeded only once in over fifty years.
Of course, we don’t decide presidential elections through a national popular vote, but via contests in 50 states and the District of Columbia. State-level polls give the same picture of the Clinton vs. Trump contest. I have examined all available state-level polls that were taken since February. If the election followed those polls on average, Clinton would get about 350 electoral votes, and Trump would get about 160 electoral votes. This pattern of states is about what would we would expect for a 7-percentage-point popular vote win — exactly where Clinton and Trump are in national polls. [See this map, which contains my estimates.]
There is yet another telling piece of information: President Obama’s approval rating, which is a measure of public approval of his party and is predictive of the election outcome. Currently he stands at a net approval of +1%; that is, slightly more people approve of his performance than disapprove. That is just two percentage points worse than the day he was re-elected in 2012. Net approval can also move, but not as much as Presidential polls. So while this information does not tell us directly about Hillary vs. the Donald, it does tell us that Democrats are likely to have the advantage this fall.
The Presidential race is worth watching not just for who will occupy the White House, but also for down-ticket races. In 2012, I found that Senate races followed Presidential polls up and down. In the Senate, Republicans are at a disadvantage because they are defending far more seats than Democrats. Larry Sabato’s Crystal Ball currently suggests that Democrats are likely to end up with at least 47 seats, with at least 5 more seats in play. Because Clinton would have coattails, a Trump loss might very well take Republican leader Mitch McConnell’s Senate majority down with him.
On the House side, Democrats have a steeper climb, thanks to an equal combination of how their voters are arranged geographically and Republican-driven gerrymandering. I estimate that to have a chance at retaking the House and putting the speaker’s gavel back in Nancy Pelosi's hands, Democrats would have to win at least 5 percentage points more votes in Congressional races nationwide. This would be tough, but then again, Democrats won the national Congressional vote by 8 points in 2008, the year of Barack Obama’s first election.
Publicar un comentario