Ever since June, former US secretary of state Hillary Clinton of the Democratic Party has been leading television personality Donald Trump of the Republican Party in the race for the US presidency in the Nov. 8 election. Both Clinton and Trump are very far ahead of the two other candidates, businessman Gary Johnson of the Libertarian Party, and physician Jill Stein of the Green Party.
The data cited here are mainly from“fivethirtyeight.com,” the site established by numbers maven Nate Silver, which operates, not by doing its own polls, but by compiling and averaging as many polls as possible, making its own adjustments for research quality.
The US presidency is won by getting at least 270 of the 538 electoral college votes (ECV). The number corresponds to the 435 representatives and 100 senators, plus three more for the District of Columbia. It is the ECV, not the popular vote, that determines the winner. Getting less than 270 will send the election decision to the House of Representatives, with each state (not each representative) having a single vote.
FiveThirtyEight uses three models to forecast the Nov. 8 election outcome: a) Polls-Only, or based on polls alone, which appears to be its default model; b) Polls-Plus, based also on the economy and other historical data; and c) Now-Cast, or “who would win if the election were held today.”
Polls-only odds: 60.3/39.7. Election odds are the probability that a certain party (by convention, the Democrats—i.e., Hillary Clinton) will win, divided by the probability that it will lose. Odds can go from infinity (sure to win) to zero (sure to lose); odds of 1 mean equal chances to win or lose.
From about 70/30 at the beginning of June, the odds grew to 75/25 or more in the first half of July. After mid-July it fell rapidly, reaching 50/50 by month’s end. (Note: The Democratic party convention was on July 25-28 in Philadelphia, Pennsylvania, a soft blue state. The Republican party convention was on July 18-21 in Cleveland, Ohio, a soft red state.) The odds rose to 80/20 by mid-August, dipped to 75/25 by the start of September, and dipped some more to 60/40 by mid-September. Hillary’s lead has indeed fluctuated, both upward and downward, but it has held so far.
The polls-only national overview of ECV is: Clinton 289.3, Trump 248.3, Johnson 0.4; this 41 ECV lead means that Trump needs to convert 21 votes in order to win. Its national overview of the popular vote is: Clinton 46.5 percent, Trump 44.3 percent, and Johnson 7.9 percent.
The states where the odds are close are called “battleground states.” Depending on the number of its ECV, such a state might be crucial to tipping the entire election if the underdog beats the odds.
Here are the five most important battleground states, ranked by chance of tipping the election, followed by number of ECV, and vote percentages as of 9/16.
Florida, 16.7 percent tipping chance, 29 ECV, latest vote 46.4 Trump, 46.1 Clinton, 6.2 Johnson. If she wins here, Clinton would raise her EVs by a decisive 58 (29 times 2).
Pennsylvania, 11.5 percent tipping chance, 20 ECV, latest vote 47.4 Clinton, 44.0 Trump. If he wins here, Trump would almost overcome Clinton’s lead.
Michigan, 10.5 percent tipping chance, 16 ECV, latest vote 47.6 Clinton, 43.7 Trump.
Ohio, 9.2 percent tipping chance, 18 ECV, latest vote 45.6 percent Trump, 44.4 percent Clinton.
North Carolina, 8.3 percent tipping chance, 15 ECV, latest vote 46.6 percent Trump, 46.0 percent Clinton.
Polls-plus odds: 59.6/40.3. With this forecast model, the odds started at 65/35 in June, but were reduced to 60/40 in late July. It became 75/25 by mid-August, and then 70:30 in early September, but recently fell to 60/40 again.
The polls-plus national overview of the ECV is Clinton 284.9, Trump 252.9, and Johnson 0.2. Its national overview of the popular vote is Clinton 46.8 percent, Trump 45.0 percent, and Johnson 6.9 percent.
Now-cast odds: 56.7/43.3. The odds started at 80/20 in June, but at the end of July had bounced slightly in favor of Trump. By early August it had rebounded to 95/5; it was 75/25 at the start of September.
The now-cast’s national overview of ECV is Clinton 281.8, Trump 255.9, and Johnson 0.3. Its national overview of the popular vote is Clinton 46.0 percent, Trump 44.2 percent, and Johnson 8.5 percent. Thus, Clinton’s lead holds in all three models, but is so narrow that it should not be taken for granted.
Other sentiments also matter. In early September, a national poll by Langer Research Associates for ABC News found that:
45 percent of likely voters say the United States is not as great as before, and 70 percent of them are for Trump. Of the majority that say the United States is just as great if not greater now, 72 percent are for Clinton.
24 percent say people like them are worse off than their parents; Trump wins their votes by 58-29. But Clinton wins by wide margins among those who feel steadiness or improvement in their quality of life.
33 percent feel they have no influence on the government; they vote 55 percent for Trump, 24 percent for Clinton, 12 percent for Johnson, and 5 percent for Stein. People with a little influence are split between Clinton and Trump; those with more influence are 70 percent for Clinton.
Such personal sentiments affect the votes in the ways expected. But, being general in nature, they are unlikely to change in the course of the campaign. More crucial would be the three forthcoming presidential debates, on Sept. 26, Oct. 9, and Oct. 19.
* * *
Contact mahar.mangahas@sws.org.ph.