Reader question:

Please explain “exit poll” in this sentence: At least one exit poll has Hillary Clinton losing….

My comments:

That is to say, at least one opinion poll right after votes were cast pointed towards Hillary Clinton losing on election night.

Literally, “exit poll” means a poll at the exit of the polls.

Sorry about the confusion. The polls (plural) refer to a polling station or a polling place where voters cast their votes at the poll booths. But the “poll” in “exit poll” refers to an opinion poll, a survey of public opinion via asking people questions.

Anyways, a poll at the exit of a polling station works like this. After each person comes out of polling station, having cast their votes, pollsters ask him or her who they have voted for – in the case of this year’s American election, Clinton or Donald Trump.

After gathering a sufficient number of answers, pollsters will announce the result of their survey at their particular locale, whether Clinton has won more votes than Trump does – or vice versa.

There are many polling stations around the country, of course. Therefore, exit polls can be conducted anywhere – at any such “exit”.

By conducting the exit poll, a newspaper or TV station or some online news service is able to make a reasonably successful prediction on election results because it is done after people have actually cast their votes. Polls before the election can theoretically be considered more or less erroneous because people, for example, may not have made up their mind at that point in time. After the election, however, people are usually more willing to tell – and tell the truth.

Anyways, exit polls are tools used by news outlets to predict election results during the time lap – during the hours after votes are cast and before final results are formally announced.

That’s it. Oh, one more comment. If this year’s American Presidential election tells anything (and it tells many people different things according to their politics, religion, sex or sexual orientation, not to mention the colors of their skin), it tells us that this election-year’s opinion polls (after-the-fact exit polls excepted) are not to be trusted.

For one thing, an overwhelming number of polls in the immediate run-up to the election has Clinton winning, and mostly by a comfortable margin, too.

In the end, however, Trump was the one to celebrate. And he, too, as a matter of fact, won quite comfortably, judging by Electoral College votes.

How did this happen?

Well, there are no easy answers to that. So here, now, let’s read media examples of “exit polls”:

1. Social and intellectual conventions are supposed to settle slowly, but conventional wisdom can congeal instantly and without much wisdom. That’s what has happened over the past several weeks with a prevailing interpretation of this year’s presidential election -- the great “moral values” theory.

The Big Political Idea of the 2004 election goes something like this: “Moral values” turned out to be the most important issue to voters, not the economy or the Iraq war or terrorism. President Bush won because a legion of “values voters” -- whose growing numbers escaped the attention of an inattentive media -- preferred him. The Democrats are doomed until they can woo the voters who belong to this new political force.

It’s a neat theory -- but wrong. How it came to be regarded as the real story of George W. Bush’s victory is a fascinating and sobering example of journalism’s quest for freshness and surprise.

Here’s the simple fact: The evidence that moral values determined the election rests on a single dodgy exit-poll question. And it’s not at all clear that more voters are preoccupied with moral values now than were fretting about “family values” on election day 1996, when exit pollsters included that phrase in a question about “priorities for the new administration.” But in the often arid and repetitive arena of American political ideas, fun new contestants can be hard to disqualify. The myth of the moral values election is proving hard to snuff out.

The mantra was in full hum on election night. Television commentators were understandably struck by the results of the question asked of almost 7, 000 voters as they left their polling places: “Which one issue mattered most in deciding how you voted for president?” The most cited issue on the list of seven options offered to those surveyed was “moral values” at 22 percent; 80 percent of these voters went for President Bush, 18 percent for Democratic nominee John Kerry. “Economy/jobs” came next on the list at 20 percent, followed by terrorism (19 percent), Iraq (15 percent) and then health care, taxes and education in single digits.

Brian Healy was the CBS News producer covering the exit polls, something he has done in many elections. He recalled that everyone was surprised that moral values topped the list as the numbers came in, but it wasn’t until about 4 a.m. that someone quite innocently asked, “What exactly are ‘moral values’?”

Too late. The story line was already set. And the surprise nature of the moral values result boosted its allure for the commentariat. When the newspapers could finally write definitive headlines, the notion that moral values were a synonym for various conservative positions became a given -- as did its decisive effect on the outcome of the contest. “Faith, Values Fueled Win,” reported the Chicago Tribune. “Moral Values Decide Election,” the Tri-Valley Herald in Northern California told its online readers.

From the modest experiment of one exit-poll question, a Unified Theory of Election 2004 was hatched. Pundits began to spread the word.

Some reporters were even apologetic for missing the big story. “Somewhere along the line, all of us missed this moral values thing,” said CNN’s Candy Crowley in a speech to a Florida audience.

- How story of 2004 election hinged on one exit poll, SFGate.com, December 6, 2004.

2. Donald Trump has beaten Hillary Clinton in the race to be president of the United States.

Much of the narrative ahead of the election had been that Mr Trump was supported by angry, white men. To get an insight into which groups actually voted for him, you can look at the exit poll conducted across the country by Edison Research for the National Election Pool, a consortium of ABC News, The Associated Press, CBS News, CNN, Fox News and NBC News.

It is very difficult to get a genuinely representative sample of how more than 120 million people have voted. It is a big survey - of almost 25,000 voters - and they are the best figures available, but they should be used with caution.

It throws up some odd results, such as that 10% of people who support the idea of a wall along the Mexican border nonetheless voted for Mrs Clinton, while 5% of people who thought the next president should continue the policies of Barack Obama voted for Mr Trump.

Bear in mind that the proportions are unlikely to add up to 100%, because not everybody answered all the questions and there were other candidates standing in the election, who received about 5% of the votes.

The poll suggests that 53% of men voted for Mr Trump, with 41% voting for Mrs Clinton - those proportions are almost exactly reversed for women.

Among white voters (who made up 70% of voters), Mr Trump won 58% to Mrs Clinton's 37%, while the Democratic candidate won the support of a huge majority of black voters - 88% to Mr Trump's 8% - and Hispanic voters - 65% to his 29%.

Looking specifically at white women, they favoured Mr Trump, with 53% supporting him compared with 43% for Mrs Clinton.

- Reality Check: Who voted for Donald Trump? BBC.com, November 9, 2016.

3. Many observers thought this presidential election would be decided by Donald Trump’s polarizing rhetoric, his history of behavior toward women and his questionable qualifications for the office.

Instead, CBS News exit polls suggest Trump’s win was in large part a repudiation of Hillary Clinton by a substantial number of white voters. While Clinton did win big majorities of minority voters, she did not get the level of support from those groups that she needed to overcome her deficit among white voters.

There are also indications that Clinton’s gender was a factor in the outcome. The gender gap was substantial. Trump beat Clinton by 53 percent to 41 percent among men while Clinton won among women by 54 percent to 42 percent. Four years ago, President Obama won 45 percent of men’s votes and Mitt Romney won 44 percent of women’s votes.

More telling is the gender breakdown among white voters: Trump beat Clinton among white women 53 percent to 43 percent. This was close to Romney’s margin in 2012. While Mr. Obama won 35 percent of white, male voters in 2012, Clinton lost to Trump among this group by 63 percent to 31 percent.

As expected, Trump did best among white voters without a college degree, beating Clinton by the enormous margin of 72 percent to 23 percent. Trump also won among white, non-college women 62 to 34 percent and white college-educated men, 54 to 39 percent. Among white voters, Clinton only won among women with a college degree by a 51 to 45 percent margin. Interestingly, among white voters, there is no evidence in the exit poll that income affected the likelihood that they supported Trump.

Clinton needed extremely strong support from African-American voters to try to offset Trump’s margin among whites. She did win 88 percent of the black vote to just 8 percent for Trump. However, this was significantly lower than the 93 percent of black voters Mr. Obama won four years ago. The falloff in her share of the black vote was entirely due to black men. Clinton won among black women by a 93 percent to 4 percent margin. Among black men she won by 80 percent to 13 percent.

Many political observers thought a significant number of Republicans would either vote for Clinton, one of the third party candidates, or stay home rather than casting their votes for Trump. According to the exit polls, Republicans stayed loyal to their presidential candidate. Some 89 percent of self-described Republicans voted for Trump; 91 percent of white Republicans did. In contrast, only 84 percent of white Democrats voted for Clinton. She did win 86 percent of white Democratic women, but only 81 percent of white, Democratic men voted for her.

Surprisingly, given all of the attention to Trump’s attitudes and behavior toward women, he did virtually as well among white, Republican women (91 percent support) as he did among white, Republican men (92 percent). Clinton was more competitive among white independent women than men, losing to Trump by a 49 to 41 percent margin among independent women and by 57 to 31 percent among independent men.

- CBS News Exit Polls: How Donald Trump won the U.S. presidency, CBSNews.com, November 9, 2016.

本文僅代表作者本人觀點，與本網立場無關。歡迎大家討論學術問題，尊重他人，禁止人身攻擊和發布一切違反國家現行法律法規的內容。

About the author:

Zhang Xin is Trainer at chinadaily.com.cn. He has been with China Daily since 1988, when he graduated from Beijing Foreign Studies University. Write him at: zhangxin@chinadaily.com.cn, or raise a question for potential use in a future column.

（作者：張欣 編輯：丹妮）