評論
打印
=====精彩回顧===
Why You Shouldn't Trust Polls
為何民意測驗結果不靠譜
Polling titan Gallup has decided to withdraw from the US presidential primaries and may even sit out the general election. Perhaps it’s not too surprising, given that they predicted Mitt Romney would beat Barack Obama in 2012 and instead he lost by about a 4 point margin. Over the past few decades, polling has become increasingly inaccurate. With the upcoming 2016 election it’s important to know, why do polls get it wrong so often?
民意測驗巨頭蓋洛普決定退出美國總統選舉初選，甚至可能面對整個大選袖手旁觀?；蛟S這并不意外，因為2012年，蓋洛普預測米特·羅姆尼會打敗奧巴馬，最終他卻以4%的差距敗于奧巴馬之手。過去幾十年，民意測驗的準確性越來越差。隨著2016年選舉即將來臨，有一點很重要，我們想要知道，為什么民意測驗如此經常出錯？
Well, there are a number of reasons why polling accuracy has declined. For one, cell phones are on the rise, while fewer and fewer people use landlines. Why would this be a problem? Cell phones can’t be autodialed according to the 1991 Telephone Consumer Protection Act. In order to call someone’s cell phone,a live person needs to be the one to do it.For every thousand respondents, it takes about 20,000 calls to random numbers, most of which are non working numbers. This is incredibly expensive and as a result, some polls don’t include cell phones. And since younger people,lower income Americans and minority groups often only use cell phones, they are extremely under-represented, compared to older, white people who still rely on landlines.
民意測驗準確性下降有幾個原因。其中一個原因就是，手機用戶不斷增加，而越來越少的人使用座機電話。為什么這會是一個問題呢？根據1991年通過的《電話消費者保護法案》，手機是不能自動撥號的。為了撥打某人的手機，必須有專人撥打。要獲得一千名受調查對象，必須隨機撥打大約2萬個手機，而其中大部分是無效號碼。成本非常高昂，因此，一些民意測驗并不包括手機用戶調查。而由于年輕人，低收入美國人和少數民族經常只用手機，與仍然使用座機的老人，白人相比，前者的民意幾乎沒有得到代表。
Another problem with polling is the reticence to even state an opinion. Polls now have extremely low response rates. Back in the late 1970’s, an 80% response rate was considered acceptable.By 2012, the Pew Research Center reported that that number had dropped to 9%. Some say this is due to fears over privacy and confidentiality. It’s pretty tough to determine the pulse of a nation when relatively few people seem interested in sharing their opinions.
民意測驗的另外一個問題是，受調查對象不愿發表自己的意見。現在，民意測驗的回應率極低。追溯到上世紀70年代，80%的答案是可接受的。到了2012年，皮尤研究中心報道該數字已經降低到9%。一些人說，這是由于擔心隱私和機密性。由于相對來說很少有人愿意分享自己的意見，很難掌握國民動向。
But polling inaccuracy extends past the phone line. Internet polls come with their own set of problems. Most significantly, they do not account for a relevant group of responders. While 93% of 18-29 year olds in the US use the internet, in the 2014 midterm elections,only 13% of that age group showed up to vote.
但是民意測驗的不準確性不僅僅表現在座機上?；ヂ摼W民意測驗也存在自己的問題。最重要的是，他們沒有代表相關的群體。盡管美國93%的18歲至29歲的年輕人使用互聯網，而在2014年的中期選舉中，這個年齡群體只有13%的人投票。
Meanwhile, more likely voters have considerably lower rates of internet use, rendering online polls a poor representation of public opinion.
與此同時，更有可能投票的人使用互聯網的比率卻很低，導致在線民意調查很難代表公眾民意。
Another huge issue is the influence of early polling. Early polls tend not to correlate with final results, as most respondents have not had the time to learn about new candidates or issues. But those early polls are frequently the only source of information available for preliminary media coverage. So uninformed polls leading to glorified coverage causes a cyclical interest level. Potential respondents see the media talking about early poll frontrunners,which influences them to voice a stronger opinion, and thus possibly skew future polls.
另外一個很大的問題是早期民意測驗的影響。早期民意測驗與最終的結果關聯性很低，因為最開始的時候，大部分受調查對象沒有時間了解新的候選人或問題。但是早期民意測驗卻是媒體早期報道的主要消息來源。所以，消息不靈通的民意測驗導致媒體大肆報道，引起人們的興趣。潛在的受調查對象看到媒體報道早期民意測驗的領頭羊，影響他們表達出更強烈的意見，因此可能會歪曲將來的民意測驗。
We’ve already seen this happening in the 2016 primaries. For example, online polls showed that Bernie Sanders was the winner of the first Democratic debate. However in more traditional phone based polls, Hillary Clinton came out ahead.
我們在2016年的初期選舉中已經看到這一點。例如，在線民意調查顯示伯納德·桑德斯是民主黨第一場辯論的獲勝者。然而，在更傳統的電話民意測驗中，希拉里·克林頓脫穎而出。
But inaccurate polling isn’t just a problem in the US. The most recent election in the UK predicted that the conservative party would barely win by 1% which could have wreaked
havoc on parliament, but election results saw them winning by about 37% to the opposition’s 31%. With so many issues surrounding polling methods, it’s best to take them with a hefty grain of salt.
但是民意測驗結果不準確并不是美國獨有的問題。英國最近的選舉預測保守黨只能以1%的輕微優勢獲勝，這在議會引起了混亂，但是選舉結果顯示他們以37%的選票獲勝，而反對派只有31%。由于民意測驗方法存在諸多問題，最好對測驗結果有所保留。
While polling isn’t as reliable as it used to be, you can make up your mind for yourself about upcoming presidential candidates by watching our playlist.
盡管民意測驗并不像過去那么可靠，你也可以通過觀看我們的節目來了解未來的總統候選人。
