11月3日，新闻受众很可能会接触到一种独特的、也是典型选举之夜体验组成部分的信息：出口民调。与选举前进行的民调不同，出口民调在选民投完票后、通常是在他们离开投票站时，询问他们投给了谁。
On Nov. 3 news consumers are likely to come across information unique to and part of the quintessential election night experience: exit polls. Unlike polls done before the election, exit polls ask people who they voted for after they cast their ballot, often when they exit their polling place.
出口民调能为选举结果提供早期洞见。但其涵盖范围远不止“你投给了谁？”这一个问题。出口民调的问题涵盖从经济等宏观议题到候选人经历等州级具体话题。如果说选举前民调提供的是选举前竞选态势的快照，那么出口民调则让记者能够实时剖析选举结果背后的“原因”。
Exit polls can provide early insight into the outcome of an election. But they cover far more ground than just the question “Who did you vote for?”Exit poll questions range from big-picture issues like the economy to state-specific topics like a candidate’s experience. Where pre-election polls provide a snapshot of where a race stands leading up to an election, exit polls enable journalists to unpack the “why” underlying the election outcome in real time.
它们能确定驱动选民的关键议题，以及不同选民群体倾向的候选人。例如，它们能告诉我们，在2026年中期选举中投出选票的人是受经济还是伊朗战争驱动，以及女性、拉美裔或年轻选民的表现是否与以往选举有所不同。
They identify the key issues driving voters, and which candidates different segments of the electorate favor. They can tell us, for instance, if those who cast ballots in the 2026 midterms were motivated by the economy or the Iran war, and whether women, Latinos or young voters behaved differently than in previous elections.
为了生产新一代出口民调数据，MS NOW正与其他多家新闻机构合作，对人们的投票决定进行广泛评估，该项目名为“选民民调”。在选举之夜，我们的选举团队将处理来自全美50个州的代表性数据，了解人们投给了谁，以及他们在投票时考虑了什么。
To produce the next generation of exit poll data, MS NOW is collaborating with several other news organizations in an extensive evaluation of people’s voting decisions in a project we call The Voter Poll. On election night, our elections team will be processing representative data from all 50 states about who people voted for and what was on their mind as they cast their ballot.
我们计划在实体投票站、网络和电话上对约10万名选民进行民调，从而获得超过7.5万名受访者的全国样本，以及全美50个州的样本。
We plan to poll approximately 100,000 voters at physical polling locations, online and by phone, providing us a national sample of more than 75,000 respondents, as well as samples in all 50 states.
选民投票数据将在选举日分批陆续到达，并实时处理。结果随后将在选举之夜分阶段发布。最早可能在选举日美国东部时间下午 5 点看到任何选民投票估算。首批发布可能包含选民特征和新出现的观点，但不会包含指示某场特定竞选结果的问题。重要的是，在所有投票站关闭前，随着更多数据的到来，结果可能会略有变动。
The Voter Poll data will arrive in waves throughout Election Day and is processed in real time. The results are then released in stages on election night. The earliest you might see any Voter Poll estimates is 5 p.m. ET on Election Day. That initial release may include the characteristics and emerging views of the electorate, but it will not include questions that indicate the outcome of a particular race. And importantly, results can shift slightly as more data comes in before all the polls close.
有关选民选择哪位候选人的问题，要等到某州预定投票结束时间到达且投票已结束后才会发布。例如，如果您想了解不同群体的得克萨斯州选民在民主党人詹姆斯·塔拉里科和共和党人肯·帕克斯顿的参议院竞选中如何投票，请留意美国东部时间晚上 9 点发布的选民投票结果，那时得克萨斯州全州投票站已关闭。该项目的运作方式如下：选民投票是由七家新闻机构——MS NOW、ABC、CBS、CNN、福克斯新闻、NBC 新闻和美联社——以及民意调查研究公司 SSRS 合作开展的。
Questions about which candidates voters chose will not be reported until polls are scheduled to close in a state and voting has ended. For example, if you want to see how different groups of Texans voted in the Senate race between Democrat James Talarico and Republican Ken Paxton, look out for The Voter Poll results at 9 p.m. ET, when polls are closed statewide in Texas. Here’s how the project works: The Voter Poll is a collaboration between seven news organizations — MS NOW, ABC, CBS, CNN, Fox News, NBC News and The Associated Press — and the survey research firm SSRS.
选举前，各机构讨论方法论问题，并决定在每个州采用的数据抽样方案和调查问题。SSRS 是一家独立、无党派的研究机构，负责实施民调，包括从全美 50 个州经核实的具有代表性的选民样本中收集和处理数据。
Before the election, the organizations discuss methodological issues and decide on the data sampling plan and survey questions to ask in each state. SSRS is an independent, nonpartisan research organization responsible for conducting the poll, including collecting and processing data from a representative sample of verified voters across all 50 states.
选民投票不仅询问受访者选择了哪位候选人，还询问他们的决策过程，包括投票理由以及何时决定支持谁。所有地区的选民都会被问及是什么驱动他们在联邦众议员选举中的投票，特定州的选民则会被问及联邦参议员和州长竞选，甚至某些投票议题。
The Voter Poll asks respondents not only which candidates they chose, but also about their decision-making process, including the reasons behind their vote and when they decided whom to support. Voters in all locations are asked about what drove their vote for the U.S. House, while voters in specific states are asked about their U.S. Senate and gubernatorial races, and even some ballot measures.
最初的回答是更长调查的一部分。选民还会被问及对政府及其他重要议题的看法，如生活成本、移民和外交政策。他们还将被问及一些个人背景问题，包括年龄、种族、性别和教育程度。这些人口统计指标使我们既能考察不同群体的投票情况，又能确保基础数据能代表整个选民群体。
The initial answers are part of a longer survey. Voters are also asked for their views on government and other salient issues like the cost of living, immigration and foreign policy. They will also be asked some questions about their personal background, including their age, race, gender and education level. These demographic measurements enable us to both examine how different groups voted and ensure that the underlying data is representative of the entire electorate.
在几个关键州，“选民调查”紧贴传统出口民调的做法。访问员将驻守在投票站外，请选民在投完票后填写一份简短、保密的问卷。并非所有人都会接受调查。为公平、一致地选定受访者，并确保样本具有随机性，访问员会在一定数量的人投完票后调查一人（例如每隔五人调查一位）。
In several key states, The Voter Poll closely mirrors a traditional exit poll. Interviewers will be stationed outside polling locations, and they will ask voters to fill out a short, confidential survey after they cast their vote. Not everyone is surveyed. To be fair and consistent in who is selected to participate, and ensure the sample is random, interviewers survey someone after a set number of people have cast their ballots (for example, every fifth person).
随着提前投票人数增多，民调方法也随之演变，以便在选民所在之处接触他们。这意味着选民现在会通过电话、邮件和电子邮件被联系，并可选择在线作答或通过电话与现场访问员作答。目标是通过多种方式接触选民，并给他们提供一个舒适的作答选项。无论选民选择哪种方式参与，其回答都将保持匿名。
As more people vote early, the poll methodology has evolved to meet them where they are. That means voters are now contacted by phone, mail and email, with options to take the survey online or on the phone with a live interviewer. The goal is to reach voters in a variety of ways and give them a comfortable option for taking the survey. Whichever way voters opt to participate, their answers are kept anonymous.
对于已经投过票的人，经核对州政府记录确认后，其候选人偏好将相应标记。尚未投出选票的人将在投票结束前几小时收到一份后续调查，以确定他们实际投给了谁以及如何投票。
For those who will have already voted, which is confirmed by reviewing state records, their candidate preferences are marked accordingly. Those who haven’t yet cast a ballot will receive a follow-up survey in the hours just before polls close to determine who they actually voted for and how they cast their ballot.
现场受访者经核实为投票站的选民，而提前投票者则通过将受访者与州选举办公室维护的选民档案进行匹配来识别。尽管我们无法亲眼看到这些人离开投票站，但我们可以通过与州政府官员的信息交叉核对，验证他们已投出选票。被联系但无法投票的人，例如未满18岁者或非公民，将无法参与调查。
Respondents interviewed in person are verified to be voters at the polling place, while advance voters are identified by matching respondents to voter files maintained by state election offices. Even though we can’t physically see these people leave the polls, we can verify that they have cast their ballot by cross-checking their information from state officials. People contacted who cannot vote, such as those under age 18 or people who are not citizens, will not be able to take the survey.
《选民民调》经过精心设计，旨在准确估算全国及各州的美国选民情况。数据通过上述多模式设计收集，随后利用关于选民的最佳可用信息进行处理，包括选民登记档案和选举日投票率信息。这些复杂的方法建立在调查研究的最佳实践基础上，由该领域的顶尖专家开发。
The Voter Poll has been carefully designed to accurately estimate the American electorate nationally and in each state. The data is collected through the multimodal design explained above, and is then processed using the best available information about the electorate, including voter registration files and Election Day turnout information. The sophisticated methods are founded on best practices in survey research and developed by leading experts in the field.
也就是说，《选民民调》附带误差范围，因为调查民众的过程固有不确定性。其报告的数字均为估算值。在选举夜早期、尚不了解实际选举结果之前，这些数字可能与最终结果略有出入。最终，各州数据将依据该州的选举结果进行加权，以使数据更加准确。
That said, The Voter Poll comes with margins of error, as there is some uncertainty inherent in the process of surveying people. The numbers reported from it are estimates. Early in the night, before we know the actual results of the election, the numbers may differ slightly from the final results. Eventually, each state’s data is weighted to the election results there to make the data more accurate.