政治 · 中期选举民调持续报道 · 第 2 篇,共 2 篇
经济不满情绪强烈,民主党在众议院民调中领先8个百分点
ABC/Post/Ipsos调查显示,特朗普支持率为33%,经济不满情绪普遍。民主党在全国范围内的优势对共和党构成警讯,但并非众议院控制权的预测。
原语言视频

文字实录 · 正在加载
译自英文原文。
持续报道第 61 天
2 篇报道47 家媒体76 个来源3 个视频
完整报道自上次报道以来的变化
60 天后
当时
临近中期选举,特朗普支持率降至33%当时尚无定论
- 现有材料中没有经核实的样本量、误差范围、加权方法或完整人口统计分类。
- 没有经核实的来源支持47%的任期初支持率、下滑14个百分点、71%的共和党人预期数据或一般选票倾向领先5个百分点的说法。
我们说过将关注的
- 民主党在经济信任度上的优势是否会延续到11月的竞选期间。
- 汽油价格和对战争长期化的预期会促使支持率进一步下降,还是使其趋于稳定。
现在
ABC/Post/Ipsos调查显示,特朗普支持率为33%,经济不满情绪普遍。民主党在全国范围内的优势对共和党构成警讯,但并非众议院控制权的预测。
一项最新公布的ABC News/Washington Post/Ipsos民调显示,共和党人面临的局面已变得多么艰难。在可能投票的选民中,民主党在全国众议院投票意向上以52%对44%领先共和党;唐纳德·特朗普总统的工作表现支持率在成年人中为33%。这项于10月18日报道的调查结果显示,民主党在受访者表达的意向上优势显著,但并不能决定众议院将由谁掌控。1 22 28
关键事实
不满情绪并不只针对总统。中心调查发现,72%的人认为国家正朝着错误方向发展,约四分之三的人对经济持负面看法,65%的人认为对伊朗动武是个错误。这些结果使共和党在全国范围内处于守势,但不能证明每一位不满的选民都会选择民主党,也不能证明反对军事行动导致了众议院投票意向差距。1
33%的支持率也需要结合其历史背景来看。8月19日,HeyDay News报道了另一项Reuters/Ipsos调查,该调查前一天公布,显示特朗普在成年人中的支持率为33%——在那项调查中,这是他本届任期内当时的最低值。10月的相同数字来自另一项调查。这是一次新的测量,并不能证明自8月以来民意没有变化。这里的新进展,是10月的众议院投票意向结果与总统支持率再次处于低位同时出现。
全国领先幅度说明了什么
民主党领先8个百分点是全国层面的衡量结果,不是选区结果,也不是席位预测。众议院选举有各自的候选人和选民。支持者居住在哪里、是否投票,决定了全国优势会如何转化为国会权力。这项民调让民主党有理由对政治环境感到乐观,但并不意味着多数席位已属民主党。
该调查通过基于概率抽样的Ipsos KnowledgePanel,于10月9日至10月13日采访了2,545名美国成年人,其中包括2,032名登记选民。为得出可能投票选民的结果,研究人员为登记选民分配投票概率,让被认为更可能参与投票的人具有更大影响力。现有方法说明没有提供单独的可能投票受访者人数。19 20
这一区别很重要。众议院投票意向结果针对的是经过模型估算的可能投票选民;总统支持率结果针对的是成年人。两者衡量的是不同群体的不同判断。《今日美国》对同一项调查的报道显示,成年人中有66%不支持特朗普,33%支持。ABC和《华盛顿邮报》是同一项调查的赞助方,并非两项相互印证的独立民调。22 28 6 5
公布的抽样误差范围为:成年人正负2个百分点;登记选民和可能投票选民正负2.2个百分点。向半数样本提问的问题涉及1,271至1,274名成年人,误差范围为2.8个百分点。这些数值包含加权效应,但现有方法说明没有指出哪些相关问题使用了半数样本,也没有说明这些误差范围对应的置信水平。19 20
某个百分比的误差范围,并不自动等于两党差距的不确定性,也未涵盖无回应、措辞或投票率模型可能造成的所有误差。支持本文叙述的发布摘录中,没有原始众议院结果表和确切的问题措辞;详细的加权程序及各题的具体样本基数也仍不清楚。应当认真看待报告中的领先幅度,但不应为其赋予尚未得到证实的统计显著性检验结果。19 20 22
不同民调,类似警讯
据NBC和《今日美国》报道,另一项由NBC News Decision Desk与SurveyMonkey合作开展的调查发现,71%的成年人认为国家正走在错误的轨道上。这进一步印证了普遍悲观的景象,但这并不意味着中心调查中的72%需要更正,也不表明这一比例随时间下降了一个百分点。7 8
Political Wire报道称,NBC调查显示,成年人中有34%支持、66%不支持特朗普;在登记选民中,民主党的国会选举意向为55%,共和党为45%。这些数字仍是经二手报道的结果:此处无法获取完整的原始结果、调查日期、样本量和问卷。报道还为成年人和登记选民的结果分别给出了不同的不确定性数值,但没有说明其计算依据。31 33
福克斯的一项调查呈现了另一种不同结果:登记选民中,36%支持、64%不支持特朗普。该调查于10月9日至10月12日采访了1,202名登记选民,这些受访者从全国选民名册中随机抽取。Beacon Research和Shaw & Company Research进行了114次固定电话采访、786次手机采访,并通过短信链接收集了302份在线答复。其登记选民总体不同,这也是不应把这些支持率读作同一趋势中连续变化的原因。18 50 49
福克斯称其支持率结果创下历史新低,《国会山报》则称这是特朗普任一任期中的最低值;但要独立核实这一说法所需的完整历史数据表无法获取。同样,Axios所称民主党在全国民调中领先8至10个百分点,是对多项民调的综合报道,并非汇总估值或席位模型。18 50 4
另一个数字应单独看待:NBC全国国会选举民调平均值显示,民主党为51%,共和党为43%,最近一次更新是在10月14日。平均值综合了多项民调,既不是NBC调查报告的55%–45%,也不是中心调查的52%–44%。这些数据指向相似的政治趋势,但不能互换使用。36
经济愤怒遭遇不确定的时间表
MS NOW报道称,自2014年以来,登记选民首次表示,相比共和党,他们更信任民主党管理经济。这在政治上可能意义重大,但现有证据无法确认两党的具体百分比、差距大小和历史比较。对经济不满,与相信某个政党有能力解决问题,是两种不同判断;两者不能互相替代。2
财政部长斯科特·贝森特称,特朗普就任时通胀率为3%,他接受采访时为3.7%。现有报道没有说明这一比较所依据的通胀指数或参考月份,因此这些数字应继续归于贝森特本人,而不应作为特定官方序列中经核实的比较来呈现。即使通胀率降低,也意味着物价涨得较慢,并不一定意味着价格会回落到家庭记忆中的水平。2
贝森特还将能源价格缓解与战争结束联系起来,并表示:“我们会迎来转机。”当被问及价格是否会在新年之前下降时,他说:“这取决于伊朗冲突的进展。”条件不是期限。民调衡量的是当下的不满;政府的回答则取决于尚未给出时间表的事态发展。56 59
目前,证据支持一个明确但有限的结论:民主党在全国投票意向上拥有显著优势,共和党则面对对当前状况深感不满的选民。最终这是否会转化为众议院多数席位,仍要由选民在一个个选区中作出决定。
播出原貌 26 行
- On August 19, we reported Trump's approval at 33 percent. Today's survey adds an eight-point Democratic House preference lead.
- Republicans face broad dissatisfaction with the economy and the country's direction, sixteen days before the election. But how much can that national advantage tell us about who will actually control the House?
- Democrats draw 52 percent in the national House preference measure, and Republicans draw 44 percent. That eight-percentage-point gap comes from likely voters, with registered voters weighted by their estimated chances of casting a ballot.
- The survey interviewed 2,545 adults from October 9 through October 13. It included 2,032 registered voters. The likely-voter calculation uses those registered voters and their chances of voting, rather than treating every adult as equally likely to participate.
- President Donald Trump's job approval stands at 33 percent among adults, with 66 percent disapproving. That measures a broader population than the likely voters behind the House result. The two findings describe different judgments by different groups.
- The matching approval number from August doesn't tell us opinion stayed unchanged. That earlier survey found the lowest approval of Trump's current term within its own series. Today's figure comes from a different survey, so the two readings can't trace a continuous trend.
- The poll used a probability-based survey panel. Its likely-voter model gives more influence to registered voters judged more likely to vote. So turnout assumptions help shape the House preference result, alongside the answers people give.
- The poll's two news sponsors published findings from the same survey. Counting their coverage separately would double-count the evidence. And a national preference question captures party choice across the country, while each House contest has its own candidates and electorate.
- The published sampling error is plus or minus two points for adults, and 2.2 points for registered and likely voters. Half-sample questions carry a 2.8-point margin. Those margins include the effects of weighting, which adjusts how much different responses count.
- A percentage's error margin doesn't automatically give the uncertainty around the gap between two parties. Sampling margins also leave out possible errors from question wording, people declining to respond, or the turnout model. The exact House question and question-by-question sample details aren't clear yet.
- Individual district elections will decide House control. A national eight-point preference advantage can't be converted directly into seats. The voters who actually turn out, and where they live, determine how national sentiment translates into congressional power.
- The survey finds 72 percent saying the country is headed in the wrong direction. About 75 percent view the economy negatively. And 65 percent say using force against Iran was a mistake.
- Those findings put Republicans in a difficult national environment. They don't tell us every unhappy voter will choose a Democrat. And people's judgments about the economy and military action don't, by themselves, explain why they chose a party in the House question.
- A separate national survey finds 71 percent of adults saying the country is on the wrong track. It reportedly puts approval at 34 percent and disapproval at 66 percent, with registered voters favoring Democrats for Congress by 55 to 45.
- The complete results and methods for that separate survey aren't available yet. Its reported uncertainty figures also lack a clear explanation of how they were calculated. And the difference between 71 and 72 percent across two surveys doesn't show that opinion moved.
- A separate national polling average showed Democrats at 51 percent and Republicans at 43 percent on October 14. That average combines polls. It belongs alongside the individual surveys as a different product, with potentially different questions and populations.
- Another survey puts Trump's approval at 36 percent and disapproval at 64 percent among registered voters. It interviewed 1,202 people from October 9 through October 12, drawing them randomly from a national voter file.
- Two research firms conducted that poll using 114 landline interviews, 786 cellphone interviews and 302 online responses through text-message links. Its registered-voter population differs from the central survey's adult approval measure, another reason the percentages shouldn't be read as successive steps in one trend.
- The national preference results point in a similar political direction. A broader description of Democratic leads ranging from eight to ten points remains a reported synthesis. It doesn't supply a combined estimate or a model of how many seats either party will win.
- Registered voters reportedly trust Democrats more to manage the economy, described as their first such advantage since 2014. The size of that lead and the historical comparison are still unclear. Disliking economic conditions and trusting a party to improve them are separate judgments.
- Treasury Secretary Scott Bessent says inflation was three percent when Trump took office and 3.7 percent at the time of his interview. He hasn't clearly identified the inflation measure or reference month behind that comparison.
- Even a lower inflation rate would mean prices rising more slowly. Households could still be paying more than they did before those increases. The administration's explanation for fuel costs links relief to the end of the Iran conflict.
- Vice President JD Vance says gasoline prices will fall when the conflict ends. He also blames part of today's high prices on refining capacity that went offline during Joe Biden's presidency. That explanation comes without a firm date for cheaper gasoline.
- Bessent also ties lower energy prices to the war's end, saying, “we will get to the other side.” Asked about prices falling by the new year, he says that “will depend on what happens with the Iran conflict.”
- Neither official has given households a firm timetable for relief. The polling captures dissatisfaction now, while the administration's expectation of lower fuel costs depends on how the conflict develops.
- Democrats hold a substantial national preference advantage, but district-level voting will decide whether they gain control of the House. The general election is November 3. Turnout and the results of those individual contests will determine who has the votes to govern.
来源
65 个来源40 家媒体2 个视频1 篇通稿
来源地图
追溯本篇报道的依据。图谱展示其关联;下方链接可打开本文引用的来源。
来源图谱将在您阅读时加载。
译文更新于
修订记录
- r1首次发布。

