在乔·拜登总统7月21日退出总统竞选之前，民意测验专家已经注意到一个一致的趋势:黑人选民不太稳固在民主阵营比最近的选举，包括2020年。共和党注意到了这一点:在共和党全国代表大会期间，像众议员。约翰·詹姆斯和玛德琳·布拉姆为吸引这一关键群体做出了特别努力。

　　在几个州，黑人选民至少占总人口的10%最有可能决定2024年大选的州—从宾夕法尼亚到密歇根到佛罗里达。这些问题在乔治亚州和北卡罗来纳州也很突出，这两个州在最近的过去一直很接近。因此，即使美国黑人对民主党的支持发生微小变化，也可能让关键州出局。

　　副总统卡玛拉·哈里斯的候选人资格赢回黑人选民的潜力当拜登被提名时，他远离了民主党。但是这些潜在的新的黑人摇摆选民到底是谁呢?这个问题很难回答，部分原因是民意调查很少针对多样化的人群进行优化比加拿大的总人口还多。然而，通过比较2008年至2024年的高质量民调，我们可以看到年龄已经成为黑人选民支持民主党的一个主要因素——年轻的黑人选民不再坚定地支持民主党。

　　拜登失去了多少黑人支持?

　　第一位黑人总统巴拉克·奥巴马获胜不足为奇超过90%他在2012年竞选连任时赢得了黑人选民的支持黑人投票率很高，也是。因此，为了在前总统唐纳德·特朗普(Donald Trump)出现之前获得黑人选民对白人民主党人的支持，我们必须回顾2008年的初选，当时的参议员希拉里·克林顿(Hillary Clinton)是民主党提名的领先者之一。

　　那年春天，在提名确定之前，民意调查公司Knowledge Networks使用基于人口的方法——在线调查受访者，然后在线招募他们——为全国安纳伯格选举调查。仅在2008年春季的一波调查中，就有9888名样本，这项调查有一个大样本(962名)的黑人受访者。总体而言，75%的成年黑人表示，他们将在假设的大选中投票给克林顿，与当时的参议员约翰·麦凯恩(John McCain)对决，而后者在成年黑人中的支持率仅为8%。这是67个百分点的差距。

　　然而，在2024年初，这些数字看起来完全不同。今年2月和3月，政治学家Gall Sigler和我与NORC合作，对2024年的选举进行了一项调查，其中包括834名黑人成年人，我们发现拜登在这些受访者中以61%至20%的微弱优势领先，差距仅为41个百分点。(那离三月/四月不远了YouGov对2004名黑人或非裔美国人的调查显示作为全国黑人选民项目的一部分，拜登的支持率上升了62%至15%。)可以肯定的是，这仍然是一个巨大的亲民主倾向。但是这比民主党人最近享受的要少。

　　年轻的美国黑人现在可能是摇摆不定的选民

　　2008年到2024年之间发生了什么变化?按年龄来看黑人的支持度就能找到答案。正如你在下面的图表中看到的，在2008年，各年龄段的黑人对克林顿和麦凯恩的支持相当一致。在所有年龄段的美国黑人中，克林顿的支持率在67%到84%之间。

　　但到2024年，美国黑人之间出现了明显的年龄差距，至少在考虑拜登和特朗普的对决时是如此。年龄较大的黑人选民仍然是压倒性的民主党人:例如，65-74岁的黑人受访者支持拜登70个百分点。但正如你在下一张图表中看到的，年轻的黑人选民对拜登的支持率要低得多。事实上，25岁至34岁的黑人选民对拜登和特朗普的支持大致相当。

　　为什么年轻的美国黑人不那么支持拜登?

　　当然，关键因素可能不是年龄本身，而是与年龄相关的东西，比如投票倾向。回到4月，使用相同的数据集，我发现特朗普表现得更好，而拜登表现得更差不经常参加投票的黑人选民。在2018年、2020年和2022年投票的黑人选民中，拜登对特朗普的支持率为81个百分点，而在这三次选举中均未投票的选民中，拜登的支持率仅为10个百分点。

　　年轻选民往往不那么一致，所以也许拜登在年轻选民中的表现不佳对于不经常投票的选民来说确实是一个弱点。但是投票历史并不是故事的全部。当我们拟合一个线性回归模型来同时解释先前投票率和选民偏好之间的关系时，黑人选民中基于年龄的选民偏好差异减少了，但没有消除。

　　另一个因素可能是黑人选民的个人政治经历。理论上，民权时代的年轻黑人选民可能会特别忠于民主党，因为该党在通过1965年《选举权法》等里程碑式的立法中发挥了作用他们最容易受影响。但如果这是真的，我们可能会看到70多岁的人不连续，因为今天77岁的人在1965年是18岁。相反，年龄组的实际趋势看起来更像是一条直线，表明其他过程也在起作用。

　　事实上，还有其他因素也随着年龄的变化而变化，这也许可以解释我们所看到的年龄差距。30年前，政治学家麦可·道森出版了《论在骡子后面该书认为，黑人选民是一个异常团结的投票群体，因为有一种叫做“命运相连”的东西。“在很大程度上，道森写道，美国黑人认为他们自己的机会与黑人社区的机会息息相关，所以他们以压倒性优势投票给他们认为对该社区最有利的政党。

　　在2024年初的调查中，我们通过询问黑人受访者来衡量这一想法，“你认为这个国家的黑人普遍发生的事情与你生活中发生的事情有关系吗?”总的来说，我们的受访者倾向于同意这一观点，75%的人表示，黑人的普遍遭遇与他们自己的生活有“很多”或“一些”关系。但这里也存在年龄差距，尽管差距较小:55-64岁的黑人受访者对命运相关的感觉最高(79%)，而18-24岁的受访者最低(60%)。因此，年轻黑人选民对命运联系的看法减弱有助于解释总统支持率的年龄差距，这似乎是合理的。

　　但是，让我们精确一下这种关系的大小。根据线性回归模型，随着受访者从最低的联系命运感(“完全没有”)转向最高的联系命运感(“很多”)，他们对拜登的支持度在0-1的范围内增加了约0.11。如此紧密相连的命运是解释的一部分，但也不是全部。即使我们在统计模型中考虑了相关的命运，年龄差距仍然是有意义的。*

　　这对秋季竞选意味着什么

　　多年来，民主党竞选团队投入了大量时间和精力来动员黑人选民。今年也肯定会是这种情况，因为即使拜登水平的表现保持不变，民主党人仍将受益于高黑人投票率——但一些快速计算表明，与前几年相比，这些好处可能会受到抑制。

　　如果黑人选民作为一个整体将超过90%的选票投给了民主党，那么100名新的黑人选民将导致80张民主党选票的净收益——不过，这是假设民主党在黑人选民中获胜的几率非常高，而且只是偶尔投票的黑人选民与那些持续投票的黑人选民拥有相同的支持水平。但是，如果那些本来不会投票的黑人选民只支持民主党以47%对29%的优势胜出，100名新的黑人选民平均只为民主党带来18张选票。因此，如果民主党失去了年轻黑人选民或参与度较低的黑人选民的支持，他们可能会发现，他们的投票努力不会像过去几年那样取得同样的成果。

　　Young Black voters might be swing voters now

　　Before President Joe Biden exited the presidential race on July 21, pollsters had noticed aconsistent trend: Black voterswere less solidlyinthe Democratic campthan in recent elections, including 2020. The Republican Party took notice: During the Republican National Convention, speakers like Rep.John JamesandMadeline Bramemade special efforts to appeal to this key constituency.

　　Black voters are at least 10 percent of the population in several of thestates most likely to decide the 2024 election— from Pennsylvania to Michigan to Florida. They also loom large in Georgia and North Carolina, states that have been close in the recent past. So even small changes in Black Americans' support for Democrats could take key states off the table.

　　Vice President Kamala Harris's candidacy has thepotential to win back Black voterswho shifted away from Democrats when Biden was the nominee. But who exactly are these potential new Black swing voters? That's a hard question to answer, in part becausepolls are rarely optimized to target a demographic that is diverseand larger than the entire population of Canada. By comparing high-quality polls from 2008 to 2024, though, we can see that age has emerged as a major factor in Black voters' support for Democrats — with younger Black voters no longer as solidly in Democrats' camp.

　　How much Black support had Biden lost?

　　It's unsurprising that the first Black president, Barack Obama, wonmore than 90 percentof Black voters during his own reelection bid in 2012 and enjoyedstrong Black turnout, too. So, to get a baseline of Black voters' support for a white Democrat before former President Donald Trump came on the scene, we've got to look back to the 2008 primaries, when then-Sen. Hillary Clinton was one of the front-runners for the Democratic nomination.

　　That spring, before the nomination was set, the polling firm Knowledge Networks used population-based methods — surveying respondents online after recruiting them offline — to conduct a panel survey for theNational Annenberg Election Survey. With a sample size of 9,888 in its spring 2008 wave alone, this survey had a large sample (962) of Black respondents. Overall, 75 percent of Black adults said they would vote for Clinton in a hypothetical general election matchup with then-Sen. John McCain, who got only 8 percent among Black adults. That's a 67-percentage-point gap.

　　In early 2024, however, the numbers looked quite different. Political scientist Gall Sigler and I partnered with NORC in February and March of this year to conduct a survey of the 2024 election that included 834 Black adults, and we found Biden led among these respondents by a smaller margin of 61 percent to 20 percent — a gap of only 41 points. (That wasn't far off a March/AprilYouGov poll of 2,004 Black or African American respondentsconducted as part of the National Black Voter Project, which found Biden up 62 percent to 15 percent.) To be sure, that was still a massive pro-Democratic lean. But it was less than Democrats have enjoyed in the recent past.

　　Younger Black Americans may now be swing voters

　　What changed between 2008 and 2024? A look at Black support by age reveals the answer. As you can see in the chart below, in 2008, Black support for Clinton and McCain was reasonably consistent across age groups. Clinton's support was between 67 and 84 percent among Black Americans across all age categories.

　　But by 2024, a clear age gap had opened up among Black Americans, at least when thinking about a Biden-Trump matchup. Older Black voters remained overwhelmingly Democratic: For example, Black respondents aged 65-74 supported Biden by 70 points. But as you can see in the next chart, younger Black voters favored Biden by much smaller margins. In fact, Black voters between 25 and 34 gave just about equal support to Biden and Trump.

　　Why were younger Black Americans less supportive of Biden?

　　Of course, it's possible that the key factor wasn't age itself, but rather something correlated with age, like propensity to turn out to vote. Back in April, using the same data set, I found that Trump performed much better — and Biden much worse — amongBlack voters who turned out less frequently. Biden's margin over Trump was 81 points among Black voters who voted in 2018, 2020 and 2022, while it was only 10 points among those who voted in none of those three elections.

　　Younger voters tend to turn out less consistently, so maybe Biden's underperformance with young voters was really a weakness with infrequent voters. But vote history wasn't the whole story. When we fit a linear regression model to account for the relationship between prior turnout and voter preference simultaneously, age-based differences in voter preference among Black voters were reduced but not eliminated.

　　Another factor could have been Black voters' personal experiences with politics. In theory, Black voters who were young adults during the civil-rights era might be especially loyal to the Democratic Party because of the party's role in passing landmark legislation like the 1965 Voting Rights Act whenthey were most impressionable. But if that were true, we might expect to see a discontinuity for people in their late 70s, since someone who is 77 today was 18 in 1965. Instead, the actual trend by age groups looked more like a straight line, suggesting that other processes were at work.

　　Indeed, there are other factors that also vary with age and might explain the age gap we saw. Thirty years ago, political scientist Michael Dawson published the book "Behind the Mule," which argued that Black voters are an unusually cohesive voting bloc because of something called "linked fate." To an important extent, Dawson wrote, Black Americans see their own opportunities as tethered to those of the Black community, so they vote overwhelmingly for the party they see as best for that community.

　　In the early 2024 survey, we measured this idea partly by asking Black respondents, "Do you think what happens generally to Black people in this country will have something to do with what happens in your life, or not?" Overall, our respondents tended to agree, with 75 percent saying what happens generally to Black people has "a lot" or "some" to do with their own life. But there was an age gap here too, albeit a more modest one: Black respondents aged 55-64 reported the highest sense of linked fate (79 percent), while those who were 18-24 reported the lowest (60 percent). It's plausible, then, that weaker perceptions of linked fate among younger Black voters helps explain the age gap in presidential support.

　　But let's be precise about the magnitude of the relationship. As respondents shifted from the lowest sense of linked fate ("not at all") to the highest ("a lot"), their support for Biden increased by about 0.11 on a 0-1 scale, according to a linear regression model. So linked fate was a part of the explanation, but it was not the whole story either. Even when we accounted for linked fate in statistical models, the age gap remained meaningful.*

　　What this means for the fall campaign

　　For years, Democratic campaigns have devoted a lot of time and energy specifically to mobilizing Black voters. That will surely also be the case this year since, even if the Biden-level performance holds, Democrats will still benefit from high Black turnout — but some quick math shows us that those benefits may be dampened compared to prior years.

　　If Black voters as a whole cast more than 90 percent of their ballots for the Democrats, turning out 100 new Black voters will result in a net gain of 80 Democratic votes — that's assuming, though, a very high margin of Democratic victory among Black voters, and that the Black voters who only vote occasionally have the same levels of support as those who vote consistently. But if Black voters who wouldn't have otherwise voted are only with theDemocrats by a margin of, say, 47 percent to 29 percent, turning out 100 new Black voters produces just 18 votes for Democrats on average. So if Democrats are losing the support of younger Black voters or less engaged Black voters, they may find that their get-out-the-vote efforts don't bear the same fruit as in past years.

　　Sophia Leung and Gall Sigler contributed research.