https://projects.fivethirtyeight.com/2024-election-forecast/
Those polls — most notably from Emerson College, Marist College and The New York Times, three of the most prolific pollsters in America — show Harris and former President Donald Trump tied in the northern battlegrounds and Harris trailing heavily in Florida. But our forecast came to a different conclusion:
这些民调——最著名的来自爱默生学院、玛丽斯特学院和《纽约时报》，它们是美国最活跃的三大民调机构——显示哈里斯与前总统唐纳德·特朗普在北方的关键战场持平，而哈里斯在佛罗里达州大幅落后。但我们的预测得出了不同的结论：
不论Emerson民调结果如何，算哈领先一分。
The polls from Emerson in particular have tended to underestimate Harris compared to polls with the same methodology (a mix of robocalls and online interviews) in the same geographies. This means our starting point when looking at these new polls is that Harris is doing better than they indicate, by about a point on vote margin, and that she is either holding steady or gaining ground.
尤其是爱默生学院的民调，往往低估了哈里斯的表现，与其他采用相同方法（混合使用自动电话和在线访谈）的民调相比，在相同地区结果有所偏差。这意味着我们在解读这些新民调时的起点是，哈里斯的表现比数据所显示的要好，投票差距大约领先一个百分点，而且她的支持率要么保持稳定，要么有所提升。
给川普去掉一个最高分。什么，纽约时报的？纽约时报的也得扣除。
But The New York Times/Siena College poll in Florida should worry Democrats. It single-handedly moved the 538 poll average in the state from Trump+4 to Trump +5, and could indicate broader weakness of pollsters to sample Republican voters properly. However, The Times’s surveys in Texas and Montana came in right on target with our forecast, prompting our model to think the race is more of the same than worth a dramatic shift in estimates. So this is a good place to remind you: More of the same in forecasting today is still forecasting a toss-up race.
但《纽约时报》/锡耶纳学院在佛罗里达州的民调应该引起民主党的担忧。它单独将538网站上佛州的民调平均值从特朗普领先4个百分点推高至领先5个百分点，这可能表明民调机构在正确抽样共和党选民方面存在更广泛的问题。然而，《纽约时报》在得克萨斯州和蒙大拿州的调查结果却与我们的预测非常吻合，这使得我们的模型认为这场竞选的整体态势没有发生显著变化。因此，这是提醒你的好时机：今天的预测仍然显示竞选结果难分胜负。
