With the national conventions (such as they were) in the rearview mirror, a sizable batch of initial polling shows a presidential contest where Joe Biden is still in the lead by roughly the same margin as before. Because of the timing and the close proximity of the two conventions, it’s unclear whether both parties both got offsetting “bounces” or if the conventions essentially had no impact on the race.
In the RealClearPolitics polling averages, Biden’s lead stood at 7.7 percent the day before the DNC began (August 16), and is at 7.5 percent today. According to FiveThirtyEight, which weighs polls for accuracy and adjusts them for partisan bias, Biden’s lead was 8.0 percent on August 16 and is at 7.4 percent now.
Trump fans can cherry-pick one national poll taken since the conventions to claim Trump has all but eliminated Biden’s lead: an Emerson College survey showing Biden up 49/47 among likely voters (the same outlet gave Biden an unusually low four-point lead in July). But it looks like an outlier, particularly since it showed Trump’s job-approval rating at a sky-high 49 percent. Other polls with post-convention data include the gold standard Selzer poll (for Grinnell College), which has Biden up 49/41 among likely voters; a Suffolk/USA Today poll of registered voters showing Biden up 50/43; an IDB/TIPP survey giving Biden an identical 49/41 lead among a sample of registered voters; an Economist/YouGov survey showing Biden up 51/40 among registered voters; and the first national likely voter survey from Quinnipiac, which gives Biden a 52/42 lead.
Trump’s average job-approval rating is at 44.1 percent at RealClearPolitics and 43.3 percent at FiveThirtyEight, a bit higher than it has typically been since COVID-19 really kicked in, but within the narrow band his popularity has exhibited, uniquely, throughout his presidency.
Keep in mind that most analysts continue to believe Trump will do better in key battleground states than his national numbers suggest, which means he could again lose the popular vote and win the electoral college. This estimate from Nate Silver suggests his advantage amounts to 2–3 percentage points:
Speaking of battleground states, the best polling news today for Team Trump comes from a Pennsylvania survey by the high-quality pollster Monmouth, which showed Biden with a double-digit lead in the Keystone State in mid-July:
Among all registered voters in Pennsylvania, the race for president stands at 49% for Biden and 45% for Trump. Another 2% support Libertarian Jo Jorgensen, less than 1% back the Green Party’s Howie Hawkins, and 4% are undecided…
The contest tightens when different likely voter models are applied. A model based on a somewhat higher level of turnout than 2016 puts the race at 49% for Biden and 46% for Trump, while one reflecting lower turnout has it at 48% for Biden and 47% for Trump.
It looks like Trump made a good investment when he campaigned in Scranton during the DNC:
Trump has increased his support in ten counties where the vote margins were closest in the 2016 presidential election. In these swing counties*, which are concentrated in a swath that runs from the Philadelphia suburbs into the northeast region of the commonwealth, the race stands at 46% for Trump and 44% for Biden. Just over six weeks ago, Biden had a sizable 54% to 35% lead among voters in this key county grouping.
The national polling contains one particular warning sign for Trump, aside from Biden’s steady lead: The Democrat’s support level is close to a majority in nearly every survey. And at the same time, Trump’s “very unfavorable” ratings — which indicate voters who have tuned him out — are unusually high (at 49 percent in the Quinnipiac survey (compared to 35 percent for Biden), and 45 percent at Economist/You Gov (compared to 33 percent for Biden)). So Trump has quite a bit of ground to make up and a fairly narrow group of persuadable voters. On this day in 2016, he trailed Hillary Clinton by 3.9 percent in the RCP polling averages, which was only 1.8 percent greater than her ultimate margin in the national popular vote.