New Delhi: Democratic presidential nominee Joe Biden has a 3.3 percentage point lead over US President Donald Trump across six key swing states, according to latest data from Chicago-based polling data aggregator, Real Clear Politics (RCP).
These states include Wisconsin, Michigan, Pennsylvania, North Carolina, Florida and Arizona. Based on an average of local polls in these states, Biden has 49.1 per cent chance of winning and Trump 45.8 per cent chance, according to the RCP data published Monday.
However, a candidate does not need a national vote to win the US election, but rather 270 electoral votes that can be won across separate state-to-state contests, which is why there is more focus on polls in swing states.
According to the RCP, Biden has a solid 118 electoral votes in his pocket and Trump has 57, which leaves 95 either in the “leaning” or “toss-up” categories.
Despite Biden’s lead in swing states, there is usually a margin of error of plus or minus 3 percentage points or even double in some cases, suggesting that the race could be more narrow than expected.
Last week, two prognosticators, Inside Elections and Cook Political Report, unexpectedly shifted Texas from a ‘leaning Republican’ state to the toss-up category with speculations that it could go Democratic for the first time since 1976.
Despite Trump’s performance in recent polls, a Gallup survey conducted from 16-27 October found his approval rating to be roughly the same as about four years ago, in fact one percentage point higher.
The RCP has similar findings for the president’s ‘job approval‘, which is 45.3 per cent now, and was 43.8 per cent in January 2017.
Also read: Biden leads nationally, in key states but Trump still has a ‘narrow but viable’ path to win
Where Trump, Biden are safe and where they aren’t
At the start of October, around the time Trump contracted Covid-19, Biden was leading by 3.7 percentage points across the six swing states, according to the RCP. His lead rose to 5 percentage points on 13 October, hovered around 4.5 for a while and then started to decline a week ago.
Though Biden is maintaining his lead in states like Wisconsin, Michigan and Pennsylvania, the race is tighter in Florida and North Carolina.
A Washington Post-ABC poll published Sunday found Trump leading by 2 percentage points in Florida, the iconic “bellwether” state known for backing the winner of every presidential race since 1964 except on one occasion in 1992 where it sided with incumbent President George Bush instead of the presidential winner that year, Bill Clinton.
The underreported Trump vote
Though pollsters correctly predicted the national popular vote in 2016, they were off in upper mid-West states where Trump won by razor-thin margins.
For example, polls showed Clinton winning Florida by about 1 percentage point, North Carolina by 3 percentage points, Pennsylvania by 5 points and Wisconsin by 6-7 points. But Trump ended up winning in all these states by 1.2, 3.6, 0.7 and 0.7 percentage points, respectively.
Doug Schwartz, director and vice-president of Quinnipiac University Poll, argued that “late deciders” who ended up supporting Trump could have contributed to this but more importantly, pollsters underestimated white voters without a college degree.
“It’s significant because these voters supported Trump and they were undercounted … in important states like Wisconsin, Michigan, and Pennsylvania, where Trump won by a very narrow margin,” said Schwartz in a briefing last month.
Pollsters have since attempted to correct methods that underestimated Trump’s support in 2016.
Daron Shaw, co-director of various polls like Fox News Poll and the University of Texas/Texas Tribune Poll, named three ways pollsters are trying to counter the “shy Trump” phenomenon in which a person is unlikely to admit they are voting for Trump as they perceive it as socially undesirable.
First, they ask respondents how they think people in their neighbourhood or community are going to vote. “It’s possible for people to say: ‘Well I’m not going to vote for Trump but yeah, everyone around me is going to’,” said Shaw during a briefing last week, adding that some “paranoid” Democrats may also have a similar line of thought.
Second, in order to identify response bias, pollsters compare party registrations and poll responses, said Shaw.
Since people register by party in many states, a sample could show 45 per cent Republican, 45 per cent Democrat and 10 per cent ‘declined to state’. A response bias is clear if polling shows 65 per cent Democrat and 35 per cent Republican. “But we haven’t seen much of that,” added Shaw.
Finally, they ask people how they voted in 2016, which has a strong recall measure. “It’s possible that somebody would admit they voted for Trump in 2016 and not in 2020,” said Shaw.
Also read: Three words that haunt Joe Biden — ‘Dewey Defeats Truman’