The polling is narrowing in the US presidential race. There is no doubt about that. The leads that Biden enjoyed in swing states two months ago, in some cases in the double digits, have come back and now sit either just outside or neatly within the pollsters’ margins of error.
How reliable is the polling this time around? The obvious answer to that question is we’ll know in seven weeks.
It goes without saying the polling companies got a great deal wrong in 2016. They did get the popular vote more or less right, but it was in key swing states where support for Donald Trump was understated. In Wisconsin, the polls were out by 6.5 per cent and in Pennsylvania and Michigan between four and five points.
Much has been made of a Trump polling effect, where it is said potential voters when polled won’t admit to supporting Donald Trump. The theory goes that this was one cause for underestimating Trump’s chances of victory, especially in face-to-face polling.
There is little evidence of this however, according to a post-mortem of polling in the 2016 presidential campaign conducted by the American Association of Public Opinion Research.
Rather, the AAPOR analysis points the finger at polling companies inadvertently skewing results by including too many college educated and postgraduate educated respondents in their polls.
Prior to 2016, education levels were not considered to be powerful voting determinants. Then along came Donald Trump and turned that assumption on its head.
The major polling companies now say they have rectified the imbalance and they expect their swing state polling to reflect the outcome on November 3.
At this stage of the campaign, the national polling figure has become almost trivial, something to glance at and move on.
Joe Biden will almost certainly win the popular vote as Hillary Clinton did in 2016. But it is in the swing states where the real battles will take place.
A presidential candidate needs to win 270 of the 538 electoral college votes to win the election. Almost all states are winner take all with Nebraska and Maine the exceptions. In 2016, Trump won 304 electoral college votes, including Florida’s 29.
Florida is the most crucial of all swing states. When Australians sit down on the morning of Wednesday, November 4 (AEDT) to watch the vote count, Florida, officially known as the Sunshine State and informally as the Penis of America will have our undivided attention as it will across the world.
Trump has to win it to stay in the game. A loss in Florida still gives Trump a pathway to 270, but it narrows considerably.
If Florida turns blue, and any one of a number of swing states do likewise including Pennsylvania where Biden leads by four per cent on Real Clear Politics polling averages, Michigan (3.3 per cent), Wisconsin (5.0 per cent), North Carolina (0.6 per cent), Ohio (2.4 per cent) or Arizona (5.0 per cent) or states where Trump holds narrow leads in RCP average polling — Iowa (0.7 per cent) and Georgia (1.3 per cent) and it’s all over for the Donald.
Biden would dearly love to win Florida but if he doesn’t, he can still win the presidency with gains in Pennsylvania, Michigan and Wisconsin, for example. What will be of concern to the Biden camp is his polling leads have diminished in those states, including Florida. Biden held big leads in Michigan a month ago that seem to be dwindling by the day.
On the polling, the one state that might turn red is Minnesota. In July, Biden looked to have it in the bag. The Trafalgar Group, a GOP polling company, put the race in Minnesota neck and neck last week while other polling has Biden around four points in front. Ominously for Biden and the Dems, the Trafalgar Group got their polling right in the neighbouring states of Michigan and Wisconsin in 2016 when all the other polling companies got it so wrong.
Other blue states that Trump’s re-election team were eyeing off include Nevada and New Hampshire but on the polling at this stage, Biden continues to hold big leads. Virginia seems to be a lock for Biden. It’s hard to see any other blue state turning Trumpian red.
One might think this is all hollow necromancy at this stage of the race and that may be right. But there is one big difference in polling between 2016 and now. The number of undecided likely voters has been shaved down to around six per cent. That figure stands in national polling and swing state polling.
In 2016, in three of the most crucial swing states — Florida, Wisconsin and Pennsylvania, 13 per cent of voters were undecided in polling two weeks out before the election. Seven weeks out, that figure has been halved. Those undecideds may still have a powerful influence on the outcome of the election, but it is clear there are less of them.
It’s not quite true to say that minds are made up, but more minds are made up than was the case in 2016 where the election turned Trump’s way in the last two weeks.
This makes the big shocks of 2016 less likely. Perhaps more importantly, it confirms that politically the US is more divided, more polarised than it was in 2016.
If the polling is more accurate than it was in 2016, what we can see is that Trump has got a lot of ground still to make up if he is to beat Biden but momentum has now swung his way and we can all but dismiss the Democrat wet dream of a Biden landslide.
More Coverage
Read related topics:Donald Trump
Join the conversation
What a Harris or Trump victory means for Australia
An analysis paper by AMP chief economist Shane Oliver released on Wednesday says that the US election has ‘significant potential to impact markets’.
Read more
Israel is right to fight but I just can’t trust Netanyahu
The only response to October 7 was to destroy Hamas, but any peaceful aftermath will require an act of imagination on Israel’s part.
Read more