Why pollsters got the election wrong… again

Just two predictions could come close to results, British firm JL Partners’s election forecast and The Telegraph’s own poll

Once again, it appears the polls have underestimated Donald Trump.
Results are still coming in across the country but what was presented as a race with one point in it within key swing states is quickly turning into a rout.
In North Carolina, where 97 per cent of the vote is counted, Trump was predicted to have a 1.3-point lead in the final polling averages collected by Real Clear Polling. He has exceeded that by three points.
Pennsylvania polling saw him 0.2 points ahead. It appears he has won by around three points with 95 per cent of votes in. In Wisconsin and Michigan, he was predicted to lose to Kamala Harris, but is currently on track to win both.
On average, Trump performed around three points higher in the swing states than polls suggested.
Nationally, he is on track to win the popular vote as well, surpassing his 2016 performance when he fell short on this metric. Around 20 per cent of the vote is still being counted.
Pollsters will, however, take the outcome as a victory. The real results were within the margin of error – generally three points plus or minus – and unlike 2016, they did correctly call the majority of swing states for Trump.
Lessons have continually been learnt from 2016. Then, pollsters predicted Clinton a seven-point lead in Wisconsin, but Trump went on to win comfortably. Similar errors were made in Pennsylvania and Michigan.
The main issue in 2016 was that certain groups were far easier for pollsters to reach, in particular college-educated voters, who were more likely to vote for Clinton. Republicans in general were less likely to reply to surveys.
Pollsters attempt to adjust to these gaps by weighting responses, effectively making up for lack of response by portions of society.
While the polls represent a clear improvement, they still largely got it wrong. There were just two predictions that could come close, the British firm JL Partners’s election forecast and The Telegraph’s own poll.
The presidential race saw an abundance of election forecasts as well as polls, which use polling data and model it against a number of different variables, such as differing turnout or demographic shifts.
They then produce a state-by-state estimate for the final electoral college results.
The Economist forecast a narrow victory for Ms Harris. Polling accumulator FiveThirtyEight forecast a Trump victory with 276.
JL Partners emerged as one of the most bullish in favour of Trump, with a project of 287 Electoral College Votes.
If, however, Trump does claim all seven swing states, The Telegraph’s election forecast will be the most accurate of all major polling or media outlets, with their final forecast showing 289 for Trump.
Some demographics have clearly been missed in their support for Trump.
Latino voters, for example, saw their support increase by 13 points for Trump, something that no polls have picked up on. Some groups, such as women and African Americans, were predicted to swing towards Ms Harris and Trump alike – something that also didn’t happen.
However, fears that in this election pollsters were overweighting for Trump supporters have been proven wrong. Once again, the MAGA vote has been allusive to pollsters.
A few days ago, a single poll from Ann Selzer suggested a four-point lead by Ms Harris in the state of Iowa. Trump has received 56 per cent of the votes there – higher than in both 2016 and 2020.
This figure is now laughably wrong.
On the other hand, The Telegraph’s election forecast had Trump on 289 Electoral College votes – a comfortable victory. As things are going, he is likely to surpass that.