In the US presidential election, many were surprised by Donald Trump‘s clear victory. But was the victory ultimately a surprise for the fact- and data-based election follower? writes Tuomas Kanervala.
Many media houses in the United States were clearly democrats Horrible Harris position both in his reporting and in his comments. Perhaps this came out most clearly on election night when watching the results, when many “experts” and journalists spoke as the vote counting progressed: “Yes, we are doing badly” or “we can still get up from this.”
As for the opinion polls conducted in the United States during the election, some were naturally more accurate than others.
The most accurate polls for the 2020 elections were made by the Brazilian company Atlas Intel. According to their polls for this year’s election, the Republicans Donald Trump led in all the scales of the last weeks.
In addition, in the support of the entire nation, Trump was ahead of Harris in their polls in all measurements from June to Election Day. With these results, Atlas Intel was again clearly the closest to the election result compared to other forecasters.
Opinion polls commissioned by many liberal, “mainstream” media companies showed significantly lower support for Trump both nationally and in the states that decided the election.
This was repeated in a similar fashion four years ago, when the accuracy of opinion polls published by CNN, ABC and the New York Times (NYT) was significantly below average. Now, these ranked in the weakest third when looking at predicting the winner of the Libra states.
There were also differences in the accuracy of companies that collect averages from several different polls and make their own predictions from them.
Pollfair (6/7 correct) and Real Clear Politics (5/7) were clearly more accurate than the more widely used 538 (3/7) swing state forecast. The 538 site is currently owned by ABC, so the reasons for the weak predictions of the last three presidential elections can also be considered from that point of view.
In the simulation made by 538, Trump was in the lead for two weeks before the election. There was no big change in the various polls during this period, but in the last simulation Harris emerged as the most likely president.
538 bases its forecast on evaluating the quality of various opinion polls. According to them, “higher quality polls showed more support for Harris”.
This methodology seems to have reached the end of the road, because the accuracy of the polls from one election to the next is of course more important than the transparency of the polls’ methodology or the results that are tilted in a certain direction.
Now, for example, Rasmussen Reports polls do not appear in the averages or forecasts, because they did not “meet the quality requirements”. However, they have been one of the most accurate predictors of support for the presidential elections in recent elections – as well as this time.
The “highest quality” polls of 538’s own instrument panel, NYT and ABC, were also far from the top spots this time, as already mentioned earlier. And the measurements of the other two highest quality ones (Yougov or Marquette University) were not much better.
So why are the numbers of 538 so important?
Based on the accurate forecasts made before 2016, they have also gained a strong position in the Finnish media, which has used their support averages and forecasts without much criticism. This was one of the reasons for the wrongly colored bias in the Finnish debate as well.
The numbers supporting Trump from more accurate polling houses were largely ignored, but even the small results that boosted Harris’s support were widely reported.
In Iowa, for example, one poll predicted a 3% lead for Harris a few days before the election. The end result was a victory of more than 13% points for Trump.
The Finnish media reported on the “expected measurement of the oracle”, but did not bother to find out the background of the measurement. It used the RDD (Random-digit dialing) method, where the telephone numbers of the respondents are selected randomly. The result can be very large fluctuations from one measurement to another.
Voter registration data is widely used in the USA. Already during early voting, but especially on election day, the data showed a future victory for Trump. The numbers of women registered to vote and those living in urban areas were several hundred thousand lower in many states compared to 2020. At the same time, the numbers of those living outside cities were clearly higher.
Harris’ support was highest among women and those living in cities, so it didn’t take a great analyst to realize that the numbers for him were somewhere between bad and disaster. In the end, Trump apparently increased his support in these groups as well compared to 2020, but during the election they were especially Harris’ target groups.
Betting markets, which generally predict very different events, also showed a more likely victory for Trump in the last month. Just before election day, the probabilities were on average 60%-40%.
The coverage of the betting market mainly followed the same pattern as the opinion polls. Trump’s better odds were downplayed or attributed to individual players who bet big.
In the betting market, the price (=probability) is formed by the “agreement” of two different players, so for every bet made on behalf of Trump, there is someone who is ready to make a bet on behalf of Harris with the inverse probability. In this way, the influence of a single player on the odds in a betting market played with billions is marginal, if at all.
The human mind often plays tricks, and action based on fact is primarily based on the use of faded information to support one’s own opinion, instead of examining all available information objectively. This happened to many media representatives and other political followers both in the USA and in Finland.
Why then did the people vote for Trump?
In the United States, the economy has been the most important theme from election to election. Now, many low- and middle-income people felt that things were worse than when Trump was president the last time. They knew what they were getting when they voted for Trump. But besides the economy, a few other reasons come to the fore.
The USA is not a nation state like Finland, whose unifying force is, for example, language and culture. Rather, the USA is a nation based on the ideology of freedom and individual rights.
Now, many felt that the recent development of the country is taking these basic pillars of the nation away from them. If the position of the Finnish language were to be weakened in our own country, even at the expense of English, the backlash would surely gain widespread support.
The stricter control of social media – especially X and TikTok – which the Democrats also widely raised after the election, would sit poorly with the constitutional freedom of speech.
Especially when the credibility of the liberal mainstream media is close to record lows. 38% of Americans do not trust the media at all. Likewise, the restrictions decided by the current administration or the provision of woke ideology in schools, for example, do not seem to enjoy wide support in light of the election results.
Even before this vote, many Americans have voted with their feet. California and New York are solid support areas for Democrats and at the same time have represented liberal ideology at many levels of government.
In 2023, the largest interstate net migrations in the country were from New York to Florida and from California to Texas. In both target states, the Republicans are firmly in power and the citizens’ constitutional rights are protected to the last.
Reflecting on this development, it is not surprising that the Republicans under Trump’s leadership seem to take not only the presidency, but also the majorities in the Senate and the House of Representatives for at least the next two years. For the US voters, the most important thing was the policy that will be pursued, not necessarily the person who will pursue it.
The author Tuomas Kanervala is the CEO of Accuscoren Oy, the analytics company that makes Iltalehti’s election forecasts. Accuscore also made predictions for Iltalehti about the US presidential elections.
#signs #Trumps #victory #visible #Valtamedia
