Link to the article in swedish here
No one has escaped the exciting election campaign between Harris and Trump, an election that could decide the fate of the West in one of the most sensitive geopolitical situations in a very long time.
First, opinion polls are not forecasts, but a current picture of what it would look like today. The future cannot be predicted by even the best fortune teller. However, there are a few things to keep in mind when we are bombarded by American opinion polls in our news feed.
The one with the most votes (might) not win. Clinton received more votes than Trump, 66 million to Trump’s 63 million. The United States has an electoral system in which each state has a number of votes that are not proportional to the inhabitants of the country. The election is about a majority of the 538 electoral votes.
Many polls look at “popular votes”, who gets the most votes nationally, not per state, i.e. who gets the most votes. These were relatively accurate in the elections, but they are misleading. For example, Clinton received the most votes, but she still did not win.
Demobilization, the election’s “spin doctors”, have focused on getting the opposing side not to go to vote. It is faster to undermine the opponents’ voter base than to win your own voters. The Republicans had the demographics against them before Trump and put most of their energy into reducing voter turnout for Democratic voters. We also see it in this elections, for example, that you have to be able to show a passport or birth certificate to register to vote. The fact that you have to register to vote is also a threshold.
The election is on a Tuesday, and early voting has been a complicated affair in many states. This negatively impacts voter turnout, especially for those with the lowest incomes who may not be able to afford or be able to vote.
Who will vote? In the United States, it is often more difficult to know IF you vote than on WHO. The records of those to register to vote are often available to the campaigns, htey often state what party you registered to vote for, the campaigns are largely about getting them to go and vote. Obama’s campaign manager Jim Messina told me when we met just after the election that Romney’s team failed to properly analyze “likely voters”, missing that in a campaign is devastating.
Who will actually vote is at the heart of any US election, this is a big differrence to most European election where it is focusing on what they vote for more than voter turnout. It is expensive but the campaigns have the money and spend huge sums on this. The core of the campaigns strategy is based on reliable research. A big difference from the news media’s surveys, who in practice get the polls for free from research companies consicdered good PR.
Unscientific methods. Do you remember “shy Trump voters”, an explanation that did not hold really up in the analysis afterwards, they missed white first-time voters with lower education who voted for Trump. That more of these voters suddenly voted. These are underrepresented in the self-recruited (open access) unscientific cheap web surveys that are used so often. I probably also missed that they were going to vote. A report from Stanford showed that the states they had missed had only published unscientific studies, instead of scientifically reliable methods.
In Sweden, unfortunately, the Schibsted new media group has switched to the same unreliable, unscientific surveys as the basis for its news evaluation. However, we at Novus (and Gallup Nordic) always work on a scientific basis, even in media collaborations.
Are you missing out on new voters? Harris’ voters are particularly interesting from this perspective. Her campaign seems to be mobilizing low-educated non-white first-time voters. Even more difficult to reach, esspecially the young, in self-recruited, panels and who do not usually vote and can also fall out of the “likely voter” models they work with. There may be a double effect if Trump’s voters are still bossted because they know they are underrepresented.
However, this possible underestimation would benefit Harris, because if she were to lead big before the election, it would reduce the willingness to vote for her. It is already a bid fouc to make it harder for Harris voters to vote in key US states. Should it look like she was winning big, the voter turnout would go down for the winner and may go up for the loser, in that case Trump. That’s probably why she is clear in saying that they are the “underdog”. But this method difficutly would explain why there has not been any major change in support after the congress and the debate.
But if the polls are way off which I am afraid they are, it threatens the confidence in opinion polls all over the world. All of the problems mentioned can be managed, but it cost money, money that is not in the media’s survey budget, even though they define an entire survey industry.
The responsibility is not only the media’s, but just as much my colleagues who release unreliable surveys with the explanation that the customer cannot afford real surveys. Imagine a new car where the brakes don’t work because the customer couldn’t afford working brakes?
So my advice to everyone, the US surveys are not representative of our industry. On the contrary. Don’t rely on the equivalent of a car without working brakes.
To follow the United States, I look more at the energy of the campaigns, how many people give money, how many volunteers they mobilize to get a sense of how things are going, than trying to interpret proven unreliable cheap opinion polls in the United States.
The parties have more than a $1 billion to follow and influence voters in the presidential election, that rece is not over until after the election day. The campaigns get the best surveys money can buy, they can’t afford to guess. Unfortunately, the news media is happy to guess. the news often think their job is guessing.
But reliable opinion polls is not guessing, it is knowing, not about the future but about the present.
Here at home in Sweden, the EU elections once again showed that scientific opinion polls give a good picture of the current situation, Novus’ EU poll was for the 4th consecutive EU election the closest of all surveys in Sweden.
Torbjörn Sjöström
CEO Novus
WAPOR representative Sweden
Chariman Gallup Nordic
Board member Gallup International
