Why you need more than the Wahl-O-Mat

Wahl-O-Mat and its clones want to show which parties we agree with. Unfortunately, it is wrong to assume that such an algorithm would lead to a meaningful voting decision. Shortly before the decision on Sunday, we look at the election-o-matisation of democracy.

Of the Wahl-O-Mat is now an institution. When it is activated for a federal election, the demand is so great that one or the other server breaks down under the load for a short time. It’s tempting too: answer a series of questions and I’ll tell you which party to vote for.

On closer inspection, this promise is untenable, which the Federal Agency for Civic Education – the publisher of Wahl-O-Mat – also does writes in their FAQ: The tool is used for orientation and for dealing with the content of the parties and for thinking about one’s own point of view. It is particularly recommended to read the reasons given by the respective parties.

But even with these restrictions in mind, there is a lot to criticize. For example the choice of questions. Journalist Sham Jaff complains that gender is an issue, but not sea rescue or police violence.

gender is at #WahlOMat topic, sea rescue and police violence not.

– Sham Jaff (@sham_jaff) September 3, 2021

The problem: The content is based on the statements in the party programs, which are sorted and filtered by working groups until the 38 questions are put together, which the parties answered with “agree”, “disagree” or “neutral”. The PDFs of the election manifestos of all parties represented in the Bundestag are 942 pages long. This makes it difficult to judge whether the choice of topics really only reflects how the parties shed light on the topics.

In terms of content, a lot falls by the wayside in the selection process. Digital and network policy is reduced to the question of taxation of Internet companies and the right to work from home. Broadband expansion? Upload filter? Smart Cities? Startup funding? Copyright? Cyber ​​war? State trojan? IT security, especially after the many ransomware attacks? Doesn’t matter anything.

The situation is similar in health policy: The corona pandemic and its consequences, which are also reflected in the election programs, do not appear in Wahl-O-Mat. Not a word on the subject of care shortages, absolutely nothing about the inclusion of people with disabilities, citizens’ insurance or the many other health policy issues that have been debated more than ever in recent years. This list could easily be expanded to include other policy areas.

For the Federal Agency for Civic Education this is a problem that is difficult to solve, after all, the preparations take months. Numerous other offers try to fill the gap: The Social-O-Mat was put online by the Diakonie to shed light on social issues; Wahltraut provides an overview of feminist issues; and the Tax-O-Mat asks not about content, but about salary. It shows how it would affect the net if one party were able to enforce their tax model.

And then there is Deinwal, which examines the question of how the parties voted in the current Bundestag, which among other things suffers from the fact that, of course, only the parties that were actually represented in the Bundestag are visible. Overall, however, the number of Wahl-O-Mat clones seems to have decreased. In 2017, the Digital-O-Mat asked about topics related to digitization, the Science-O-Mat specialized in science and education, and the Euromat asked questions about Europe. These Wahl-O-Mat clones are no longer included or not finished yet.

However, it is the tiny party for health research that takes the Wahl-O-Mat principle ad absurdum – albeit involuntarily. The transhumanist party is only interested in one point: to research diseases and their treatment for the purpose of human enhancement and life extension. In a dreamed-of coalition government, the party apparently wants to go along with everything that the coalition partner has on the agenda, as long as the billions are spent on health research.

Consequently, the party answered all questions in all seriousness with “neutral” – with the astonishing result that I achieved 54 percent agreement with it, which is a purely statistical effect. Regardless of whether a party that is even neutral to questions about CO2 emissions and climate catastrophes can even begin to be eligible for election, this result shows what the principle of Wahl-O-Mat suffers from: election decisions simply cannot be easily identified a couple of questions and reduce an algorithm.

The basic idea of ​​Wahl-O-Mat fits perfectly into our time. We constantly feed algorithms with all sorts of data and they then spit out results that are intended to serve as a decision-making aid. Which is extremely helpful when a navigation system is to calculate the shortest route, but is always a problem when we first have to interpret the results: Have all the important aspects been taken into account? How exactly do I have to look to notice that I seem to have a very high level of agreement with a party, but which then takes positions on one or two points that I absolutely do not share and make this party ineligible for me?

A good example was the 2017 Tegel-O-Mat referendum to keep the now closed Berlin airport open. A yes to the question of whether Berlin needs an alternate airport in addition to BER earns you a point for keeping Tegel Airport open, even if you think it is a crazy idea to have such an airport in the middle of the city.

The interpretation of Wahl-O-Mat is still relatively simple: The Federal Agency for Civic Education itself points out that the result should not be taken as a recommendation for an election. Nevertheless, it is relatively difficult to avoid the suggestion of the percentages. The psychological effect has long been known: the number on the screen exudes a certain authority, especially when it appears on the monitors of clerks, who would be more likely to get stressed with their superiors if they would question them even though it is the job of People should be able to check the results of algorithms again for their plausibility.

Note: The topic of digitization was often neglected in the election campaign. We therefore explicitly asked the parties for their opinion on the most important aspects of digitization. Click here for our big digital check for the 2021 federal elections.

Whether at Wahl-O-Mat or elsewhere: The human factor remains far too underexposed. What is in party programs does not have to survive coalition negotiations by a long way. And what a coalition writes in its treaty can be very different from how it is actually governed in the end. The old saying “Topics instead of heads” is overrated. Except for a few epidemiologists, nobody saw the corona pandemic coming and in the next four years we will be faced with questions that we are not even thinking about. And then it comes down to which minds we are most likely to trust to deal well with these questions. So it depends on which people they are who are on the lists and who are available for election as direct candidates.

Algorithms are still amazingly bad at mapping this human factor. Regardless of whether it is about the Wahl-O-Mat or a film recommendation from the streaming service. One thing, however, has to be left to Wahl-O-Mat and its clones: Posting a screenshot of the result repeatedly leads to lively and fruitful discussions on social media about why you should vote whom and who not. With real people.

T3n has shed light on the background of Wahl-O-Mats in this article.