context
First the farmers’ protests, then the demonstrations against the right: In the past few weeks, many people in Germany took to the streets. The organizers and the police usually disagree about exactly how many.
According to police reports, 100,000 people took to the streets in Berlin last weekend to demonstrate against the right. Or was it even 350,000? At least that’s what the organizers of the protest claimed. In other cities the numbers also varied widely: in Munich, according to the police, there were also around 100,000 people, according to the organizers 320,000. In Hamburg the police counted 50,000 people, the organizers 80,000.
The fact that information about the participants in a demonstration or rally differs is not a new phenomenon. There are always discussions about how many people were really on the streets. The numbers are usually collected by the police and the organizers of the demonstration – but there are no uniform procedures.
“The numbers are often estimated more or less from the gut, both by the police and the organizers,” says Sebastian Haunss from the Institute for Protest and Movement Research. For example, past events on the site would be used as comparative values. “The effort required to give the most accurate estimate of the number of participants is relatively high,” says Haunss. It is therefore very rare for something like this to happen.
Row counting and aerial photography
Different methods can be used to count people at a demonstration – depending on the size and type of protest. In smaller demonstrations, people walking next to each other in a row can be counted. The number of rows is then counted and multiplied by the first number.
“In principle, this is not a stupid idea,” says Stephan Poppe from the Institute of Sociology at the University of Leipzig. “But the problem is that people don’t line up like an army.” In addition, the events last weekend were often rallies. “So people don’t just walk past you so they can be counted.”
In such cases, aerial photographs can be used, says Haunss. “You know pretty exactly how big these places are. And then you can estimate, either on site or from pictures, how many people are standing in one square meter and then extrapolate that.” There are weaknesses here too: On the one hand, the density of people is not uniform everywhere at a rally. In addition, rallies in one place are not static: new people are always coming and others are leaving.”
If carried out professionally, such methods can be used to estimate the number of participants with an accuracy of ten to 20 percent, says Haunss. Artificial intelligence (AI) could also be used in the future to evaluate aerial photographs and count people. However, this is not currently being applied.
Is the police a neutral party?
According to the experts, the fact that there is often a significantly greater discrepancy in the information between the organizer and the police also depends on other factors. “We know that organizers naturally have a systematic interest in appreciating the event,” says Haunss. “That’s the central point of a demonstration, that you take to the streets with a large mass of people.”
From Haunss’s point of view, however, there are reasons for the police to tend to estimate demonstrations as smaller. Because the police are often not a neutral party, both in their own perception and in the perception of the demonstrators. “The police are often seen as opponents and the police also often see the demonstrators as opponents. From this position, the police also have an interest in portraying the demonstration events as smaller than they actually are.”
From a formal point of view, there could also be reasons for the police to intentionally underestimate the number of participants. “It may be, for example, that the police report a lower number of participants in order to avoid admitting that there may have been too few officers on site.”
However, Poppe considers this thesis to be worthy of discussion. “If the police were to present the numbers smaller in order to disguise the fact that they had too few personnel, then they would sometimes have to represent the numbers larger if they had deployed too many personnel,” he says.
In addition, systematic under-recording by the police can only be assumed if the real number of participants is known. “If the organizers systematically overstate the numbers, then the police information for the comparison number is always too low. The police information that we checked was unsystematically too high or too low. I can’t really see a trend there.”
AfD demos in Erfurt as Research basis
There is little research on the topic. Statistically valid results in particular can only be obtained with great effort, say the experts. There are therefore mainly individual random surveys from which no general statements can be made.
For example, the University of Leipzig looked at a total of eight AfD demos in Erfurt in 2015 and found them Figures from the police and the AfD compared with their own estimates. While the AfD was each time significantly higher than the scientists’ estimates, the police’s numbers were significantly closer overall. There was only slight under-recording in two cases, while the police estimated five times more participants than the scientists.
This study is not sufficient evidence because of the few demonstrations examined, says Poppe. In addition, it is also possible that the police might have put more effort into the estimates because it was known that the scientists were also counting.
Exact estimate is hardly possible
Overall, from Poppe’s point of view, it is simply not possible to determine the exact number of participants. “The question is: What do you actually count? The total number of participants in a demonstration? Or, for example, the number of people at a certain point in time?”
Theoretically, it is conceivable that 10,000 people will be there in the first hour of a demonstration. Then they all went home and 10,000 new people would arrive. “Then we would have had a total of 20,000 participants. But at no time were there more than 10,000 people there. What is the number of participants then?”
Poppe therefore advocates specifying a possible range of participants instead of an exact number. “If I make a precise statement, there’s a very high probability that I’m wrong. But if I say, for example, there were 14,000 to 16,000 people at a demonstration, then it’s probably more accurate.”
“Important form of expression”
Even if estimates of the participants in a demonstration are always subject to a certain margin of error, Poppe and Haunss still demand that the police count them as accurately as possible. “Demonstrations are a very important form of expression. And that’s why it’s also relevant to know how many people support a particular cause,” says Poppe.
Haunss also sees it this way: “Demonstrations are the most important means of communication for social movements. A demonstration with 20,000 participants makes more impression and gets more attention than a petition signed by 150,000 people. Because taking to the streets – possibly even when things are bad Weather or in a situation where one’s physical integrity is at risk – is a very special way of standing up for a cause.”