“Thinking outside the box” scene: survey with dubious results

A large-scale survey is intended to underpin the criticism of the corona measures in Germany. However, the methods used are largely unsuitable, and the results are therefore more than doubtful.

It sounds impressive at first: “52,929 people of all ages and professional groups” took part in the “Germany Survey 2021/2022” by the “Corona Initiative for German SMEs” (CIDM). In the “lateral thinking” milieu, the results of CIDM are quoted particularly often – they apparently underpin the criticism of the corona measures and the reporting.

The CIDM is an initiative founded in 2020 which, according to its own statements, offers “actual analyses, monitoring, trend forecasts and advice in the Corona crisis”. In 2020, they made a name for themselves in the “lateral thinking” scene for the first time with a critical “exit strategy” for the corona pandemic. The CIDM website states Wolfgang Romberg and Moritz Graf Brühl as responsible persons; there is no imprint. Moritz Graf Brühl is an insurance broker and assessor in the Frankfurt CDU district association Westend.

Wolfgang Romberg is an entrepreneur and founded the “Wir2020” party together with the radical opponent of vaccination Bodo Schiffmann. According to the imprint, Romberg is currently party chairman while Schiffmann gave up his party post.

Calls to Participate

Participation in the “Germany Survey” was pointed out in various Telegram channels and Facebook groups. For example, the participation link was shared in Facebook groups with names such as “Entrepreneurs against Corona”, “I distrust the government!”, “Prevent compulsory vaccination – citizens for justice” or “We NEVER vaccinate against Corona” – in many Cases with the request: “Generously distribute the survey among friends and acquaintances on all channels”.

The survey is also circulating in AfD circles – the AfD local association Altkreis Melsungen shares it, as does the “Alteenative (sic!) Freunde Berlin”, in which AfD politicians such as Gottfried Curio and Hugh Bronson publish posts. The “Wir2020” Facebook group was the first to share the call at the end of December.

“Sensational” survey?

The results of the CIDM survey spread massively in the scene at the end of January. In addition to various groups on Facebook, high-reach actors such as Oliver Janich, Eva Herman or “Compact Magazin”, which has been classified as right-wing extremist by the Office for the Protection of the Constitution, refer to the “sensational” online survey on Telegram. Schiffmann even presents the survey in a video on his Telegram channel that has been clicked almost 400,000 times.

The result of the survey: 65.8 percent of those vaccinated consider the vaccination requirement to be disproportionate or very disproportionate. With a number of participants of more than 50,000, the result is “more than representative”, claims Schiffmann – and continues: “A survey is representative with more than 2000 participants.”

Self-recruitment in online surveys

“That’s nonsense,” says Nico Siegel, managing director of infratest-dimap. “The number of respondents in itself is not a criterion for representativeness.” Rather, a survey is representative “if the selection of the respondents shows as many interesting characteristics of the group of people to be researched on a reduced scale”.

Josef Holnburger from the Center for Monitoring, Analysis and Strategy (CeMAS) says: “The CIDM is not representative just because they claim it is.” If you were to do a poll in a full FC Bayern stadium to find out which football club is the best, you would probably get 60,000 votes – with a result that you can imagine.

Data not representative

Several requests from tagesschau.de, how the subjects were selected, where participation was called for and whether there were measures against multiple voting remained unanswered in terms of content. However, it can already be seen from the published results that the selection of respondents is anything but representative: the age structure of the participants deviates massively from the actual one in Germany.

“People up to 40 years of age are slightly underrepresented, people over 80 years of age significantly. The latter is to be expected since it is an online survey. However, essential information is missing, for example by whom the survey was carried out, how exactly the participants were recruited and whether a weighting has taken place,” explains qualified statistician Katharina Schüller to the ARD fact finder.

Expert: Data can hardly be evaluated in a meaningful way

A weighting of the answers according to demographic characteristics can lead to a representative opinion if the response behavior depends exclusively on these characteristics and the response is sufficiently high, Schüller continues. “But if there is no careful control over who is recruited for the survey, all conceivable statistical correction mechanisms will fail.”

The proportion of the – not precisely defined – “vaccinated” among the respondents is also significantly lower at a good 40 percent. At the end of the survey period, 73.5 percent of the total population had been fully vaccinated against Covid-19, and 47.4 percent had also received a booster vaccination.

Polarizing and inappropriate questions

There were also massive methodological errors: For example, in the case of closed questions, individual, mostly very pointed and suggestive possible answers were given, which only cover a small spectrum of possible opinions. All participants who had different opinions could only indicate “Other”.

In the case of polarizing surveys, there is always the risk that people with a negative attitude will answer more often than those with an affirmative or neutral attitude, explains statistician Schüller. “If the raw data are available, certain conclusions can be drawn from the time course of the responses as to whether such distortions are present. However, these are almost never correctable.”

Already refuted data as a basis

However, the creators of the site also use other data as a basis and name apparently reputable sources. They use alleged numbers from the deployment statistics of the Berlin fire brigade to want to prove a massive increase in suicides for 2020 – and “exaggerate” these numbers to the German population as a whole. However, the Berlin fire brigade had already pointed out in June 2020 that they did not statistically evaluate suicides at all. The interpretation made of the deployment numbers was not permissible, a spokeswoman said at the time Research portal “Corrective”.

tricks in presentation

Last but not least, the makers use simple tricks: Apparently in order to show the low excess mortality from the corona pandemic, they set the daily number of deaths in relation to the total population – whether this means the corona deaths or the overall mortality remains open. According to the Federal Statistical Office, 1,020,431 people died in Germany in 2021, which is an average of almost 2,796 per day, i.e. 0.0034 percent. Even with a hundredfold increase in the number of deaths, the graph would remain unchanged with the chosen scale, so it is unsuitable for supporting any claims.