10 years of Unstatistik: How we overcome number blindness

Unstatistik des Monats has been around for ten years now. Our first book became a Spiegel bestseller and the second will be published this year. But there's still no reason to sit back and say: We've made it - Germany thinks statistically. On the contrary. We still have a long way to go.

"Why Being Fat Doesn't Make You Stupid and GMOs Don't Kill" by Thomas Bauer, Gerd Gigerenzer, Walter Krämer - Campus Verlag

With a view to dealing with numbers and statistics in the Corona crisis, one might rather believe that the "Unstatistics Virus" is spreading in ever new variants across traditional and new media. But that doesn't necessarily mean that the lack of statistical thinking is greater today than ever before. Perhaps we're just looking more closely, have more precise diagnostic tools, and thereby illuminate ever larger parts of what used to be in the dark field of bad statistics.

This is evidenced not least by the fact that more and more media outlets are relying on unstatistics. Be it by disseminating our contributions further or by consulting us unstatisticians on various topics. A small chronology of selected examples should illustrate this positive development.

1. July 2019. Capital writes about "false conclusions about vegan nutrition" and has the "experts of unstatistics" explain why their positive effects on the risk of type II diabetes are significantly lower than a study suggests.
   

2. March 2020. FOCUS Online reports on the commitment of unstatistics for better data in the pandemic. Over the following two years, the news portal asks unstatistician Katharina Schüller for assessments in almost a hundred articles and crowns her the "Corona explainer." Even BILD and BILD am Sonntag regularly ask us to explain the COVID-19 numbers. This should not disguise the fact that these influential platforms regularly provide us with unstatistics themselves - but it is a small step towards enlightening numerous people who otherwise have little exposure to a scientific approach.

1. March 2020. Horizont explains why statisticians should have a say in COVID-19 testing. The current data does not allow conclusions about the probability of spread or death rate, Horizont quotes from the unstatistics of March 2020. ZEIT invites the unstatisticians to publish this contribution as a guest article, which receives much attention and commentary. Capital immediately picks up the topic and cautions against predictions about the COVID-19 pandemic. The subsequent unstatistics in April on why the reproduction number has its pitfalls also make it onto Capital's website.

2. June 2020. The STERN asks the unstatisticians for an assessment of the COVID-19 blood type study. Blood type A equals high risk? This is a topic that the unstatistics themselves did not explicitly address. However, the magazine proactively wants to avoid producing unstatistics. Shortly thereafter, a differentiated article on the blood type study follows, reflecting our expert assessment - on RTL.de, a platform that is otherwise known for its sensationalist presentations.

3. December 2020. A special highlight for the unstatistics: Among other things, we criticized the Bayerischer Rundfunk (BR) for its misleading statements about the effectiveness of the COVID-19 vaccine. The reaction: BR Faktenfuchs reaches out to us and shortly thereafter publishes a self-critical analysis.

4. February 2021. In a Telegram post, a "massive dismantling" of hospitals "at the expense of the healthcare system" is discussed using a graphic. The fact-checking initiative Correctiv asks the unstatistics and reports our assessment of why the data is manipulatively presented. Two weeks later, Correctiv again relies on the unstatistics. This time, the occasion is false and misleading claims about potential side effects and the effectiveness of the Covid-19 vaccine from the companies Biontech and Pfizer in several blog articles.

5. June 2021. Even if it's not about COVID-19, the unstatistics are increasingly attracting attention. Horizont clarifies why Green Party members don't necessarily have an SUV parked outside and comments:

1. July 2021. The Weser-Kurier publishes a general article about the unstatistics as a "teaching example for dealing with data." In November, an interview follows regarding the background of our initiative. Subsequently, the newspaper adopts several installments of the unstatistics.

2. January 2022. The magazine Cicero consults the unstatistics for its cover story on the "Pandemic of Knowledge Gaps." Because we have "questioned newly published numbers and their interpretations since the beginning of the COVID-19 crisis" and "highlighted the problems in easily understandable terms" and "early on."

This is just a small glimpse of the significant resonance that our initiative has experienced, particularly during the COVID-19 pandemic, in social media. We are making an impact - in addition to the examples mentioned, WELT, Ruhrbarone, FAZ, numerous podcasts quote and mention us. There are even contributions in international media such as the Latin American edition of the Chicago Tribune (May 2020) or the Turkish edition of Hürriyet (December 2021), which refer to statements made by us unstatisticians.

Indeed, everyone has their own worldview. Additionally, it seems that cases where manipulation arises from a lack of data literacy are more common than those driven by malicious intent. If this admittedly optimistic assessment holds true, then at least quality media outlets can contribute to restoring people's trust in both the media and science through a careful handling of statistics.

However, for this to happen, they must also communicate the uncertainty associated with every statistic. Much of the distrust may have arisen precisely because, especially at the beginning of the COVID-19 crisis, data and conclusions were presented as established knowledge. When it later turned out that some developments unfolded differently and corrections were necessary, this did not necessarily strengthen trust in the media and science. It is therefore crucial that media outlets are transparent and communicate the dynamics and uncertainties involved in interpreting data to strengthen public trust.

It's certainly not an easy task to make it clear to people without scientific training (and even to some scientists) that correct decisions can also be derived from uncertain data.

It takes experts to interpret numbers and statistics. We unstatisticians, who have learned to properly contextualize statistics, feel responsible for pointing out correct and less correct interpretations - and for situations where it really matters what is right and what is not.

And just because the answer often is "it depends," it doesn't mean we can do away with statistics simply because it doesn't completely eliminate uncertainty. After all, people don't forgo medical assistance even though medicine has yet to make any of us immortal. Medicine, when applied correctly, prevents premature death or excessive suffering from diseases.

And to make it very clear where statistics are being misapplied, so that we can do better in the future, we believe unstatistics to be an exceptionally suitable tool.

Background

The Berlin psychologist Gerd Gigerenzer, the Bochum economist Thomas Bauer, and the Dortmund statistician Walter Krämer launched the "Unstatistik des Monats" (Unstatistic of the Month) campaign in 2012. They question both recently published figures and their interpretations every month. The campaign aims to contribute to handling data and facts sensibly, interpreting numerical representations of reality correctly, and describing an increasingly complex world and environment more meaningfully. Since August 2018, Katharina Schüller, Managing Director and Founder of STAT-UP, has been a member of the Unstatistik team.

#Covid19 #dataliteracy #RWIEssen #Unstatistik