causality or correlation – wishful thinking in science

Satire and tragedy in science …. an example.

Please believe us: We are not running a satiric blog!

Today, we shall visit the lowlands of scientific publication culture. With more than 30,000 published papers on COVID-19, inevitably there will be junk papers as well. Unfortunately, their number is not a small one. This is damaging the reputation of all sciences. One such example reached us several days ago, and we are dedicating today’s blog post to this piece of work.

Nutrition is playing a key role in our health. Everyone of us may decide on what to eat. Some foods are nourishing, some are gratifying, some are harmful, and some are all of the above. It is your choice. Now, however, we have detected another role for food: A European consortium believes it has found a link between food and COVID-19 death rates. In two papers (one published 1, one on a preprint server 2) they compare country-specific diets with the countries’ death rates.

Their premise was the assumption that politically mandated protective measures (such as physical distancing and face masks) cannot be the single most important factor for lower infection numbers and lower death rates. This is astounding since the disease is spread by a respiratory virus through droplet infection mainly.

Instead, the authors took a look at the differences in diets in several European countries. This revealed astounding things. Here are some of the findings:

The Latvians are consuming 25 g of cabbage per day and person. Latvia has got one of the lowest COVID-19 death rates.

In Cyprus, per day and person, 30 g of cucumbers are consumed. The death rate is comparable to the one in Latvia.

In Eastern Europe, there is generally a higher consumption rate of cabbage and fermented milk products. There, the death rate is lower than in Western Europe. Thus: cabbage, cucumbers, and kefir protect from COVID-19.

The contrary finding was made for two other vegetables: Consumption of broccoli and lettuce is correlating strongly with an increased COVID-19 fatality rate (examples were Great Britain and Spain, respectively).

We do not intend to point out all the weaknesses of these publications (arbitrary selection of countries and single vegetables, differences in data sets and so on), since this would go beyond the scope of this blog, and would cause us bodily pain.

There is one single important point: The described findings are not an example of causality, which means the consumption of single foods is not the cause of a better or worse COVID-19 prognosis. It is simply a great example of nonsense correlation.

Unfortunately, the authors do not tell us if this is a satirical contribution. Is it simply a “hoax”? We would consider both inappropriate at times when thousands of people are dying each day.

Both such cases have been demonstrated in great publications in the past. The 1998 book “Fashionable Nonsense” 3 described how it was possible to publish a nonsense article – a hoax – in a reputable journal.

A modern classic is an article by Franz Messerli 4 who in 2012 demonstrated a statistically significant correlation of country-specific chocolate consumption and the number of Nobel Laureates coming from the respective countries. Never was mere correlation debunked more beautifully. Highly recommended reading.

We are certain that a severe malnutrition will not be helpful in the case of COVID-19, but please go on eating cucumbers and cabbage, and do drink kefir, without shunning broccoli or lettuce.

Please keep a polite distance, for the virus remains to be a pathogen that is spread by droplets through the air. We shall inform you, should this change.

Yours,

Jörg and Sabine

SAJO – for a healthy and better future!

P.S.: One of the articles 1 points out that Swiss cantons that are German speaking do show a lower death rate than those speaking Italian or French. Unfortunately, the authors are missing out on the opportunity to indicate the protective nature of the German language. The globally acting Goethe-Institutes would have been grateful, indeed.

(We could not help to make this satiric comment.)

A heartfelt thank you to Oliver Hoogvliet for the great cartoon. ollihoo (https://hoogvliet.de)

1 Bousquet et al.: “Is diet partly responsible for differences in COVID‑19 death rates between and within countries?”; Clinical and Translational Allergy, 2020 https://doi.org/10.1186/s13601-020-00323-0

2 Fonseca et al.: „Association between consumption of vegetables and COVID-19 mortality at a country level in Europe“ https://doi.org/10.1101/2020.07.17.20155846

3 Alan Sokal and Jean Bricmont: “Fashionable Nonsense“; Picador, New York, 1998

4 Franz H. Messerli: “Chocolate Consumption, Cognitive Function, and Nobel Laureates”, New England Journal of Medicine 2012