Modelling and Model Building

…using the example of our study: “Identification of different factors associated with Covid-19 deaths in Europe during the first pandemic wave”

A large group of statistical techniques designed to explain past data and also to predict future data is statistical modelling. This means that for a given data set with very different variables, one finds a mathematical structure that represents this data set as well as possible, firstly in a purely formal way. This procedure can be used to examine the influence of different variables on an outcome variable. In the language of modelling, the variable that one wants to explain is the dependent variable or criterion or outcome variable, and the different variables that are supposed to contribute to the clarification of this one variable are several independent variables resp. predictors.

I use our recently published modelling study [1] as a concrete example. It was conceived by me, I calculated the first analyses, then my colleague Rainer J. Klement got involved, who as a physicist is much more nimble in dealing with such models than I am.

Read more

Face Masks Lead to Dangerously High Levels of Carbon Dioxide in Children’s Inhaled Air

The long version of our mask study has been republished

A workshop report and some thoughts on it

Our mask study measured carbon dioxide levels in the inhaled air of 45 children wearing face masks. It found that the inhaled air under children’s face masks contained unacceptably high levels of carbon dioxide, about 1.3% to 1.4% by volume, or 13,000 to 14,000 parts per million. Normal outdoor carbon dioxide levels are 400 ppm or 0.04% by volume. The Federal Environment Agency and various protective regulations have determined that 2,000 ppm or 0.2 vol.-% is the upper limit above which damage to health cannot be ruled out. For children, such high values, as we measured after only 3 minutes, are absolutely unacceptable. Especially against the background that children are neither at high risk of corona infections and Sars-CoV2 nor are they important spreaders of infections.

Read more

Why Do Germans Get Vaccinated…

…And What Distinguishes those Who Are Willing To Get Vaccinated from Those Who Are Hesitant?

Our representative survey is available

I was interested in why people in Germany get vaccinated, what their most important motives are, and also why people do not get vaccinated. In addition, I am of course interested in the question: what exactly distinguishes these two groups?

For readers in a hurry: The most important reason for getting vaccinated is fear of the Covid-19 disease (for 60%). The second most important reason is the desire to lead a normal life again (for 30%). The most important reason not to be vaccinated is that they did not want to be treated with substances whose long-term effects are unclear (for 40%), and the second most important reason is fear of side effects (for almost 40%). The two groups can be separated very well with a logistic regression model. Those who are willing to be vaccinated differ from the reluctant: they score higher on an “orthodoxy scale” that I developed specifically for such purposes and validated in our immunologist survey. They tend to not read the original scientific literature and rather follow conventional media. This model has a relatively good accuracy and is able to correctly match 78% of people.

Read more