The Unstatistik of the Month for July is a press release from the Technical University of Munich (TUM) titled "Corona fuels another pandemic" and its reception in the media. The "other pandemic" refers to the increase in weight among adult Germans. "Germans have gained just over five kilograms in the past year," writes ZEIT online, expressing concern: "How do we get that weight down again?" In the following interview with four personal trainers, two of them confirm their impression that Germans have visibly become heavier on the streets.
Various studies have examined the extent to which the COVID-19 measures have led to changed lifestyles, which could also be associated with weight gain. However, according to the Robert Koch Institute (RKI) based on the "Health in Germany currently" study (GEDA 2019/2020-EHIS) with over 23,000 nationally surveyed individuals aged 15 and older between April 2019 and September 2020, only an average weight gain of 1.1 kg was found. Does this mean that Germans have really put on a lot of weight in the past three quarters?
No, that would be a misconception. The 5.6 kg average weight gain, reported by the market research institute Forsa as part of a study for the Else Kröner Fresenius Center for Nutritional Medicine (EKFZ) at TUM, only applies to those respondents who have gained weight. So, it's a conditional mean. Eleven percent reported an average weight loss of 6.4 kg, and 48 percent said they maintained their weight during the pandemic. Calculating the overall average from this data yields an increase of just under 1.5 kg. (By the way, the presentation of the study results mentions "only" 5.5 kg as the average additional weight among respondents who gained weight.)
The claim on ZEIT online is clearly incorrect.
The contributions from mdr.de ("The respondents stated that they gained an average of 5.6 kilograms since the beginning of the pandemic."), Sonntagsblatt ("According to a new study by TU Munich, Germans have gained 5.6 kilograms during the corona pandemic."), and Passauer Neue Presse ("5.5 kilos more: How Corona made Germans fat") are also misleading, to say the least. However, the press release from TUM may be partly responsible for this. It states: "On average, weight gain is 5.6 kilograms," without indicating that this is only the average of those respondents who gained weight. Nevertheless, numerous other media outlets reported correctly and (almost) verbatim: "According to the study, around 40 percent of respondents have gained weight since Corona - an average of 5.6 kilograms."
Serious insight: Those who already weigh a lot are also more likely to gain weight.
According to the TUM study, however, some correlations can also be identified.
„The higher the Body Mass Index (BMI) of the respondents, the more likely they are to report having gained weight since the beginning of the pandemic."
"This simply means that with a relatively high weight in the past, one can predict weight gain in the future quite accurately. This is as unsurprising as the found correlation between psychological stress and 'eating more.' However, one can indeed question whether respondents eat more because they feel stressed or whether they are stressed because they are gaining weight.
However, dealing with statistics regarding German measurements (or masses) is not a straightforward matter. In August 2018, as well as in November 2014, the Federal Statistical Office published a special report based on 2017 data, according to which almost 53 percent of the adult population were overweight. Separated by gender, over 62 percent of men and over 43 percent of women were overweight. The figures come from an additional survey to the microcensus. Height and weight were surveyed, from which the Body Mass Index (BMI) as a measure of underweight, normal weight, or overweight can be calculated. Apparently, the mass of Germans has increased significantly. In 2013, slightly more than 52 percent of respondents were overweight, and in 1999, 'only' just under 48 percent. Men are consistently more overweight, at least according to the criterion of a BMI above the threshold of 25.
The BMI is calculated as the body weight in kg divided by the square of the height in meters. This sounds like a somewhat strange formula, but it can be visualized well: Leonardo da Vinci drew the famous picture of the 'Vitruvian Man,' the man standing with arms and legs outstretched in a circle. Now, the area of the circle is in a square relationship to the radius, and a cube of 1 kg of water, from which we humans are mostly composed, has an edge length of 10 cm or 0.1 m. So, the BMI is roughly the height (or thickness) that an average human body would have when lying flat on the ground – roughly speaking, how far the belly protrudes when lying down."
The BMI is not without controversy, and respondents are not always honest.
The Germans seem to be getting thicker or at least heavier over time. However, the BMI is not without controversy; after all, muscles weigh more than fat, so some very athletic people may be mistakenly classified as overweight according to this simple criterion. But it seems things were worse in the past when looking at another national survey: the National Nutrition Survey II (NVS II), whose data was collected in 2005/06. According to this survey, 58 percent of all Germans were already overweight 15 years ago. At that time, 66 percent of men and 51 percent of women were overweight. However, according to another RKI study based on GEDA 2014/2015-EHIS, five years ago, only 54 percent were overweight overall, with just under 47 percent of women and just under 62 percent of men. How can this be?
Average BMI of the adult population; studies 1998 – 2017. Blue: Survey, Orange: Measurement. Image source: Unstatistik/Katharina Schüller
One part of the change measured by the Federal Statistical Office in the Microcensus can at least be explained demographically. Among the 70- to 74-year-olds, almost three-quarters of men and nearly three out of five women are overweight, while among the 20- to 24-year-olds, every third man and nearly every fifth woman is overweight. As the population ages and people tend to gain weight with age (but usually do not grow taller, rather shrink), statistically, more overweight individuals are included.
Percentage of the adult population with a BMI greater than 25; studies from 1998 to 2017. Blue: Survey, Orange: Measurement. Image source: Unstatistik/Katharina Schüller
Indeed, the most revealing aspect is the difference between the Microcensus, GEDA study, TUM survey, and National Consumption Study. In the first three investigations, people were asked how much they weigh, while in the Consumption Study, they were weighed and measured. Interestingly, the Consumption Study indicates a much larger problem of overweight in Germany compared to the survey studies.
Hence, the numbers may indicate less about how the weight of Germans has changed and more about how both men and women tend to lie equally in surveys when it comes to their weight.
A look at the pure survey data also makes the results of the TUM study appear much less spectacular: A weight fluctuation of less than 1.5 kg corresponds to a BMI change of less than half a point. This falls within the range of statistical variation.
With the "Unstatistik of the Month," we, the Berlin psychologist Gerd Gigerenzer, the Dortmund statistician Walter Krämer, RWI Vice President Thomas K. Bauer, and I, question both recently published figures and their interpretations every month. All contributions can be found at www.unstatistik.de and on the Twitter account @unstatistik.