Stalin and Dieting

What on earth could the Communist dictator have to do with dieting? First, absolutely nothing with the Russian branch of Papa John’s Pizza who gave free pizza’s away in 2016 if people in Russia could name a person that had a Soviet tank named after him (above picture). But I do want to make a point about individuals and averages (means), so I have a perfectly ridiculous poll to explain it at the end. 

Let’s start with dieting. In 2018, Christopher Gardner et. al. wrote a paper on the effects of a low-fat versus a low-carbohydrate diet on the weight loss of adults. There are two ways to think about the result: the mean (average) result and what happened to each individual.

The authors had 609 obese participants in what was called The Diet Intervention Examining the Factors Interacting with Treatment Success (DIETFITS). They must have worked hard to get that title just right. It went on for 12 months. They assigned people randomly to either a healthy low-fat diet or healthy low-carb diet for 12 months. They were looking to see if they could predict weight loss based on genotype (your genetic makeup) or insulin secretion (insulin is a hormone that helps the cells absorb glucose, if it absorbs too much, your body converts it to fat).

In this study, the mean results were that the participants in both diets lost weight, on average. The low-fat group lost and average of 11.7 lbs. and the low-carb group lost and average of 13.2 pounds

For us though, that’s not the important point. Let’s go back to the mean versus individuals. In the two charts representing the individual trials, each individual is represented by a vertical line.

What’s really important is that weight changes were drastic in the tails of both groups, from 70 pounds lost to 24 pounds gained (the chart lists kilograms instead of pounds). So the fact that people lost some weight on both diets is nowhere near as important as what happened to each individual. What you would want to know is, what would happen to me? 

The same thing is true of clinical trials. They randomize groups for treatment of, for example, drugs, and some people get better, some have no change, and some get worse. They compare what happened in the test group (given the drug) and the control group (not given the drug). The results are reported as, overall, the drug works or doesn’t on the people in the test group compared to the control group. Of course, what would really be nice to know is who does it work for, why, and how efficacious is it for those individuals.

For an extreme example of why you want to know about individuals, let’s suppose that the former Premier of the Soviet Union (1941 to 1953), Joseph Stalin, wanted to take a poll to see how people reacted to one of his policies. He turned to his pollster, Peter, and said, “Pete, how am I doing with this latest policy?” 

One of his many policies to keep himself in power was to kill anyone who disagreed with him. Just one of these (out of 20 million deaths he was responsible for), were those who were tortured and executed under the “Great Terror” for disagreeing with Stalin’s policies. Pete asked the question “Do you agree or disagree with Papa Stalin’s (they literally called him ‘Papa’) policy of killing people who disagree with him?” 

There were about one million people who obviously knew they disagreed with the regime or at least knew that someone would accuse them of disagreeing. They would actually fight to stay alive. Let’s suppose that there were also an equal number of true believers who agreed with the policy (and were even willing to help).

Everyone else just wanted to keep their head down and said they didn’t care what he did. Let’s call them “indifferent.” Using the same type of chart as above, let’s take a look at the poll where each line represents 100,000 people.

If Pete, Stalin’s pollster had reported the average response to the question, do you agree or disagree with the policy?, he might have said something like,

“Comrade Stalin, I’m happy to report the people are indifferent.”