How a Pizza Pie Chart Can Be Misleading

The Washington Post tells us that pizza deliveries could “predict military action.” As an example of a correlation, it is about more than food.

False Correlation

Using Google Maps data, the Pentagon Pizza Report monitors Defense Department pie deliveries. On June 12, at 7 p.m., they cited a spike in activity:

That same evening, soon after, the U.S. attack on Iranian nuclear sites began. Sometimes called the Pizza Index by its advocates, there was an uptick in Pentagon pie deliveries before Operation Desert Storm during 1991. An Economist magazine data journalist even implied that “pie charts” were useful. Others though remind us that delivery options have multiplied since the 1990s. And, it’s tough to make any delivery to our defense bastion.

Indeed, looking closely at the facts makes us question the connection between pizza and late night military planning.

Our Bottom Line: Correlation or Causation

Similarly, as economists, we need to be sure to establish causation rather than a blurrier correlation. At the website Spurious Correlations, the list is long. Ranging from butter consumption and U.S. wind power generation to a most amazing statistical similarity between a divorce rate and runs scored by the Chicago Cubs, data can be misleading.

Butter and wind power:

Connecticut divorce rates and Chicago Cubs runs:

From less consequential topics, we can leap to why causation matters. Only then can we suggest the policy changes that lead to sought after results. For example, at Our World in Data, they investigate the connection between democracy and better health.

Below, they say the data suggest correlation:

But then they move to other more precise variables. They cite the public service delivery that comes from elevated spending, reduced biases, less corruption, and more administrative effectiveness. These, they say, establish a more causal connection between democracy and public health.

Similarly, we can look more closely at pizza deliveries for more causal insight.

My sources and more: Thanks to this Washington Post article for today’s inspiration. From there, Spurious Correlations came in handy as did Our World in Data.