Healthcare expenditure -v- life expectancy

A smart way to reduce the risk of confusing correlation with causation, especially when comparing different countries, is to plot the two variables of interest over time.

Here’s a nice example, showing that it’s likely the US has a less efficient healthcare system than other rich countries, and that the low life expectancy is not due to other confounding variables (e.g., having a higher homicide rate – something which has actually decreased over the time period shown).

More here by Lane Kenworthy