A headline in the San Jose Mercury News caught my eye the other day: “State’s COVID-19 case rate explodes.” One would get the impression that half the population of California is sick with COVID. While the media does tend to overdramatize events, the data they reported is correct. The latest weekly rate of new infections in the state increased by a significant amount over the previous weekly rate. And that is certainly cause for concern. The issue, however, is that the numbers they are reporting are averages of data with wide disparities. In such a situation one is apt to draw misleading conclusions.

Let’s drill down a bit to see what’s really happening. As of this writing there have been 67 cases per 100,000 residents across the entire state over the last 7 days. (Using a standardized measure such as cases per population makes it easier to compare with other locations). But some counties are doing much better and some are doing much worse. Los Angeles County has 87 infections per 100,000, considerably worse that the state’s average, not to mention the fact that its large population translates into a large number of infections. On the other hand San Francisco County has only 26. Clearly residents of L.A. have cause to worry, while those in S.F. could actually be congratulating themselves for having done such a good job of controlling the spread of the virus.

It might also help mitigate citizens’ concerns by comparing the rates of infection in their county with those in other states. If the residents of L.A. are worried, they should be glad they’re not living in Dewey County, South Dakota, where the weekly case rate per 100,000 is a staggering 560. Or even in Nantucket County, Massachusetts, where it is has climbed to 201.

The same issue can be seen when reporting investment performance. As I have written before, investors commonly use averages to make decisions. The problem is that averages mask extremes. For example, over the last 45 years, a portfolio comprising a mix of large U.S. company, small U.S. company, and international stocks would have gained on average about 10% annually. A $10,000 investment in such a portfolio would have grown to almost $730,000 today (excluding taxes). However, in 2008 the value of that portfolio would have dropped by about 40%. That would have represented a loss of over $150,000. Investors expecting an average annual 10% return on such an investment would have been shocked.

You should be wary of reaching conclusions from an average of a large number of data points. Without additional measures such as variance or standard deviation or, for investments, Sharpe or Sortino ratios, it can be extremely difficult to understand the potential extremes. Whether we’re talking about COVID-19 infections, investment performance, or even weather phenomena, be sure to dig into the numbers as much as you can. Your decision-making will be all the better for it.