Regression to the Mean

09 Dec 2022

Regression to the mean is a statistical phenomenon where extreme values or events tend to be followed by less extreme values or events, particularly when a similar event occurs repetitively but is impacted by random factors each time (for example, the length of time it takes you to drive from point A to point B daily, like a commute to work).

One classic example is students who score high on a test–their next score tends to be lower, even if they studied just as much. Another example: a stock that has a high return in one year will likely have a lower return the next year, closer to a “typical return.”

The best illustration of regression I’ve ever seen is when a short father has tall sons or a tall father has short sons. What causes “regression”? Humans have an average height, and those who fall outside of the average–a tall or short father–are “outliers.” The tall father will likely have a son closer to the average than he is–thus, shorter. The short father will likely have a son closer to the average than he is–thus, taller. The average of the two of them then drifts, or regresses, toward the average height of the overall population.

Regression isn’t cause-and-effect–I did well this time, so I must do less well next time. Rather, it’s a phenomenon of perception. It occurs because some events randomly “drift” around a central average–like the heights of individuals, days with earthquakes, rain patterns in particular cities, or flips of a coin. In the long run of “normal events,” extreme events do pop up, as the unlikely result of random chance.

Imagine flipping a coin 500 times. At some point, you might get a string of 5 or 10 “heads” in a row. It’s a pretty unlikely, and therefore notable, event. And when such a thing happens–a string of coin flips, an unlikely win in a game (or an election), a major disease–our pattern-seeking brain sees the events and thinks, why? Has the coin had suddenly developed a malfunction, or transformed spontaneously into a double-headed coin. Is this the new norm? Every time you flip the coin, it comes up heads? Has the coin–or the world–permanently changed (like a disaster movie)?

But then–we find out life hasn’t changed after all. You flip the coin and it comes up tails. Politics returns to “usual,” albeit with some slightly different tendencies. People rebuild after earthquakes. We return to post-pandemic “normal life,” and companies whose stock did well during the pandemic plummet “back to reality.” We regress to the mean, the average, the typical. Regression is us perceiving this return to the average, regular, run-of-the-mill situation.

Think about this in the context of eschatology. Regression is when we realize we are still in the “not yet” period. There’s a lot of talk from time to time about eschatology, signs of the times, and an inexorable arc toward the “end of the world”—apocalyptic thinking. We think we see a pattern in events, much like the patterns we see in the clouds. Then, the pattern doesn’t emerge. One day, the clouds will part - but this is not that day. And so we regress to the mean—almost with a sigh of despair.

Regression applies to the weekly Roundup, too. Events seem to be running pell-mell in a trend toward a conclusion, hurtling like a train toward a trainwreck. But then–people draw a breath. The train is brought to a stop. There is no revolution, no overthrow, no overnight change. The frog hops out of the pot. The government quashes the protest, offers an unforeseen conciliatory olive branch, or opens negotiations–because people in power for a long time know a thing or two about staying in power, and can endure a bit longer.

Regression to the mean is why I’m not really surprised the Tigray war finally ended—or that it is ending with a bit of a “whimper,” with lots of isolated data points and horror stories. Or that the balance of power in Iran seems to favor the government, but with some concessions being floated in the wind. Or that the Chinese protest seems to be ending the way it is. I was surprised China offered such seemingly massive conciliatory gestures, and I wonder how it will play out. Regression reminds us “the story is not yet done”—it’s still being written, and tomorrow may show today’s plot twist, which feels so huge, to be just another wrinkle.

The challenge with regression seems to be determining when a series of events is a “black swan outlier” and when it really is a game changer. The 1914 flu and the 2020 pandemic were not as world-changing as, say, the plough, or the Internet, or the iPhone. Some things do change the arc of world history (at least from our perspective).

Two two interesting items in this week’s Roundup fall into the curious category: the long-term climate issues affecting both Libya and Somalia. Governments use resources (people, money, technology, etc) to do things for people or to people, to cause people to do something, or to prevent them from doing things. But if a change in the availability of resources leads to either (1) no resources for the government or (2) people to leave—as in Somalia—it can create a fundamentally different situation, or even to the fall of a government.

The CBC piece on Somalia said the people see the droughts as merging. There was an “old normal” in which droughts were expected and common but there was also fruitful space between them. Now, people seem to think “normal is changing.” After five years of drought, people doubt the rains will return. They are leaving and making no plans to come back. That could fundamentally the situation in the Horn.

The Roundup is partly about trends that affect access to unreached peoples. The Somalis and the Libyans, who have been for years very unreached people, are seeing a major shift in their situation. This shift may alter their lives, and alter the equation of access, in generationally-defining ways.

Or, the rains may return, and we may regress to the mean. I’m not comfortable with making an estimate on that, although I think the probability leans more toward “the future has changed.” We watch and see, but now may be a time to be planning for a future where far more Somalis and Libyans are outside their country, in an enduring way, than inside. Nations, as Ralph Winter reminded us, aren’t geographic locations, any more than churches are. They are people.