Columbia's Extremely Civil Unrest
Was the Columbia protest a big deal? It’s a rounding error in America’s long and bloody history of violent protest, so let’s not get too worked up about it
The past few weeks saw large antiwar protests erupt on college campuses across the country. The protests culminated in the occupation of Hamilton Hall at Columbia University by students and “professional” agitators, resulting in 112 arrests.
I’ve had a few conversations since the Columbia events with friends who have been deeply worried by the protests, from both the left and the right. Both sides seem to be worried that the protests are emblematic of how deeply divided the country is.
Are these protests the latest in a series of rapidly escalating conflicts that will ultimately result in the country’s split? Between the George Floyd protests in 2020, the January 6th Capitol riot in 2021, and the current Israel protests, how long can the country be expected to stay together with so much civil unrest?
The 1960’s are well-known as a turbulent decade, and there are frequent comparisons between our current civil unrest and the unrest of the 60’s. How accurate are these comparisons? And is it true that from a civil unrest perspective, the 60’s (or now) are the worst it’s ever been?
What does the data say?
Luckily, Wikipedia has a great list of US civil unrest incidents going back to 1783. I scraped the data to create a graph showing the number of deaths related to civil unrest by decade going back to the founding of the country.
The colored bars represent the number of deaths by type of unrest event for each decade, and the dark red line represents the raw number of unrest events for each decade.
The big takeaway is that deaths from protests and civil unrest have dropped dramatically since peaking in the late 19th and early 20th centuries. From 1860 to 1929, the US averaged 458 deaths from civil unrest per decade (median 330 deaths). Since the start of the 21st century, the average has fallen to just 15 deaths per decade, representing a 97% decline in average deaths.
Some quick notes on methodology
Civil unrest is defined here as domestic incidents involving a crowd or mob in which public order fails to be maintained, resulting in property damage, injuries, arrests, and/or deaths. This data obviously excludes the US Civil War, but it also excludes assassinations, non-mob related instances of domestic terrorism (1920 Wall Street bombing, 1995 Oklahoma City bombing, etc.), instances of non-ideological “frontier justice”, and non-violent protests. Included are violent or destructive protests, rebellions, mob violence, violent labor strikes, and other similar violent incidents involving crowds.
I chose deaths as the metric to compare over time because they are the most widely reported, least subjective, and most easily verifiable of the various data points associated with gauging the severity of civil unrest. I know it isn’t a perfect measure of severity, but for the purposes of this post I’m assuming it’s a good proxy of civil unrest severity and broader societal discontent and anger. The angrier you are about a cause, the more willing you’ll be to kill or die for that cause.
Finally, even though deaths are the most verifiable of the metrics available, there are still estimate ranges for some events. For these, I took a simple average of the high and low estimates. Most of the estimates are pretty tight, with the major exception being the NYC draft riots in 1963. The graph shows a death toll of 660, which is the average of the official toll of 120 and a high estimate of 1,200, so the 1860’s bar above has a larger margin of error.
Back to the commentary
The 21st century has seen a 97% decline in civil unrest deaths since the 1860 – 1929 average. Over the past 10 years since 2015, which is when many people consider our current polarization trends to have hit high gear, the US has had 35 deaths related to civil unrest1, which is still a 92% decline from the 1860 – 1929 high-water mark.
So, was 1860 – 1929 just unusually violent? This was definitely a violent period, but even when taken in totality, the rest of US history has been extremely violent compared to the 21st century. From 1780 – 2009, the average deaths per decade was 181 (median 76). Even compared to this longer-term average, the past 10 years is still 82% lower.
In fact, there have only been two decades in the entirety of the 19th and 20th centuries that have had fewer deaths than we have had in the past 10 years – the 1950’s (with 8 deaths) and the 1820’s (with 0 deaths, assuming this older data is complete). Every other decade from 1800 to 1999 has had more deaths related to civil unrest than we have had in the past 10 years.
The dark red line above shows that even as protest deaths have been declining, the number of civil unrest events has been trending somewhat upwards, peaking in the 1960’s with another local peak in the 2010’s.
The graph below combines these two trends to show civil unrest deaths per unrest event.
Over the past 10 years, the death rate has been 1.1 per event, which is lower than every decade since the 1820’s, including the 20th century low point of 1.6 deaths in the 1950’s. The rate for the 21st century as a whole is even lower, coming in at 0.6 deaths.
Even if you think the past 10 years is too generous a time frame and would rather just focus on 2020 – 2024, the death rate has only climbed to 2.0, which is higher than the 1950’s, but still lower than every other preceding decade going back to the 1820’s.
So, what has driven this decline? There are a multitude of answers, including more effective and non-lethal protest policing, the glorification of protest, and Steven Pinker’s "Better Angels". Another obvious explanation is that the things people are protesting today are simply not as worthy of dying for as they used to be. Nat Turner revolted in 1831 to free himself and others from slavery. The Great Railroad Strike of 1877 was launched to secure a literal living wage for 100,000 workers across the country in the midst of a ruinous economic depression. 1967 and 1968 saw thousands march for basic civil rights for African Americans. As Americans’ lives have gotten dramatically better over the past 200 years, and especially over the past 30 years, the number of reasons to be willing to die in a protest has fallen dramatically. Are there still problems? Yes. Are there as many problems worth dying for? Probably not.
Aren’t recent protest deaths increasing?
Some might argue that the increase in civil unrest deaths from 1 in the 2000’s to 7 in the 2010’s to 28 in the 2020’s is cause for concern. They’re right. We should be concerned about this increase.
At the same time, we should remember that in the three decades leading up to the year 2000, there were 194 civil unrest deaths, more than five times the number we’ve had since 2000.
Every death is a tragedy, but the recent increase in civil unrest is simply not cause for proclamations about the end of the country. The United States has survived hundreds of protest deaths per decade for much of its history. A bit of broken glass at Columbia shouldn’t jeopardize things now.
Why all of the doom and gloom?
So why do negative narratives about our civil unrest feature so prominently? Shouldn’t we be glad that we’re nowhere near where we used to be in terms of political violence? Shouldn’t we celebrate a 97% decline in civil unrest deaths?
As many of you might guess, negativity bias plays a massive role here. Our drive to seek, internalize, and be consumed by negativity causes our perception of the world to be warped. Our warped perceptions stubbornly persist even when the data contradicts this pessimistic outlook.
Can it be the case that even though people are less likely to die in civil unrest, we are still just as negative and polarized and as close to coming apart as we’ve ever been? Yes, I’m not denying that we are polarized.
I am arguing that based on the data, we should be much less negative and polarized than we are. Things have been much worse in the past. The extent to which we think things are bad now is more so a function of our out-of-control negativity bias than it is a realistic reflection of the facts on the ground.
A long-term historical perspective is a great tool for dispelling pessimistic narratives. The historical narrative here is clear and worth repeating:
Deaths from civil unrest over the past 10 years are 92% lower than the 19th and 20th century historical average
Deaths from civil unrest since the year 2000 are 97% lower than the 19th and 20th century historical average
The total civil unrest deaths we’ve seen across the entire 21st century is less than the number of deaths we saw in every single decade for the past 200 years except for two: the 1950’s and 1820’s
Why does this matter?
When overly pessimistic narratives are repeated, we run the risk of self-fulfilling prophecy. Fears that we are seeing more political violence now than ever before are unfounded, but unless these pessimistic narratives are pushed back against, these fears may become reality.
Which brings us back to Columbia: Zero deaths. Minimal property damage. Less than 115 arrests. From a historical perspective, this protest is a drop in the bucket. People are divided, but things have been much worse. Remembering this historical perspective is a good way to ensure we don’t continue to get more divided.
Last 10 years total includes 6 deaths in 2016, 1 in 2017, 21 in 2020 related to the George Floyd protests, 5 in 2021 related to the January 6th Capitol attack, and 2 more separate deaths in 2021.