The following article, “No Kings Day” protests turn out millions, rebuking Trump: Our unofficial estimate is that around 4-6 million people attended a protest event yesterday. Anti-Trump resistance is outpacing 2017″ by G. Elliot Morris, is cross-posted from gelliotmorris.com:

Dear readers,

The “No Kings Day” nationwide rallies against Donald Trump/for democracy on Saturday turned out millions of people.

That’s per a collective crowdsourcing effort led by Strength In Numbers, and involving many members of the independent data journalism community. We systematized reports from official sources, accounts from the media, and self-reported attendance from thousands of social media posts into a single spreadsheet. (Researchers, please take our data!)

As of midnight on Sunday, June 15, we have data from about 40% of No Kings Day events held yesterday, accounting for over 2.6m attendees. According to our back-of-the-envelope math, that puts total attendance somewhere in the 4-6 million people range. That means roughly 1.2-1.8% of the U.S. population attended a No Kings Day event somewhere in the country yesterday. Organizers say 5m turned out, but don’t release public event-by-event numbers.

Of course, crowdsourcing data isn’t perfect; some local reports may be inflated, and others undercounted. And the formula we use to project attendance in places where we don’t have data assumes they are similar to the places where we do. That’s a necessary assumption, but an assumption nonetheless.

So this is by no means an official tally. But we do think it’s the most comprehensive tally currently available. Hundreds of data-gatherers have been compiling accounts of event attendance and checking them against available sources since Saturday morning. From a journalism perspective, this approach at least standardizes measurement and provides references to check our math, even if it doesn’t completely avoid the usual pitfalls of estimating crowd size (or the assumptions above). But in this case, we’re interested in speed and thoroughness, not perfection.

According to organizers, over 2,100 events were held under the No Kings Day banner on June 14. Some events appear to have had well over 250,000 people in attendance. Officials report 1m people in downtown Boston yesterday, but some of those were attending pride festivities. There are reports of nearly 100,000 attendees in both San Diego and Minneapolis-St. Paul, and multiple hundreds of thousands in New York.

While no one can produce official data on the number of people attending yesterday’s protests (that would require some sort of controlled entry and check-in system), we do have nearly official counts of the number of protests being held.

We have that information for yesterday, and we also have it for every day since January 1, 2017. That’s thanks to data gathered and published regularly by the Crowd Counting Consortium, a joint project of Harvard Kennedy School and the University of Connecticut.

According to the CCC, there have been over 15,000 political protests since Donald Trump’s second inauguration this January. Over the same period in 2017, during Trump’s first term, there were barely over 5,000 protests.

Protests have been broad, and large.

With our preliminary counting, the turnout at yesterday’s No Kings Day events nationwide rivals, and may exceed, turnout for the 2017 Women’s March. The 2017 Women’s March drew between 3.3 and 5.6 million people, depending on the estimate, making it the largest single-day protest in U.S. history. Our early numbers suggest No Kings Day may be in that range.

Total turnout in the No Kings Day protests is likely to fall short of the famous 3.5% population threshold for forcing action via mass protest. But the cool thing about that work is that the scholars find that smaller mobilizations of 1-1.5% of the population still have a 40-60% chance of accomplishing their goals.

Both the number of protests and their massive size are warnings for the Trump administration, which has routinely trampled the limits of public opinion during the president’s second term. On immigration, deportations, Medicaid/social spending, and democracy, the president has pushed policy much farther right than sanctioned by the U.S. public. The mobilized resistance across the country on Saturday is a real-world sign of backlash to his unpopular agenda.

Trump’s approval in our polling average is 44% today, the worst for any president at this point in their term (except Trump during his first term) going back to 1935.

If this is what resistance to Trump looks like now, not even 5 months into his term, he’s in for a world of hurt in next year’s midterms.

Thanks to everyone who helped us collect the initial data for our tracking and estimates. Like I said, they are free for anyone to use, and I hope they are helpful to people doing formal research on this subject.

Elliott