NYPD Internal Protest Arrest Data
A look at two competing data sources within NYPD show wild discrepancies which shroud more mystery on last summer's unrest than it uncovers
Politics is the greatest enemy of truth. Yet we are a nation of political mythology. Just like the ancient religions, modern America’s political religions demand faith and fealty. It’s a phenomenon that doesn’t bode well for society but which gives comfort to the peon classes enough to never raise above their false gods. A famous German philosopher once said, “religion is the opium of the people.”
The year 2020 has become one of the most important mythologic years of the American canon. Although tomes can be written on that subject, this study will focus on last summer’s civil unrest in New York City and the actual ways in which the government obfuscated what really happened there. (As I can only speak to events I witnessed myself, this analysis should only be understood to describe New York City. I have no basis to know any version of the truth from any other city.)
Today we analyze some of the NYPD’s internal data, which unfortunately is extremely flawed. The data is inconsistent, is incomplete, is distorted, and can not answer basic questions. For instance, there is no distinction between protesters and looters, of which there is ample evidence to suggest they were different actors. Necessarily, any examination of the data conflates the political protesters with the non-ideological rioters, effectively hiding one set of data behind the other. The data is confounding, most likely by design. A government investigation into last summer criticized the NYPD for its shambolic data.
For instance, the data is so flawed that we can’t even say for certain how many people were arrested outside a range of nearly one thousand, let alone get any understanding of the scope of the riots. The NYPD – and the city government – are content, it seems, to hide behind their routine brutality of political protesters in order to conceal the public safety disaster they presided over during the riots.
Although there are multiple internal data streams out of NYPD, now we have access to two of them directly. The first describes arrest and injury data from the NYPD commissioner’s Office of Management Analysis and Planning, OMAP. The next source provides a competing take on the arrest data, it’s exact origin within NYPD is not known, it simply bears the title “2020 NYC Demonstration Arrest and Discon Summons Analysis.” We’ll call that one the discon sheet here. There are additional sources of NYPD data we do not have direct access to yet, but which were described by government investigators. These are the Mass Arrest Processing Center, MAPC; the Joint Operations Center, JOC; and the Intelligence Division, their data called “sitreps,” or situational reports.
But rather than provide more clarity, the data raise more questions. Each source is inconsistent with each other, even when purporting to describe the same thing. Sometimes the various data sources are inconsistent by a factor of four or more. For instance, the May 31 OMAP data recorded 108 arrests. The discon sheet recorded 116, and the Intel sitreps recorded 78. But the MAPC and JOC data recorded 325 and 349 respectively – the JOC data recorded nearly four and half times more arrests than NYPD’s own intelligence apparatus. Which count is more accurate? Why are they so different?
A graph from the Department of Investigation report showing the discrepencies between the various NYPD data sources.
The reasons why are nuanced but still speculative in part. For starters, what is an arrest? There are three levels of detainment: violations, which do not meet the criminal threshold; misdemeanors, the lowest criminal level offense; and felonies, the highest-level crime. Some of the data sources only measure criminal offenses, therefore excluding violations, which account for the greatest number of protest arrests. Some of the data sources combine all three. Other sources are ambiguous. And even with this in mind, attempts to decompile and cross-reference the sources still lead to inexplicable and wide discrepancies. Further more, some sources disaggregate the data by day, while other sources provide only an aggregate total of arrests for the entire 11-day period between May 28 and June 7.
Beyond simple arrest counts, differentiating arrests by charge proves another frustrating task. The OMAP data provides a daily break down of arrests by violation, misdemeanor, and felony. But where the OMAP data breaks down the specific charges, for instance burglary and assault, it only provides grand totals for the entire 11 days. To provide burglary arrest counts – in other words, looting arrests – in this fashion is practically useless, and police administrators would know this. To tell us that 475 burglary arrests were made in 11 days tells us nothing about how many looters were actually arrested during the riots, which spanned two nights. The data was reported this way not because the NYPD lacks granular details, nor because they lack competent statisticians, but most likely because the department infamously did not enforce burglary laws at all during the riots. To provide documentation of that would be to admit guilt of the highest failure.
Reporting problems like this riddle the entire data. But the data suffers other blind spots, either intentional or ignorant. For instance, the OMAP and discon sheet break down the arrest counts by race and gender. But there is no examination of class. What zip codes did the arrestees live in? How many were employed and what were their income brackets? How many of the subjects had prior arrests? These are all questions we would have answers for if the year was 1977. When riots broke out during the 1977 blackout, the government and scholars had the curiosity to ask these questions. In 2020 we have no idea to what extent class played a role in the rioting, and the time to collect that data may have passed forever. Furthermore, NYPD training states clearly that the name and address of any arrestee is public information. And yet we still lack any information of arrests by resident zip code, which could have provided a basis for a class analysis of the arrestees.
For all the creative ways the NYPD was able to mask the riots in the data it provided to government investigators, there was one data source the NYPD could not obfuscate: the weekly crime statistics, called COMPSTAT. This data is made public to all, and yet remarkably, none of the multiple government investigations cited this data. None of the government reports attempt to quantify the mass looting, instead leaving it an uncurious mystery. But the data is the closest thing we have at all to understanding the scope of the riots. Although the COMPSTAT data itself is not perfect, we can safely say that at least 619 stores were looted across Manhattan and the Bronx during the peak of the riots. Why didn’t any of the NYPD data incorporate basic crime statistics in any way? Why did none of the government investigators analyze the NYPD’s most visible public data stream? Maybe the investigators were ignorant, but the consequences are that the riots are being minimized – not even recognized as riots by the government – and one of the most important sociologic events in modern city history is being distorted, ignorantly or politically, with the outcome of suppressing the truth, recasting in its place a political caricature.
But why should the truth about the riots be hidden? In whose interest is it to pretend that there were no riots? How can it be in the modern era that the government’s analytic capability is less now than it was in 1977? Of course, we can only answer these questions speculatively, but the simple answer is: politics. There is no political interest to analyze the riots, let alone honestly. Instead we are left with competing sectarian myths about what happened, both the “left” and “right” versions of the story ignorant and subservient to their respective political dogmas. But not the truth.
Stay tuned tomorrow for the final day of NYPD Docs Clearinghouse week.
Illustration by John Bolger. Fair use imagery from the American Broadcasting Company, and Warner Brothers.