Skip to content
Link copied to clipboard

Marijuana arrests reveal smoke and mirrors with public data

A slip of the finger on a police administrator's keyboard can have huge consequences in marijuana arrest reports.

Chris Goldstein/ Philly420

A slip of the finger on a police administrator's keyboard can have huge consequences in marijuana arrest reports. Simple mistakes are usually the culprit, but on rare occasions there can be an agenda behind a sudden change in the numbers.

Because public data sets involving marijuana arrests are a frequent haunt of mine for research I invariably uncover a few discrepancies, two stood out in 2017 as some insight into the background work for this column.

Just a mistake at PPD

This fall, a data glitch presented in the online Pennsylvania Uniform Crime Reporting System (UCRS) seemed to reveal a major spike in marijuana and other drug arrests in Philadelphia during April 2017. The mysterious jump was reflected only in arrests for drug possession and manufacture. The numbers returned to normal in May.

After years of looking at the monthly trends, it was obvious that things were definitely out of place. Since the biggest jump appeared to be in arrests for selling marijuana, I contacted Philadelphia Police for comment.

Deputy Commissioner Dennis Wilson agreed that the data was confusing. He said there were no major narcotics operations in April by his department or other agencies such as the High Intensity Drug Trafficking Area (HIDTA) task force, which includes federal law enforcement like the DEA.

Wilson said the data left his team "scratching our heads."

During the first call Wilson offered two explanations: Perhaps it was just a particularly successful month for the narcotics division of PPD or there might have been a "new commander" spike related to a change in leadership in April at the Narcotics Division.

He also noted the 22 arrests that happened at a marijuana party that month in a speakeasy-style raid that I wrote about earlier this year.

But, even with those potential explanations this was an earthquake in years of arrest trends, so I planned to gather more information and comments.

The very next morning, Wilson called me back to say it was all just a mistake. According to the Deputy Commissioner, the new narcotics commander had simply typed in the drug arrests wrong.

After a closer look the numbers Wilson said the public UCRS database didn't match up with the Philadelphia Police department's internal, real-time arrest mapping. Digging further, Wilson said they found the error: The year-to-date totals were entered as the monthly totals for April.

The public UCRS data was fixed within 24 hours after Wilson's second call. Now the public (and wonky writers like me) can see that April's arrest numbers for PPD are back to normal.

Let's never forget that each one of those digits is a person; someone put into handcuffs and facing the criminal justice system for less than 30 grams of cannabis.

Every illicit drug possession arrest means money spent, and for our flawed criminal justice system, the arrests can mean money earned through government appropriations and referrals to private incarceration or even drug treatment.

Wilson thanked me for catching the anomaly. No biggie, nobody's perfect. Still, hopefully those errors would have been caught in a year-end audit before 2017's data was submitted to the FBI.

New Jersey

What happened to the numbers in the city of Camden falls into an entirely different realm. This year some local county employees went about rewriting history.

New Jersey does not have a public, online arrest database like Pennsylvania. The only source for general information is the annual Uniform Crime Report compiled by the NJ State Police. The most recent data available covers 2015; not exactly real-time.

Journalists can file Open Public Record Act (OPRA) requests to individual N.J. police departments, or simply call and ask.

That's what I did last March with the Camden County Police Department. After several weeks of getting bounced around in phone calls and emails my inquiries were directed to Dan Keashen, the spokesperson for Camden County government.

Keashen agreed to provide the data and (after several weeks) emailed a spreadsheet showing a full three years of drug possession arrests. It revealed that police were most commonly arresting people for cannabis in Camden and I wrote a column saying as much.

Eight weeks later, Keashen released a different spreadsheet for the same three years to a reporter in New Jersey and I noticed that Camden's arrest numbers had drastically changed.

Suddenly, in the new data sent by Keashen, the opiate and cocaine category had nearly doubled. When presented with the discrepancy Keashen explained it as an approved modification performed by an unnamed employee.

On Twitter, Keashen claimed that between March and May of 2017 a county employee shifted nearly 1,000 arrests in the 3 year data set from the "Synthetic" to the "Cocaine/Opiate" category because they had involved hypodermic needles.

The ad-hoc alteration by the  county employee of audited Uniform Crime Report data – numbers already submitted to the FBI and included in annual reports – presents a host of serious legal, ethical and procedural issues.

Keashen's claim that the arrests were re-categorized based on use of hypodermic needles is also odd because police officers should be listing the arrests by drug type, not the delivery device.

This brings up a broader issue: Each individual police department has their own method of listing the wide variety of drugs that cops encounter every day.  Meth and common pills like valium or Xanax, along with and opiate pills like Oxycontin, are often tossed into the "Synthetic" or the catch-all "Other" category with no uniform process.

Powder drugs, like some synthetic opiates or methamphetamine, are injected. Camden police may have been correctly classifying them.

There are more than 25,000 marijuana possession arrests recorded every year in New Jersey. This embarrassing, costly and often racially-biased reality is a big factor behind the current drive toward full legalization.

We should, at least, count them all correctly. After prohibition ends these are all the records that need to be cleared.

Chris Goldstein is a cannabis consumer advocate and writer. He formerly served on the Board of Directors at several NORML chapters, helped with Philadelphia decriminalization and now teaches the "Marijuana in the News" class at Temple University. Contact: Twitter @freedomisgreen or