New at HuffPost

Saturday, March 31st, 2012

I have a feature-length piece up at Huffington Post looking at the myriad issues to spill out of the Collinsville, Illinois traffic stop of two Star Trek fans last December. It delves into asset forfeiture, the effectiveness of drug dogs, and how difficult it is to get rid of a bad cop.

I also wrote an accompanying sidebar in which I examined the data of one Illinois State Police K9 unit over an 11-month period in 2007 and 2008. The dog had a false alert rate of 28 percent. It rose to 74 percent if you include cases where it alerted to drugs, but police could only find “residue” or “shake.” (Which of course wouldn’t have been tested for verification.)

I suggested “Set Phasers To Violate” for the main article’s headline.

Alas, I was overruled.

Digg it |  reddit |  del.icio.us |  Fark

26 Responses to “New at HuffPost”

  1. #1 |  Mannie | 

    It rose to 74 percent if you include cases where it alerted to drugs, but police could only find “residue” or “shake.”

    I would argue that those are valid alerts. The dog smelled something and alerted.

    My objection to drug dogs is that you cannot prove what they are alerting to. Dogs are incredibly responsive to handler input, so they are suggestible. I know of no way to counteract that or proof against it. That’s what dogs do. So a drug dog can’t be demonstrated to be much better than a blind hunch.

    I;m more sanguine about explosives dogs. At least, if you’re using dogs to find explosives, you’re searching for something that is going to blow someone up. There, I’d be more worried about the miss rate.

  2. #2 |  Radley Balko | 

    Except that the dog also had a high false alert rate when there were no drugs at all.

    So I’m not sure we can be certain that the officer was correct every time he claimed to have found shake or residue. In the Huff stop, what Reichert claimed to be shake didn’t look like pot at all.

  3. #3 |  Pablo | 

    No doubt in my mind this was profiling–based on out of state tag and the make/model of the vehicle. LOTS of Chrysler 300s get pulled over for fishing expeditions like this. Also Dodge Charger/Magnums and Ford Mustangs.

  4. #4 |  (B)oscoH | 

    The only way this could be better is if it were an illustrated story book. Nice work Mr. Balko.

    So here’s the thing that just confounds me. Any private company that was collecting the treasure trove of live performance data that the dashcam video is would be using it for ongoing quality control purposes. New employees and those with a recent history of problems would be subject to more scrutiny. It’s not an excuse to fire people, but to help them improve and conduct themselves consistent with organization goals and values.

    Which begs the question… Do police agencies do ongoing quality control using collected field data? If not, are police unions the problem or is management just not in this millenium yet? If they are, and this passes for consistent, we’re pretty screwed.

  5. #5 |  AlgerHiss | 

    And once again, the Collinsville Chamber of Commerce needs to be advised of what is thought of their shitty little town:

    http://discovercollinsville.com/

  6. #6 |  Arthur | 

    #4

    This is why you CANNOT conflate public and private sector employees/entities. Where exactly is the incentive for police agencies to do ongoing quality control using collected field data?

    Until the true supervisors, the public, take a supervisory role there will be no accountability or improvement.

  7. #7 |  freedomfan | 

    Excellent work, Radley! And, the sidebar is more ammo for those who want to make the legal case that drug dog sniff alerts are insufficient cause for a search.

    BTW, Mannie,

    It rose to 74 percent if you include cases where it alerted to drugs, but police could only find “residue” or “shake.”

    I would argue that those are valid alerts. The dog smelled something and alerted.

    That seems like a poor assumption, as there is no evidence for it. In the Huff case, it’s clear that some officers are willing to classify basically any debris in a vehicle as drug residue to justify the alert the K9 gave, probably in response to an officer’s perceived desire for an alert. In practical terms, the whole notion “shake” (which might more honestly be termed “shake-down residue”, given that it’s a by-product of the asset forfeiture scam) is no different from claiming the dog alerted to magic fairy dust.

  8. #8 |  (B)oscoH | 

    #6: Well, I’m not conflating them. I am saying that if this kind of behavior is widespread, that it is simple and ultimately in the best interest of these local departments to measure and correct. You don’t have to resurrect William Edwards Deming or implement Six Smegma to do basic quality control. The local shift manager at McDonalds probably has enough insight to curb what looks like flagrant and callous abuse by at least one officer parking his ass on this Interstate.

    One filmmaker (I’m guessing) used an FOIA request or some evidentiary request to get the dashcam video of his encounter. Most people who have this kind of experience would just get over it and move on. It is not feasible for the public to grab a meaningful sample, review, and hold a department accountable. At the same time, quality control needs a transparency and accessibility to instill confidence from the public.

    I’ve been listening to Penn Jillette’s podcast lately. In last Sunday’s episode, he talks about getting pulled over in his Nissan Leaf by a Vegas cop recently. He makes the point that we should not be afraid of cops, nor cops of us. They are a necessary piece of keeping the peace in society, and for most people, there should not be an adversarial relationship. It’s a lesson he learned as a kid who had several interactions with the Five-Oh. And Penn’s libertarian credentials are beyond question. I think he is right. And I think Radley is right to bring light to so many of these awful abuses caused by bad officers and bad public policy. Ultimately, the way to Penn’s ideal is through sunshine.

  9. #9 |  MTB | 

    Radley, if you don’t mind my asking, how have your ratings/viewings been at huffington? I was excited to see you get a much larger audience, no offense to theagitator.

  10. #10 |  JThompson | 

    It rose to 74 percent if you include cases where it alerted to drugs, but police could only find “residue” or “shake.”

    I would argue that those are valid alerts. The dog smelled something and alerted.

    Anyone that’s ever had “shake” found in their car knows what it means: Any dust, powder, dirt, sand, gravel, or dried organic matter (leaves, grass clippings, dog crap, etc), that looks vaguely like residue or seed.

    It doesn’t mean drugs were found. It doesn’t even mean drugs were ever anywhere near that car. It just means the car’s owner needs to vacuum their floorboard.

    Seriously, I’ve seen a cop ask a friend why he had “shake” in the car while pointing at what was clearly the better part of an oak leaf and I’ve had one ask me about “seeds” while indicating pea gravel. It’s to see if you’ll freak out and scream “Oh god the drugs are in the trunk!” I guess.

  11. #11 |  Burgers Allday | 

    The Trouble With Popos.

  12. #12 |  Jerryskids | 

    The 28% failure rate is of course only the false positive rate. There is no way to know what the false negative rate might be. I suspect that dogs aren’t much more reliable than chance – assuming the handler has a reasonable suspicion in the first place, there is a good chance the dog will alert.

    Explosives dogs in some situations might be more reliable assuming the handler has no idea where explosives might be. As explosives dogs become used more for detecting trace residues on people, I suspect the alert rate for brown-skinned young males with funny accents will skyrocket.

  13. #13 |  JOR | 

    “Do police agencies do ongoing quality control using collected field data?”

    Almost certainly.

    If it doesn’t seem like it to you, your error is almost certainly in your assumptions about police organization goals and values.

  14. #14 |  Aresen | 

    @ Jerryskids | March 31st, 2012 at 9:08 pm

    With explosives dogs, I am fairly certain that a false negative would soon become obvious. ;P

  15. #15 |  Bob | 

    #13: JOR

    “Do police agencies do ongoing quality control using collected field data?”

    Almost certainly.

    If it doesn’t seem like it to you, your error is almost certainly in your assumptions about police organization goals and values.

    What? Is it not blatantly obvious that Police Agencies not only DO NOT execute any ongoing quality control in these areas, but actively strive to prevent it?

    They don’t want or need accurate data or quality control… they want and need mechanics that allow searches and seizures.

  16. #16 |  susan hawke | 

    Radley–I’ve been having similar problems in Illinois because I recently moved here from Washington State. I blogged one experience that’s been published on a Web site. I titled it, “Interrogation Nation.” You definitely have a story here (I’m a former journalist).

    My recent civil right affronts story is the same: Getting pulled over for bogus reasons (“not speeding up fast enough” once out of a SPEED TRAP construction zone; not changing to right lane even though a semi-truck was blocking me–we would’ve crashed, etc. etc.). Then I comply with officer’s request, then comes the “Have you ever smoked marijuana?” and “Isn’t medical marijuana legal in WA State?” and other interrogations that have no probable cause nor anything to do with initial traffic stop. There have been other inappropriate (at best) nuances to these stops, as well. Contact me for more details. Ohio & Indiana are bad, but Illinois is the worst. I discovered this on a recent Nov. 2011 trek from Seattle to New York City. Susan

  17. #17 |  JOR | 

    If they want mechanics that allow easy searches and seizures, and they manufacture and continually improve such mechanics, then there’s a lot of things they’re doing wrong, but quality control is not among them.

  18. #18 |  jmcross | 

    A nice pair of articles. Some interesting comments on the dogs story. Several folks with tales of their own K9 cop drama.

  19. #19 |  AlgerHiss | 

    An aside, Collinsville seems to be a really crappy place. Here they are banning saggy pants:

    http://www.videoandcomment.com/v/iUoFZVhz48M

  20. #20 |  Jerryskids | 

    #14 Aresen With explosives dogs, I am fairly certain that a false negative would soon become obvious. ;P

    That’s a good point. Since the military uses dogs, I am sure it would be easy to find stories of dogs finding IEDs and probably some figures as to how many they have found. My guess is that you won’t find the figures on how many times they have falsely alerted or failed to alert, but given the military’s penchant for record keeping I would bet you that the Pentagon knows exactly. Good luck getting that information, though.

  21. #21 |  Other Sean | 

    Something else to think about with this, is the strangely reliable instinct police and police departments have when it comes to picking targets who are less able to fight back. They picked the wrong two nerds in Collinsville, but that only serves to show how rare it is for anyone to do anything other than suffer and keep quiet (as BoscoH already said).

    Just looking at the incentives for asset forfeiture money, you’d expect to see a long list of incidents where the greed for pirate loot tempted the cops into fights they couldn’t win, with clumsy pre-text stops against high profile businessmen, doctors, lawyers, junior league women, and the like. Yet other than the odd racial incident every few years, that doesn’t happen enough to make an impact.

    Maybe that’s what the dogs are really trained for: sniffing out people at or below the lower bound of the middle class, who don’t have the political access or the funds to lawyer up properly, and who by themselves won’t make a terribly convincing witness against the state.

  22. #22 |  picachu | 

    “I suggested “Set Phasers To Violate” for the main article’s headline.

    Alas, I was overruled.”

    That would have been epic! Or maybe this “Stardate 2012, While traveling through the land of the free…”

  23. #23 |  Christ on a Cracker | 

    I was crossing the border from Tecate to US about a year ago. The car right ahead of me (white, mid-size, four-door sedan, probably a rental, unremarkable) was “alerted to” by a drug dog. The handler kept patting the fender and doors, causing the dog to jump on the areas where he was patting. He was absolutely showing the dog where to jump, just like when I want my dog to jump up on the couch. The handler glanced up several times towards the cameras. Clearly he was putting on a show.

    They pulled the car into the “further investigation” areas. I don’t know what happened.

  24. #24 |  Christ on a Cracker | 

    Re: (B)oscoH @#4

    Quality Control, from the cops. That’s funny.

    What incentive would cops (and prosecutors) have for QC? If there’s a conviction, they they win. If not, the “SOB got away with it. We’ll get him next time”.

  25. #25 |  Deoxy | 

    #15 – I think I have to agree #13 JOR – you’re misunderstanding their definition of “quality”.

  26. #26 |  Sinchy | 

    This requires a media sting operation.

Leave a Reply