Dog Scent Lineup Gets Wrong Guy, Junk Science?

In Texas, police used a dog scent lineup to get the perp in a murder case. The dogs got it wrong.

According to the dogs, the perp was one of the cops. He was under suspicion for five months. Then, DNA revealed the real murderer who pleaded guilty.

The Innocence Project of Texas calls the practice "junk science that's being used by prosecutors and judges to convict people." The nonprofit group... wants state governments to ban the use of dog-scent lineups. It says an unknown number of people have been wrongly accused or convicted from the dog-scent lineups.


How the lineup works:
Dog-scent lineups are similar to visual lineups; but instead of a witness picking a suspect from a group of people, bloodhounds walk along a line of tin cans containing individual scents from possible suspects.

Investigators get the scents from rubbing a gauze pad on someone's body or clothes, and that gauze pad is then placed in a tin can. The dog handler gives the bloodhound the scent they're looking for, and then the handler and animal walk down the line. If the dog matches the scent, dog handlers say the animal will give a "sign," which is usually stopping at the can or barking.

< New Report on Plea Bargains | A Taste For Bluster >
  • The Online Magazine with Liberal coverage of crime-related political and injustice news

  • Contribute To TalkLeft

  • Display: Sort:
    I think that people are so amazed (5.00 / 1) (#2)
    by eric on Mon Oct 05, 2009 at 11:25:38 AM EST
    by what dogs CAN do with their noses, that they begin to think they can't make mistakes.  Let them stick to tracking and locating buried avalanche victims.

    My dog will find food and treats (5.00 / 1) (#3)
    by scribe on Mon Oct 05, 2009 at 11:31:41 AM EST
    almost infallibly.  The only problem is that the treats might have been in the pocket of the person she sniffs out - like two weeks ago.  That does not deter her from going into her "absolutely certain of food being present" mode, slobbering and sitting looking all pretty and pretty-please.

    But she's still wrong.

    And, she's a bird dog who pays no attention to birds, but does point cats.

    Dogs respond to cues their master gives them, which cues might be wholly unconscious on the master's part.  And they will alert to please their master.  Or to wheedle a treat.

    Keep the dogs to finding people buried in rubble and missing kids in the woods.

    I think it depends on the team (5.00 / 1) (#5)
    by nycstray on Mon Oct 05, 2009 at 11:43:30 AM EST
    are these team trained specifically for line ups or are they scent dogs from other areas brought in to do a line up?

    And while there might not be a treat present in the pocket, the scent is still there, so your dog is in fact right that the scent of a treat is there ;) And you could take that skill and put it on cue. I actually did put my dogs nose on cue. Came in handy in the park. She would lose her brain over all the scents, so it was easier to train her after we got that under control. I swear she was a hound in her former life . . .


    You pointed something important (5.00 / 1) (#8)
    by Fabian on Mon Oct 05, 2009 at 12:59:32 PM EST
    out.  Dogs can detect old scents - which means that they can alert on scents from the past.  Anyone who had significant contact with the alleged perpetrator could have been picked out by a trained dog!

    Good point. (none / 0) (#12)
    by scribe on Mon Oct 05, 2009 at 02:32:53 PM EST
    My setter is good for up to 2 weeks when it comes to biscuits, and maybe more for Beggin' Strips (TM).

    I have no idea how long she would be good for real bacon....


    And when she's doing her (none / 0) (#14)
    by scribe on Mon Oct 05, 2009 at 02:38:16 PM EST
    "pretty sit give me treats" you get something that looks like this, only she'll stare right through you.  When she's straining at the leash to get at someone's pocket, you get something that looks like this.

    beautiful pup! (none / 0) (#18)
    by nycstray on Mon Oct 05, 2009 at 04:05:35 PM EST
    I was wondering how they handle that (none / 0) (#19)
    by nycstray on Mon Oct 05, 2009 at 04:10:41 PM EST
    since it should be pretty obvious if a person had contact with a victim through friendship, family etc, there's a good chance their scent could be around. Identifying their scent present should just be one piece I would think. If you didn't have any items from the actual murderer, the dog wouldn't signal that that scent was present also . .

    Almost seems like it would be better for a dog to eliminate suspects (who had other possible evidence against them) vs Id-ing . . .

    Was this guy ever charged?


    Sounds like a Roman Auger. (none / 0) (#1)
    by Salo on Mon Oct 05, 2009 at 11:15:44 AM EST
    How was this allowed into law? Like ever?

    Junk science or human/canine error? (none / 0) (#4)
    by nycstray on Mon Oct 05, 2009 at 11:35:34 AM EST
    Critics of dog-scent lineups say the problem is that dog handlers aren't certified or regulated and that there isn't a system in place to check a bloodhound's track record.

    Are there any stats to back up the "unknown number of people have been wrongly accused or convicted from the dog-scent lineups"?

    Talking about junk science (none / 0) (#6)
    by caramel on Mon Oct 05, 2009 at 12:38:54 PM EST
    I sent you about 3 articles about the Willigham situation (last week) that is turning into a political scandal in TX but you didn't seem interested. Gov. Perry removed three experts from the Forensic Science Commission less than 2 days before the hearing on Todd's case was to take place - including the head of the commission. Talk about junk science...

    Sounds like a bad idea (none / 0) (#7)
    by Fabian on Mon Oct 05, 2009 at 12:56:19 PM EST
    from the start.  Dogs have an amazing sense of smell, but they can't talk - so they can only alert based on their training.  Who knows what training they get, or who trained them, or how good their handler is!  

    teach the cop (none / 0) (#9)
    by Capt Howdy on Mon Oct 05, 2009 at 01:16:56 PM EST
    to have milk bones in his pocket

    I could tell (none / 0) (#10)
    by Fabian on Mon Oct 05, 2009 at 01:45:53 PM EST
    how many pets you had by how thoroughly my dog went over you.  A quick pass?  No pets.  A two minute, centimeter by centimeter inspection?  Multiple pets including at least a couple of dogs.

    yep (5.00 / 1) (#13)
    by Capt Howdy on Mon Oct 05, 2009 at 02:36:03 PM EST
    3-5 minutes or until you drag them away means a cat or a ferret a rabbit or some other "food animal"

    "what 's that lassie?" (none / 0) (#11)
    by Salo on Mon Oct 05, 2009 at 01:52:16 PM EST

    "arrrfff arrfff."

    "Bricktop Bob here chopped their heads off and threw 'em down the well? You are nicked old chum."

    "arrrff arrff."

    "yes Lassie a milk bone as a reward is in order."

    "arrrfff arrfff."

    "Yes bacon flavor."


    The last TL on dog sniffing, iirc, (none / 0) (#15)
    by sarcastic unnamed one on Mon Oct 05, 2009 at 02:50:33 PM EST
    focused on what seems like a more obvious uses like sniffing out pot hidden in a car.

    I would much more easily question the use of dogs as described in today's thread.

    I am surprised that dog-sniffing 'evidence' would be used in a trial to convict somebody, or, rather, I am surprised a jury would accept such "evidence" - but apparently they do.

    Here's similar recent case where a mistaken conviction was due to reliance on dogs.

    It also could have been a ruse. (none / 0) (#21)
    by scribe on Mon Oct 05, 2009 at 07:23:59 PM EST
    Remember, if you can, one of the early episodes of Homicide:  Life on the Street.  The cops want to get some mope (who doesn't want to invoke Miranda and would rather try to talk his way out of it) to confess.  So, they load the paper tray in the copier with paper on which they have pre-printed "Truth" or "Lie" in big block letters.  They take the mope over to the machine, tell him it's really a lie detector (they might have rigged some wires or something), then tell the mope that all he has to do is put his hand under the cover and answer the questions, and the machine will tell them whether he's being truthful.

    They persuade the mope it's safe, and then proceed.

    First question, they ask him his name, he answers and they press the "copy" button.  Out comes a sheet "Truth".
    Address?  He answers.  Paper says "truth".
    Were you at the scene.  Answers "no".  Paper says "Lie".
    Mope starts sweating.
    Another question, another truthful answer, "truth".
    "Did you shoot the dead guy?"  Answers "no". Paper says "Lie".

    After a few more iterations, Mope breaks and confesses.

    This dog stunt could have been the same thing.  Not so much as to get the guy with admissible evidence derived from the dog, as to get the "guilty" guy - the one the dog "designates" - to confess more or less spontaneously through the application of "technology"/


    I have respect for the value of dogs... (none / 0) (#17)
    by BJohnM on Mon Oct 05, 2009 at 03:48:56 PM EST
    ...in law enforcement, but they have to be properly handled.

    Just this weekend we watched one of the Cops shows on Spike. It was one we'd seen before and it infuriates me. I don't remember the jurisdiction, but some under-cover cops stop a guy on a bicycle who was supposedly reported to be selling drugs. When patting him down they found a nice stash of crack or something.

    But they also found like $1,500. Probably all the guy had to his name. They told him if the money had the scent of drugs on it, they could seize it. So they called in a K-9 unit, and stuck the money inside the gas cap bay. Then walked the dog around the car. The K-9 officer never waved or touched the car until he got to the gas cap. The dog didn't react until he tapped on the cover, and...you guessed it, Fido "alerted."

    Combine that with the study by some academic institution that something like 80% of currency has traces of illegal drugs on it (with D.C. having the highest percentage), I hope the guy had a decent attorney, who had access to that film, but I'm not betting on that.

    Make no mistake, these dogs want to please and get their treat, so it doesn't take much for them to learn to pick up even subtle signals from their handlers.

    I saw some footage of trained military (none / 0) (#20)
    by nycstray on Mon Oct 05, 2009 at 04:21:56 PM EST
    bomb dogs. Lil' different from your Cops as far as handling goes. You would think for a sniff-up they would need to be handle more along those lines.

    I didn't know they had dogs sniffing money in that way. You'd think a test to see if it was the same drugs the guy had on him would be more accurate. Especially in light of the study you mentioned.