Extraordinary Claims Require

A great search dog is honored by the truth

It's annoying being a pretty rational, empirical person who is frequently regarded as a sort of witch doctor.

Depending on the particular religious leanings of the person in question, I'm either able to invoke cryptic Majikal Powers to Save the Day, or I'm a complete charlatan, on par with Jon Edwards and Miss Cleo.

Unfortunately, both believers and disbelievers are frequently in a position to make decisions that could save or kill someone.

It's hard to remember, after nineteen years of partnering with four different dogs' minds to exploit four different dogs' noses to accomplish feats that seem inexplicable to people who have not made a study of such things, that other people have no idea how this works.

So Perfesser Chaos and I -- and hundreds of our colleagues, in various ways -- make yet another part-time unpaid career out of studying, documenting, and explaining both the science of canine olfaction and the capabilities and limitations of real-world dog teams.

Part of that project is the collective work of establishing standards of competence. Because when you tell the chief of police that an operational dog team should be able to stay on task for x hours in y conditions, and cover that much ground to this probability of detection, it would be nice to know that the next "operational" dog team that shows up at his command post can do just that.

At this point, I cannot. A grim topic for another day. But those who don't meet basic standards of competence -- especially those who claim to have done so but have dodged any legitimate and objective quality controls -- are the #1 reason that responsible authorities (RA's) cast a jaundiced eye on those handlers who have actually done so. It's always nice to have been preceded by a slacker whose mouth wrote checks that his ass bounced.

Unfortunately, a related perennial duty of the thoughtful, legitimate, educated, careful, self-skeptical, rigorous search dog handler is to explain what a competent dog team cannot be expected to do.

Part of that is self-protection.

No Sheriff, given that our subject, an endurance speed-hiker with early-onset Alzheimers, has been missing for four days and we have a theoretical search area the size of Delaware, calling in one airscenting dog and handler, no trained search management, and no other field resources is highly unlikely to fix your problem at this point.

Part of it is an attempt to inoculate decision-makers against bullshit.

The bullshit follows a gradient.

I remember a search about fourteen years ago, one that dragged on for days and was stumping a good DCNR incident commander (IC) and the several trained searchers who stuck it out. We were having difficulty getting additional trained resources, especially canine teams, from the immediate area. The IC asked me whether it would be worth the expense to bring in an "expert" cadaver dog team from a four-hour drive away by helicopter. Because the handler and dog were, he'd been told, sooper speshul.

I had driven three and a half hours to respond to the search. No one had offered to pay for my gas. It had not occurred to me to demand an airlift. I knew who the handler was. I'd known her since her dog -- her first dog -- was a tiny pup. In the ensuing few years, she'd devoted plenty of energy to self-promotion in her own locale. I advised him that a celebrity dog team was unlikely to solve his problem, and that a sooper speshul "cadaver" dog was not the relevant resource for an ever-expanding wilderness search area. The IC ultimately decided that one small-area cadaver dog did not merit 800 gallons of helicopter fuel. (The missing man was found, very much alive, by a ground team on a Hail Mary task at the outer limits of the search area the next day.)

Now, the would-be jetsetting team was not by any means incompetent. As far as I know, there was nothing fishy about their unit's procedures for declaring them operational; the unit was generally regarded as legitimate and as employing the usual training and certification practices for the time. The handler wasn't claiming that the dog had supernatural powers -- just that its training was "better" than that of the dogs already deployed, so much better that the team merited star treatment. She was making that claim based on an unsupportable (and ultimately falsified) opinion about the nature of the search, and an unsupportable opinion about the nature of her dog and its training relative to everybody else.

Bullshit-meter reading, about a two* out of ten.

At around the same time, the canine SAR community was graced with the collegial presence of Sandra Marie Anderson, aka Sandy Anderson, international celebrity cadaver dog handler.

There were those of us who frequently cautioned about dogs and trainers who seemed too good to be true. And there was the zombie army of Anderson's students and partisans, who attacked all skeptics with accusations of "jealousy." We were engaging in "politics," see.

Anderson's claims about her Majick Dog became increasingly far-fetched. The dog, a Doberman mix named Eagle, never missed. He found tiny scraps of remains where other dogs had "failed." This was due both to his own inherent Majickalosity and Sandra Anderson's sooper speshul training methods.

She became the darling of the FBI. Always a bad sign in dog-handler land.

She jetted to other countries to deploy the Majick Dog in high-profile, politically-charged searches for murder victims.

Took quite a while for someone to definitively swing out the mirrors and reveal the flim-flam.

A Michigan woman once recognized as one of the nation's best trainers and handlers of cadaver-sniffing dogs was sentenced yesterday to 21 months in prison for planting bones and other fake evidence in cases she worked.

Yeah, when you are finding body parts that are later discovered to be still attached to the whole bodies, when none of the DNA of the bones you find matches that of the missing people -- but it matches stuff you got at your house, when the bloody saw blade at the putative murder scene has your blood on it -- well, it would have been nice if the authorities had bothered to run some DNA panels a little sooner. Most people frown upon framing someone for murder just to shore up your own delusions of grandeur.

She got a sweetheart deal in the plea agreement.

She's out of Federal prison now, I believe past the term of her supervised probation. After getting out of prison, she married a former SAR dog handler in southwest Virginia, Dan Crumrine**, and disappeared into the identity of Sandy Crumrine, or Sandra Crumrine, Cassondra, Cassondra Cummins, Ms's "C," or "Ms. Sande" -- or, if you know any other aliases, please add them in the comments. She's apparently directing a scheme to "train service dogs for disabled children" that both fundraises and charges families, and seems to be involved a "search and rescue unit" in SW Virginia and this organization, whose mission seems a little dissipated. As the local resident who uncovered most of the post-prison aliases and connections points out about the "credentials" on the various websites:

If you look at the groups listed they either do not have a web presence, are not really a group you join ( like AKC), or are websites that she appears to own.

(edit: I was able to verify by phone that a Sandra Crumrine is listed as an obedience judge by ASCA, though the person who answered the phone indicated that she had been inactive since 1995. ASCA was, however, not aware of Ms. Anderson-Crumrine's criminal history. Ms. Anderson-Crumrine did take the trouble to update her name on the ASCA rolls after her release from the penitentiary and marriage to Dan Crumrine; the small matter of the federal felony conviction must have slipped her mind at that time.)

It has not been my experience that serial perjurers, habitual frauds, and people whose own mothers tell a judge "she's a pathological liar" are in the habit of reforming themselves and treading the straight and narrow. Give me an alley-mugger for an ex-con with prospects any day. Considering the many aliases, gravitation towards vulnerable populations, and bogus credentialing, things sure aren't looking good.

Bullshit-meter reading, This one goes to Eleven.

So that kind of defines the ends of the spectrum of SAR handler capability-inflation.

When the Level Two Majick Dog Handler fails to produce up to his own hype -- wastes all that helicopter fuel for an ego trip -- he contributes to an incident commander's suspicion that dogs may not be all they are cracked up to be.

That IC may not have the background to distinguish between one handler's baseless claim that his dog has a better nose than all the other dogs, and a more conservative handler's caution that he be careful about deploying that uncertified handler who has no training logs. It all gets dismissed as "dog handler politics."

More important, to the uneducated IC, all claims about the useful olfactory powers of trained search dogs may seem equally implausible.

Can you spot the implausible, wild-assed claims among those below?

This dog has detected a person's scent on the wind (no ground trail), and found her at a distance of over 200 meters in good atmospheric conditions.

This dog spontaneously alerted on graves that were over 30 years old.

This dog can distinguish between the individual scents of a pair of identical twins.

This dog reliably distinguishes between animal remains and human remains.

This dog can indicate the location of human remains under 50' of water.

This dog will reliably signal the absence of the lost person's trail.

This dog works confidently and independently on unstable rubble right next to bulldozers and jackhammers, and only searches for and indicates people who are buried, not workers out on the pile.

This dog has accurately followed a 48-hour old trail.

Give up?

Although any of these claims may, indeed, be utter bullshit about any given untrained or poorly-trained dog, they are all standard findings or expectations for SAR dogs properly trained in various disciplines. Some are codified in standards -- the minimum performance expected for a dog to be operational in its discipline.

While it's important to be skeptical-until-shown-otherwise of all unknown handlers who claim to have met these standard expectations (that are the subject of broad agreement about essential competencies), it's even more important that IC's and other public servants -- not to mention politicians, journalists, and the families of missing people -- understand when they are encountering an actual wild-assed claim.

The claim that one has met ordinary, industry-standard, levels of competence should require that one present ordinary proofs. Back in the day, it was training logs and some token that one had been declared operational by a unit that had published standards. Nowadays, an external (presumed to be more objective) certification by a third-party organization, to standards that are NIMS-compliant is an increasingly common additional expectation, depending on the specialty.

What if the claim is extraordinary? The handler maintains that his dog can perform feats that are orders of magnitude superior to the industry standard?

The dog is never wrong.

The dog can follow scent trails that are months old.

The dog can track bullets.

The dog can track anthrax.

The dog responded to an average of more than one search a day for fourteen years, with 2.25 finds per week over her whole life.

The dog's ID is not just a reasonable means to establish probable cause, but is sufficient evidence to convict someone of a crime in the absence of any other evidence ...

These are the claims that bring reporters running, shock and awe the fatuous and desperate, and inexplicably seem to qualify the claimants as go-to guys in the eyes of the FBI and sometimes other "top level" law enforcement.

Just as Sandra Anderson's claims that Eagle was capable of things that no other dog could do made her into a celebrity -- and skeptics into pariahs -- these extraordinary claims are polarizing.

It's not because they threaten the egos of lesser handlers with mere mortal dogs.

It's because non-experts frequently can't or won't distinguish between them and the "industry standard" SAR dog abilities I listed above.

So when the "extraordinary" handler's deliberate fraud or unconscious delusion is unmasked, it's the legitimate SAR community that is left damaged. Those who perform due diligence in both their training and proofing and in their claims about their capabilities know that they will be fighting to recover from utterly unfounded guilt by association.

So what need we ask of anyone who makes extraordinary claims about his scent-detecting dog?

Presentation of extraordinary evidence.

Let's say we have a dog who we claim not only can, but HAS followed an individual human's 24-mile trail that is six months old, and was primarily left while the subject was being driven on a highway in a car, in an arid climate, in an area of moderately high vehicle and human traffic.

The dog has done this in part by being let out at off-ramps and indicating whether the car continued or exited.

Now, in case you were wondering, yes, this is an extraordinary claim.

Here is how I would test it.

An experienced, professional SAR unit that has no connection to the handler making the claim designs a double-blinded trail task.

When the trail is laid:

The trail-layer is a visitor to the area who will not return in the next six months -- a friend or relative of a team member. This is to ensure that the trail-layer will not inadvertently cross his trail in the course of normal activities and travel while it ages. The SAR team member acts as driver and selects the route. Both driver and trail-layer sign agreements that they will disclose nothing about the trail's location to any person until the trail is run.

The test administrator collects three scent articles from the trail layer on the day of the trail, and preserves them in whatever manner has been requested by the handler.

A test administrator designates a starting point. This point is selected so that there is easy, close access to more than one limited-access roadway. Three nearby highways with regular on-and-off ramps offers six possible initial directions of travel on limited-access roads. There may be other roads, not limited access, leading away from the designated start point. The driver can choose any road he wishes.

The driver and trail-layer are provided with the following instructions:

The trail layer may get into the car at the start point or walk some distance -- either on or off the road -- and be picked up to continue. Or he may lay the entire trail on foot. He is not to mark the trail in any way.

Please determine how many options there are for initial direction of travel, and use a random process (dice throw or random number generator) to choose one of them. After that, you may choose a route that suits you.

Do not double back or cross your route. Do not drive or walk parallel to a previous leg of your route any closer than 300 meters. Do not make turns that bring you closer than 300 meters to any part of your previous route. Be sure you don't double back, come within 300 meters, or cross it as you leave the area after laying the trail, either, even if it means taking a long detour.

The trail should be between ____ and ____ miles long. (The test administrator will generate these numbers with a random process. The handler will not be told the ranges, and only the trail-layer and the driver will know the actual length.)

The trail-layer will carry a top-of-the-line consumer model GPS (e.g. Garmin 60csx, Garmin Oregon, Magellan Triton) and record the track from the start point to the end point. While the trail-layer is in the vehicle and at all times, he will ensure that the GPS is positioned for optimum reception and has adequate satellite acquisition for accuracy within 20 meters.

At the end point, the trail layer will save the track and will start a new track, the purpose of which is to document that the driver and trail-layer did not cross or impinge on the trail-layer's track as they left the area.

The driver will leave an appropriate marker at the end point, something that is not likely to be removed or disturbed over six months of weathering. This may be spray paint on the roadway berm, or whatever is most likely to be durable at the end point. The marker will be symbolic/coded only, and the mark used will be known only to the test administrator, the driver, and the trail-layer. The driver will then leave the trail-layer at the end point and place four similar coded markers at similar points no less than 500 meters from the end-point or any part of the trail, and no less than 500 meters from each other, preferably further. The real and false end-markers will be waypointed on the GPS.

The trail-layer gives a scent article to the driver at the completion of the trail, and the driver preserves it in whatever manner is specified by the handler.

Upon returning home, the driver downloads the GPS tracks, labels them, and preserves them as electronic files on a thumb drive and imports them into the appropriate mapping software, where they are both preserved electronically and printed out onto maps of appropriate scale. The driver seals the maps and the thumb drive in an envelope, signs across the seal, and maintains it in a secure location along with the trail-layer's scent article.

The trail-layer retains a map that indicates the area of the trail, so that if he must travel in the vicinity, he can avoid crossing his own trail inadvertently.

Six months pass.

On the day before the team attempts to run the trail:
The driver returns to the end point directly without reiterating the trail and conceals the trail-layer's scent article within 10 meters of the end-point markings. The article is concealed from view, but marked in such a way that it can be definitively identified as the trail-layer's once revealed. The driver then returns to the "dummy" end markers and conceals an identical article that has never been in the same building or vehicle with the trail-layer at each of those locations.

The driver transfers the sealed envelope containing the records of the trail to the test administrator.

On the day the team attempts to run the trail:

The trail-layer submits a signed affidavit attesting that he or she has not been in the area of the trail since laying it.

The trail-layer and the driver submit signed affidavits attesting that neither has disclosed any information about the route of the trail to any person.

The test administrator, the handler and dog, and at least three observers convene at the designated starting point. The handler may designate up to two observers, the unit providing the testing may designate at least two observers, one of whom will serve as videographer and one of whom will ensure that the GPS is operating in good order for the duration. The observers do not need to be members of the testing team, but should be qualified SAR personnel.

A vehicle sufficient to accommodate handler, dog, and observers will be provided by whomever has such at hand. One observer will be designated the driver. The driver will stay with the vehicle, while all other observers will remain with the team.

The test administrator will bring the preserved scent articles, the same or substantially similar model GPS used by the trail-layer with adequate batteries and media for 12 hours of use, maps of the search area, the sealed envelope containing the electronic and printed definitive record of the trail, a computer and peripherals adequate to immediately download both the old and new tracks and overlay them on the appropriate mapping software, appropriate radios for communications, and a digital video camera or cameras with adequate batteries and recording media for 12 hours of use.

The start time will be early morning. The team will have 12 hours to complete the trail. The handler may take breaks as needed. Each person is responsible for his or her own food and water and personal needs for the duration. The handler will be briefed, and cautioned that accepting or soliciting help in determining the direction of the trail from any outside party or test observer will invalidate the test. The observers will be briefed and cautioned that any interference with the handler, to aid or hinder, will result in their expulsion as observers and may invalidate the test. All participants will be instructed to mute their cell phones and place or receive no calls or text messages, with the exception of emergency communications with the test administrator.

The test administrator will provide the handler with the preserved scent articles and the maps of the area when the handler states that she is ready. The clock starts at that point.

The videographer will begin recording, keeping the camera at all times on the team, and as much as possible keeping both dog and handler in-frame. He will continue recording for the duration of the test, including breaks, and will not stop or pause except to change batteries or media. He will endeavor to remain close enough to record all audio of conversation between persons or between handler and dog during the test.

The GPS-keeper will attach the GPS to the handler in a manner designed for optimum satellite acquisition, begin track-log recording, and will be responsible for monitoring it for accuracy and battery/media needs throughout the test. The GPS will never be turned off except to change batteries/media.

The handler may use her own GPS for navigation purposes, but agrees that the definitive record of the trail will be the one recorded on the test-GPS.

The test administrator will remain at the start point or a location nearby for the duration of the test.

Communications will be limited to hourly status reports and emergencies or technical glitches. Status reports will be simple check-ins, and will not include the team's location or current activities. The test ends when the handler identifies the end-point based on finding the scent article or the handler decides to stop.

After 12 hours have passed, if the team has not either arrived at and identified the end-point or given up, the test administrator will instruct the team and all observers to return to base.

The GPS-keeper will save the current track and start a new one for the return to base.

The administrator will then open the envelope with the printed maps and the electronic records of the trail before the handler and all observers, and while being videotaped. The administrator will then download the GPS tracks of the team's movements that day and overlay them on the map of the original trail.

All participants will receive a copy of the map with both sets of tracks, and electronic copies of all the records, including the full video record.

All participants agree that the map with both sets of tracks, and individual accounts of the test by the trail-layer, driver, observers, handler, and test administrator may be published.

What constitutes success?

Before initiating the process, all parties must agree that --

The team must correctly identify the initial direction of travel at the start point.

If the team follows the trail accurately and maintains overall progress of 1 mph (total time, not moving time), then it need not reach the end point (assuming a trail longer than 12 miles). "Accurate" shall be defined as following the actual roadway on which the trail-layer was driven, and following within 100 meters of any off-road path taken by the trail-layer.

If the handler chooses to use the vehicle to "jump track" between exits or intersections, then the team must not miss a turn or exit, nor follow a false trail more than 300 meters at a turn or or exit ramp.

Trailing the driver in places where his trail and the trail-layer's diverge does not meet the criteria for success.

Finding the end-point by any means other than following the path of the trail-layer does not meet the criteria for success.

Identifying any of the dummy markers and articles as correct does not meet the criteria for success.

Continuing to trail past the real marker -- as long as the team follows the driver and trail-layer's route -- does meet the criteria for success, i.e., the dog does not have to indicate the concealed article.

So -- sound reasonable?

I'm sure my SAR colleagues and other good minds can find ways to improve the protocol, make it tighter and fairer and more objective, and find places where I forgot to specify something that does not "go without saying" when we are being so formal.

If you were claiming that your dog could for real follow the trail of an abduction victim for 24 miles, mostly on highways, without any hints from people already working the case, is there anything in this protocol that you would find inapplicable or unfair? Anything you would want added or removed?

If someone making this extraordinary claim would not undergo this kind of test when it was offered, claiming that it is about "SAR politics," how would that reflect on the credibility of the claim for you?


______________________

* Why two? Because just about every handler with enough ego to tackle the job at all is operating at one.

** Authorities had found Anderson holed up at Crumrine's house in Virginia when she skipped out on her sentencing hearing in Michigan.