TEACHING THE REFIND TO AIR SCENTING SAR DOGS
Shirley Hammond, CARDA & FEMA
Gail McCarthy, MARK9 & FEMA
The training of an air scenting search and rescue ("SAR") dog
(1) teaching the dog that finding people is an acceptable thing for
the dog to do; and
(2) teaching the dog what to do once the dog has found
As a dog is born knowing how to use its nose, SAR training does not
technically teach the dog "how" to follow scent so much as it
communicates to the dog on what specific thing we want it to focus its superior
scenting abilities. In the process, SAR training raises the dog's
"skill-levels" by helping it work through varying wind patterns (like
Crazy-Ivans and lofting); ensuring that the dog feels comfortable moving in
different terrain and footings; and showing the dog that the source of human
scent can be traced not only to people lost in the woods but to people hanging
in trees or buried in the ground or under snow or rubble.
Teaching the dog "what to do once it has found someone" is the
hardest step in the training of a SAR dog. Typically, the beginning air scenting
SAR dog is taught one of the following two "communication" methods:
(1) the dog is trained to stay with the subject and bark (ie, the "bark
(2) the dog is taught to return to the handler and then lead
the handler back to the subject (ie, the "refind").
variations on the refind range from the dog simply returning to the handler and
the handler reading the subtle changes in the dog's behavior to tell the handler
that the dog has found (ie, the "natural refind," "passive refind"
or "nuance refind") to the dog performing some special signal in order
to indicate that it has found (ie, the "active refind" or "parlor
trick refind"). Although there are an infinite number of trained or
behavioral indications, aka "parlor tricks," that
one can teach the dog to use in signaling to the handler that it has found, the
most common are:
(1) the bringsel alert;
(2) grabbing a fuzzy handing off the
(3) barking at the handler;
(4) jumping up on the handler; and
(5) coming up and sitting or lying down in front of the handler.
It is best to select and
condition a trick-behavior which the dog performs naturally without prompting
from the handler. Both the "natural/passive refind" and the
"parlor trick/active refind" require that the handler learn to
recognize the signs each dog typically exhibits when it is working a scent (ie
the "alert"). (For example, although each dog will develop its own
unique search style, Dobermans generally work a scent plume with their heads up
leading to our breed being nicknamed "headhunters"). In short, the
handler must learn to "read" the dog and the dog and handler must
learn to work together as a team.
The pros and cons of the "bark alert" versus the "refind"
and the merits of the "natural refind" compared to the "parlor
trick refind" is beyond the scope of this article. Suffice to say there is
no consensus in the SAR community on these issues. Each method has its positive
and negative aspects and each has been used successfully over the years. One
should select the method that makes sense to the handler . . . and it should be
a method that is accepted by one's SAR team!!!
Set forth below is one method that has been used successfully to train the
"parlor trick refind" to air scenting SAR dogs. Keep in mind, however,
that there is as many methods to training a "parlor trick refind" as
there are people to think them up. What is important is to keep to the forefront
the following training principles:
- Dogs tend to repeat whatever behavior is
- Train for one behavior at a time.
- Teach the last step of any
behavioral chain first.
- Although a specific behavior can initially be lured
or cued in order to communicate to the dog what the target behavior may be, the
dog must ultimately offer the desired behavior voluntarily so that it can be
shaped and reinforced.
- If the dog is not offering the target behavior, then
break the desired behavior down into however many small units as are necessary.
- Raise criteria in increments small enough so that the dog always has a
realistic chance for reinforcement.
- Polish each behavior before chaining the
behaviors together and do not ask the dog to execute the complete behavioral
chain until the dog thoroughly knows each step.
- When training anything new,
or training in a new place, temporarily relax the performance standard of the
existing behavior until the behavior has become generalized or is
- Know the steps in your training plan so that you can
stay ahead of your dog.
- If an existing behavior deteriorates, go back to
the basics and encourage the dog with easy reinforcements.
- Try to avoid a
deterioration in behavior by keeping a higher ratio of easily solvable problems
to those problems in which the dog works for long periods of time with no
reward. (Another way of avoiding behavior deterioration is to put the behavior
on a variable system of reinforcement).
- If one element of a behavioral
chain is weak, take that element out of the chain and perfect it before asking
the dog to perform the element as part of the chain.
- If one way of teaching
isn't working, try something else or break the training down into smaller
- And, finally, have fun! SAR work is based on positive
reinforcement so quit if you feel yourself getting frustrated! Conversely, if
your dog does something perfectly or has a breakthrough in understanding, stop
training and reward your dog with a big jackpot of whatever is meaningful and
rewarding to the dog.
The training principles referred to herein, commonly
referred to as the "Laws of Learning," are based on the "Ten
Laws of Shaping," as set forth by Karen Pryor in her excellent (and
highly recommended) book, "Don't Shoot The Dogs" (Bantam Books 1984).
"PARLOR TRICK REFIND"
AS PART OF A REVERSE BEHAVIORAL CHAIN*
*Please note that the following back-chained behavioral training system is
based on PRIMARY reinforcement as opposed to CONDITIONED reinforcement. For
those that understand the difference, a conditioned reinforcer can easily be
introduced into the training system set forth below.
TEACH THE PARLOR TRICK
Step (1) Teach your
parlor trick-behavior (ie, jump up; bark, sit; grab fuzzy
or bringsel, etc.) in a distraction-free area. Put the trick-behavior on cue.
Once the dog performs the trick-behavior reliably in a distraction-free area,
practice the area in many other areas so that the behavior becomes reliable in
the face of distraction.
TEACHING FIRST LEG
(Subject within visual of the dog)
Step (2) Go into your living room, your backyard, or wherever you can have
someone (ie, the "subject") run 10-15 feet and still have control over
the dog. Area again must be distraction-free.
Step (3) Have your dog watch the subject run away (like ARDA but this step is
being performed only to get the dog to go to another person (and also tapping
into prey drive) but NOT to find at this point).
Step (4) Release the dog and have the subject play tug-of-war or ball or give
food or praise. It is immaterial what type of reward is used so long as the
reward is meaningful and pleasureful for the dog. Thus, the reward will differ
from dog to dog and may even have to be changed at regular intervals to keep a
particular dog motivated.
TEACHING SECOND LEG
(Subject still within visual of the dog)
Step (5) When the dog has performed sufficient runaways so that it is
apparent that the dog knows what it is expected to do (ie, as evidenced by the
dog "rocketing" out to the subject with great gusto), give the dog's
verbal cue for its trick-behavior after the subject has rewarded the dog for
properly performing the first leg. If the dog does not come back to you to
perform the trick-behavior, then have the subject temporarily hold the dog by
its collar and throw whatever the subject was using to reward the dog back to
you (which is why it is best not to use food to reward the dog). For example, if
the dog's reward is tug-of-war, then the subject should throw the tug toy back
to the handler. (Yes, you are cueing the dog, but the dog will be weaned off the
cue so that the behavior will eventually occur independently). Release the dog
when you have the reward in your hand.
Step (6) Reward the dog with the toy, etc., when it runs back to you and
performs the trick-behavior. If the trick-behavior fails to occur, go back to
step (1) and solidify the trick-behavior so that the dog will reliably perform
the behavior on cue.
TEACHING THIRD LEG
(Subject still within visual of the dog)
Step (7) After a few repetitions of practicing the first and second legs
together, the dog is asked to "show me" after being rewarded for
performing the second leg. The handler initially encourages the dog to run back
to the subject by actually running with the dog back to the subject. (For people
applying pure operant conditioning principles, the cue "show me" would
not be used until the third leg-behavior is in place). Everyone rewards the dog
when dog, subject and handler are all together. It is important that the biggest
reward come from the subject at this stage in order to encourage "subject
loyalty" in the beginning SAR dog (as "subject loyalty" is one of
the foundation elements of a seasoned SAR dog). Therefore, while the handler may
also reward the dog, the "big bucks" should come from the subject.
GET THE BEHAVIORAL UNIT TO OCCUR
(Subject still within visual of the dog)
Step (8) Practice the above behavioral chain until the dog rockets back and
forth performing all three legs in one smooth behavioral unit -- ie, the dog
will cease thinking about each separate leg but will independently (and with no
cueing) perform all three legs as if there were no separation. How long it takes
to get the total concept across to the dog will be SPECIFIC to each individual
dog but, until the dog masters the total concept, you will need to reward each
leg or reward alternate legs (although the biggest reward should be reserved for
when everyone is united at the end). This will make the process feel "hiccuppy"
and disjointed but eventually the light will go off -- the dog will, at some
point, in the training process "perk up" when they realize, "Hey,
this ain't hard!!!," as they bounce between the handler and the subject,
eagerly awaiting the big reward at the end (this is about the time some
food-reward dogs start to drool because they KNOW there is something wonderful
in the food container in the handler's pocket.).
Until the dog begins to perform confidently and independently, do NOT
increase distance for beginner dogs, do NOT add the "find" element and
STILL work in controlled environments. The goal is to merge all three separate
legs into a behavioral unit in the dog's brain. (And, since the intensity of a
trained-refind tend to diminish when the (longer) finds are tacked on, the
refind at this point should be crisp and strong and intense. For example, no dog
maintains the rocket-speed of the reinforcing runaways when longer problems are
practiced. Similarly, the dog may only jump up a little bit when this
trick-behavior is added onto the search element, but the bodyslam some dogs give
at this training-stage (and even during the routine reinforcement of this
trick-behavior) is so strong that it can literally bowl the handler over (and is
the main reason why handlers of dogs as large as Dobermans should think
carefully about using jumping up as their dog's parlor trick!).
Step (10) Once the handler is ABSOLUTELY CERTAIN that the behavioral unit is
completely reliable in a distraction-free environment, then start to perform the
behavioral unit in new areas with increasing levels of distraction, new people
Step (11) Once the behavioral unit is reliable under new and differing
circumstances, start to add the find (finally!!!) to the behavioral unit and it
is here that you can fall back on the wonderful ARDA/Sryotuck scenting and
search problems (the half-c's, etc). The Doberman's high energy, love of
running, good nose and natural curiosity will be assets at this training stage
(although some effort will be needed to encourage some members of our breed to
use their nose as opposed to relying on their strong visual acuity). Gradually
increase complexity, distance and length of time during which the dog must
search before it gets its reward but ALWAYS go back to basics (ie, routinely
(and often!) and practice the basic behavioral unit if the behavior starts to
weaken (and it will fluctuate)).
It is at this point that the true art of dog handling comes into play.
Learning to read your dog means more than just recognizing when it is working a
scent, but also means knowing when to reward and when to encourage and knowing
when the dog is ready to move onto more complex learning problems or when it
needs a series of easy reinforcements. One should strive at being fluid in one's
handling but always make sure that you know what behavior you want to condition
and always have a training-plan to ensure you get there!!
Good luck and Good training!!
Copyright © 1999