Category: Advanced Stuff ¤
  Author: Elizabeth Naime ¤
   Title: Impact of Variable Schedule of Reinforcements ¤

No, I think the idea of the VSR is to strengthen behaviors -- look at how
many unwanted behaviors are completely cemented into place with an
unintentional variable schedule of reinforcement:  begging at table,
barking for attention, jumping up for attention, and I am sure there are
many many more.  In fact that is what caught my interest in "clicker
training" and operant conditioning in the first place, realizing that I
was getting 100% better attention from the dog sitting in the dining room
hoping against hope for me to give in and "drop" something than I was
getting in formal training with continuous reinforcement.  I am not sure
why it is so, but the belief that THIS time I'll hit the jackpot is more
powerful than the certainty that I will get x amount of treats each and
every time.  And, it is very true for humans as well!

The best analogy I have heard for this is the soda machine vs. the slot
machine.  With the soda machine, you put in your coins and get a soda, and
you expect this to happen every time (well, unless your vendor doesn't
maintain the machines very well, in which case you will start to see a lot
of intense and often superstitous behavior around the coke machine!).  In
consequence, most of us only bother with the soda machine when we are
thirsty, and we don't put a lot of ourselves into the interaction unless
something goes wrong.  If something does go wrong, and it's not common in
the machine user's experience, there will be a short burst of activity as
s/he pushes buttons and pulls the coin release and maybe whacks the
machine a couple of times, then the soda-machine related behaviors cease
as s/he walks away grumling.  With a slot machine, you will see intensity,
people putting a lot of themselves into their actions;  people who like to
gamble will continue to feed coins into the machine after a jackpot, when
they can't possibly be "thirsty" any more; time after time the slot
machine gives nothing but the slot-machine-operating behaviors do not
extinguish.

I don't want my dogs to go as totally crazy as some slot players I've
watched, but I do want intensity and I do want the behaviors I reinforce
to be persistent.  Variable reinforcement does this for me, unless of
course I wait TOO long between reinforcements and the dog gives up.  This
is called "ratio strain" and has been studied, too.  The idea of a
variable schedule of reinforcement is to hit the golden mean between
certain reinforcement (the "fixed schedule", in which the subject will put
forth the least effort required to get the reinforcement) and ratio
strain.

For behaviors like "sit", I think *habit* has a lot to do with it.  If you
ask for a sit very often, after you have intentionally put it on a VSR,
you will find yourself asking for a sit in conditions where you may not
have a reinforcer or you may be to busy to reinforce just then.  Example
when you are out on a walk and stop to talk to another person, when the
dog is in the way.  If you're like me, long after you've stopped clicking
sits you will still try to appreciate them when you remember -- to say
"good dog" or reach down and pet the sitting dog.  That's reinforcement
(social reinforcers, also may be secondary reinforcers if you have said
"good dog" in the past before giving a cookie).  Directly food-related
situations just seem to naturally call for a sit:  many dogs sit before
their dinner dishes are being put down, mine sit for treats usually.  It
gives the dog something to do and sure beats jumping for the treat/food
dish or running back and forth for you to trip over.  In consequence, the
dog who was taught to sit before being given dinner or cookies will
habitually volunteer sits in the hopes of getting food -- in some cases
you were going to give a treat anyway, so you do -- poof! an extremely
long variable schedule of reinforcement AND the situations in which it
happens become cues in themselves!

I seldom intentionally reinforce Jackie for sits anymore, and I almost
never intentionally reinforce coming when called with any of the dogs --
but they get a lot of unthinking, very random reinforcement as I have
described above.  So as far as what I'm intentionally doing, I am not
trying to put coming when called on any reinforcement schedule at all, but
an outsider watching me call my dogs would see a variable schedule of
reinforcement.  Handlers act out of habit as much as their dogs do! :-)

Elizabeth Naime
and the menagerie