Gradual Desensitization and Counter Conditioning: Too little, too late

Author: Gary Wilkes
From a presentation at the Central Veterinary Conference, San Diego 2010

In the world of modern behavior therapy, there are two popular mainstays – gradual desensitization and counter conditioning. If a dog is terrified of thunderstorms, one plays back a sound recording of thunder at very low volume and “desensitizes” the dog to the thunder over a long period of time. Counter Conditioning refers to replacing a currently objectionable behavior with an acceptable behavior. Though widely used and recommended, these tools have very limited ability to control unacceptable companion animal behavior.

Gradual Desensitization:
Gradual desensitization is more properly termed “habituation.” This protocol is recommended for everything from inordinate fear of thunder to violent fearful reactions to strangers or other dogs. The process starts by applying faint replications of various stimuli that are associated with the dog’s unacceptable behavior. This could take the form of CD recordings of thunder or the sight of another dog at 100 yards. If the mail man is the trigger for the dog’s aberrant behavior, someone must attempt to recreate the sights and sounds of a mail man to reduce the dog’s fearfulness. Most suggestions of this method include offering palatable treats to affect a change in the dog’s attitude. The use of food in this protocol is actually an example of respondent or classical conditioning and would be more accurately described as such. The food triggers an alimentary response which includes various effects caused by the parasympathetic nervous system, such as reduced heart-rate, blood pressure and respiration. There are several reasons why this methodology is unlikely to work under normal circumstances.

  1. Terrified dogs rarely take treats.
  2. Recreating a tornado or thunderstorm with CD recordings invariably lacks all the stimuli present in a real tornado or thunderstorm. The shear variability of stimuli that occur during a simple thunderstorm is in itself an observable stimulus. i.e. Thunder patterns are highly variable – CD recordings are not. If you teach the dog that one pattern is safe and a real storm rattles your windows at an unpredictable moment, the entire conditioning program falls apart. Ultimately, failing to vary the lights, sounds and barometric pressure changes associated with bad weather is unlikely to fool the dog.
  3. A dog that goes crazy over the community garbage truck or the UPS van is almost impossible to gradually desensitize or counter condition. By definition, these methods require many repetitions under controlled settings. Unless you rent a very large truck (one that includes hydraulic noises and the sound of banging trash cans) you will be stuck with only the regular occurrences of the offending stimulus. Having one repetition per week to teach the dog to ignore a large truck would take a lifetime to correct, if at all.
  4. The high level of concentration and dedication required to use this modality is uncommon in pet owners. Many pet owners have difficulty giving simple medications in a timely fashion. It is unlikely the average pet owner will do much better with practicing realistic pseudo-storms during thunder season.


Counter Conditioning:
Several years ago a group of graduate student behavior analysts studied how to make shelter dogs more adoptable. After several months of study, they concluded that if you could train a dog for nine days you could dramatically increase its chances for adoption. Their methods were based in teaching acceptable behavior with positive reinforcement. This synopsis sounds wonderful. Unfortunately, it doesn’t do much, if anything, to improve the process of adoption. There are two critical flaws to this suggestion.

  1. The time a dog stays in a shelter is, on average, no more than three days and is often less. For every 100 dogs available for adoption, there are about 15 homes. After the first three days of training, 8.5 dogs died to make room for the one you are training. From day four to day six, an additional 8.5 dogs die, daily. The same is true for days seven through nine. In all, 25.5 dogs died in the time it took to train one dog to an arbitrary standard that may or may not achieve a permanent adoption. The bottom line is that if you increase the number of trained dogs past the 15 homes per hundred dogs, you do not increase the number of dogs who are adopted. Unless a way is discovered to increase the saturation point of dogs per community, this is a pipe dream. Even if you could increase net adoptions, overloading dog owners with multiple dogs usually makes collective behavior problems more likely. The end result is not an improvement.
  2. Positive reinforcement for acceptable behavior does not inhibit normally occurring behaviors. EG: Virtually every normally friendly shelter dog jumps up to get attention and affection. They are taught this behavior as infants by humans who pick them up and cuddle them. This early reinforcement causes the behavior to become firmly established in the dog’s repertoire. Counter conditioning the dog to sit instead of jumping up requires that a highly desired reinforcer be applied to the new behavior. This reinforcer can indeed drown out the old behavior – under ideal circumstances. If that reinforcer fails to appear, there is nothing to prevent the dog from returning to its original “successful” behavior. Maintaining the new behavior requires perpetual vigilance. Pet owners are rarely vigilant.
  3. Giving a dog treats to stop it from biting a vet tech is a questionable suggestion. The situation that caused the dog to bite in the first place is not abated merely because you offer a treat for “sitting” or “down.” Since loss of appetite is a common side effect of terror, asking for a behavior that was created and maintained through an alimentary reinforcer is an improbable solution. Note: The majority of vets have treats available during exams. If counter conditioning was a viable fix for unacceptable behavior, it would already be working.


The majority of “at-risk”dogs are dead, long before gradual desensitization can lessen the intensity of their unacceptable behavior. Most pet owners are unwilling to accept thousands of dollars in destroyed property from a panicked animal or hundreds of dollars in fines because their dog jumped a fence and landed in the county pound. Complicated, slow processes fail because the owner’s resolve has usually been taxed to its limits, long before they ask for help. Likewise, using “bait-and-switch” counter conditioning to prevent a dog from offering a long-standing and oft reinforced behavior is like paying an addictive poker player to deal Black Jack. As adjuncts to other forms of behavioral control there is a place in any practitioner’s tool box for these gentle, slow acting tools. As the primary solution for fixing serious and long-standing behaviors these two avenues are usually, literally, dead ends.

One thought on “Gradual Desensitization and Counter Conditioning: Too little, too late

Leave a Reply

Your email address will not be published. Required fields are marked *