Punishing dogs does not work, it deteriorates your relationship with them, and it damages them mentally. Punishment does not achieve the goal of changing or removing a problem behavior. Differential reinforcement is a proven method of changing dogs behavior, it is non aversive and minimally invasive.
Differential Reinforcement is a preferred method to behavior change programs with dogs than using punishment.
Punishment has sinister and subtle side effects, it drives problem behaviors underground and they tend to then be displayed in the absence of the punisher. Punishment also activates the dogs’ emotional system which has a negative impact on the dogs’ ability to think and learn making it almost impossible to teach the dog new more acceptable behaviors. Because punishment can never be applied consistently and predictability this leads to learned helplessness, a state resistant to rehabilitation, as the dog cannot control nor have confidence in its environment (O’Heare 2004).
A successful behavior change program must make the problematic target behavior ineffective, irrelevant and inefficient. The program must also establish new behaviors and skills while decreasing the frequency, intensity or duration of problem behaviors. Since punishment can create many problems, alternative methods of modifying unwanted behavior, such as differential reinforcement, are preferable (Miltenberger 2004)
Differential reinforcement is an operant conditioning technique that can be used to both increase and decrease behaviors. For many problematic operant behaviors during differential reinforcement, the problematic behavior is targeted for extinction while another behavior is simultaneously positively reinforced. There are several differential reinforcement protocols and each has its application and suitability based on the behavioral problem.
When choosing a differential reinforcement protocol we must first establish whether the target behavior is already present and requires a dimension of the behavior to be reduced, such as barking, or whether the goal is to place the old behavior on extinction and reinforce a new behavior, such as jumping on guests when sitting would be more acceptable.
When considering which differential reinforcement protocol to use for teaching new behaviors we need to assess if the preferred behavior is already occurring, can it be captured or lured or will it need to be shaped. We also need to establish if we will use extinction trials or not and whether given the behavior problem they may create frustration or cause an animal to aggress, both causing frustration during training or a trigger of aggression are considered training failures and should be avoided at all costs (O’Heare 2009).
Differential Reinforcement of an incompatible behavior is best used when we can identify an incompatible behavior, a behavior that is physically incompatible with the problem behavior and one that commonly occurs so it can be captured or prompted easily without exposing the animal to the problem stimulus. We must have access to a reinforcer that can be delivered after the new behavior and the ability to withhold reinforcement for the undesirable behavior.
The new incompatible behavior is trained in a different context to the problem behavior and then exposed gradually to the problem stimulus to avoid extinction trials as this can result in aggression and extinction bursts. If a dog is jumping all over people as they come in through a door we would first teach a solid sit for greetings away from the door and then gradually move the behavior back to the problem location. If an incompatible behavior cannot be identified then, using the same methods, we can differentially reinforce a specific alternate behavior (O’Heare 2009).
When working with an animal that is highly excitable or has a propensity to engage in frustration behaviors, such as nipping or mouthing, then one should adopt the differential reinforcement protocol of “other behavior”. This is a simple procedure where the reinforcer is contingent on the absence of the problem behavior. Reinforcement is easily accessed by the animal and the procedure is easily carried out by the owner. Because there is no specifically targeted behavior to reinforce the technician must remain cognizant that differential reinforcement of other behaviors can create superstitious behaviors (O’Heare 2009).
With behavior that cannot be reduced to zero (barking) or some level of the behavior is acceptable and/or it is difficult to train an incompatible or alternate behavior, then using differential reinforcement of low rates is used. It would first be determined whether the goal of the behavior change program is to reduce the intensity, frequency or duration of the baking behavior and the protocol would then focus on that behavior dimension. Reinforcement would then only be available to the animal if the specific target behavior occurred at a specified rate during a specified period of time (O’Heare 2009).
If the goal of the behavior change program is to establish a new replacement behavior and the animal being trained lacks social confidence or offers ridged behaviors as a result of previously harsh or aversive training protocols then using differential reinforcement of successive approximations to a terminal behavior (shaping) would be most appropriate.
Shaping can also be used to develop operant behaviors that are not initially visible, need to be trained in small approximations or the conditions of the training need to support empowering the animal to experiment with new behaviors. Shaping procedures can also be used to teach alternative or incompatible behaviors without the use of extinction trials. A key benefit to using shaping and differential reinforcement protocols is that they also positively impact the animal’s respondent behavior (O’Heare 2009).
Differential reinforcement protocols that focus on operant extinction are designed to decrease behaviors by withholding or preventing any reinforcement for the problem behavior. If it is necessary to use protocols that focus on extinction trials because other less invasive or aversive methods have been tried and failed the extinction procedures should always be used with a reinforcement procedure of a desirable behavior
Operant extinction trials for aggressive behavior should only be used when the behavior is positively reinforced, i.e. the positive reinforcer is no longer delivered after the behavior, most aggression behaviors are negatively reinforced, and the dog exhibits the behavior to remove the stimulus from its environment.
© Niki Tudge June 2009
Miltenberger (2004) Behavior Modification Principles and Procedures Third Edition, Thompson. USA
O’Heare, J. (2008) Behavior Change Programming and Procedures 2009, CASI Course Notes,
O’Heare, J. (2004) Canine Separation Anxiety Workbook 5th Edition DogPsych Publishing Canada