Regardless of individual training methods and techniques, effectiveness (as measured by time and trials to a reliable criterion) depends on a thorough understanding of learning theory. Unfortunately, the creedal principles of learning theory do not necessarily work in practice.
Most learning theory has been derived from observing consistent computers training rats and pigeons. As such, a good 50% of laboratory research findings are irrelevant in the real world, wherein inconsistent humans try to train dogs and other family members. Specifically, most reinforcement schedules do not exist outside of the laboratory, and punishment techniques can be notoriously ineffective.
The most common reason for resorting to the use of frequent and excessive punishment during training is a lack of understanding of traditional learning theory and especially, not understanding the many constraints on learning theory when it is put to practice. Although theoretically and experimentally sound, punishment seldom works in practice.
In just the past decade, as dog training has become more dog-friendly, response-reliability has gone down the toilet. Where are all the prompt and reliable emergency sits and downs? How about three-minute sit-stays and five-minute down-stays? Some trainers no longer teach on-leash walking but instead opt for permanent management. And snazzy heeling and off-leash reliability are quickly becoming distant memories.
Precision, reliability and dog-friendly techniques need not, and should not, be mutually exclusive.
Positive punishment and negative reinforcement can be extremely effective and efficacious, if one adheres to the “Eight Criteria of Punishment.”
Positive punishment and negative reinforcement need not, and should not, be scary, stressful, or painful in order to be effective. On the contrary, effective instructive reprimands may be delivered in a soft, sweet tone, and they may be administered in a gentle negative reinforcement format.
Our precious quadrant requires serious practical review. For the most part, consequential feedback comprises short-term rewards and punishments (praise and reprimands), i.e., only utilizing half of the quadrant. Moreover, the relative pros and cons of reward/punishment consequences are more profitably analyzed by comparing the six different types of binomial feedback. The quadrant only really applies to the use of long-term rewards and punishments. However, it is impossible to negatively reinforce a desired behavior (e.g., forced retrieve) without previously positively punishing the dog (for allowing handler to grab it). And so, with negative reinforcement, we are on very thin ice. Similarly, when positively reinforcing a dog with a lengthy reward (e.g., Go Play), the dog will be negatively punished when the reward ends. ("Come Here"—play session over.)
Learning theory terminology is laughably so abominably ambiguous, it is really difficult to know what to do about it. A simple discussion between dog trainers can become very confusing (often leading to disagreements because trainers fail to see that they actually agree with each other). Regardless of what we do within our profession, we should not inflict our flawed terminology on the unsuspecting dog-owning public...it would fry their brains. With dog owners, I suggest we speak in simple jargon-free sentences, explaining how they can teach their dogs to promptly sit on cue, to reliably stay, and to walk like a charm on a loose leash.