Can I ask why balance training is bad..? I call myself a balance trainer because while I clicker train and focus on positive reinforcement, I do use a stern voiced "no" as a negative marker and a few positive reinforcement trainers got on me and told me that disqualifies you as being +r.
“No.” means nothing. It is an antecedent like any other cue. What follows is what separates r+ and “balanced” trainers. If the “No” is followed by a leash pop, a swat, a shock, you are forming that word into a punishment. The lowered tone is threatening, which does nothing but create distrust. The dog is learning that you are dangerous when you use that tone or that word and their natural inclination is to avoid confontation, to avoid YOU. Imagine trying to get a recall out of that. Good luck.
Now alternatively, if “No.” is paired with a time out, with the play stopping, with you leaving the room and toys going away, it creates an entirely different working atmosphere in which you are no longer the source of bad things, and training itself does not result in punishing consequences and the likeliness of having a shut down, anxious dog significantly decreases.
Now as far as being “balanced” goes, I would suggest you do some research on poisoned cues. This is primarily where balanced training fails. The idea is, simplistically speaking, to teach via the carrot or the stick method. A balanced trainer would say “sit” and if the dog sits it gets a treat, if it doesn’t, it gets a leash pop or pressure on the bum until it sits. It makes sense to us, however, in the dog’s mind, you have then made the cue “sit” to be ambiguous. Which SUCKS! You said “sit” and afterwards both rewards AND punishments come! Clarity is absolutely necessary for the learner. Ambiguity is more than frustrating, it is damaging. It is shown to lead to what is called experimental neurosis in lab animals. It can be even more confusing than if the cue simply led to either good or bad things.
And by the way, it is not just words. Objects, places, even people can become poisoned. And you can’t entirely undo the damage you have done by poisoning a cue. It is recommended to drop it entirely and retrain with a different word or gesture. You can’t change you, so take care not to make yourself a bad guy in your dog’s eyes.
A couple things here…
1. Using the word no does not mean you are “disqualified” from being an R+ trainer. In fact, most of the trainers I know that consider themselves positive reinforcement trainers deliberately use the other quadrants occasionally or even regularly. “R+ only” is a really stupid term because when you actually get down to the science, it is really impossible to only use one quadrant. And no, i don’t mean difficult, I mean IMPOSSIBLE. when speaking about operant conditioning Positive means the addition of a stimulus, Negative is the removal of a stimulus, Reinforcement means the behavior increases, Punishment means the behavior decreases. and you literally cannot only add stimuli all the time, occasionally stimuli will be removed, you will not feed, you will not praise. Also most shaping requires you to eventually extinguish lower criteria responses. So most R+ trainers will use at least positive reinforcement, negative punishment and extinction.
2. “No” can be a punisher in and of itself due to the tone or volume of the word. Most often, no is associated with other unpleasant consequences leading it to become a positive punisher through classical conditioning. How can you tell if no is a punisher? Examine the behavior it follows. If that behavior decreases, it is a punisher. If that behavior increases, it is a reinforcer. If that behavior stays the same or you see an extinction burst before that behavior decreases, then “no” probably doesn’t mean anything.
3. “No” can be an antecedent and/or a consequence. Often times its both. What a lot of people don’t understand is that antecedent only hold power when connected with consequences for particular behavior. Thus, antecedents often take on the reinforcing or punishing effects of their associated consequences.
4. This leads us to poisoned cues. What really makes poisoned cues awful is not that the behavior has been trained with more than one quadrant. The trouble with poisoned cues that those behaviors can no longer be used as reinforcers in behavior chains because they don’t retain their reinforcing value. In a behavior chain, each response acts as or creates an environment which acts as both a reinforcer for the previous behavior and as an antecedent for the next behavior. However, if you attempt to place a poisoned cue behavior in the middle of a chain, it will fall apart because it doesn’t reinforce the previous behavior, thus breaking the reinforcement chain and leading to extinction and ever degrading behavior.
5. NRMs signal that a behavior will not be reinforced. I usually don’t use them because my training should be clear enough in most scenarios that my animal can tell if did the behavior correctly or not. Clicker training is binary, you did it and you get a click and a reinforcer or you don’t hear a click and you get nothing and then you try again. So I don’t use NRMs because 1. its redundant and 2. studies have shown that it can add additional stress to the animal without providing any benefit to the training. The one scenario where I do use NRMs is when working a behavior chain, where it may be unclear which behavior was performed incorrectly, and in these scenarios that NRM usually just serves as a LRS/Recall so I can reset the behavior. If I have to use a NRM multiple times, it means I need to break down my shaping a little more or possibly do some retraining of the individual components.
6. Being a balanced trainer doesn’t just mean one thing and it doesn’t always mean that you are a terrible or ambiguous trainer. How one applies principles is what makes one a good or bad trainer. For example, there are individuals who use punishment to teach behaviors. This can be extremely stressful. Basically this is the opposite of clicker training. You get corrected for anything except the exact right behavior. How are you supposed to know what that is? ¯\_(ツ)_/¯ On the other hand, a lot of people use punishment only to proof behavior which has already been trained. Ones skill at applying behavior principles is what makes you a good or bad trainer, not which principles you use or don’t use or think you don’t use.
