Positive reinforcement to diminish barking

Question

This is a question for my friend who has a (female) Chihuahua (8 years old) named Maple. She has a tendency to bark at pretty much every person who enters the house, even if it's family and friends.

Now I'm pretty sure I remember from my AP Psychology class in high school that reinforcing negative behavior with positive reinforcement will simply increase the likelihood of the negative behavior [severe lack of source] (and this is, of course, not desirable).

Backstory

We're about to move into a house together with 4 other people (6 total), and we have both agreed that we should try and train her (if possible) to only bark at strangers instead of every person that walks in the door. We don't know if she's barking out of fear of the person, anger that someone disturbed the atmosphere, or just generally because she wants to let us know that someone entered the house (friendly or not). We thought that maybe positive reinforcement would be a good idea, even going against basic psychological principles [still no source].

Edit: Maple is usually right next to the front door when she is barking, usually (but not always) accompanied by my friend (the owner). To get her to stop barking, he usually calls her into his room. She is never aggressive; she simply barks, hops back, barks again, hops back, etc. Sometimes she barks while dancing (not happy dancing) in circles around you. That said, she always barks, whether or not the owner is around (I have gone to his house while he is/isn't home and she always barks).

Goal

Train Maple not to bark at friends and family. - This is technically a punishment, because we want the behavior to disappear.

Method

1. A friend or family member walks in the door;
2. Maple starts to bark;
3. We grab a small treat from a treat bowl near the door and give it to Maple (with lots of affection);
4. Maple runs off happily to eat her treat.

---

What we don't want to do is punish her for barking. We don't want to use a spray bottle, we don't want to yell at her, etc. We believe that reinforcement is the way to go.

Theory

By giving Maple positive reinforcement (a treat), she will learn to like the friends/family and not feel the need to bark because she knows the person, and knows that she gets a treat when she sees them.

Fears

We have a sneaking suspicion that Maple will take the positive reinforcement as "If I bark a lot when someone comes in, I get a treat!", which would of course not be the desired behavior. Because we would be giving her a treat during a negative behavior that we want to correct, the reinforcement (treat) may increase the negative (barking) behavior.

Other Potential Problems

1. We all go broke from buying so many treats.
2. Too many(?) treats may negatively affect Maple's health.
   • We could regulate this, of course. There are 6 people in the house, so treats may not be the best method if there are health repercussions.

---

Questions

1. Are there risks to this method of behavior training?
2. Will Maple's age affect her ability to be behaviorally-trained?
3. Is there a better way to go about correcting her barking behavior?
   • None of us know much about psychology nor animal training, so any help from either perspective would be much appreciated!

Answer 1

First off, I want to applaud you for looking for a better way to fix a behavior that so many people would fix with positive punishment. I also agree with most of mattgiltaji's answer, but there are a few extra points I wanted to add or clarify.

Operant Conditioning

	Reinforcement (increases behavior)	Punishment (suppresses behavior)
Positive	Add something good	Add something bad
Negative	Remove something bad	Remove something good

It's important to note that you don't necessarily know which behavior your dog will associate with the reinforcement / punishment. Having very clear criteria will help show the dog what you want.

Why stopping barking is so hard

Dogs bark for a variety of reasons, but it's a very difficult behavior to stop, and depending on breed can go against their entire being. Don't expect to get new behavior overnight and don't expect to get new behavior unless you're very consistent.

Barking is inherently self-reinforcing for many dogs. This means that doing nothing is the same as rewarding. To add to this, consider the following scenario:

Fido looks out the window and sees someone begin to walk by. He barks to alert his owner and tell the stranger to back off. The stranger continues walking until he's out of view.

What just happened in this scenario? Not only did Fido enjoy barking, but the stranger left when he was barking! It must have worked. So he continues barking when strangers appear. Because strangers move at different speeds, he learns to be persistent in his barking.

Methods

When moving from positive punishment to reinforcement, we sometimes need a change in mindset of what we want. By asking how can I stop this behavior, you've already put yourself in a punishment mindset. Instead think about what you want your dog to do when someone visits. Now that you have something to encourage, you can use positive reinforcement to accomplish this. I'll use the following desired behavior and then explain some options to help accomplish it:

1. Doorbell rings / Knock at door
2. Dog goes to a special place (mat, bed, kennel, anything with clear boundaries)
3. Dog assumes control position (stay - down / sit / stand / pretty / tall)
4. You open door and allow guest in and greet.
5. You release dog to greet the person and sit awaiting pets.

Separating Events

If you notice the items above, you can work on almost all the things separately. You can use a doorbell as a cue to go to bed. You can work on staying in the bed with other high-value distractions so he's used to staying when excited. You can teach your dog how to greet people nicely.

Creating a Pattern

Now you can create a pattern using backchaining or some other method to teach your dog that when the doorbell rings (or somebody arrives) he should rush to his special spot and wait to be released. This gives your dog something to do when guests arrive, but you might find your dog has discovered he can do just about all those things while barking. Be sure that not barking is always part of your criteria to prevent this.

Encouraging Silence

It's definitely better to go and get your dog instead of calling him to you. A recall should be extremely reinforcing for your dog and so you could actually be reinforcing your dog's barking by calling him to you. But this still gives your dog opportunity for reinforcement and ideally we only want them to have the opportunity to fail without getting reinforced.

Instead, I recommend setting up a physical barrier (such as a gate) where your dog stays in place until he's quiet. Make it very clear that any movement or noises before you release causes the gate to close and stillness and silence allow the gate to open. This prevents any unwanted reinforcement the dog gets before you can pick him up and opposition reflex caused by using a leash. Don't forget to reward your dog when he's quiet behind the barrier too. The barrier shouldn't be a punishment, but a way to show the dog clear boundaries so he can be successful. He sees clearly that he only gets to greet people when he's quiet. If he starts barking after you've released him, make sure your guests completely ignore him and then put him back behind the gate.

You can definitely add additional reinforcement when your dog greets quietly, although meeting new people may be very reinforcing already.

Fear Barking

This is where I'm going to disagree partially with mattgiltaji. If your dog is fearful of people or too aroused then dropping treats the minute people arrive (ideally before he begins barking or at least at the same time) can be very effective and does not reinforce barking. This is because the dog is not in a proper mental state to learn behaviors. In fact, spreading a variety of treats on the floor can distract the dog enough that he stops barking and returns to a state where you can teach him proper behavior. You're also not reinforcing barking because you're rewarding whenever people arrive regardless if he is barking or not. He may even start looking for food when people arrive on his own instead of barking.

Putting Barking on Cue

Another option to help reduce barking is to put barking on cue, and then don't give the cue. This is probably the least effective option, but teaching a dog a new trick is fun so why not? The theory behind this is that by adding reinforcement every time you cue a bark, but not when he barks on his own, he will only want to bark when cued. Of course, if barking is already reinforcing, he may not care much.

---

Edit

Further information about why dogs might bark: New Thoughts about Barking - Particia McConnell, PhD, CAAB

Answer 2

Psychology Terms

When talking about positive/negative punishment/reinforcement from a psychological perspective, it can be easy to get tripped up by the emotions around the words themselves. (You have punishment and reinforcement laid out properly in your question, I'm just adding extra details for future visitors)

The way to look at it is that the positive/negative and punishment/reinforcement are two completely separate categories. Positive/negative indicates how you are trying to change the behavior. Adding something is positive and removing something is negative. Punishment/reinforcement indicates the end goal of the behavior change. Stopping or reducing the behavior would use punishment. Increasing or encouraging the behavior would use reinforcement.

Positive/Negative and punishment/reinforcement are not necessarily good or bad or mean or fair, they just describe how you are trying to adjust the behavior.

Examples:

When you turn on a car, the car makes annoying beeping sounds until you put on your seat belt. The car is trying to encourage you to wear a seat belt and removes the annoying beep once you do the desired behavior. This is negative (removed beeping) reinforcement (want more seat belt wearing in future).

When a team of employees successfully completes a challenging project, the company throws a party for them. The company is trying to encourage the employees to complete challenging projects and adds the party after the project is successfully completed. This is positive (added party) reinforcement (want more challenging projects completed in future).

When a child throws a tantrum at the playground, the parent takes them home immediately. The parent is trying to discourage the tantrum and removes the child's access to the playground when the tantrum occurs. This is negative (removed playground) punishment (want fewer tantrums in future).

When a driver drives faster than the speed limit, they are given a traffic ticket by a police officer. The officer is trying to discourage breaking the speed rules and adds the traffic ticket when the driver breaks the rules. This is positive (added ticket) punishment (want fewer occurrences of breaking the speed limit).

Applying to your situation

All the various forms of behavior change will work, provided that you are consistent when you enforce them. You also do not need to stick to just one, you can use more than one punishment/reinforcement to affect the same behavior. (In fact, you'll find that most of the time you are naturally doing both a punishment and a reinforcement.) With that said, let's dive into the specific challenge presented here.

We want to discourage Maple from barking at friends and family. Because we want to reduce the barking, this will be a either a positive punishment or negative punishment. It does not need to be mean or vindictive, but it is a punishment nonetheless. Simultaneously, we want to encourage Maple to be calm when friends and family arrive. This will be a reinforcement of the calm behavior.

The effectiveness of the punishment and reinforcement system will depend on what motivates Maple. We have a pug that loves both treats and attention and hates being ignored, but other dogs will vary based on their own personalities, so feel free to substitute the rewards and punishments for whatever works with your own dog. With that said, DO NOT use any sort of physical or verbal abuse as a punishment, it does much more harm than good and teaches your dog to be afraid of you more than anything else.

If Maple is excited to see your visitors and wants to interact or play with them, pick Maple up and take her to another room and don't let her come out until she stops barking. This is negative (removing Maple from the exciting visitors) punishment (want to reduce the barking).

After Maple stops barking, bring her back into the room and allow her to play with the visitors as long as she doesn't bark. This is positive (adding Maple back to the exciting visitors) reinforcement (want to encourage interacting with the visitors calmly, without any barking). If Maple starts to bark again, repeat the punishment until she stops, then bring her back and try again. It may take several cycles for her to understand and break out of old habits, especially if she has been barking before without any consequences.

Eventually, when friends and family arrive and Maple does not bark at all, go all out with the reinforcement. Give treats, praise, petting, ear rubs, tummy rubs, whatever she appreciates the most. (If you don't want to give her too many treats due to her small size, break up a treat into little pieces and give them to her one by one. It makes the single treat seem like more treats, even though the end result is the same total volume of treats eaten.) You want to make it clear as night and day when Maple is doing what you want, and lots of reinforcement the first time she does this correctly will help her link the positive outcome to the behavior. After she is consistently behaving as expected, you can start to lower the reinforcement, and eventually switch to variable rewards, where you give the higher value rewards only periodically. It is still important to provide at least some reinforcement (such as verbal praise or a quick pat) so that Maple keeps up the behavior and doesn't revert to her old ways.

tl;dr

Your fears are correct. You should not give her a treat when she barks at visitors, because that will encourage her to repeat the behavior. You should instead give a treat when she does not bark at friends and family at all.

To use fewer treats, break the treats into smaller pieces. As long as Maple understands that she is getting a reward when she behaves properly, the size of the reward is not as important. (Caveat being that the first few times she behaves properly, you want to lavish her with rewards to really drive the point home). As Maple gets better at understanding what she needs to do, you can decrease and/or randomize the reward and still have her behave.

As long as you are not physically or emotionally harming her (and you seem to be a caring pet parent, so I don't see that as being a problem), the only risk is that you accidentally encourage the wrong behavior. You can always adjust how you set up your rewards and punishments to correct that.

Maple's age does not affect how well this will work. If she has developed a habit of barking at friends and family, it will take longer to break that habit than if she was a new puppy, but she will still get there eventually. Being consistent in applying your rewards and punishments is key. If you only punish her every three times, she will learn that she can bark 2/3rds of the time without consequence.

Alternatives: If Maple is not barking, you can have your friends and family be the ones to give her a treat (replacing the treat you would give her), helping her associate friends and family with good things. It is very important to not give any treats if she barks, you want her to associate the treats with being calm and not barking.