A Teaching Aside
The other day my wife and I were reminiscing about our time doing animal research and she was describing my animals’ behaviors while running different experiments. We looked at each other and she suggested the following blog post would be fun to write.
I thought it would be interesting to explain a bit of background on why I approach behavioral management of students the way I do. I seem to have somewhat of a unique perspective because my background focused on rodent AND human behavior. This means I actually spent 15ish years of my life doing the type of research that many of the behavioral methods used in classrooms and ABA are actually based on, particularly as related to reward schedules.
My Research Life
When I was an undergraduate student, I started to work in a rat behavior laboratory as part of a work study program. I ended up spending almost 8 years there. What we were interested in were the roles of different parts of the brain for learning and memory. However, I am actually going to focus on a more mundane part of rodent (and mouse) training: acclimatizing the rats to dealing with humans, other rats, and new situations.
Before we allowed any student to start a behavioral experiment with a rat, they were required to spend 5 days handling the rat and feeding it Froot Loops for 15 minutes a day. To do this, there was a very simple shaping process. We started by moving the rat’s cage over to the table the new student researcher was sitting at. Step one/Day one was having the researcher reach into the rat’s cage and put their hand in the cage away from the rat. The rat was allowed to come over and sniff the researcher’s hand and cuddle up to the warmth. Step two/Day two was the researcher reaching into the cage and gently grabbing the rat around the midsection, lifting the rat an inch or so, and then letting the rat go. Step three/Day three involved picking the rat out of the cage and placing it on a lab coat on the researcher’s lap. The researcher then kept contact with the rat with at least one hand and basically pet the rat for 15 minutes. Step four/Day four was identical to Day three. Step five/Day five involved the researcher picking up the rat and walking around while holding the rat and petting it. Note: The same process works for mice!
We found over time that this was an essential step to the success of our research program. If anyone skipped this habituation stage, then the rat would never perform optimally during the experiment. The rats would be impulsive and show a tendency to attack or try to attempt an escape from the researcher. These rats would react poorly (i.e., flip out) when they were touched and would react negatively to noises in lab. Out of curiosity, we quantify this by looking at the mazes and tables after unhabituated rats were finished, the mazes had significantly more rat droppings that stunk something awful and the droppings were not as solid as normal. There was also increased amounts of urine on the mazes as well. We concluded that unhabituated rats were stressed and unable to meet expectations during the behavioral tasks.
I distinctly remember that others in lab were always trying to figure out why my rats were always so smart. They also wondered why I would often perch a rat on my shoulder like a parrot and walk around with a rat nuzzling into my neck. I think it was always a shock when I explained that I had smart rats because they were not anxious or stressed out – in more anthropomorphic terms – they trusted me. At least that was until I had to use aversive stimuli as part of my research. But we will get to that later.
Trust translates into motivation
We had three types of behavioral experiments: 1. Experiments that used Froot Loop rewards to motivate behavior. 2. Experiments that used a rat’s natural tendency to explore their environment as intrinsic motivation. And 3. Experiments that used aversive stimulus to condition fear (the shocks were the same experience as touching a 9-Volt battery to your tongue – it was aversive but not painful).
Interestingly, the first two types of experiments, those relying on either positive reinforcement (Froot Loops) or intrinsic reward/motivation (exploration) required the researcher to develop a good relationship with the rat. That is, the 5 day habituation protocol (I called it habituation or snuggle training) was required in order for rats to perform optimally during the tasks. If that period was skipped, even control rats (those that did not have any experimental manipulation done to them) had a very hard time learning the tasks. However, if we helped the rats to trust us, then they were able to do remarkable things.
With trust, we were able to motivate rats to jump across an increasingly large gap between tables, learned to tell the difference between things that were extremely similar, wait a rather long amount of time to respond for reward, and learn rather complicated rules. If interested, here are a few links to some of my more interesting behavioral experiments that required rats to learn using rewards (Click Here and scroll down to “Behavioral Dissociation of Brain Function”). If you clicked any of those links you have guessed that these rats were required to put forth a fair amount of effort to accomplish these tasks. Often times we would watch them and feel a weird sense of pride in the rat stopping at a decision point and looking back and forth before acting – they were thinking. They were working hard. And they were rewarded with Froot Loops for their hard work.
For rats that were part of these first 2 types of experiments wherein the rats either engaged in intrinsically rewarding behavior (exploring) or I rewarded them, there were never any issues with rats trying to escape from the maze or their home cage, brutal self-injury, or attacking other rats, etc. I asked my wife if she had this same experience and she said she did, her rats were sweethearts that never tried to even bite her. But my wife and I were fastidious about giving each and every rat those first 5 days of habituation and in giving them individualized attention while we were weighing and feeding them. The rats treated us as friends.
Distrust translates into … ?
When we had student researchers that neglected to handle the rats as we instructed, there was an entirely different profile. The rats were jumpy – as in they would jump out of the cages, the researcher’s hands, off mazes, etc. When people walked by the home cages, the rats cowered in the back of the cage and made anxious squeaking noises. These rats were notorious for biting researchers, so the students had to use thick gloves to handle them. These animals left diarrhea and urine all over the experimental apparatus. These rats tried to jump off the mazes to hide in corners. They attacked other rats and engaged in serious self-injury. In other words, these rats were treating humans as, for lack of a better word, predators. And they were doing what was necessary to intimidate us and to get us to back off.
Often times, I would have to step in and rehabilitate those rats. We had a strict policy in the research lab that rats were not to be put down for aggressive behaviors or in any other way “wasted”. So, tough rats came onto my docket, because I found out early that I had the patience and ability to work with the “tough” animals. My first approach was to do the 5 step habituation session. Even though I was bitten quite a few times during this process, I never got mad at the rat and I definitely never gave up on them. I let the rats realize I was not going to do anything bad to them. Most came around at this point. If there was a rat that did not come around with that shaping, I would more formally write a plan to tame the rat. This usually involved a lot of noncontingent reward. I would crush froot loops in my hands and let the rats eat out of my hand and lick the sugar dust off of me. This was irrespective to them biting me. I fulfilled one of their basic needs, and it worked.
This is not rats, but the corporal cuddling segments seem relevant to how I helped rats learn humans are not too bad…
What about that third kind of experiment … aversive
This is actually the point of this post. Fear conditioning sucks. Royally.
Here is a typical fear conditioning experimental protocol:
- A rat is placed in a clear box with a metal rod floor and allowed to explore for 2-3 minutes.
- At the end of this period, a 10 second tone comes on, and the last 2 seconds of this tone is paired with a shock. The shock is not painful, but it is rather unpleasant and super-duper scary.
- Then there is silence for 1 minute and the tone begins
- … and so on for 8-10 trials.
- For the next 2 days the rat gets either put in the box in silence for 10 minutes or put in a new box with the tone for 10 minutes to see if they freeze or show fear across days.
Clearly, there is absolutely no rewarding component to this experiment. It is pure punishment, and used to study fear pathways and how fear responses develop. I think it goes without saying that these rats are not happy at the end of the experiment.
Well, these rats knew exactly who put them through this cough me. These rats would jump against the cage top and dislodge it to escape. They would aggressively self-mutilate to the point of actually amputating their own limbs!! They would posture and attack the side of the cage and lunge at me as I walked by. When I had to get a hold of them, it took all of my skill in handling rats to not get bitten. And many of them left me with scars on my hands I have to this day.
Many of these rats were loving and calm while running rewarded tasks. When I changed and started providing punishing stimulus, they changed. I did not change what I was doing with regards to giving them attention and calm touches – in fact I increased the noncontingent attention and rewards – but it was never enough. The fact that I was providing or administering punishment was enough to change the researcher-subject dynamic. I punished the rat, so the rat was going to defend itself. I was no longer associated with reward.
Let’s bring this back to the classroom
I am quickly becoming an outspoken critic of response cost and other behavioral management methods that amount to positive and negative punishment because they just do not work in my opinion. I hear over and over that if there are enough positives then the negatives of a level system are not that bad. I can tell you from experience in rat research (which all the ABA methods are fundamentally based on), there is no amount of reward I can give (including literal handfuls of Froot Loops in their home cage) that makes up for punishment. What I mean by that is that punishment is infinitely more salient than reward. So, we can give a million gold stars in a day, but all the kid is going to remember from the day that he was moved down from green to yellow.
These ideas are actually supported by research in classroom management by Robert Marzano among many others that demonstrate for “typical” kids there need to be at least 4-6 positive statements for every negative statement used in a classroom (and that is per child, not per class, so if you correct a student, you need to praise that student 6 times). For special education students, the research keeps expanding upon that number, requiring greater and greater numbers of positive interactions for any negative or aversive interaction (the last time I checked the research articles, it was 16:1 ratio). This clearly demonstrates just how influential any negative or punishing comments can be.
Negative systems hold kids back. It teaches them that they have to passively conform to adult demands or they get punished. Unfortunately, it also teaches them that the teacher is to be feared. Just like my rats, kids will fight back by yelling, screaming, punching, biting, spitting, running away, hiding, and so on. Kids cease to be kids and draw into an almost primal state of self-preservation. And it is not their fault. We did it to them. Just like I did my rats.
However, if one runs an exclusively positive classroom management system (which is very, very hard to do), then students will learn. Positive systems provide an atmosphere in which the students are capable of a great many things: 1. They trust teachers. 2. They try new things. 3. They feel safe enough to ask for help. 4. They accomplish what they thought was impossible. 5. They believe in themselves.
All of these capabilities combine to make learning a positive experience, and that should be our goal as teachers in the end.