Operant Conditioning (warning, this post is a beast!)
0 comment Monday, June 23, 2014 |
Some time ago I approached GrAdvisor about a particularly problematic data set. I proposed that the technique that we had been beating the dead horse with was just not going to yield any results no matter how long we flogged it. He agreed. He asked what I wanted to do with this data set instead. I proposed some other fairly well established techniques to get at the same question, but he didn't like any of these...and wouldn't say specifically why not.This was a bit frustrating. All the other techniques I proposed were perfectly valid ways to approach said question so what gives? Too expensive? No current expertise in the lab? It's not like we haven't gotten around these obstacles before so WHAT exactly is wrong with them? I asked him these questions and the guy completely blew me off.He said, "AA, you're restricting yourself. Think harder." And then ended the discussion. Um...what? Think harder? Like I don't think hard enough about my dissertation project as it is?I took this to mean that he felt that my proposed approaches were too conventional, so I did some more research and then proposed some newer, more novel techniques with the caveat that they had not been proven in our system."Do you think these techniques would be better?" I asked."No, that's not it either," he said. I paused, hoping he'd then enlighten me in his infinite wisdom as to what, pray tell, it might be. Nuthin'."OK, I guess I'll have to think more 'outside the box'.....""It's not that far outside the box," he snapped.Right. So I spent another week or so wracking my brains (and those of my colleagues) trying to figure out what the hell magic number he was thinking of. I did eventually come up with one more out-of-the-box technique, but I still to this day have no idea if it was the one he wanted me to guess.WTF, right? I mean, it was a total waste of time for me to spend my days trying to read his mind if there was indeed something very specific that he wanted from me. I don't expect him to hold my hand and I don't want him to tell me what to do all the time. But I think that it's reasonable to give specific requests if he actually expects me to meet his specific demands and expectations.I see this kind of set-up-to-fail communication going on all the damn time lately. What is up with this?I really think that it just comes down to convenience. It is easier to tell people that there is a problem and just leave it at that, than it is to specifically define our expectations. Plus there's the added benefit with the convenient option of congratulating ourselves on being soooo smart and having it alllllll figured out while Flunky over there flounders around because they don't know what to do since they are not as clever as our wonderful selves.Why the fuck do we do this? I mean really, isn't the goal of most interpersonal communication to get to a point where we are in agreement or at least on a similar level of understanding? I would argue that this is precisely the point of most communication. And yet, so often we're going about it as if it's some kind of contest in clairvoyance. I bet you can't guess what I'm thinking! No? Yay, I win! Except I didn't win if the communication didn't move the larger picture forward. Me being clever enough to hide what I'm thinking from someone else means that NOBODY wins, because whatever we're talking about just doesn't. get. done.Is it really sooooooo bleeding difficult to put our egos aside in the interest of effective communication?I mean sure, we all live in our own heads to some extent and if we're really being honest with ourselves we are each the most important things in our own lives. As such, it's pretty easy to get wrapped up in the need to protect our egos. It's easy to assume that any criticism that comes our way is a character attack rather than a suggestion about our work. Sometimes the two are so intertwined that it's hard to let them stand alone.I've been thinking about this a lot lately. How do we communicate most effectively so as to accomplish our goals? Are our goals the real motive in determining how we choose to communicate with one another?I've been thinking about this for a number of reasons:I'm involved in a protracted process of training DangerDog to respond to other dogs with something (anything) other than fear-aggression. Interspecific communication is difficult for obvious reasons.I've been frustrated by my complete and utter lack of ability to read GrAdvisor's mind, and his inability or unwillingness to see the problems in our communication.I've had some conversations with a friend regarding her frustration with her significant other's behavior, in which she wants him to do what she wants, but doesn't want to tell him what she wants. Does anyone else think that this is messed up?JLK and I have been having some interesting discussion via email about education and public science literacy (which I may post excerpts from in the future), which has me thinking more about science education (I topic that I am passionate about, but do not discuss much here for reasons that I will make clear in the long awaited posts about my ambivalence...they're still coming, I swear).I've also been following some rather snappy exchange going on in the blogosphere which is riddled with miscommunication from all sides.All of these things have sort of converged recently, so I want to blog a little bit about operant conditioning (this is not my research field so all you neuroscientists out there please bear with me), positive reinforcement, and constructive criticism.I have a fair bit of personal experience in implementing these techniques. I've trained a fair number of working animals to do their jobs successfully (mostly horses, a few dogs). I taught horseback riding lessons to more people than I can possibly remember, and in college I earned spending money by tutoring kids in various subjects, elementary through high school. So I've done a fair bit of teaching/training, and while I am by no means an expert, I like to think I am fairly good at it.Because I see effective communication as essential to doing my job (scientist, but also teacher, trainer) well, and because I think that poor or misguided communication is the absolute number one contributor to my frustration and my advisor's when I don't meet his expectations, I want to talk about good communication.Let's start with a simple example. If you want someone to execute a specific behavior, what is the best (i.e., most efficient) way to achieve a successful outcome? Maximizing the rate and probability of success is the foundation of operant conditioning (OC). OC is the technique used by behavioral scientists to train animals to perform a specific task that can later be used in an experiment. Most lay people are familiar with OC if they've ever been to SeaWorld. This is how they train dolphins and killer whales. It also works on any other species with a neural network, from sea slugs to humans. Basically, it goes like this:Subject performs desirable behavior.Subject receives reward.Subject associates reward with desirable behavior, and so continues to perform desirable behavior.You can also use operant conditioning to teach avoidance - in this case, substitute an undesirable behavior and subsequent punishment. I'll focus on positive reinforcement because motivation based on reward rather than fear is better at producing patterns of desirable behavior.The key to successfully achieving the desired result as a behavior pattern is setting the subject up for success, so that they get the reward early and often. This quickly co-opts reward pathways and establishes the desirable behavior as the default action. Failing to elicit the desirable response early and often usually leaves everyone involved very frustrated, and frequently fails to pattern the desirable behavior.Let's start with some simple illustrations of this principle:Scenario 1: Trainer teaching dog to sitTrainer: Sit! Dog: [lies down] Trainer: No! Bad dog! Trainer: Sit! Dog: [stands up] Trainer: No! Bad dog! Trainer: Sit! Dog: [Oh, fuck this!]What did the dog learn here? Nothing! He still doesn't know how to sit. Why? Because there are a thousand ways NOT to sit, and only one way to do it right. The dog has to essentially guess at what the trainer is asking and what are the odds he's going to guess right? One in one thousand.Let's try again:Trainer: Sit! [Guides dog's butt to meet ground, then gives reward.] Dog: [Huh. Treats! How do I get more?] Trainer: Sit! [Guides dog's butt to meet ground, then gives reward.] Dog: [Right on!] Trainer: Sit! Dog: [Puts butt on the ground himself, collects reward - Sweet!]What did the dog learn? Putting my butt on the ground gets me a treat! Sweet! I'm going to do that all the time! Success - it's easy.Scenario 2: Teacher wants child to write her name. Teacher: Can you write your name? Child: Hmmm...I know that names look like a bunch of lines [scribbles lines on page]. Teacher: No, that's wrong. Try again. Child: [Scribbles lines on page with a different colored crayon.] Teacher: No, that's not right either. Try again. Child: I don't want to anymore.What did the child learn? Certainly not how to write her name! The child learned nothing, except maybe that it's hard and stressful when the teacher asks you to do something.Teacher: Can you write your name? It looks like this [Prints child's name on paper.]Child: [Makes attempt to copy what she sees.]Teacher: Good! You wrote all the right letters...do you see how the "R" I drew has the round part on the this side and the "R" you drew has the round part on the other side? Can you write your name again with an "R" that matches mine?Child: [Writes name again, with the "R" in the right direction.]Teacher: Great job! That's exactly right! Do you want to practice writing your name some more?Child: Sure! That was fun!That did the child learn here? Not just how to write her name. Also, that she doesn't need to be afraid of making mistakes, and that the person asking her to perform this task is going to help her out. Success all around!I'm not going to bore you, dear reader, with any more examples. I think you can probably see how clearly defined expectations, positive reinforcement for desired behavior, and constructive criticism to improve the things that aren't quite there yet are essential components to effective communication, and achieving desired results. I think it's also not hard to extrapolate the benefits of applying this kind of communication to conversations with our colleagues, mentors, mentees, manatees, whathaveyou.But lest you think that I am being too hard on mentors, teachers, and trainers...the responsibility for implementing effective communication is not theirs alone. It gets complicated in these and other relationships in which the balance of power and authority is not equal. In part, because one person has something that the other one needs and this places the less powerful person in a vulnerable position. It's easy to become defensive, to assume that any criticism is a character attack or an unreasonable demand, when you're vulnerable. The trouble is, that's when we most often miss out on the real gems of constructive criticism. We're too wrapped up in our own hurt feelings and protecting ourselves from these powerful critics to hear some really useful shit.But who's responsibility is it to make sure that everyone's behaving like adults in these situations? I say, nobody's but our own. Because let's face it, that's the only person that we've got any control over anyway. On the other hand, when I take on the role of the more powerful person in the exchange (either by being the keeper of information, or by trying to solicit change) I do think that I have a responsibility, both to myself and to whomever I'm interacting with, to make my expectations clear, to give positive reinforcement where it's called for, and if offering criticism to ensure that it's constructive. Why does the responsibility fall on me? It's not like I should single-handedly be the communication police.JLK have been talking about this sort of peripherally in our email exchanges -- whose job is it to educate the public about science? Whose responsibility is it to increase public science literacy? I would say, that since I have taken on the role of "scientist" (in this interaction that makes me the more powerful holder of information) that places some responsibility on me. Does that mean I am obligated to engage and respond to every Joe Public out there on all matters scientific? Of course not. The idea that one person should be obligated to educate everyone else out there is preposterous. However, I am invested in science education because a more scientifically literate public means that they get what I'm doing, they support more research funding, and I can do my job better. Everybody wins. This is why I talk to creationists and IDers about evolution, if I think that we're actually having a conversation rather than a contest. [There are of course some people who won't engage on this level, and these are the exceptions. I don't talk to them because I know that they're not really interested in what I have to say, and nothing I say is going to make a damn bit of difference so I don't waste my time and energy.]I don't owe it to anybody but me. Because ultimately I'm the one who stands to gain, whether I've gotten the information I need or whether I've changed someone's mind for the better. At the same time, I owe it to myself to have the maturity and humility to hear constructive criticism for what it is even if my ego objects, because if I don't I've missed out on a big opportunity to make things better for myself.So I guess my point is that, in spite of the fact that we're the ones who stand to gain from an ego-free communication, maybe we should all make more of an effort to realize that it's not all about me and my ego all the damn time, whether we're in a position of power or on the other end of things. It's not about me being the almighty scientist who is always clever and right. It's not about making sure that every else acknowledges my poor put-upon student self. I am ultimately the one who stands to benefit from an effective exchange so I owe it to myself to check the ego at the damn door and focus on the information rather than the pissing match.