One of our hopes in creating PEA Soup was to provide a forum for discussion about certain issues that may crop up in teaching moral philosophy. I suppose, then, that this is the first post on that topic. For several years now, when introducing Hobbes to students, I run a version of the Hobbes Game, which I believe was created by John Immermahr and published in Teaching Philosophy over ten years ago. The idea is to introduce the Prisoner’s Dilemma (an interpretation of one aspect of Hobbes’ state of nature) to students in a dramatic way, one that forces students to really figure out for themselves what it’s individually rational to do in cases of strategic interaction. So this is how I present the game. At the beginning of the class period I say that 10% of their grade in the class will be determined by the grade they request during the class period. The twist is that they’ll write down the grade they want (either an A or a B), and they’ll be brought up in pairs to submit them simultaneously. If they both request a B, they’ll both get a B. If one requests an A and the other requests a B, the first will get the A and the second will get an F (and vice versa). And if they both request an A, they’ll both get D’s. I then have them think about their strategy for a few minutes (without talking to one another), and then we begin. Some students at first will try to “cooperate” by asking for B’s, but soon enough someone will request an A, resulting in an F for the other, and inevitably the strategy mostly becomes asking for an A (which then results in lots of D’s). Some students invariably get rather distressed by what’s going on, and I’ll occasionally offer students who got F’s another chance at it (and it’s always suprising when they ask for B’s yet again and get burned). As soon as the exercise is over, I immediately announce that the grades don’t count (much to their relief), and we then discuss what the best strategy was, and why, before talking about the direct relevance to Hobbes.
The issue is this: most universities have in place a policy against psychological experimentation on human subjects without their permission. I’m interested in hearing from others who have used this game in their classes (or exercises like it) about whether or not they think it constitutes “human experimentation.” Clearly the game wouldn’t have its intended effect if we were to get students’ permission beforehand to run it. On the other hand, there is some distress involved during the game itself (which is very quickly replaced by relief and laughter, once the truth has been revealed), but is this enough (a) to think of it as experimentation, and (b) to undermine it as a legitimate teaching tool? It is amazingly effective: there are very few other examples I use in class that are burned into students’ minds as well as this one, and that’s precisely the impression I want to leave. But it remains a serious question: is this kind of manipulation (however brief) appropriate in class? And this is related to a larger issue: what precisely are the limits to getting across a serious point in class? I’m very interested to hear what others thinks about this.
Dave,
The question you ask is whether your strategy amounts to an unacceptable form of experimenting on a group of people without their permission. In order to answer this question we must ask why performing experiments on people without their permission is objectionable in the first place.
Perhaps the Kantian answer applies here: in performing an experiment on a person without her permission, you are using her simply as a means and not as an end in herself. What you are doing is not subject to this criticism. You are attemtping to teach the students about a difficult concept in a unique and hands-on way; the goal is to allow students to better themselves. In doing this, you are treating them as ends and not merely as a means.
Since your goal is to teach, I hesitate to call what you do an experiment. After all, the goal of the exercise is not for you to observe and quantify the responses of the students and then reach a conclusion on the basis of that. So you are not performing an experiment but merely teaching well.
However, even though you are not using someone merely as a means, you are lying to your students. Perhaps that raises another issue. However, I lie to my students sometimes. For example, when I give tests in very large classes I will tell my students there are four separate tests, A, B, C, and D, when in fact tests A and B are identical to each other and tests C and D are identical to each other. This makes it easier for me to create and grade the tests, and makes the students less likely to cheat, which is in their own long terms interests. I see no moral problem with a lie of that kind, and I think your lie is similar to it.
In short, keep up the good work in the classroom.
Thanks for the post, Scott. You make some very good points. I think the larger topic is indeed whether or not lying to our students is occasionally justified — it certainly may be on consequentialist grounds, but I often don’t feel comfortable on consequentialist grounds; they’re subject to tremors!
I’m not so sure this is experimentation at all. Experiments are usually done as a means to get information from the experimented body. How they react to stress, or what they think about such-and-such movie. There seems to be no element of that in this “experiment.” As a professor you’re not trying to get some bit of information from your students. It is of little or no concern to you in what way they respond to your grading system; your concern is one for them doing a certain amount of critical, prudential thinking. They’re being traumatized (to whatever diminished degree it may be) is merely a result of them walking that path. Compare this to watching a graphic video on the plight of the poor in Third World countries, as I have several times in various classes. The experience is traumatizing indeed, but it is not anything close to an experiment. Again, the distinction being that the professor’s goal is not to get information from her students (as a Psychologist might seek when sending electrical charges through a rat’s brain), but to teach them something.
Since your interest in your student’s is didactic and not inquisitive, it seem that this Hobbesian game is not an experiment at all; but, as you said, a game.
I’m not so sure this is experimentation at all. Experiments are usually done as a means to get information from the experimented body. How they react to stress, or what they think about such-and-such movie. There seems to be no element of that in this “experiment.” As a professor you’re not trying to get some bit of information from your students. It is of little or no concern to you in what way they respond to your grading system; your concern is one for them doing a certain amount of critical, prudential thinking. They’re being traumatized (to whatever diminished degree it may be) is merely a result of them walking that path. Compare this to watching a graphic video on the plight of the poor in Third World countries, as I have several times in various classes. The experience is traumatizing indeed, but it is not anything close to an experiment. Again, the distinction being that the professor’s goal is not to get information from her students (as a Psychologist might seek when sending electrical charges through a rat’s brain), but to teach them something.
Since your interest in your student’s is didactic and not inquisitive, it seems that this Hobbesian game is not an experiment at all; but, as you said, a game.
Dave, two comments re: the Hobbes game. 1) It is my understanding that pre-experimentation review (IRB clearance) was not originally intended to be applied to research in education. I am uncertain whether the laws have changed since, judging by my faculty paycheck, I am no lawyer. However, that might be the start of an excuse if you are ever called before the MAN. 2) I do NOT think that use of the game is a form of ‘research’. (a) technically research usually involves a specific question and a specific sort of intervention. The Hobbes game has neither (in a flat-footed sense). (b) If you say: “imagine if…” you avoid the problem, but students fall back into their static-filled inner life. The use of vivid experiments, some of which can startle or cause fear, are common tools in teaching. A chemistry teacher who explodes a hydrogen filled balloon MIGHT cause a student to wet themselves of have a heart attack, but it is permissible (or was the last time I checked). The Hobbes game is more like that–you expose the students to a contrived event –not because you want to study their reaction, but because you want them to have a certain kind of experience.
Isn’t our job to upset our students, if only a little?
The question of professorial intent is an important one here. It’s certainly true that the intent is to make students think critically about the prisoner’s dilemma and it seems quite effective in this. But quite apart from the issue of intent is this: what knowledge has been generated through this classroom intervention? It seems to me, based on the description provided, that some knowledge has indeed been generated. For example, we all now know that: (1) some students will try a cooperative strategy; (2) some will go for the “A”; (3) others will go for the “B” and get burned; (4) some are distressed by the classroom scenario; and (5) this is a pattern of behavior that repeats itself from year to year [I assume that this is true, but it may not be]. We could all try to repeat this in our own classrooms and we’d have several hypotheses at hand, namely, (1) through (5), or something like them. And, if we investigated further, we might all find out that certain strategies are correlated to, say, gender or ethnicity. Described in this way—and I’ve described it in this way to make my point—this classroom intervention is starting to look like experimentation, namely, the production of generalizable knowledge, regardless of the pedagogic intent of the professor.
So I’m not sure that the prisoner’s dilemma game isn’t experimentation. Whether something unethical is afoot here is a separate question, one to be answered only after attending to the harms (minor, transient anxiety among some), the wrongs (having been lied to) and the benefits (an interesting learning experience) involved. And while it may be true (although this is an empirical question) that telling them to simply “pretend” that 10% of their grade depends upon their submitted choice will not engage them as much as if something concrete depended on it, I suspect that they’d be engaged enough to satisfy the professor’s pedagogic intent.
Wow, fantastic way of teaching the prisoner’s dilemma. I don’t see that you’re testing a hypothesis, or harming them for that matter. So I hope you don’t mind if I steal your idea and use it myself!
Just a precaution: As soon as Susie, believing the grades are real, becomes distraught after picking “B” twice and getting burned with two Fs, and works herself into an anxiety attack… you’re gonna have a problem on your hands. Although, as Robert pointed out, you aren’t necessarily testing a hypothesis, your withholding the nature of the experiment could be viewed as manipulative by a loudmouthed parent. Conceded, this is an excellent way to teach the Prisoner’s Dilemma, but it’s my opinion that you should at least clear this with the Chair or Dean first. All in all, though, very creative stuff…
“Games” such as this is what constitutes good teaching. Because you tell the students by the end of the class that it really isn’t going to count as a grade, you have not caused undue anxiety or suffering. Whatever the students DO experience during the class is exactly what makes the lesson stick with them. If only more such games were played! However, a line is crossed if the students are allowed to continue believing something stressful and upsetting once the class has ended. There is enough in some college classrooms that is stressful and upsetting and at the same time legit standard procedure without being falsely so! I doubt if playing the “game” you described has ever caused any students to complain. Rather I would imagine that you have impressed them, taught them, and made them think, and for that, they are praising you.
Thanks to everyone for their comments. Just a quick reply here to a few of you. First, Robert, who lists the various bits of information gleaned from this “game” and then suggests that it may constitute experimentation because it produces generalizable knowledge. Not to go all Doctrine of Double Effect on you, but given that such knowledge isn’t the intended effect of the game, that it’s only a foreseeable side effect, does it really seem as much like experimentation (where the intended effect *is* the generation of knowledge)?
Oh, thanks also to Bob Barnard — I too lean to the view that our job is indeed to occasionally upset our students, and I’m not so convinced that it should be only a temporary upsetting. I remember my first philosophy class in college as being deeply troubling — it was completely messing with my parochial little world view, and I couldn’t stop thinking about what was going on in that class. And thank goodness for that! (A long overdue shout-out is owed to that professor, Dr. Brian Sayers, who left philosophy to be a landscaper about ten years ago because, as he put it, he felt he was having no effect on his students — au contraire, mon frere.)
I borrowed this technique from Dave, first using it in 2001. However, I hoped that Robert’s suggestion would be right, that if I asked the students to merely *pretend* their grades were at stake, rather than actually deceiving them into thinking their grades really were at stake, this would have the desired pedagogical effect, without having to engage in any temporary deceit or anxiety-production. Unfortunately, my hope was not fully realized. To be sure, the students did get *something* out of the pretend exercise, surely including the main point, which is to understand the Prisoner’s Dilemma. But it also seemed to lose a bit of its vividness, which might better convey the force of the Hobbesian psychological premises behind the dilemma. So, on my limited experience, I’d say that the pretend version is somewhat effective, but less effective than the version in which the students are temporarily deceived into thinking that their grades really are at stake. None of this comment on effectiveness, though, speaks to the separate question of whether the exercise is ethical.
David — I guess I would respond to your double effecting me like this: When one invokes the intended/foreseeable side effects distinction, the subject upon which one focuses is typically the same, e.g., I intend to ease the suffering of this person by administration of morphine, although I foresee that the morphine will hasten this same person’s death. You could invoke this in your situation: you intend to teach your students about the prisoner’s dilemma, although you foresee that this might cause some of them minor and transient anxiety. This may or may not be sufficient to justify the classroom exercise. But I don’t think you can invoke the distinction in this way: you intend to teach *them* about the prisoner’s dilemma, although you foresee that this will create generalizable knowledge for *you* (and all of those with whom you share your experiences). The intended effect occurs in the minds of your students, while the foreseeable effect occurs in your mind (and those of your interlocutors). I’m not sure that the intended/foreseeable side effects distinction travels across subjects in this way. Put differently, I think that the experimental flavor of your classroom exercise is a structural feature of the situation, one that is triggered at the point when you say “10% of your grade will depend upon…”
Robert
I’m not completely sure about this, but the general policy at my university is that human subjects approval/IRB clearance is required only if you intend to publish the results of such an activity. This, of course,completely begs the question of the effect on the students.
BTW, my reading of game theory suggests you’d get different outcomes in a repeated game, but that’s another issue.
Thanks, Donald. Yeah, iterated games yield a different result, but I’m just trying to introduce a very limited amount of game theory to get students to see and think about the strategic interaction in a new way (for them). I ACTUALLY think the best representation of the Hobbesian state of nature isn’t a prisoner’s dilemma at all. Rather, it’s what I and Graham Dodds (following an unpublished work by Greg Kavka) call an Assurance Dilemma (see our “Why We Can’t All Just Get Along: Human Variety and Game Theory in Hobbes’s State of Nature” in one of last year’s Southern Journal of Philosophy issues). It’s a combination of the Assurance Game and the Prisoner’s Dilemma, and it actually takes note of the fact that Hobbes discusses two types of people in the SN: moderates and dominators, each with different preference orderings (the dominators prefer unilateral non-cooperation, while the moderates prefer mutual cooperation). The PD can’t capture this aspect, assuming as it does that both parties have the same preference orderings. Nevertheless, even with the Assurance Dilemma structure, the war of all against all would still arise.
Roshi,
This is not experiment — it is a koan.
All the best,
So-Called Austin Mayor
From a regulatory viewpoint: Probably not research, so not governed by human research protection procedures (e.g., IRB review, informed consent). The regulation most often applied in universities (the Common Rule, adopted by 17 federal agencies that fund the research upon which universities depend) has this definition: “Research means a systematic investigation, including research development, testing and evaluation, designed to develop or contribute to generalizable knowledge.”
Generally universities apply the same set of policies and procedures to all their research, not just the federally funded research, so yours most likely has adopted this definition. As Robert commented, intent is paramount in deciding whether an activity is research.
Regulations and ethics, of course, do not always correspond exactly, so the question of whether it is ethical to induce this type of distress in students is certainly open to debate.
The problem is a very specific one, as the game concerns an aim and a demonstration of optimal strategy to achieve that aim. You couldn’t have this type of dilema in teaching principles of natural science. The crux of the problem, and the power of it as a teaching tool, lies in the difficulty of determining the optimal strategy. If you are using the student for research, gaining permission assures in some way that they agree with the aims of their manipulation. If you are teaching them moral philosophy, you are in a position, and are probably trying, to change their fundamental value and judgement structures. Do you need their permission to do this? One could argue that their presence in class is the permission you need. But the more general tension is that between the subject as developing judgement through education, and the subject as independently participating in a widely accepted method of evaluation of claims. How do individuals participate in the construction of the latter? Scientific methods are fine tuned not by direct tinkering, but by extended and aggressive application. But application of “optimal” strategies for action can never be independently evaluated in the way that claims about the world can. It seems that their persmission would be meaningless in the moral philosophy class as they wouldn’t really have a basis on which to give it in that their sense of judgement with respect to these issues is not fully developed. Also, the concern probably arises from considering the experiment in light of a demonstration of a concept rather than something to be directly evaluated as a useful characterization of moral choice. Consider whether or not it would be unethical to use a student as Pavlov’s dog. It seems a wrong thing to do, but it is not really an experiment. But the separation of the individual from the lesson of the demonstration is clear.
Maybe this was already mentioned it, and I missed it, but I’ve used a variant of the Hobbes game in class where I told the students that it would count toward their final grade, but then when it was over, I told them it didn’t. This of course involves some deception as well, and maybe there are ethical problems with this approach.
I’ve also done the game where it counted, but for a very small fraction of the course grade (about 2 or 3 points out of 500. My experience is that students still take it seriously – they want every point that they can get.
Another possibility is to have rewards that are “external” to the grading system in the course. I know folks who’ve done the game with jelly beans or candy. In the Locke game (also by Immerwhar and some other folks), they suggest that the winner get a copy of Locke’s Second Treatise on Goverment…) By the way, the Locke game is really anti-Locke, so if you want your students to be skeptical of Locke’s political philosophy, it is a good way to get them thinking…
Incidentally, the Rawls game is also great fun…
Best,
Chris
No offense, but you aren’t experimenting on the students, as you already know the results. This isn’t a data gathering exercise, but a teaching one.
Any teacher who had used this in a class in order to teach predation as necessary and natural, and who included it in a grade, would have gotten appropriate teacher evaluations.
I still remember collating 50 written essays on why a teacher was worthless from a class that was particularly irritated. That teacher lost tenure.
Which is an appropriate lesson people like Chris need to learn as they have their fun. State and group action are always interesting when invoked by external irrational threats.
The research compliance officer helpfully gives us the CFR definition of research: “Research means a systematic investigation, including research development, testing and evaluation, designed to develop or contribute to generalizable knowledge.” However, it’s important to focus on the important part of this definition, namely, the development of or contribution to generalizable knowledge. The development of or contribution to generalizable knowledge, as I’ve mentioned before, is a structural feature of the classroom situation David describes. As such, the Prisoner’s Dilemma exercise has an experimental flavor. The other parts of the CFR definition shield this from view. In particular, a focus on “systematic investigation,” “research development,” and “testing and evaluation” all imply experimental intent. But this is a fault of the definition; intent underdetermines a proper evaluation of the situation. When one has the intent of contributing to generalizable knowledge, one then develops a protocol that outlines one’s systematic investigation of a question, is thus part of development in research, and will involve testing and evaluation. If one does not have this intent, however, then one will not put together a protocol for IRB review. Again, the CFR definition hides this; in particular, the “designed to…” locution, which puts intent at the forefront. But regardless of intent, the classroom exercise will generate knowledge about a specific question; there’s no way around this. And this is the feature upon which one should, in part, focus.
Now, one might respond to this by claiming that any teaching style or method, when shared with others, contributes to generalizable knowledge, namely, the knowledge teachers need to achieve their pedagogic goals. (The existence of the journal, Teaching Philosophy, attests to the importance of sharing methods that one may use to achieve pedagogic goals.) So, e.g., if a colleague suggests that to break up lectures and let students do small group work on a focused question for 15 minutes during class is helpful in getting students to think through and understand the material, then this too must count as experimentation. And, surely, to count a teaching method as being on a par with experimentation is absurd. And I would agree. But the difference between this situation, a teaching method, and David’s example, is that there is no lie, no manipulation, involved, however minor the lie or manipulation, however minor the consequences. And so I’m inclined to think that the initial question, “Teaching or experimentation?,” is not altogether on target. It is both. It’s clearly a great teaching strategy; and regardless of whether there’s a formal protocol involved, regardless of whether most would think that this classroom exercise should trigger IRB review, it’s experimentation. It seems to me that the question to which one should attend is this: Are the harms, however minor, and the wrong of being lied to, justifiable?
I have no opinion about whether your version of the Prisoner’s Dilemma game is experimentation or not. Regarding the appropriateness of it as a teaching tool however, I have a strong opinion.
Perhaps my reactions are atypical, but I’m surprised at the number of comments that have endorsed your game as excellent teaching. If a professor played this game (which involves lying, and causing the student momentary but completely unneccessary distress) with me, I would never trust anything that professor said to me again.
Furthermore, I would feel (rightly or wrongly) that the professor’s own ethical landscape was most likely seriously flawed. I would be very angry, and would have a difficult time getting past that anger to evaluate later teaching in an objective fashion.
Perhaps this strong reaction is a personal quirk, or perhaps it comes from immersion in the professional ethics of my own field (medicine), in which causing unneccessary distress and lying are both viewed as serious ethical violations.
In law school, we played an extremely effective version of the Iterated Prisoner’s Dilemma, involving oil pricing. We read pieces on the Prisoner’s Dilemma beforehand, as homework, so we all knew the score going into it. Then we split into 4 person teams and negotiated oil production and pricing within a classic iterated Prisoner’s Dilemma framework. Then we compared the various teams’ strategies and how well they had worked. The cooperating teams, as predicted, did best, and yet, even having read ahead of time that this would be the likely outcome, many teams still chose not to cooperate (and subsequently did poorly). It was an extremely dramatic lesson, involved no lying, and was respectful of the students.
Like some above, I myself wouldn’t do this because of the lying involved. (And I wouldn’t do it for real, because, for instance, some A-requestors will receive As, while others receive Ds, though they’ve played equally well.) Many have mentioned the stress caused to the game losers before it’s revealed that the 10% of one’s grade was a lie. Not mentioned yet, though (unless I’ve missed it): What of those who “won” the As? They’ve been promised those As will count as 10% of their grade, but that promise is broken.