NB: Peter Sellers trying to set an example of good humor in The Party, Blake Edwards
The frosts of December usually brings two calamities , especially for the university that I am, the weather (snow, ice and other climatic merrymaking) surveillance and investigation.
Maybe I should not destroy the illusions of our little darlings, brown and other colors of hair displayed by our student (s), but, no, I'm not going to review my subscriptions with a light step and serene, happy to do my duty as a teacher and a sleep without sleeping troubles in post-tests which I know they happened in peace, in order that my students have the best chance to get away with gusto and can spend Christmas holidays with peace of mind.
[if the above sentence seems incomprehensible to you, reader, used the ax with which thou hast certainly pledged to kill one or two Christmas trees to cut, please, where you good fit. Advance, thank you]
So in the middle of last week, I went as usual at this time to a monitoring review. Then I come to my supervision, delayed for reasons that it is unnecessary to explain here, and I see that we are two monitors to one hundred students, which is roughly consistent with the standard. The other person was not a teacher-researcher (at least in business, universities sometimes using extra pensioners out of necessity, the number of teachers can not ensure their regulatory oversight is not negligible) and was paid for it. In my case, surveillance is part of my service obligations, and, like many academics in France are not remunerated.
I then started to perform routine tasks at the beginning of trial in such circumstances: card check students pointing present and verification of the total in each event. Then came the most difficult period, that of surveillance "pure", in which the supervisor goes through the examination room, or do not travel, or traveling from time to time (delete as appropriate depending your personal experience). In my case, I still tend to get around frequently, but not always and above all, guided by the interests of fairness, I make sure to watch "serious" to minimize the risk of fraud.
The other supervisor was at that moment sitting in the back of the room and placed me on the other side of the room. I present some administrative documents required to complete and endow me the usual package of leaves and other brouilllons review copies and inserts that will ask me, in all likelihood, students now compose. Then I start to walk the room with a firm step.
At this point, the supervisor who stands with me, then also starts to go (but with no less determined than I am seems to me there) the examination room.
After one or two rounds of room, I sat on the chair and sit. I see that my colleague is then the back of the room and sits well.
Moments pass, time to connect my computer to access my mail, then I take back my package of copies, and get up in the idea of a new route to the amphitheater. Ten seconds later, my fellow sufferer gets up and is in turn a journey as I continue my own. Then, once completed my course and some students have satisfied their urge to white paper and another copy, I sit back in the pulpit. My colleague then also sits. Having noticed this little game, I wait a few minutes to make sure it is not a mere coincidence, he did not simply lifted along with me a few seconds apart, chance that we have modeled upon one another our cycle surveillance. I decided to wait a little longer in the sitting position. He remains silent, concentrated in the reading I do not know which magazine. I set it and then get up again with a determined air. Five seconds later he gets up too and did the same route as above. Then again, I sit and it sits well. This game was repeated as at least five times before I am busy with other pursuits relating always to the same supervision.
After the test, amused by the experience that I had to live (he does not take much to liven up this boring exercise, as you see reader), I tried to discern the behavioral foundations, the outcome of this game like me seemed particularly clear.
How to describe the situation? A monitoring review is more effective (to minimize fraud) and just that the level of effort and cooperation of supervisors is important. More supervisors roam the examination room, students can be more focused on their work and even get a copy or another when they need it. In fact, the situation is that once the public good. Every supervisor who exerts a costly oversight in fact benefit the students and all supervisors in the room, then he would prefer to read either France or West Matthew Rabin, or his favorite blogs economy.
In this situation, we find the usual behavior of a position of contribution to public good: there are contributors fans, who will go through the room in all directions to deter fraud and effectively to meet student demand. There are also riders who will do as little as possible, sitting down to read them newspaper and / or eat their sandwiches while others do the job, and he will not do you move an inch. There are also those who discuss and gather between free riders "like attracts like" the saying goes).
Then there are conditional contributors, those who are waiting to see what you do to decide.
In my case, depending on my mood and level of occupation, I'm a contributor unconditional or conditional. But if the other decides to do nothing or do very little, I think it's impossible for me to do the same, where the race will turn the show and at that time, I exercise a significant level of effort to compensate for the free riding on the other. Moreover, if the other is very far, I break a bit until I found it relaxes (sometimes, very rarely, that never happens) and now here I am then relay.
In these circumstances, the role of a "leader", that is to say a person who chooses to work to show what he thinks it is good to be decisive, as I have had some sort of evidence from the monitoring mentioned above.
After all, why this behavior on the part of the supervisor? He thought that since I was a teacher-researcher and her supervisor simple, I had a kind of ex post checks on him, the risk is that I indicate insufficient effort monitoring? I do not know too much and like what was happening was too good to be true (it was a sort of natural experiment), I tried to remember what was said by the experimental literature on the topic of leadership. More precisely, a leader is simply someone whose departure we observe the behavior, this observation help us form our decision.
Imagine for example a sequential prisoner's dilemma, a player playing in the first and one second. The second therefore observed the action of the first, and say he is the follower, the first player being the leader.
If the first player cooperates, the second player can decide to be strictly rational to defect, which brings in the maximum gain. Moreover, it may also decide to cooperate and get the Pareto-optimal gain. But the argument of backward induction advocates show that the balance of the game is the defection of players successively, as in the extensive form game given below:
The effect of leadership has been extensively studied in experimental economics for ten years, the theoretical reflection on this theme is initiated by Hermalin in 1998.
In the experiment conducted for example by Guth et al., 2007, to study the influence of leadership on the levels of cooperation, the chosen game is a classic game of contribution to public good. One of the treatments used as a benchmark (benchmark) and consists of a simultaneous game of contribution to public good (VCM: Voluntary Contributions Mechanism) under standard conditions (see here for new readers or Readers who have forgotten). In other treatments, a randomly selected entrant (the leader) determines its level of contribution to public good first, which is announced to the other three members of the group (followers) before they determine their own level simultaneously Contribution (VCM With leadership). From a theoretical viewpoint, the fact that a player announces his first contribution leads to the same theoretical predictions as in the case of simultaneous contributions: the dominant strategy is to pick a contribution level equal to zero, we fall back on classical problem of free riding behavior. These authors also treatment they call "strong leadership": the leader chooses its contribution, announced the three followers, who determine their own contribution, then after that, the leader becomes aware of these levels of cooperation and may decide exclude one of the followers for the next period (each participant can player up to 25 times that game). In this case, the excluded player earns its endowment and can not take this opportunity related to the public good.
Furthermore, the leader is determined in two ways: either he was drawn early in the session and remains the leader during the first 16 periods of the repeated game of contributions (there are 25 repetitions in all), or the 4 participants will have the opportunity to be a leader for the same number of periods (4 times each then), the sequence of successive leaders being announced in early Thurs
In fact, it's a little more subtle like design (that's why I always liked what was Werner Güth, it is both rigorous and subtle an experimental point of view ), the first 16 periods of the game determine the leader exogenously, and the periods from 17 to 24 give participants the opportunity to each group to choose endogenously their leader. In fact, during the period between 16 and 20, participants have the opportunity to vote, or that they want to keep their leader (or not) or to say who they want to have as a leader among the 4 members of group. The graph below shows the results on the average contribution of the participants in the benchmark treatment (control), a leader and strong leader:
Source: Guth et al., 2007, JPE
The results are both expected and intriguing: the average contribution level is all higher than the power of the leader is strong: it can announce its contribution and exclude (punish) a member of the group, the average contribution is twice as strong as in the case of simultaneous game without a leader, and represents approximately 80 % of the total allocation given to each participant (25 tokens). For treatment Leadership "simple" (without exclusion power), the average level of cooperation is a bit higher, but the impact of leadership tends to wane over time to converge to the average contribution level without leadership. The power of exclusion is a clear difference in terms of level of cooperation!
They reported no significant effect on the determination of exogenous leader: having always the same leader or there is rotation does not influence the levels of cooperation. The most important result is that they measure a significant correlation between the contribution to the Leader and the contribution of followers, which runs counter to theoretical predictions of the simplest (mutual defection).
course, they also find that the probability of being excluded for a follower is more important than its contribution is small compared to the average contributions of his group. On average, the leader has ruled out a follower in 20% of cases, exclusion is even stronger than the deviation of the contribution of the follower relative to the mean was significant (only the negative deviations, my contribution is lower the medium-increased the probability of exclusion).
In short, leadership increases the level of cooperation so essential, that remains an enigma as a theoretical point of view. In another recent experimental study, David Levy, Kail Padgitt, Sandra Peart, Erte Xiao and Daniel Houser (2010) show that this positive effect is real only leadership if the leader is human and not a robot (not a joke, it was published in the Journal of Economic Behavior and Organization).
And I thought to send my copy robotic within a few years to monitor my place neither seen nor known. Blood and Guts, I'm like a rat! Unless the other supervisor can not guess this is my copy ...
PS: the picture that illustrates this post has nothing to do with the connection, but the temptation make a fitting tribute to Blake Edwar, who left to join Peter Sellers in paradise geniuses last week, was too strong.
