Principal Pac-Man 1.0 (Chasing Behavior)

Disclaimer: The following is an idea I have been thinking about this week. I have absolutely no idea whether it would work. I don't know if anyone anywhere is approaching principal work in this way. Some of this has been daydreaming or thinking during a long drive to a conference. These ideas may seem disconnected, but I will try my best to explain the relationships I see between these ideas.

We had some professional development days to start this week. I enjoyed two presentations Monday by David Webb from the University of Colorado Boulder and the Freudenthal Institute. His morning and afternoon sessions focused on formative assessment in mathematics. When he saw many morning participants planned to stay for the afternoon session, he quickly talked about something he and his colleagues use to teach early computer programming concepts to middle school students.

Dr. Webb posed this question to the audience: how do we design intelligent ghosts that will actually chase Pac-Man? The mathematical process, known as collaborative diffusion, describes a possible method for programming ghosts to effectively chase Pac-Man. Here's a link to an academic paper on collaborative diffusion by Alexander Repenning. A screenshot from the paper appears below.


Think of the spaces around Pac-Man as the yellow mountain above. The ghosts want to climb the mountain - and effectively destroy Pac-Man - by climbing to the top of the mountain as quickly as possible. I was thinking about this idea of how the ghosts are chasing down Pac-Man. Then, I thought about how we often in school try to chase down behavior. For example, when a teacher is in the hallway greeting students, sometimes amorous couples try to hide from the teacher's line of sight. If the teacher has to help a student in the class, and cannot man the hallway post, then the threat of punishment is gone. Speeding tickets then came to mind. I thought about how punishment rarely works well as a behavioral deterrent. Drivers may choose not to speed when a police officer is nearby, but once the police officer leaves, look out.

To tie this stream of consciousness back to teaching, think about how often teachers must identify, on the fly and while making mental decisions regarding content delivery, students misbehaving in the classroom. Proximity works well as a deterrent - walking near the student, pointing at the open book on the student's desk as the teacher walks by - but this technique also has its limitations. As soon as the teacher walks away, the student may misbehave again.

Then I thought about how tough it can be to be a principal. Here's a great post on how to navigate the frequent interruptions a principal faces. The principal position can sometimes be very similar to the function of a police officer - a deterrent. But, as a principal leaves, so does the threat of getting into trouble, and the idea is the same as the teacher that walks away from the student's desk. How do we address this behavioral piece while teaching? How do we keep students on task?


The possibility of being called on randomly. 

While thinking about police officers, I thought back to another article I read in a grad class about The Santa Cruz Experiment. The article, which appeared in Popular Science, described predictive police work. <think Minority Report> A mathematician designs an algorithm based on data for allocating patrols. Though random phenomena may be wildly unpredictable in the short term, long terms trends and patterns emerge.












Tying this back to the principal idea... if the ghosts chase Pac-Man... doesn't the principal chase the behavior? Suppose we try to incorporate a random mechanism into the principal's behavior in an effort to make chasing this behavior - just like the patrols in Santa Cruz - more efficient. Let's set up an imaginary simulation. We will declare the following events as things the principal could do.

0 = observe 1st floor hallways
1 = observe 2nd floor hallways
2 =  observe 3rd floor hallways
3 = observe 1st floor classrooms
4 = observe 2nd floor classrooms
5 = observe 3rd floor classrooms
6 = observe school entrance / parking lot exit
7 = monitor stairwells
8 = monitor cafeteria
9 = monitor library

Then, we could use some sort of random process to generate a random behavior for the principal.


Looks like today's focus will be first floor classrooms. Because all outcomes are equally likely, we now have a mechanism like the Popsicle sticks in the classroom. This will be a more efficient approach to deterring negative behaviors among students as well as teachers. This would also give the impression to students that the principal could be anywhere. Thinking back to collaborative diffusion, and Pac-Man emitting a scent that can be chased down by ghosts... the metaphor places data in the role of the scent. We have plenty of sources of data on student misbehavior. Also consider certain events more likely given certain days of the week and months of the year. Isn't a student more likely to get a discipline referral close to a vacation, after a long block of no days off from school, because teachers' behavioral tolerance is lower? Isn't a staff member more likely to violate dress code on a Friday? Aren't students more likely to be off-task close to passing periods? We could use data (and a different random digits assignment scheme) to make an attempt at 'predictive principalship' much like the predictive policing in the Santa Cruz Experiment article.

It would be interesting to see whether this is a viable strategy for administrators to use.

P. S. If you've made it this far in the article, please be sure to read the disclaimer at the top of the article a second time.

P.S.S. I know the title doesn't quite jive with what was discussed here... since the metaphorical principal is the ghost and the metaphorical Pac-Man is the behavior... but the title is way more catchy this way.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.