What Math Teacher Do My Students Want?

At TMC16 one of the impressive experiences was listening to the keynote speech by Tracy Zager (you can find it here ) and I remember briefly mentioning it in a post soon after the conference. Sadly, I never went back and really dissected my positive feelings about it, but I do know that one of the habits I developed after her speech was to communicate pretty regularly with our elementary school teachers more regularly when I ran across interesting math tidbits. I also know that Tracy has written a well regarded book called Becoming the Math Teacher You Wish You’d Had. I have not read this book (yet!) but the title of the book has been haunting me. I am not speaking with any deep knowledge about the strategies that she is advocating and I may be doing her a disservice, for which I apologize in advance. But as I am sitting here at the beginning of my spring break, I am thinking about what has gone right and what has gone wrong so far, 2/3 of the way through my 2018 – 2019 academic year. I also think about the phrase from Dan Meyer’s TED talk, ‘Be Less Helpful’ and this morning I want to try and wrestle with each of these phrases that are resonating in my head.

I will admit up front that my interpretations of these two phrases are my interpretations. I also need to fill in a little background to explain how I understand Prof. Zager’s phrase. I had two high school teachers who were enormous influences on me. I had the same English teacher, Mrs. Myra Schwerdt, for my junior and senior year. The way her typical class went was this – we came into class, she tossed out a question prompted by our reading assignment from the night before and then she moderated the discussion. Tossing out another question if the conversation ran dry or probing someone on an opinion/interpretation that they made. She was sort of first among equals in these conversations. I never felt that she forced us to interpret anything in any way but I also felt that we better have some backing for what we had to say. In my senior year I took AP Calculus BC from Mr. Barry Felps. A typical day in that class started off with a quick look at a new idea with Mr. Felps offering an example or two of ‘how to’ after explaining some new idea. He would sometimes go over a HW question or two from the recent past (often after his go to joke, we’d ask ‘Mr. Felps, can you do problem #12?’ He’d look at the book and nod while saying ‘Yep’. Still cracks me up a bit thinking about this…) After this, typically 15 – 20 minutes total, he’d say ‘Okay, you have work to do and so do I’ He would sit and we would work with our neighbors and friends. If we got stuck, we’d go ask him but most of the time we felt that we wanted to figure it out ourselves. We also had a study group that would periodically meet on Sundays for a combination of football, pizza, calculus, and physics. So, here I have in my mind what ‘Be Less Helpful’ looks like (to my 17 year old self for sure) and an image of the math teacher I wish I’d had (which, I feel enormously fortunate to say is a math teacher I did have)

Fast forward from 1981 – 1982 academic year to the 2018 – 2019 academic year. This year I am teaching four different classes with pretty different groups of students. I teach Geometry with the text I wrote four years ago to mostly 9th and 10th graders, I teach a Discrete Math elective to mostly seniors, I teach a non-AP Calculus class to a mix of juniors and seniors, and I teach AP Calculus BC to a mix of juniors and seniors (with one brilliant sophomore in the mix) These classes all have different needs and different inherent investments from the students involved.

It is, of course, unfair to generalize too broadly, but I think it is fair to say that the general needs/wants of these classes differ. I try to account for that, but I know that there are some tendencies in my teaching that appear in all four subjects. I ask more than I tell. I redirect questions to other students to get their input. I have student in groups of three facing each other. I randomize these groups so that a group of three is together at most for five classes. I write problem sets that dip into past knowledge and ask some questions for which we have not been explicitly prepared. My quizzes are narrowly focused on recent information and they are weighted less than tests. My tests are all cumulative in nature but they are about 70% focused on what has happened since the last test. I want at least one test question to feel novel, to ask students to put together information in a way that feels new – a problem rather than an exercise, I guess I’d say. We have a test correction policy that we adopted this year (you can read about that here, here,  and here  ) and this policy seems to have helped reduce anxiety a good bit. All of this is to provide a little context into my classroom and my vision of what these phrases ‘Be Less Helpful’ and ‘Becoming the Math teacher You Wish You’d Had’ mean.

What I have struggled with, this year more than in the recent past, is the discrepancy between the math teacher I wish I’d had and the math teacher that my current students wish they had. A good number of my students seem to be happy wth how our classes run. I have posted a number of entries this year about student success. I have received some lovely emails from parents of students in Geometry, some nice remarks passed along from colleagues about students in Calculus Honors, some terrific conversations with students in Discrete Math, and an abundance of energy and creativity from my AP Calculus BC gang. In all four of my classes I feel I am reaching some students and making a positive impact. What concerns me, and what prompts this post, is the fact that a number of students are not buying what I am selling. They are frustrated when I respond to a question with a question. They think it is unfair to have a test question that does not look like something they have explicitly practiced. They feel that I am off loading my responsibility as the teacher when I ask them to work in groups to figure something out instead of lecturing and telling them how to figure it out. I have students across a wide range of abilities and a wide range of reactions to what I am trying to accomplish. It is important that I recognize this and do not simply bask in the glow of the students for whom this approach really clicks. Where I struggle, is trying to reconcile what I think I understand about teaching research, what I understand about NCTM’s recommendations, and what I valued as a student with the discomfort and unhappiness that I see in some of my students. I also struggle with balancing what is clearly working with some students in each subject. I don’t want to lose that energy and motivation that I see in students who value what feels like a different way to experience math. I know that this is not an either/or situation. I have two weeks of spring break to think and reflect. I know that there is a way to reach the frustrated kids without giving up the the facets of my class that are valuable to some of their colleagues (and to me!), I know that there is a balance to be struck between asking kids to step out and feel challenged and making sure that they still feel supported.

I think back to a comment from a student abut 6 years ago. About three weeks into the school year she asked for a personal conference to talk about her struggle in our Honors Calculus class. By the way, struggle for her meant that she had a B average instead of her typical A. She said to me ‘I thought that I needed to learn formulas and how to use them. That isn’t working here.’ She and I had a lovely conversation and in the years since I have run into her dad a number of times. He always tells me how important my class was for his daughter. She has sent me lovely thank you notes updating me on her progress. She is an example of a student I was able to help cross a certain threshold. I want that feeling for all of my students in some way or another and I have solid evidence that this is not happening often enough right now. I have some serious thinking to do.

A Guest Post

The last time I wrote, I was talking about how spoiled I am by my students. I have two students who did an enormous amount of thinking about a random walk problem I proposed. Just for your reference, the problem was

Starting at the origin, a bug jumps one unit either left, right, up, or down. He jumps once each second. List the possible locations (and probabilities associated with those locations) that the bug could be after 6 seconds.

After working on it for awhile with one of my students I waived this one off and told my students to feel free to ignore it. A number did not. Two in particular, Bobby and Matthew (both of whom seemed happy to have me name them!) collaborated over a chat line of some sort and delivered a lovely presentation on their work. I asked them to guest post for me and below is their post. I cannot emphasize enough how impressed I am by their ingenuity and determination. There are two other files they asked me to make available.  A 3d representation and a 4d representation generated by their code.

 

Bobby: Last week, a blog post referred to a random walk problem that our Calculus class worked on, and two students that took a coding approach. We are those two students, and we’re here to discuss how we solved the problem, and the far more interesting work that came after it. For those of you who don’t know, the problem was a 2-dimensional random walk of 6 steps.

Matthew: When I first looked at this problem in class, I thought it would give way far more easily than it actually did. My first attempt was to rewrite the problem as choosing from the four cardinal directions a set of six steps, e.g.  {↑ ,→, ↓, ,←,}. By doing so, my hope was to reduce every position on the grid into a set of the bare minimum moves needed to reach it, and pairs of blank spaces which I could fill with a pairs of steps which added to the zero vector. Since the grid had 8-fold symmetry, I was unafraid of solving each point individually, however when the time came to actually do the problem, I found a number of oversights in my initial work which resulted in some pesky double counting.

Bobby: For lazy Calculus students (read: myself), the obvious solution for any random walk problem is to make a computer do it, so i set out to approximate probabilities by running a billion trials of a random walk and counting up the results.

However, this brutish method is relatively unsatisfying, and by the time I finished, Matthew had abandoned his set permutation approach and decided to make a program that iterated through each possible set of 6 steps once and counted up the destinations. This clearly being the better approach, I did the same thing, and we exchanged our results (which matched) and exchanged our code to make improvements: I took some of Matthew’s ideas to make the program easier to scale up and down, and I’d like to think he took some of my ideas and stopped naming variables “DONK,” but I’m sure he didn’t.

This grid shows the number of ways to land at any given point. After a short time looking over our grid of results, we noticed that each edge of the square was the 6th row of pascal’s triangle. A little more poking around and Matthew noticed that the grid was in fact a multiplication table of that row of the triangle with itself, meaning every entry could be written in the form (nCa)*(nCb), with n being the number of steps and a and b being the relative coordinates of the entry in the table. Finally, we made the observation that this is the 6th layer of Pascal’s square pyramid, a 3d structure much like the triangle where each number is the sum of the four above it. Recalling that a 1d random walk is the nth row of Pascal’s triangle where n = # of steps, a pattern seemed to emerge that we hoped might scale up in dimensions.

Bobby:  We did a lot more work, particularly in 3 and 4 dimensions. Our program logic started by defining a variable to track which number walk we were on, and keyed in on the base-four version of that number to define a set of moves for our bug. For example, test 5 translates to 000011 in base 4, and the bug reads each digit individually, right to left, using it as an instruction for motion: 0 = right, 1 = up, 2 = left, 3 = down, translating 000011 to U, U, R, R, R, R.

This logic makes it easy to change the number of dimensions or steps, so we went to 3 dimensions, expecting to get 3-dimensional cross-sections of some 4-dimensional solid for each particular solution. The results did not disappoint, as we got the expected octahedron as a result. Obviously, this output was a little tough to format, but I think what we settled on is easy enough to read. Each grid separated by the others from a row of asterisks is a 2d cross-section of the solution octahedron. LINK DATA Stack those layers on top of each other and the visualization is pretty easy, I think.

At this point, we had to do four dimensions, and had a good idea of how it would look. The 2d solution is the shape of the 1d solution stacked on itself, the 3d solution is a stack of the shapes of the 2d solution, so we the 4d solution would obviously be a series of 3d octahedrons in the same x, y, z, but translated along our fourth spatial axis. It took me a moment to wrap my head around this data, but I’ve become very comfortable with it. LINK DATA Each section under the rows reading “NEW 3D SOLID…” is a series of grids that are stacked into an octahedron, and then the octahedrons are stacked on each other in the fourth spatial dimension. This, of course, is also a 4d cross-section of a 5d Pascalian solid, each nth 4d layer of which is a solution to the 4d random walk of n steps, but that’s about as far as I can go envisioning these shapes. Matthew tells me the 5d structure is “Pascal’s Orthoplexal Hyperpyramid” but I’m not sure I can trust that.

Matthew: Of course, the idea of layers of Pascal’s square pyramid as multiplication tables of the rows of Pascal’s triangle is a fascinating one, and one I sought to prove. Though my initial effort to prove it algebraically did meet with success, I was unsatisfied with my clunky proof and did some quick googling on the subject. My search turned up a blog with a simple proof by induction, far more elegant than anything I had done. Still, I find that proofs by induction are rarely enough to understand a result intuitively, and so I continued working with Bobby, looking for a better way to interpret the formula in terms of the problem, as well as potentially a way of generalizing the result to higher dimensional walks.

Eventually, I took a step back from the symbol soup and searched for a more intuitive interpretation of the formula we had discovered in terms of the original problem. After some while, I had an epiphany while placing the problem back into the language of sets. My initial idea of viewing the problem as selecting a set of 6 steps along cardinal directions could be reformulated in a way which made the formula obvious. Instead of considering steps along the cardinal directions like the original problem had stated, I realized that every unit vector along a cardinal direction could be rewritten uniquely as the sum of two diagonal vectors of magnitude √2/2, e.g. {} = {↖,↗}. With this realization in hand, the original problem can be rephrased as a random walk in six steps of magnitude √2/2 along the line y=x, followed by another random walk of six steps perpendicular to the first. Or, to give an example in my initial interpretation of the problem as sets of steps, the two dimensional walk {↑ ,→, ↓, ,←,} is the same walk as {{↖,↗},{↘,↗},{↘,↙},{↖,↙},{↖,↙},{↘,↙}}, which is the same as the union of the two one dimensional walks {↖,↘,↘,↖,↖,↘} and {↗,↗,↙,↙,↙,↙}. Once the problem is reformulated in this way, the reason why the table obeys such a simple multiplicative relationship in two dimensions becomes satisfyingly clear. Though I am pleased that we were able to transform such an originally dense and unapproachable problem into a setting where it is obvious, we have unfortunately as of yet been unable to find an application of this method to higher dimensions.

I’m Spoiled

This post is inspired by my AP Calculus BC class. At the beginning of each of our seven day rotations, I give them a problem set. These problem sets are pretty wide ranging, some Calculus questions, but mostly wide open fun problems to explore. Here is a problem from the set that was due today (actually due yesterday, but we were snowed out):

Starting at the origin, a bug jumps one unit either left, right, up, or down. He jumps once each second. List the possible locations (and probabilities associated with those locations) that the bug could be after 6 seconds.

I have given them variations of this where the intrepid bug could only move left or right. I thought that this would be an intriguing extension. During my library duty nine nights ago one of my students spent about 15 minutes with me bouncing ideas around. We came up with the GeoGebra sketch below:  

This seemed pretty daunting and I admitted to him that I may have overreached with this problem. We shared our conversation the next day with the whole class, including showing this intimidating graph. Trying to figure out how many of the 4096 pathways could lead to each point felt overwhelming. A couple of students started tossing out ideas but I kind of encouraged them to let this problem slide and blame me for poor planning.

When the problem sets came in today three of my students discussed some coding attacks that they took. Two of them were collaborating late into the night working on an approach to the problem and the output from their attack is below:

They took over the class conversation today pointing out that they recognized a pattern based on a multiplication table where the column header AND the row header were each the 6th row of Pascal’s Triangle. The numbers in their printout above are the products of this table. A long explanation invoking a three dimensional Pascal’s Triangle (I thought of layers like oranges in a conical pile) was presented with great enthusiasm. Some of the kids in class kind of glazed out, but a number were fully engaged. I was engaged, awed, and slightly confused. I have asked the two who collaborated on this if they would be willing to be guest bloggers here and they seem amenable to that idea. I hope to have a follow up, in their words, in the next day or two.

I am SO spoiled to be able to sit back and learn from students who could have easily left this problem alone, but were unsatisfied with that notion.