Wisconsin Teaching Fellowship Project Question
I’m participating in the Wisconsin Teaching Fellows and Scholars program. This program aims to foster real scientific research on teaching and learning, an area that is typically snubbed by tenure committees at schools bigger than my own. Each fellow or scholar proposes a research question, deliberates with a team of others, spends a year studying the question, and then shares the results.
For my project, I am using Madeup. I’m currently preparing for a week of deliberation, during which I’m to present a brief summary of my project idea. I share my notes here.
At this point, what is your primary question? Is there an intervention?
Compared to traditional programming exercises, does programmatically generating 3-D models improve student performance and retention in an introductory computer programming course?
We teach a lot of abstract concepts in computer science. However, I don’t think “abstract” necessarily means impractical or theoretical, at least not in the courses I teach. Rather, the sense of abstract that really gets students is that we teach things that are many layers above obvious. We don’t have a model of the physical world to relate to, students don’t have a lot of prior exposure to software development, and each generation of computer scientists has left a legacy of decisions that haven’t ever really gone away. This drops a high cognitive load on a new programmer, which leads to “drops” of a sadder kind.
Therefore, I propose incorporating a simpler language in our introductory programming course that reduces cognitive load in two respects: 1) it has less historical baggage and 2) it involves navigating three-dimensional space. This language is one I am writing. It’s called Madeup.
What have others found? How can you use that work?
There’s a growing body of evidence that:
- spatial reasoning skills predict success in mathematics
- spatial reasoning skills can be improved
See Wai, Lubinski, and Benbow, Cheng and Mix, and Newcombe.
I haven’t found many papers examining spatial reasoning specifically with computer science. That makes this project important, I guess!
However, the idea is not entirely novel. Seymour Papert heralded his twist on constructivism, which he called constructionism—the idea that learning is most effective when the learner produces a tangible artifact that can be consumed by others. He developed a little programming language called Logo, in which the programmer navigated a turtle through 2-D space. Papert argued that the programmer identified with the turtle and effectively envisioned her own body in 2-D space. Such “body-syntonic learning” facilitates learning, he argued. While Logo lives on in various forms, the research project fizzled out before I was born.
We largely don’t use Logo in “real” computer science courses. This project may affirm that non-practice or suggest that we reincorporate it.
The Spatial Intelligence and Learning Center may be a good resource for me to contact.
At this point, how can you envision designing your project? What will you do and how will you measure student learning?
In the fall semester, I teach our 90-student introductory programming course. These students are split up across 3 30-student lab sections. In these closed labs, students complete two short coding exercises.
My plan for assessing the impact of Madeup as a tool for enhancing spatial reasoning is to select one of these lab sections and replace the first coding exercise with a Madeup coding exercise.
I will constrast the test scores, homework achievement, and retention against the control labs to determine the impact of the treatment.
What are the main questions you want help with at this point? How can the group help you think through your idea?
Experimental design is not something I’ve done a lot and certainly not done well. I expect I’ll need some sort of pre-test, and I don’t know what that should look like.