Jason Woodard

Exploring the Olin Curriculum

1) In a few sentences, what is the main idea of the project?

The main idea of the project is to develop an empirical understanding of the Olin curriculum using historical course registration data. Although the goals are primarily descriptive at this stage, my top-secret ulterior motive (not to be disclosed to the students under any circumstances :-) is to help lay the groundwork for data-driven discussions about how to *improve* the curriculum — both in terms of solving specific problems (e.g., how to make it easier for students to complete their major requirements, study abroad, etc.) and addressing more general curriculum design goals (e.g., how to facilitate innovation, give students more choices, make staffing more flexible, etc.).

2) Which data sets will the student team use for this project?  Are there any obstacles to obtaining these data sets?

The ideal data set, I think, would contain a list of every course taken by every Olin student from, say, 2002 to 2015. In other words, we need individual student-level data (as opposed to aggregate enrollments per course), with unique identifiers for students. (Course numbers should be sufficient to identify courses, although some complications arise.) Students can certainly be anonymized, although we will probably want some additional data on each student (e.g., year of enrollment / graduation, declared major, etc.). Some additional information on courses might be desirable as well (e.g., prerequisites, whether the course is required for graduation or a given major), although it might also be interesting to try to infer some of these things empirically.

I have obtained similar data from the Registrar’s Office on a limited set of courses, so I know it’s possible — and I understand that Paul has worked on this data as well, so the data set we need may well exist in usable form already. In any case, I think we’re aiming for low-hanging fruit here — let’s get whatever data is easy to obtain and build on whatever analysis has already been done.

3) What are the key research questions you would like the student team to answer?

I think the team should define the specific research questions they are interested in answering. Questions I think are interesting, but would need to be fleshed out further, include: How can we characterize the variation in the academic paths people take through Olin, and to what extent does it map to the three majors? What really constitutes the “core” in the sense of courses that (almost) everyone takes? Which parts of the curriculum are “easy” vs. “hard” to change, as measured by the strength of their downstream dependencies?

It might be worth stating explicitly that I am NOT interested in student-level outcomes (e.g., grades, time to graduation, etc.) in this project. I can imagine a really interesting follow-on project related to student learning and assessment, but not this year.

4) What special skills or knowledge should the students ideally have to be successful with this project?

I don’t think the project requires any special skills, but since we’re talking about an exploratory project I think it’s important for students to be comfortable approaching it with an exploratory mindset. In particular, I think they will need to be comfortable pursuing leads, pruning dead ends, refining their own research questions, keeping track of what they’ve done and what they’ve found, and figuring out what it all means. In other words, if you expect me to tell you exactly what to do, please do not choose this project :-)

I do NOT expect this project to involve heavy-duty statistical testing. Depending on student interest, there might be some scope for machine learning techniques (e.g., try to classify students into majors without knowing their major, or discover “quasi-majors” empirically using clustering algorithms) and/or network analysis (e.g., construct and analyze a directed graph representing paths through the curriculum).

5) What form would you like the final deliverable(s) to take?

I’m imagining a fairly informal writeup, probably containing a combination of descriptive statistics (tables and/or graphs) and visualizations. I’m not sure how much data visualization is covered in the course, but I think there are some very cool opportunities for “mapping” the curriculum in a visual way; I suspect we could learn a lot just by looking at a few brilliantly constructed pictures!

I would also be delighted if the team were to see this project as part of an ongoing effort, and consider the set of deliverables to include tools that can be reused and extended by future teams. It might also be worth curating some of the intermediate products (e.g., processed data) that doesn’t end up in the final project — again in the spirit of building “giant shoulders” for others to stand on.