Grinnell’s new registration process (#1208)

Topics/tags: Registration

It’s the second half of fall semester. That means it’s time for another registration period [1]. But this year will not be the same as past years. Grinnell has a brand new registration process [2]. Don’t worry! It’s still uniquely Grinnell and still confusing. And like anything new in academia, it’s causing consternation and worry.

Here’s how it works. Or at least here’s how I think it works.

Registration is a multi-round process. There are three primary rounds, plus some intervening administrative stuff and some fairly routine post-registration cleanup.

In round one, caps for all courses are set at 99 students, and students register for up to nine credits. Nine credits permit two four-credit courses plus a one-credit thing or one four-credit course and one five-credit course [3] or one four-credit course and two two-credit courses and one more credit or …. This first round lasts for four days, Wednesday through Saturday.

On Sunday, all the courses get cut back to their true caps. The Registrar then gets to fix all the over-enrolled classes. That is, they cut students from those classes.

That may sound similar to the old process, but it differs in a few ways. First, the register for any course period is for nine credits rather than eighteen. That smaller number limits the number of conflicts we will have. Second, instead of individual faculty deciding what happens to over-filled courses, the Registrar follows an algorithm for each course. We’ll get back to the algorithms in a moment. Third, the Registrar will not balance classes. That is, if there are two sections of a course, one is over-enrolled, and the other is not, the Registrar will only cut from the over-enrolled section, they will not move those students to the other section. The Registrar also won’t mark classes as closed. I suppose there’s one other significant difference: This cutting will happen in one day rather than over a week. I do not envy the Registrar that day.

In any case, that’s what I know about round one and the post-round-one cut process. After the cut process, some students will still have all the classes they tried to enroll for in round one, and others will not. The Registrar tells us that she’s confident that every student who enters two four-credit courses in round one will still have at least one four-credit course at the end of the cuts. That’s better than our old process; in that system some students would be cut from their top two courses and some even got cut from three courses.

There are also some enrollment patterns that would prevent us from achieving the goal of students get at least one four-credit course. For example, if every first-year student selected ART-134, Drawing, and PHI-101, Logic, in round one, there’s no way we can meet that demand [4]. However, our registration history suggests that we’ll see a good distribution of student choices, and the Registrar will likely meet their goals.

On to round two! Round two is intended to help clean up issues from round one and is reserved for students who got cut from a course in round one or who neglected to register for a course in round one [5]. In round two, those students are permitted to bring their schedule up to nine credits in a first-come, first-served process. Well, it’s a bit more complicated than that. First seniors get to add courses in the first-come, first-served process. They get a day. Next third-years get to add courses. They also get a day. Then second-years. Then first-years. At the end of round two, everyone should have eight or nine credits.

On to round three!. In round three, students get to bring their schedules up to eighteen credits or so [6]. Once again, we have a seniority-based first-come, first-served process: one day for seniors, one for third-years, one for second-years, and one for first-years. In some sense, round 3 is like the old post-prereg cleanup, except that there’s a priority policy at play. That is, students must scramble to figure out what courses to add to fill out their schedules.

Wait! There’s one more new thing. In round three, students can also add themselves to waitlists. And unlike in the past, the Registrar keeps the waitlists.

Once round three is complete, about a week before Turkey Break, we’ve completed the main part of registration. After that, it’s the normal post-registration process. Students may still adjust their schedules until week two of the subsequent semester. Well, it’s a bit different because of the waitlists.

How do the waitlists work? In one of two ways as determined by individual departments. A department can ask the Registrar to use its priority policies from round one to select students off of the waitlist or a department can have the Registrar select students in a first-come, first-served order [7]. Students have a day to claim a slot if it is offered to them.

What do those priority lists look like? They provide a sequence of instructions that the Registrar follows until they get down to the course cap. There are two basic kinds of instructions: cut and keep. As you might expect, a cut instruction says to cut people in a category. If more people in the category are enrolled than need to be cut, then the Registrar randomly cuts students from that category. In contrast, a keep instruction says to try to keep students in that category. If more fewer seats than students in that category are available, then the Registrar randomly cuts students from the category.

For example, suppose our policy for a course is Keep undeclared students, keep majors, cut third-years, cut seniors, the course cap is twenty-four (24), and eleven (11) undeclared students, six (6) majors, eight (8) third-years in other majors, and four (4) seniors in other majors enrolled in round one. We would keep all the undeclared students, leaving us thirteen (13) remaining seats. We’d then keep the six (6) majors, leaving us seven (7) remaining seats. There isn’t room for the twelve (12) students left. Since we are supposed to cut third-years before we cut seniors, we cut five (5) third-years randomly and then enroll the remaining three (3) third-years and all four (4) seniors. More precisely, the three (3) remaining third-years and the six (4) seniors were all enrolled already; we don’t have to cut them.

Well, it’s a bit more complicated than that. We can also reserve a certain number of seats for particular class years. For example, in the prior system, we might have left slots for twelve first-year students. If only eight (8) of the eleven (11) undeclared students were first-years, we would still need four (4) slots for first-years available for rounds two (2) and three (3). That means we’d cut all the third-years and one (1) of the seniors.

Let me check: We start with twenty-four (24) slots. We keep the eleven (11) undeclared students. We keep four (4) open slots for first-year students. That leaves nine (9) slots. Six (6) of them are covered by the majors. That leaves three (3) slots. We cut all of the third-years. We have four (4) seniors still in the class, but only three (3) slots. So we cut one (1) randomly. Yup. My informal analysis holds.

As you might guess, this prioritization approach has some advantages and some disadvantages. One significant advantage is that the priority information is public: Any student or advisor can check out the policies for each class [8]. However, it also takes removes things from the control of faculty, which means that faculty can’t consider special circumstances not encoded in the system [9]. In contrast, it means that students are treated equally; we don’t have a situation in which some students know how to use the system while others do not [10].

Of course, we all know that equal treatment is not necessarily equitable treatment. I mourn that loss, particularly since Grinnell is small enough that we should be able to address individual cases individually.

That prior paragraph suggests that I’ve started to move from the process to some additional commentary. I’ll admit that I feel better about the system now than I did at first. It retains many of the strengths of Grinnell’s registration chaos while addressing some important issues. I’m hopeful that over the long term, we can encode some of those individual situations better or find ways to address them.

What do I see as likely positives?

• We gain much by clarifying our cut policies. I’ve been surprised reading some of the policies, particularly since I might have prioritized differently.
• It seems much less likely that students will be cut from their high-priority courses, not least because they can better indicate those high-priority courses.
• We won’t encounter situations in which a student is cut from two or three courses and has to scramble to find other courses. In part, that’s because we’re asking them to plan ahead. In part, that’s because they should have a better sense as to what’s available.

What do I see as postential issues?

• Students will still going be scrambling to add courses round 3, particularly students toward the bottom of round 3 (i.e., second-years and first-years). It will not necessarily be easy for those students to find courses that fit into their schedule by the time it’s their turn. Some people are confident that this will work out, but we’ll have to see. It may be that students will need to be more willing to experiment with new subjects.
• Some colleagues have suggested that the system may have hidden biases against students with disabilities, particularly with regard to the first-come, first-served aspect.
• There are new ways to game the system. For example, since classes don’t close, a student might be able to negotiate a strategy with another student in which one student takes a slot in a class and then frees it for their friend right at the time their friend adds it.
• Since we need to work on fallback plans during the initial registration advising meeting, advising takes more time.
• I’ll admit that I’m not sure what happens with paired courses, like BIO-251 and CHM-221. I’m also not sure how many paired courses we have.
• We’re losing some of the opportunity to change the schedule during registration. For example, in the old system, if we saw few students in course P and over-enrollment in course Q, we might consider dropping course P for another section of course Q. Or we might try negotiating with the Dean for another section of course 1. I suppose we can use the waitlists for that kind of negotiation, but it seems complicated.
• I’m not sure that all categories are adequately covered. For example, what do we do with a student who has previously failed a course and wants to retake it? Does it matter whether they need the course for their major?
• I’m not sure what the loss of balancing will do to the system. It seems like some students who might have been balanced to another section in past semesters won’t be able to enroll at all in this model (e.g., because they are a second-year and all the slots are taking by third-years and seniors). It may be that we have to handle that with other aspects of the priority process.
• As I’ve said, it’s not necessarily equitable. I have some students who need three specific classes to make appropriate forward progress. I have others who only need one. Our system treats them the same.

Some aspects are neither positive nor negative. Here’s one. Since students can change courses during round one, I expect round one will be an interesting exercise in game theory. For example, if a student sees that a course is over-enrolled and they are in a low-priority group, it behooves them to switch to another course, even if it is less important. After all, they will likely be dropped anyway. This switching seems particularly likely to happen in multi-section courses; since we are no longer going to balance students, students will have to balance themselves. Unfortunately, students will need to make these decisions with incomplete knowledge. That is, they will know how many students are enrolled in each class, but they won’t know where the other students stand on the priority list.

I look forward to CS students building a tool that helps their peers keep track of enrollment and to Econ students writing a paper or strategy guide on the game theory problem.

We’ll certainly see more benefits and issues as the system works itself out. And, in spite of the list of potential issues, I consider the new system better than our old one. I’m having trouble conceiving of a much better one.

Still, this first semester of the new registration system is going to be hard. And there will be flaws. The second will likely be better, if only because it will only involve three class years. After that, I hope we’ll be able to work out most of the remaining problems in the process, even if it means changing some parts.

Fingers crossed.

Postscript: Here are a few sample priority policies.

AMS-130-01, Intro to American Studies

• 8 slots are reserved for 1st-year students.
• Keep previously cut students.
• Keep undeclared students.
• Keep 1st-year students.
• Keep 2nd-year students.
• Cut students randomly.

CHM-129-01, General Chemistry

• 8 slots are reserved for 1st-year students.
• Cut students randomly.

CSC-151-01, Functional Problem Solving

• 16 slots are reserved for 1st-year students.
• Keep 1st-year students.
• Keep previously-cut students. (A side note indicates this only applies to first-year students.) [11]
• Keep undeclared students.
• Keep 2nd-year students.
• Cut students randomly.

CSC-301-01, Algorithm Analysis

• 10 slots are reserved for seniors.
• Cut students declared in other Major.
• Drop students who do not need the course for degree/graduation.
• Drop students enrolled in more than one major course. (There’s an exception for students who have fewer remaining semesters than remaining courses.)
• Keep graduating seniors.
• Keep seniors.
• Keep 3rd-year students.
• Keep 2nd-year students.

ECN-280-01, Microeconomic Analysis

• Keep Economics majors.
• Keep undeclared students.
• Keep previously-cut students.
• Keep students declared in other majors (Other declared majors priority to SR, JR, SO, 1st year in order) [12]
• Cut randomly.

[1] Also known as Open Registration, Preregistration, and some other things that I’ve forgotten.

[2] Those who’ve forgotten or don’t know about our old registration process can refer back to earlier musings in this series.

[3] We don’t have many five-credit courses but we have a few. They are primarily introductory language courses.

[4] No, that is not a suggestion to our first-year students. However, they are both great classes.

[6] I say or so because I think you can add one credit of music lessons, PE, or first-year=experience courses to bring the total up to 19. Some students may also choose to enroll in fewer than 18 credits.

[7] I don’t know what other departments are doing. CS has insisted on using our round-one priorities.

[8] Well, it’s semi-public. It’s available to anyone with an account on the GrinCo SharePoint(less) server.

[9] For example, some students need to have a C class or a Java class for certain graduate programs. We’ve tried to take that into account in selecting students for courses in our introductory sequence. For the time being, we can no longer do so.

[10] That’s a serious concern. I’ve talked to some students who had no idea that they could try to persuade a faculty member to prioritize them or add them to a waitlist. I’ve seen others for whom it seems like second nature.

[11] Won’t those students be enrolled according to the prior keep? Perhaps we should have revered those.

[12] Why isn’t Keep other major listed as Keep seniors; keep 3rd-years, keep 2nd-years, keep first-years? Presumably because there weren’t enough columns in the spreadsheet.

Version 1.0 released 2022-11-01.

Version 1.0.01 of 2022-11-01.