The BPC Relevance of Common Assessment in the Introductory Sequence

The importance of common assessment is high as the rapid growth in enrollments in computing increasingly necessitate that schools offer multiple sections of in-demand courses. This is particularly true for introductory classes that serve computing majors/minors and students from other disciplines. For courses with multiple sections, most computing departments in North America have a set of shared learning outcomes (usually developed by the undergraduate CS curriculum committee). Computing departments differ however, in their implementation of shared outcomes, which can be as simple as all course instructors receiving a list of the concepts a student should have learned upon successful completion. Here, we argue for the other end of the spectrum, in which departments ensure shared outcomes via common assessment through identical exams, assignments, and grading standards.

Our arguments also apply to very large classes where a single instructor teaches all students in a single section, but the students attend labs or recitations taught by other instructors and/or Teaching Assistants (TAs). While some faculty may view common assessment as overkill and potentially in conflict with the idea of academic freedom, we argue that the pros far outweigh the cons. In the conclusion, we argue that common assessment is one of a set of institutional changes CS departments can make toward broadening participation in computing (BPC) and that the changes are most effective when done together.

The Benefits of Common Assessment Far Outweigh the Perceived Drawbacks

At the Center for Inclusive Computing (CIC) at Northeastern University,^a we view common assessment as a core tenet of broadening participation in computing (BPC). Consider the following case that the CIC encountered at one of our partner universities. In an analysis of course-to-course retention and performance data, the school uncovered that students’ grades in CS2 correlated more with who the student had as an instructor in CS1 than with their own effort or mastery of the material (as measured by their final grade in CS1). The school determined that some students were consistently achieving lower levels of foundational programming knowledge, which was quickly creating problems for them later in the curriculum. This variance in preparation from CS1 was compounded for students who were less likely to have taken coding in high school. Because prior experience in computing is not uniformly distributed across demographics, poor preparation in certain sections of CS1 had a larger impact on students from identities historically marginalized in tech,^b leading to a much higher drop/fail/withdraw rate in CS2 in comparison to students from majority identities. Common assessment in CS1 ensures all students, no matter their prior background or instructor, can enter CS2 with a shared foundation.

Common assessment offers other benefits to students and faculty. For faculty, a significant benefit is a sharing of the workload of creating new assignments and exams each semester. Another advantage is that resources related to TAs and graders can be centralized, meaning there is greater coverage for help outside of class, grading happens in a more timely fashion, and faculty can delegate course management to a single faculty course coordinator.^c In addition, some universities hire full-time staff members to help manage large multi-section classes, thereby freeing up time for faculty to conduct research and meet with students. Common assessment also facilitates the onboarding of part-time/adjunct faculty, reducing the burden on these instructors of devising their own assessments and reducing the risk that they may, unintentionally, be out of sync with other faculty. Finally, there is the benefit to faculty who teach upstream from the multi-section course as they now know what was covered and no longer receive students who have had access to different levels of preparation.

For students, common assessment means they can be more confident in their level of preparation and not worry that they are less ready than their peers due to differences in instructor. Similarly, with common assessment, students can see any TA for the course, meaning they have more flexibility in their schedule and are more likely to seek help. Because any student can now see any TA, and not just the TAs assigned to their section, they can also seek out a TA with a preferred intersectional identity or teaching style.

Objections Are Easily Managed

When the CIC discusses introducing common assessment at partner universities, some faculty resist the idea because they enjoy overseeing all aspects of their course. Indeed, certain faculty will make no secret of their decision to cover material that is different from what other professors teaching the same course are covering. The CIC is quick to emphasize that no one is telling faculty how to teach the class material; rather we are stressing the critical importance of students achieving the same outcomes. One proven strategy for getting faculty on board is to let them take turns being the course coordinator so that they have the chance to influence the curriculum while also experiencing firsthand the many economies of scale that the approach affords. Another idea is to reserve a portion of the course for material that is unrelated to the learning outcomes—for example, each instructor could have the last two weeks of the term to teach whatever they want.

We conclude by addressing one final objection, that of exams. Some universities will interpret common assessment as necessitating having all students take exams at the same time in order to mitigate risks of cheating. This interpretation raises concerns about having a space large enough to fit all students from all sections. Furthermore, universities might, rightfully, balk at the idea of a single exam time given the burden this puts on students who work or have other life responsibilities. We argue that there are different ways to discourage cheating and that common assessment does not have to mean a single time slot for exams. Many of the schools the CIC works with successfully create a bank of exam questions that assess the required outcomes but vary sufficiently as to mitigate the risks of cheating. Indeed, the CIC has not found any significant limitations to this solution.

Concerns raised of common assessment are valid and legitimate. The CIC’s view, however, is that with good leadership and a clear plan, any concerns are easily overcome. Above all, the objections are outweighed by the myriad and long-lasting benefits of common assessment—to faculty and students alike.

Common Assessment Is One of Several Best Practices for Retention in CS

The CIC sees common assessment as a necessary component for the retention of students in computing from historically marginalized populations. The CIC emphasizes, however, that this approach is even more beneficial when combined with other best practices that have been proven to increase retention. As such, we close this column by briefly reviewing five other efforts that, in our minds, constitute the “ideal” set of systemic, retention-focused best practices.

The first complementary practice is the active management of the distribution of prior computing experience in CS1.² Prior experience in CS is not uniformly distributed across all genders, races and ethnicities, and, further, in 2024 CS is only offered in 57% of U.S. high schools (more often in regions of economic privilege). Thus, the individuals experiencing the first course required for a computing major are more likely to be from less privileged geographies and from genders and races/ethnicities historically marginalized in tech. In the June 2022 issue of CRA’s Computing Research News² we outlined several strategies for handling the distribution of prior experience, while also taking the context of the university into account.

The second complementary practice is to centralize the recruitment and training of TAs—both undergraduate and graduate. At many universities the CIC observes little to no training for TAs (and, if there is training, it is often the job of each individual faculty member to train those assigned to them). Because student experience and sense of belonging are often tied to their experience with the TAs, it is vital to both train and evaluate TAs (see AIICE¹ and Muzny and Shah⁶ for publicly available TA training materials).

Third, we argue that departments need to carefully consider whether the number of required courses and prerequisite structure in their CS degree(s) are reasonable.⁴ Many computing departments require in excess of 20, semester-long computing courses in addition to math and science requirements. In such cases, it is impossible for a student to discover and graduate in computing within four years if the student does not know from day one that they want to be a computing major. A good rule of thumb is that all requirements for a degree in computing can be safely satisfied in three years, which allows a student to discover computing in their second year of college and makes transferring from another institution feasible without adding significant time to graduation.

The fourth best practice is to collect and analyze the drop/fail/withdraw rates of each section of each course intersectionally^d (looking at every combination of race and gender separately) every semester to identify any issues with specific courses/instructors.⁵

The final practice—likely the most controversial—is the removal of GPA-based enrollment caps. If a lack of sufficient resources means that a department feels it has no option but to cap enrollments in the number of CS majors, the CIC suggests the department consider alternate strategies that can meet the department’s needs without sacrificing its BPC goals.³

While we believe these practices are most impactful when implemented all together, each is powerful in its own right—as demonstrated by the arguments for common assessment herein. Indeed, common assessment can be an excellent starting point for departments given the significant benefits to both students and faculty. Department leaders can monitor the results of this first step and then implement additional strategies as time permits.