After close to a decade as an educator, and a lifetime of being in classrooms, I can unequivocally say I don’t just hate tests - I find them completely pointless in my teaching style. By testing, I specifically mean those events where we compel students to recall and organize information from class or some readings in a time-restricted venue in exchange for a score which substantially impacts their cumulative grade (e.g. quizzes, midterms, finals etc.). These kinds of tests don't provide a true evaluation of how much a student has actually learned. They're great at proving that someone has either a close-to-perfect grasp of the exact things asked on a test, or they can show that a student is completely lost in the subject at hand. But these tests don’t serve the goals of educating a student, which go beyond bringing students to the level of being able to identify or explain concepts and phenomena.
In my understanding, there are roughly three reasons: (1) to provide students with direct feedback on how well they have grasped the course material, (2) to evaluate areas the professor should revisit during class time, and (3) to provide a signal to grantmakers, graduate programs, and some employers regarding which students have high scholastic ability (and effectively lock certain “low-performing” students out of some paths).
I see fundamental issues with all three approaches though. If the goal is to provide feedback to students, one need only think back to their time in college to remember that it’s rare for a student to truly look at their exam grade as anything more than an imposition on their sanity. Highly motivated students may attend office hours to clarify topics they struggled with on an exam, but it’s more common (at least anecdotally) for students to just be satisfied with a high grade, or take a low grade as motivation to study harder - or more realistically to write a course or subject off entirely. Even at top universities it’s rare for every student to care deeply about every class they take, but it may not be the worst thing in the world to mainly cater to highly motivated students.
If the goal is for professors to evaluate areas for review, it may be ok to give occasional tests, but you’d wonder why they couldn’t accomplish the same thing with discussions in class or on virtual boards, or through the evaluation of written work. Discussions and written work could also be considered a kind of testing, and also with the recent explosion of generative artificial intelligence-based cheating, this path might also be a bad way to truly evaluate a course’s students. However, cheating is still an issue with exams, and when presented properly, discussions and written assignments might not create the same pressure to cheat. There’s plenty of online discussion about how college students have been caught using AI on inconsequential assignments like digital discussion boards, but I also have seen that when students are engaged they are more likely to want to share their opinion and knowledge. Again, we do have to deal with the fact that not all students will be engaged no matter how hard a professor tries, but here too, I’ll assert that perhaps we should mainly be thinking about how to facilitate learning for motivated students.
The last reason for testing is, to my eyes, the most egregious. The idea that testing helps signal a student’s scholastic merit fails to account for a basic truth: learning isn’t static or one-size-fits-all. A student who takes less time to grasp enough of a subject to pass exams is advantaged in the current testing tradition, while students who might need more time but are still equally capable of mastering the material persistently find their GPA deflated, especially at big schools with “weed-out” courses like my alma mater, Cal.
I saw this reality firsthand as a college student - as a biochemistry student I struggled like hell to pass biophysics. However, now I have a perfectly decent understanding of how protein folding works and how those processes can be mathematically modeled. Maybe not enough of an understanding to work professionally in that area, but definitely a higher than average understanding when compared to the population of people with college degrees. Of course, I don’t use this understanding in my daily life as a social scientist, but it strikes me that it wasn’t true to say I was actually struggling in that class - it just took more time for the concepts to click. If I took the same tests today, I might not remember the exact nuances of the formulas and chemical structures germane to the class, but I would be able to place ideas in a more organized mental structure and discuss them with greater clarity.
It’s true that I wasn’t ready to pursue graduate studies or a professional career in modeling protein folds after my junior year of college. Maybe if I had kept at it, I would have been ready in another two or three years, but I was definitely discouraged to continue with biophysics after the experience of floundering through midterms and finals that made me feel dumb.
“But Anyun!” you might say, “if a student really loves a subject they can keep going even if they do poorly on exams.” That’s absolutely true, especially when it comes to research. Almost every academic I’ve spoken to has a story about how they sucked at a subject in school, but now they’re super solid at it. The common wisdom about graduate admissions follows this narrative, with people repeating ad nauseam that, while GPA matters, having research experience and publication history is way more important because it demonstrates ability to deconstruct cutting edge questions and contribute to human knowledge. But there’s too much noise in this supposed relationship between effort and outcome to claim it’s not mediated by test performance. Low test scores exclude students from opportunities to try again, especially when jobs, internships, and research opportunities at the entry level often do look at GPA. Additionally, there’s the simple reality that low test scores are discouraging - without enough mental resilience it’s very easy for students to believe themselves incapable of grasping a subject merely because they got a low grade on a test.
You could just try not having tests. It sometimes is just that simple. I love providing students with narrative evaluations based on written submissions like research papers and their in-class participation. While providing letter grades is pretty necessary in the US college system, I often structure narrative evaluations as a discussion of what I see as a student’s strengths, what areas of the class I saw them struggle with, what improvements I noticed in them, and future directions I think they ought to consider - like other courses they might enjoy, internships they could pursue, or reflections on what went wrong in the course if they got a failing grade.
A basic rebuttal here would be to assert that (a) requiring research papers or written submissions is still a form of testing, and also that (b) my approach just doesn’t work with larger classrooms (50+ people) and especially not for hard science and math courses. To the first point, I’ll clarify that I characterize testing as requiring constraints like having right and wrong answers. Open-ended research papers and written submissions which evaluate students based on their ability to apply a course’s subject matter and framework to a phenomena of interest doesn’t necessarily have these constraints, even though a time-restriction might still be necessary. But I’ll concede that written submissions could still be seen as a kind of testing, just not the kind I find immediately problematic in my own pedagogy.
As for the second rebuttal point - I agree! My approach doesn’t work in bigger classrooms. And I don’t think we should have classrooms of such a size that the professor couldn’t possibly have a 1-1 relationship with each student. We have huge classrooms, especially at public universities, because of a flawed social narrative that everyone should go to college, while simultaneously defunding public education. In reality, my pedagogy only works flawlessly when we have public universities with exacting admission standards and smaller class sizes - a complete pipe dream, I know. My teaching style might be much more well suited to smaller liberal arts colleges (SLACs), but I think it’s possibly more important to bring it to large R1 research oriented universities. My alma mater and the institution where I’m currently earning my PhD both serve college students with differing levels of college preparedness, and these institutions both serve as massive drivers of social mobility. My students deserve individual attention, and arguably need it more than students at SLACs.
As for the argument that my style of assessment doesn’t work for STEM subjects, I think STEM educators are a bit shortsighted about what makes a good scientist, engineer, doctor, etc. It’s not the ability to regurgitate large amounts of information. In fact, I regularly see my friends doing basic science research or providing healthcare using search engines and digital databases to look up information like medical symptoms or chemical structures. What makes professionals in these fields good at what they do isn’t memorization, it’s having a system of organized understanding in their minds. When I was in college, the exams I took did test on memorization, like spitting out amino acid structures, but the meat of the exams often relies on being able to break problems down into their basic components and analyze the situation step-by-step. Building and evaluating that kind of thinking doesn’t actually require tests, using research papers, take-home problem sets, longform projects, in-class presentations, or other ways to allow students to demonstrate what they’ve learned seems equally valid to me.
“Without testing,” you might say, “how are we supposed to know which students are ‘good,’ and which are ‘bad?’” You don’t - you have to take chances and find out. We have another problematic social narrative that a good GPA from a good university automatically indicates a good potential employee. Not everyone believes this, but plenty of so-called prestigious employers like consulting firms, investment banking firms, and tech companies recruit from top schools and use GPA as one of their basic criteria for evaluating candidates. While it makes sense to focus on hiring students from programs known for having strong faculty and facilities, I personally see no reason to privilege one bachelor’s degree over another. All this focus on GPA and prestige really accomplishes is a lack of true competition. We’re all kidding ourselves if we think college admissions at the top level isn’t a complete crap-shoot nowadays, and accrediting bodies (in theory) validate the notion that all bachelor’s degrees are generally indicators of the same level of scholastic ability. My take on this matter is likely to be a little less palatable, especially to people who went to prestigious schools and worked hard to earn high GPAs. I’m not trying to deny these peoples’ talents - but after having taught at a couple of perfectly good schools that don’t share the veneer of prestige that the Ivies and other top-25 programs had, I can unequivocally say that the students I’ve worked with are just as hardworking and motivated. They are often not as lucky or well-prepared coming into college, but they all turn into fine young accomplished people at the end of the process.
The last thought I’ll offer on this matter has to do with zero-sum situations like testing for nursing or medical credentials. This matter is not one that I can directly speak to, but from anecdotal discussion from friends in healthcare, I’ve become aware of the enormous pressure students are under to cheat, both at unaccredited programs, as well as well-regarded Ivy-league ones. Whether it be for the NCLEX or the medical board Steps exams, the pressure students in these programs are under just isn’t healthy or conducive to creating the best possible medical workforce. To that end, medical admissions introduced CASPR to test emotional intelligence on top of content knowledge in the MCAT, and the Step 1 board exam is now pass/fail. Even medical educators accept on some level that the testing system they have isn’t ideal - although they probably wouldn’t agree with me that we should do away with testing altogether. And I suppose for a healthcare credential, you want to know that the professionals have been able to answer whatever questions were thrown at them. But my advertising students don’t need to be held to that standard, and neither do the vast majority of college students who don’t become healthcare professionals. For most college students, what really matters is that they became better thinkers and writers, and became more well-rounded people for it. And I don’t think I’ll ever become convinced that subjecting my students to a midterm is necessary to accomplish that goal.
---