What Makes a Graduate-Level Course Different From a Senior Undergraduate Course?

• October 17, 2025
• essays  
• education  
• learning  
• university  
• teaching  
• academia  
• research  
• grad school  
• undergrad  

DISCLAIMER: My field is computer science, and I mostly did theoretical courses. I really have no idea how other fields work, but I assume it’s at least somewhat different!

Introduction

About a year ago, while I was only a few weeks into the first graduate-level course I ever took, I tried asking a bunch of grad students what the difference between taking courses at the graduate level and the undergraduate level is. Being researchers in training (researchers are, in my experience, terrible at explaining themselves), they gave me delightful non-answers such as “it’s not that different” (this is a lie) and “your professors treat you like adults” (whatever that means).

Fast forward a year later, and I have taken three graduate-level courses (well, I took two of them as fourth year courses, but they were crosslisted as grad courses and ran like grad courses) in three different departments at my university: Mathematics, Computer Science, and English. I’m going to hold off on making broad claims about English courses, since I haven’t really taken enough literature courses at the undergrad level to have a strong opinion, but I am going to make some claims about what a grad course really is based on my experience.¹

The most important distinction between a grad course and an undergrad course lies in the teaching philosophy the instructors bring to the course. Undergraduate courses are meant to teach you information, or perhaps even concrete skills. Graduate courses are meant to teach you how to do research,² or at the very least, meant to teach you how to think. Now, you might think this is a slight, insignificant, and perhaps unhelpful distinction, since it’s not uncommon for undergrad courses to involve some sort of research project, and some grad courses assess knowledge using exams and assignments. However, this shift in perspective explains everything³ and really has been the throughline between all of the grad courses I’ve taken.

I’m writing this article because no one actually explains these things to undergrads,⁴ and I really wish someone had sat me down and told me some of the things that I’m about to write. I feel like a really big factor when it comes to being successful at anything new is knowing what the hell you’re getting yourself into. So much of what goes on in universities isn’t explained. (See the concept of hidden curricula.) People just expect you to know things or figure them out yourself, which I don’t always think is great. I’m hoping that someday, someone will read this article and find it useful. If I had understood what a grad course is before last year, I probably would have approached my first two courses very differently.

At my university, as I’m sure is true in many other places, sometimes a course will be crosslisted and simultaneously be offered at both the undergraduate and graduate levels. Professors deal with this in multiple ways: sometimes you get a hybrid grad/undergrad course, and sometimes you get an upper-level undergrad course that happens to have grad students in it, but usually what you actually get is a grad course that happens to have some undergrads in it. What I’ve found is that sometimes grad students are really pissed about having to share their courses with undergrads because they’re worried that we can’t handle the courses.⁵

Of course, this is an incredibly stupid assumption. As far as I’m concerned, the fundamental gap in ability between an academically strong fourth year and a first-year grad student is small (or even non-existent in some cases). There isn’t some magic switch that flips on and suddenly makes you able to do grad level work when you start a master’s. The main difference between a random fourth year and a beginning grad student is that a grad student is being trained to become a researcher and has some idea of what research is, so a professor expecting them to have this mindset isn’t a shock to them. The main difference between me and a random fourth year student is that I got the shock out of my system early.

That’s it. All you need to do to solve this problem is make sure students know what they’re getting themselves into before they voluntarily decide to take these courses.

“Research” and the point of university

I always find myself having to explain how university works, because none of what I’m about to say makes sense if you don’t have an appropriate mental model of what’s going on in academic institutions, and I always find it shocking whenever I explain this to someone and they’ve never heard this perspective before.

Ph.D. granting universities are research institutions. They employ professors who do research and those professors then train grad students to do research. (Professors also teach and do administrative work and sit on a bunch of committees, but I think a lot of them see themselves as researchers first.) It’s true that over time, universities have become businesses, which has changed a lot of the dynamics in terms of what kinds of activities are valued, what standards are upheld, which courses are taught (and in what way), and how programs are administered. That being said, looking at universities as research institutions first is still, in my opinion, the best mental model for understanding what the heck is going on in courses at all levels.

Researchers are professional experts – they know things for a living, and more specifically, they know things no one else knows for a living. The problem is that you can’t become an expert purely by being taught things by other people, because if you know something because someone else taught it to you, it’s no longer something no one else knows. The point of doing research is to learn something no one else knows. Sure, you then publish that information and add it to the sum of “human knowledge,” but broadly speaking, research is about trying to ask questions that no one has answered yet and then answering them. In the process of doing this, you end up learning pretty much everything there is to know about the topic you’re researching.

This is the point of grad school. You learn how to be a person who knows things no one else knows, and as a side effect, you become an expert (usually in something extremely niche and specialized that no one else really cares about).

To become a grad student, you need to have an undergraduate degree.⁶ In fact, in a lot of disciplines (especially the non-professional disciplines), the undergraduate degree is still more or less designed to prepare students for grad school. This means it’s about context and foundational ideas, and not so much about hands-on skills that will be applicable in industry. You’re meant to get a lot of breadth in areas of your field that academic researchers believe are important so that you’ll be able to “speak the language” of the field once you graduate. Now, do most students succeed at this? No. I still have no idea how a computer works, and I almost have a degree in computer science. But having a set of foundational mental models ready to go in your head is definitely the goal.

This is why computer science students, for instance, spend a lot of time learning theoretical nonsense instead of learning how to write software. Knowing what a Turing machine is will probably be irrelevant in 99.9% of industry jobs, but if you’re in an area of computer science that deals with algorithms, even if you mostly deal with implementations of very practical algorithms, you will likely need to have a basic grasp of computational complexity theory. However, the vast majority of academics do not care about React/Django/Astro/whatever the latest web development framework is. Frameworks change all the time. They’re not deep or pure enough to still be relevant in 30 years.

To sum it up succinctly, undergrads are evaluated on their ability to digest and own a canon of knowledge that someone else has taught them. Grad students, on the other hand, are evaluated on their ability to demonstrate a sliver of new knowledge they’ve acquired on their own. Traditionally, the demonstration method is writing a thesis and the evaluation method is giving a presentation about your thesis to a panel of experts who then ask you a series of extremely pointed questions about it.⁷ If your handling of their questions is catastrophic, you fail.

Senior undergrad courses

Senior undergrad courses are… weird. My program actually doesn’t have that many of them – most of the courses you can take at the 4000 level are really thinly disguised grad courses – but I have taken maybe two or three real senior undergrad courses, and I have a few opinions on what they should (and often do) look like.

The first thing is that 4th year courses are usually an introduction to some sort of specialized topic. The instructor will generally be someone who is passionate about teaching that topic and has a pretty deep knowledge of it, and it is assumed that the students in the course more or less understand the foundations of the discipline and are ready to start learning more interesting and specific things. If you’re lucky, you will be in a smaller course with peers who want to be there and with a professor who also wants to be there.⁸ This is in contrast to mandatory courses at the 3000 level or below, where it’s possible that no one in the room actually wants to be there and the professor might not be able to assign anything interesting because they need to test you on the basics and there are too many students in the course. (They also might not even care about the course content, which is always fun.) I’ve found that 4th year courses tend to have significantly more interesting assignments and assessments because they’re not prerequisites for anything, and therefore the prof doesn’t really need to weed anyone out.

A fourth-year course is still an undergrad course, which means that it’s at this point still your professor’s job to teach you things, and the actual content of the course is likely the focus both in terms of where lecture time is spent and how you are evaluated. At their best, these courses are an expert guiding you into understanding the specific area that the course is focusing on: ideally you’ll spend some time learning about the big picture and the context in which the main ideas sit as well as exploring some aspect of the topic in some depth. If it’s an interdisciplinary field, the professor should approach it with more of an interdisciplinary lens and give an overview of different approaches and perspectives as well as picking one lens to go over in more detail. Undergrad, in my opinion, is still fundamentally about breadth, and so in my opinion, senior level courses should be about giving you a reasonable taste of different areas.

Another difference with a 4th year course is that a good one should give you some amount of agency in your learning. Maybe there’s an open-ended project, or maybe the assignments leave a little bit more room for creativity, but if you can get through a 4th year course without having some sort of self-generated thought or idea, something is probably wrong with it. At this level, your professors should be expecting you to be able to think for yourself to a certain extent.

In my opinion, the main difference between a 4th year course and a grad course is that a 4th year course shouldn’t be about training you to do research in that specialized sub-discipline; it should be about giving you enough broad-ish exposure to it that you can decide whether or not you actually like it. It’s not about giving you enough of a crash course to the main ideas and terminology that you can start to navigate the academic literature on your own; it’s about giving you enough context that you understand where the discipline comes from and fits in the grand scheme of things.

The main difference between a 4th year course and a third-year course is that third year courses are still broad and survey-ish and primarily about covering a large amount of foundational information. Since you’ll probably be covering some new basic-ish topic every week or two, you don’t really have time to get into the weeds of anything, the expectation of academic maturity from the students is lower, and you might not actually be expected to demonstrate much original thinking. These courses tend to also be prerequisites for other courses.

Of course, your mileage may vary. I’ve taken 4th year courses that really should have been 3rd year courses (security course, I’m looking at you), especially because of how they were taught. To a certain extent, I think the 4th year level is defined by an ethos where you have a little bit more freedom, and what you get out of the course is what you put into it, but you can still expect your professor to know things and you still get handheld a little bit.

Alright, let’s talk about grad courses

“Your professors treat you like adults”

A prof once told me that he much prefers teaching grad courses to undergrad courses, because rather than spending a bunch of time doing review at the beginning of the course like he would with undergrads, he can just point students to a textbook (or sometimes even more than one textbook) and expect them to read it to catch up on the background.

I would say that this anecdote, to a large extent, is a good illustration of the difference between how professors treat undergrads and how professors treat grad students. Professors no longer feel the need to spell out every detail, assuming that the student will be able to fill them in. They will often immediately start the course with the course content and leave the students to figure out how to get their own bearings. They will assume that the students are intelligent and competent and can do things. They will expect students to know things, or if that fails, independently find out what they need to know and learn it. Or, in other words, “this is grad school, figure it out.” If being in undergrad is like being treated like an adolescent in terms of academic maturity, grad courses are where you start to be treated like a young adult.

The bright side of this added responsibility is that at this level, professors are significantly more likely to solicit feedback from their students and actually act upon the responses that they get. They may even change the course structure or content to better fit into the students’ interests. When I took my data science algorithms course, the professor frequently added or removed course content based on what he thought would be useful for our course project topics. He also frequently asked us what we wanted to learn and made an effort to incorporate that content into the course.

Another nice thing about grad courses is that there are often significantly fewer students in them, so while you get treated more like an adult, you also get treated more like a human. A prof once agreed to end class half an hour early when I was struggling to stay awake in lecture. That never happens in a larger undergrad course, where the prof might not even know that you exist.

Content-focused grad courses

I think grad-level courses sort of fall on a spectrum of content-focused courses to research-based courses, and the two main formats I’ve seen are the seminar format and the lecture-based format. In my experience, the seminar format typically implies that it is a research-focused course. The lecture format in no way implies that it is a content-focused course. The two grad courses I took last fall were both lecture-based, but I would consider one of them to have been more research-focused and the other one was definitely more content-focused.

I think on the math side, content-focused courses are more common than research-focused courses, purely because there is so much to learn in pure math that beginning grad students aren’t really able to do original research yet, unless they’re very very strong students. There is just too much math they don’t know yet. These courses feel a lot like a typical undergrad course: there are lectures, there are assignments, and there might even be an exam. However, the main difference is that you’re assumed to be able to unpack things yourself, and the professor may be trying to give you enough knowledge to begin to approach academic literature in that area. There probably won’t be a ton of handholding or exploration, and the entire course will probably have a somewhat narrow lens rather than being a broader introduction to the subfield.

I find that at the grad level, courses are significantly more information dense than at the undergrad level. They will often go very fast and cover insane amounts of content in a short amount of time – you might cover two or three times more information than you would in an undergrad course, because the prof is just presenting key ideas and highlights to you, and unpacking them and figuring out what is going on is your problem. For example, in a math course, you might get a series of statements with no proofs, or a high-level explanation of a procedure with no examples, or you might even just be told to go read about something in the textbook. In an undergrad course, the professor unpacks the knowledge for you. It took me a while to realize this difference and get used to it. If I had expected it, I probably would have studied very differently.

Anyway, the point of a content-based grad course is to master a whole bunch of foundational knowledge so that you can have enough language to begin attempting to decipher whatever the hell researchers are talking about. It is very much a “drinking from the firehose” type of experience – you are trying to build a specific tool that you need for your research very quickly.

What I learned from taking a content-based grad course (I took it as a 4000-level, but still) is that these are really not fun to do as an undergrad, even when they’re technically also listed as fourth year courses, because in undergrad you’re probably not actively specializing yet. What you will get if you take a course like this as an undergrad is exactly the same course as the grad students, just with significantly more generous assignment and exam grading. If your primary consideration is grades, then maybe the trade-off is worth it, I don’t know. But what I wanted when I took quantum computing was a bit more of a big picture, interdisciplinary approach to learning the content, and what I got instead was an extremely formalized and theoretical introduction to the mathematical foundations of quantum computing. It was extremely decoupled from the physics and computer science aspects of quantum computing, which I was really unhappy about.

This is fine to do in a grad-level course, I think. It’s the prof’s prerogative to be a little bit more utilitarian. The approach chosen was more to teach a course that could rapidly get you up to speed in terms of being able to stumble through quantum computing literature, because we covered a lot of formalism very quickly. But the sense of the big picture was completely lacking from this course. I have no idea what a quantum computer is or how one works, or what a qubit might look like in real life, or how one might actually implement quantum algorithms, and to me those were much bigger issues than my failure to understand what symplectic geometry is.⁹

Research-focused grad courses

The goal in a research-focused course isn’t to learn a bunch of content. Depending on what kind of course it is, this might be more or less obvious. For example, in a seminar course where you mostly do readings and present and/or discuss them in class, it’s probably pretty clear that the goal of the course isn’t just to memorize a whole bunch of concepts and information. However, in a lecture-based course (like the data science algorithms course I took), the format of the course might obfuscate the fact that your goal isn’t actually to master the content, since you’ll be spending lots of lecture time watching your professor throw content at you. The key thing to remember is that the goal of a research-focused course is to expose you to a whole bunch of tools and get you to start using them.

I don’t think I really understood this until partway through my data science algorithms course, mainly because I was so caught up in struggling to understand all of the details of what was going on to step back and realize that mastering all of the material in rigorous detail wasn’t the point. In fact, at one point I started to suspect that the point of the course had nothing to do with actually learning the details of anything we were doing. My professor’s lecturing was incredibly flexible and diverse throughout the term: sometimes we would spend an entire three hour detail on an involved proof with lots of technical details, but other times, he was perfectly happy to draw some pictures, talk about the general idea of something, gloss over details, and move on. Sometimes it felt like the goal was to see one idea in multiple different contexts, whereas other times, understanding the core idea was the goal.

In hindsight, I feel like the goal was to introduce us to a bunch of different mathematical techniques and hope that we would pick up how to think about them and where we might use them via osmosis. If we did a long and technical argument, it was likely a common type of argument, and it was presented so we would be familiar with seeing that type of argument in other places. We learned about a lot of high level ideas and about concepts that might appear in multiple different places. He presented a lot of ideas from different research papers, which made the content feel more interesting and less contrived than in an undergrad course. Our assignments were more meant to make sure that we could read stuff on our own, synthesize information, and attempt to creatively solve problems. And at the end, we had a project, where the goal was to apply the research skills he’d implicitly been trying to teach us.

More than anything, what I learned from that course was how to navigate academic literature, how to think about problems, a mishmash of common techniques in algorithms, and general fluency with certain times of arguments, but that’s not really a concrete learning outcome. Most of the content didn’t stick, but that wasn’t the point. If you learn content once, you can learn it again.

Of course, the professor never explained any of this. Researchers are terrible at explaining themselves.

A few things that surprised me about grad-level courses

The course assessment schemes are… weird

This absolutely drove me insane. In one course, I had insanely long assignments that felt impossible but weren’t actually worth very much in the grand scheme of things, and they were graded very loosely, which was very surprising coming from undergrad, where typically, grading is somewhat more strict. I think my conclusion about this course was that the prof mainly wanted to make sure we were thinking about the course content enough, but didn’t care about rigor so much as long as we understood the big picture?

Actually, a lot was weird in that course. For example, we got bonus marks for contributing to Wikipedia based on what we learned from our projects, which is not something I commonly see in courses.

I think at this level, it’s a lot more common for expectations or grade breakdowns to be somewhat more ambiguously defined. For example, a lot of seminar courses have a significant portion assigned to “class participation”. What does that mean? I don’t know, the prof may or may not tell. But this is a grad course, so why on earth are you here if you won’t participate?

I swear, profs in grad courses definitely give off “you’re an adult now, go figure it out” vibes when talking to students quite a bit of the time.

Your professors no longer have all of the answers

I think the first time the difference between being in a grad level course and an undergrad level course really hit me was when I went into my prof’s office to ask for help with my final project. I think I asked him what the relationship between the two different definitions of a submodular function were, and he said he had no idea and would probably have to look it up. Then he moved on, and we never talked about it again, at which point I understood that learning the required background for my course project was mostly going to be my problem and that I was going to have to find some other way to learn it.

At the undergrad level, I think profs tend to steer students towards topics they know a lot about and have already seen before, and therefore are able to provide a lot more guidance to students who are struggling and need help. In fact, some courses require all students to do exactly the same project or write their report on the same topic, whereas in a grad course you’re pushed to choose something that is interesting and relevant to you, so long as it’s at least tangentially related to the course content. What I did not realize was that at this level, you’re starting to get specific enough that “tangentially related” might already be far outside of your professor’s area of expertise. At that point, you’re on your own.

I think the coolest experience that came out of this, however, was when I was absolutely struggling to understand a section of the paper I was reading for my project, and brought it to my professor during office hours. I was able to watch how he approached reading the text and working out the meaning, and just that was so much more valuable than the help he gave me with the actual content of the paper. We spend so much time being taught by professors, but significantly less learning about how they actually learn or think about things, and I think seeing how a professional mathematician approached a math paper really changed how I approached my reading from then on.

You’re supposed to be thinking of yourself as a researcher now

What I found particularly weird about taking a grad seminar in the English department was that it felt like we were all supposed to take on an identity of being researchers, or, more accurately I guess, “scholars” (which seems to be the term preferred in the arts and humanities). Some of the ways in which this expectation manifested itself were obvious and explicit. I remember that during the first lecture, our professor introduced herself to us and talked to us briefly about her background and her research. After she finished talking about herself, she asked us to go around the room and do the same: that is, say our names, a bit about our backgrounds, and tell the class our “academic interests”. Of course, I was thrown pretty off-guard by this question: having defined academic interests seems like the sort of consideration that mainly belongs to professors and researchers, and as an undergraduate student in a different discipline, I had absolutely no idea how to respond to this.¹⁰

I’m sure this is different in other fields. The average third year in computer science has zero experience with doing research and has likely never read an academic paper, but students in the humanities write a lot of research reports (“papers”) and read a lot of academic literature for class, so maybe it is absolutely normal to expect fourth year students to have a fully realized idea of what their academic interests are. I have no idea. But my point is that the requirements of my undergraduate degree in computer science did not prepare me in any way for the expectation to start thinking of myself and behaving like a researcher. And God, the amount of reading… I love reading, but CS undergrad does not train you to read voraciously or to gain any sort of limited expertise about anything. In this course, the professor would say things about “writing from a place of authority” and about how that could only be done if you’d consumed so much information about your topic that you’d seen everything and actively had to filter out what to include or not include in your published work. This is the only course where I’ve had to submit a works consulted section in addition to a works cited section. She really wanted to make sure we were being thorough.

This idea that we were supposed to be thinking about ourselves as researchers also manifested itself in just how much emphasis was put on the final research paper. It felt like the entire course was built around it, in some sense, and our professor made a big deal out of making us discuss our “research” with each other in small groups as well as with her and with the entire class. This brings me to my next observation, which is about these sorts of final projects…

What really matters is your final project, and it’s supposed to be good

I am a terrible student, and I was a particularly bad one during my data science algorithms course. I did not master any of the formally assessed content of the course – in fact, I am sketchy on what some of this even was at this point – but this turned out to be mostly fine, because the course had a final project, and in a lot of courses with a final project, the project is the only real thing that matters, especially if it is worth 50% or more of the final grade.

So remember when I said that grad courses are intended to teach you how to think like a researcher? These projects are where you demonstrate that you can apply your newly acquired skills. The assessments (assignments, tests) on the formal content of the course were graded super loosely, but as far as I can tell, the project was graded significantly more rigorously because it was meant to be something under your control that you learned and presented to the class. I spent so many hours working on my project for the data science algorithms course I took. I learned (read: taught myself from many, many obscure sources) considerable amounts of new math just so I could write an 8 page report and give a coherent presentation about one or two sections of an algorithms paper, and it was worth it, because that is where I learned the most. Everything else in the course was just context.

The corollary of this is that doing the bare minimum on projects in grad school is extremely frowned upon. (If you don’t want to put in effort and learn things, why are you in grad school?) In some courses, the professors even explicitly tell students to aim for publishable quality (though whether or not this is actually reasonable is highly debatable in my opinion, not that professors care about reasonable expectations).

Honestly, I like this philosophy quite a bit, because it puts control over the outcome of the course mostly into the hands of the students. At least with a project, you get some tangible representation of your effort and learning out of it. If you’re lucky, you might even discover something or make something useful.

What the heck do you get out of writing an exam?

Conclusion

I’m not sure how to end this. I didn’t realize that I had over 6000 words to say about this topic, but apparently, I do. I think part of why I felt the need to write this is because so much of what grad students do and why is hidden from undergrads, and I’m not convinced that these things are ever explained to most grad students either. Academia really seems to have this culture of ambiguity, which is mildly understandable in some scenarios, but terribly frustrating in others. If no one is going to explain things, I will try my best. With that said, please keep in mind that these are my thoughts, observations, and best guesses. If you’re a professor, feel free to let me know how far I strayed from the mark.

One thing I didn’t mention earlier, but will mention here, is that while taking a grad-level course for the first time can be a culture shock experience (it certainly was for me), once you adjust, it is really difficult to go back to doing undergrad courses that don’t have a similar amount of depth or give students a similar amount of freedom and flexibility (which is most of them). Anyone who knows me well will tell you that I am not a fan of undergrad. I find a lot of undergrad courses to be limiting and overly prescriptive, and in my case, a lot of them weren’t very academically challenging, either. It wasn’t until I was in grad courses that I came up against course content that deeply challenged me – that is, that challenged me enough that I wondered if I was cut out for it. And when I came out of those courses, I really felt like I’d levelled up. I came out of those courses with more skills than I’d had before.

So, should undergrads take grad courses before they graduate? A year ago I would have said no, but now I’d say absolutely, if you think you can handle it. You will likely learn more and differently in one of those than in the vast majority of undergrad courses, and I think it’s worth taking them when the high expectations don’t actually matter to you: A C in a grad course is a fail if you’re a grad student, but not if you’re in undergrad. Not to mention, it’s so much easier to be academically adventurous in these courses than in an undergrad course.

That being said, a grad course is not a substitute for a real fourth year course: one that manages to inspire students to explore a new area, exposes them to different perspectives, and still goes in enough depth to be interesting. My program is sorely lacking in these, and I really would have liked to take more of these instead of thinly disguised grad courses.