When I'm teaching an advanced graduate class where the source material is drawn primarily from current research papers, there isn't a canonical text per se, and a typical lecture, while loosely structured, also involves discussion that might help clarify the papers.

In such a setting, asking students to scribe lectures serves a useful purpose. Students hopefully remember the material better after writing it down, and are forced to be more precise. And classroom discussion can be captured for posterity.

But I don't think I've found a way to implement this effectively. My current mechanics include:

  • students are assigned as scribes between a week in advance, and on the day of the lecture
  • I provide a latex style file, and also an example or two of prior scribe notes.
  • students are expected to produce a first draft within a few days of the lecture (lectures are once a week), and the hope is that a good version of the notes is on the website a week after the lecture.
  • these are "seminar" classes: they are 1-credit, and the only work the student needs to do to earn a grade is one or more scribes and (sometimes) one lecture.

What happens is that the initial draft is usually abysmal, and it either takes me a really long time to wrestle the document into shape with the student, or I give up and do it myself, which also takes a lot of time.

Are there useful practices that can improve this workflow ? Is my timeline unrealistic ?

p.s To avoid confusion, when I say scribe (which is common parlance in my area) I don't mean a literal recording of minutes but a synthesis of material that if well done becomes something resembling lecture notes.

  • 7
    I guess it might depend on what is "abysmal" about the drafts? Formatting, general writing quality, typos, incomplete, poor understanding of topics, etc? Each of these might require different steps to address (and in principle some of them may be tied back to the lectures themselves). Apr 24, 2014 at 15:30
  • 3
    Formatting is a problem (but minor). The biggest problem is writing quality and basic structuring of the material as a "lecture" as opposed to 'minutes of a meeting'.
    – Suresh
    Apr 24, 2014 at 15:33
  • 4
    You must mean transcribe which, in my dictionary, means, "to write out from speech." The verb scribe means to cut (as in carpentry) to score (as in metalwork) or to engrave (as in, to inscribe upon). Or perhaps you just mean taken written notes?
    – mako
    Apr 24, 2014 at 17:51
  • 2
    Ah. I've used 'scribe' colloquially for a long time. But I stand corrected.
    – Suresh
    Apr 24, 2014 at 17:52
  • 3
    (I do think this usage has caused some minor confusion here: I am confident that Suresh does not mean literal word-for-word transcription, as some answerers have understandably thought. He means the practice of designated students typing (probably texing) nice lecture notes, a practice which is now relatively common.) Apr 25, 2014 at 3:54

10 Answers 10


You have the following goals:

(I) Every student will transcribe at least one lecture.

(II) The transcribed lecture notes will be of a sufficient quality to be useful to the other students, and ideally to be posted somewhere for others to access.

(III) You don't want to spend an unreasonable time writing or correcting the notes yourself.

In my opinion these are three worthy goals and any one of them is attainable, but as a set they are probably incompatible with each other. So you may want to choose (not necessarily once and for all, but depending on the course) which of these worthy goals is the worthiest and prepare to sacrifice one or both of the others. In more detail:

If you prioritize (I), then the emphasis is placed on creating a learning experience for the students. If you do this you will find that the variation in the quality of the note-taking will be amazing. I have seen this in students' written notes over the years. Yesterday a student left his notebook in my office. I flipped through it today to try to figure out who it was. It is one step above chicken scratch: for instance, he is writing on lined paper but his mathematical expressions and equations only stay within the lines about half the time: he can have a "written slope" which is only slightly larger than -1. (This helped me identify the student, who does not have his name written down anywhere. I noticed his characteristic negative slope when he was writing on the board in my office yesterday.) On the other hand, in my graduate course I have glanced at several students' handwritten notes at the end of the lecture and marvelled at how much better they look than my own handwritten pre-lecture notes and what is on the blackboard at the end of the leecture. At one point years ago I gave in a borderline undergrad/graduate level class what I thought was one of my best lectures of all time and at the end realized that I hadn't prepared any lecture notes and sadly reflected on the fact that it would be lost forever. Then I remembered that I had one student in the course who took great lecture notes, and I asked her for her lecture notes for this purpose. I think that what she gave me was an improvement on my delivered lecture, and using her notes I got my favorite part of what amounts to a textbook on number theory: you can see the lecture notes here. (Yes, I touched it up a bit afterwards, but not by that much.) You can also see that I thanked this student at the end of my notes.

All of the students mentioned above are "A students". I personally do feel that the gender difference between the students is playing a role here. This is not a scientific observation and I certainly mean no offense to anyone, but in my experience the difference in note-taking care between male and female students is, on average, dramatic.

Anyway, if you prioritize (I) and (III) you're going to get such uneven samples as to make it a pretty jarring experience for anyone to try to follow the entire course by reading the notes. Having the notes be texed adds other levels and brings other skills into play. Some people will naturally go back and spend time prettying up their tex files; others won't. (I started out as the latter and very slowly over the course of thousands of pages of notes am heading towards the former. But for someone who has thousands of pages of texed mathematics on his webpage, my texing skills are distressingly middling.)

Thus if you prioritize (II), then I think you need to seek out the students who will do the best job. This works against (I), and you need to look for extra compensation. I agree with the other answers that providing good latexed notes is something that may be worth paying for.

If you really prioritize (II) then at a certain point you will decide to type the lecture notes yourself. This is what I do in most classes I teach. At this point I have a bit of a reputation as a guy who has lots of typed lecture notes. When I begin a graduate course now, I often tell students explicitly that I will not be typing lecture notes as the course progresses, because that is very time-consuming. But then, at some point in the semester I usually break down and reveal to the students where my heretofore secret lecture notes can be found. I actually just did that yesterday in my graduate class this semester....with less than one week of class left.

Typing your own notes totally destroys (III): to get something that looks good you have to work significantly harder during the semester than you otherwise would need to do in order to deliver exactly the same lecture and you need to put more time into it after the semester is over. On the other hand, I have found this practice to be extremely valuable to both my learning and my career: I have gotten a lot of wonderful professional interactions out of it. Thus I would say that the real drawback is that it defeats (I).

To summarize a very long answer: I think you need to decide whether the point of this is the process or the end result. You either provide a learning experience for all the students or you provide decent-to-high quality lecture notes, or you provide both but don't get it done until after the end of the course. I don't see how to do everything at once.

  • To add to this from a student/scribe perspective: I think (I) is the condition that should be left out. It might be a good way for the lecturer to evaluate the students' performance, but it conflicts with each of (II) and (III) and creates bad incentives. (I tend to sit in certain classes merely to get an idea what they're about, possibly catch connections to subjects I am working on, and just to have an hour and a half of work time without an easy way to distract myself with reddit. Requiring everyone to scribe would scare me out of them.) Apr 27, 2014 at 23:30
  • I also think a week is too little for a set of scribe notes to "mature", and MIT OpenCourseWare notes would heavily gain in quality if they would be maintained beyond the end of the respective classes. Apr 27, 2014 at 23:31
  • @darij: Presumably some of the classes that you take or once took had actual, formal requirements. Presumably you avoided some other classes because the actual, formal requirements did not appeal to you (I took no laboratory science courses after high school). These are the standard rules of the game. I think most students would agree that scribing lectures is not the most odious of actual, formal requirements: if you do that instead of a midterm exam or a couple of problem sets, is it so bad? Apr 28, 2014 at 1:22
  • I don't recall attending a class with formal requirements since I finished high school. Some privilege and luck was involved in this (in Munich, where I was an undergrad, classes probably do have formal requirements now that they switched to bachelor/master, and at MIT undergrad classes have requirements). But that's not the main issue here. Having to scribe is a greater burden to us "course vagabonds" because failure is exposed (bad scribe notes appear in google searches for one's name) and because one is inconveniencing the lecturer (who has to "fix" the abysmal scribe notes) ... Apr 28, 2014 at 3:22
  • ... and fellow students (who might rely on scribe notes for their own learning). In contrast, my psets and exams remain between me and the lecturer or TA. I am not arguing against work requirement in general; I, in fact, wish for more (graded and returned) problem sets. I am just saying that making scribe notes a passing requirement turns away students who don't already feel safe that they will pass; and I don't think this is a good thing. (To end with an anecdote, I did pass a class by virtue of making scribe notes, despite having done barely any homework. I realize I am arguing ... Apr 28, 2014 at 3:24

My experience with transcription: Our university uses a lecture capture system (Camtasia Relay) that creates a video file with the instructor's screen plus the audio of the instructor's lecturing. We recently hired students to take the software's "best guess" captioning for a single course and create good closed captions. To our amazement, every 50 minutes of lecture took roughly EIGHT HOURS to transcribe.

So my main answer is yes, your timeline is likely too tight.

You may want to look into some sort of lecture capture system.

My socratic teacher self also thinks that students creating their own summary of the discussion is likely more helpful to deep learning than receiving a transcript. Or have students create a wiki or google doc for each seminar with the main figures (codes? equations?) provided for them to discuss. A different student could be in charge of each day's wiki setup and upload the appropriate discussion points and edit the final product and report to you on class participation.

  • 5
    As commented above, Suresh is not looking for a word-for-word transcription (and "transcribe" was not really the right verb; it has since been changed). Rather, he means a set of polished lecture notes containing the details from the lecture in technical prose, something like these. So they are not expected to record every word, but rather to digest the information, fill in details, etc. It will still be time consuming! Apr 25, 2014 at 13:52
  • Thanks for clarifying. I'm in my own little cocoon and didn't realize that my word choice could cause confusion :)
    – Suresh
    Apr 25, 2014 at 14:57
  • Despite the semantics related to the question this answer is a good guide - you might cut the time by 50% for a digested set of notes - if they're good at the material and slow at typing. If the material is a complete mystery but they can type, treating it as dictation might be quicker. It's not an easy task.
    – Chris H
    Apr 25, 2014 at 16:28

I originally wrote this answer back in 2014. I'm updating it for 2018, since I now have a bunch more experience.

I have had students scribe for me in graduate classes many times. You can read their work on the course webpages: (1, 2, 3, 4, 5, 6). I have also twice taught honors classes where I assigned students to prove a lot of results in class and assigned them to write up notes on the day's work. You can read the more recent version here (1, 2); I also did this a previous time but no longer have a digital copy of those notes. Some advice:

  • Keep the assignment manageable. For a graduate student, I think writing up one 50 minute lecture is a reasonable task. If I were doing longer lectures, I'd probably break the responsibility in half. For undergrads, a group of 2 or 3 working together to cover 50-80 minutes of classwork is reasonable.

  • To make the assignment more manageable and the notes more uniform for my grad students, I made a big LaTeX template with all the macros I thought they might want.

  • For my undergrads, at first I required them to use LaTeX, working together on Overleaf, but I found I was spending a lot of time answering LaTeX questions, so I switched to telling them they could use any medium they wanted. You should probably decide whether it is worth spending your time teaching undergrads to use LaTeX. A downside of letting them use any medium is that there wasn't a good way for me to edit those notes.

  • Impose a tight deadline. My deadline now is to have the notes done 24 hours after the class, before people forget what happened.

  • To make the tight deadline fair, let students choose what date they want to work on several weeks in advance. Ideally, warn about dates which will be unusually diagram heavy.

  • I scribe the first lecture myself, to illustrate to the students what I was looking for. I post the TeX source of the notes along with the PDF, so that students could copy techniques from each other.

  • For my grad classes, I edit the notes, both to remove typos and to add additional comments on material which I had not covered well. This was probably 30 or 60 minutes of work per lecture, but when the turn-around cycle was tight, I felt that the time doubled as helpful preparation for the next lecture by reminding me what I said rather than what I meant to say. It would be better if I edited the undergrad notes as well, although I currently am not because so many of them are handwritten or in Word.

  • I frequently say things in lecture like "I don't know the precise answer to your question, but it is something like ... and I'll put the details in the course notes", and I in other ways encourage people to treat the notes as a reference.

Some things I haven't figured out a good way to do yet

  • Incorporate this into the course grade. For grad students, I just don't; I instead praise good work informally, in public and private. For undergrads, I had a small portion of the grade which was my feeling about how well they did on class participation and note taking, but I will freely admit it was very subjective.

  • Teach students to use LaTeX better, or to write better. It is pretty easy to fix their work, but I never feel that I have time or energy to tell them how to improve it next time.

  • Thanks. There are many good ideas in this, including using a breakpoint and double scribing, as well as giving students more notice and more information about what they'd be scribing.
    – Suresh
    Apr 25, 2014 at 14:56
  • (most of the links have stopped working sadly)
    – mafu
    Sep 28, 2015 at 11:43
  • +1 for "I made a big LaTeX template with all the macros I thought they might want." :-)
    – Mico
    Jan 7, 2020 at 12:41

Pay a student who's not taking the class, and preferably has already taken it, to capture the lecture and discussion. That person will be able to focus objectively on what's happening.

Alternatively videotape or do an audio recording that the scribe can use.

Ask everyone in the class to provide their notes to the scribe.

The scribe will be focused on trying to capture questions, answers, comments, etc. which means he or she may find it difficult to acquire a deep understanding of the material. The scribe also may find it difficult to participate in class because they're too busy writing down/typing everyone else's questions and comments.

I've performed as a 'scribe' as a graduate student--I was a TA and sat in the class to help deepen my understanding of the material and to know what the students were being taught. I was then able to expand on (and correct some equations) the material during lab sessions. I had to pay close attention, but having previously studied the material made it a lot easier.

I've also taken minutes for a subcommittee that I'm on occasionally. I find it tremendously difficult to switch focus between trying to capture everything important and actually participating through providing information or asking questions.

  • 3
    I like the idea of audio/videotaping, or in more general not having to rely on real-time note taking.
    – Suresh
    Apr 24, 2014 at 17:51
  • 1
    FWIW, there is free software that can capture audio, video from a computer and/or a camera. For blackboard lectures, it's also possible to set up a high-res camera that takes periodic shots of the board.
    – Raphael
    Apr 25, 2014 at 10:31

As a student, I would send out a Google Doc to each member of my class. We would all collaboratively take notes together. This was, of course, chaotic and messy but it worked very well. The anxiety about 'missing something while writing down something else' was alleviated with more eyes and ears on the subject. When the lecture already had or could have followed a written structure, (say a review sheet or an agenda for the day) we would base the document off this, which worked even better as people would focus on one topic or bullet point and make sure that was finished and fleshed out as the rest of the class raced on.

However, this does mean giving your students an additional task. Sometimes, as has been noted, active note taking and learning don't go well together. However, student lead, collaborative note taking is another option, if maybe not the best one.


I have taught a similarly structured graduate course, and seen a similar wide variation in quality of the notes. I did assign scribes at the beginning of the class, and allowed them to swap as long as everyone ended up scribing at some point. I feel like this helped in matching students to topics they were particularly keen on.

For the future, I've considered having all the notes be in a wiki. There would still be a "main" scribe for each topic, but everybody (including me) could contribute to any page. They would be assessed on their overall contribution, which is trackable from the wiki history. The assessment structure is pretty much pass/fail anyway, so I'm not sure it makes much of a difference, but I like the idea of rewarding students for helping each other. Also, it means it's OK for each wiki page to start out as very rough notes, potentially from many hands: it can be progressively rewritten. This avoids the problem where the original scribe gets stuck at an early stage - if they missed something, or misunderstood something, then they can get help on it, but still be able to make a meaningful contribution to the end product.

This may not be suitable for you if you have a lot of mathematics or diagrams (depending on choice of wiki software), or if you would prefer things to be in LaTeX for other reasons.

  • 2
    I like this idea. And with things like sharelatex it's even possible to do collaborative latex editing.
    – Suresh
    Apr 24, 2014 at 20:46

You say it is an advanced course, so you may have met some of the students before in beginners courses, projects, etc.

Pick one or two competent and reliable students and pay them to write lecture notes¹. Repeat in subsequent iterations with the purpose of incremental refinement.

You can sell the idea to them by noting that

  1. they'd get paid for work they'd have to do in some form, anyway,
  2. they'd learn more because they'd have to do it a bit more precisely than for their own jot paper and would get feedback from other students (and you),
  3. they'd have external motivation to do said work (counter-procrastination) and
  4. they'd have the chance to make a good impression with you.

  1. I think lecture notes should be different from transcripts, but ymmv.

students are expected to produce a first draft within a few days of the lecture (lectures are once a week), and the hope is that a good version of the notes is on the website a week after the lecture.

The timing seems a bit strict. If the students have a lot of projects going on they may not (want to/be able to) make time for this. Maybe this is one of the reasons you are getting abysmal first drafts.

  • 2
    Yes. the constraint is that it's useful to have notes from the previous lecture in time for the new one. I should clarify that the students in these classes are usually 2-3rd year Ph.D students or even more seniors, so they don't have the same class load as (say) UGs or intro grad students.
    – Suresh
    Apr 24, 2014 at 15:32
  • @Suresh but clearly, PhD Students can be busy as well? Paper deadlines etc ...
    – xLeitix
    Apr 24, 2014 at 15:57
  • 5
    It could help to assign the scribes further in advance, say, at the start of the semester. Then they know in advance when they will be scribing and can try to clear their schedules accordingly. But it seems to me that timing is probably not the main problem. Apr 24, 2014 at 16:04
  • 1
    @NateEldredge Given the number of students who seem to do every assignment on the night before the deadline, regardless of when it was set, it's not clear that scheduling in advance will help anything at all! Apr 25, 2014 at 9:16

You might find a "delta-scribing" process more effective. To prepare, provide (enough copies of) a rough outline of the material to be covered in that lecture, ideally in the order to be covered, with a blank box next to each bullet point, and space somewhere (perhaps on a different sheet of paper).

Now the goal of the transcription team is to record the actual order of points presented. The order presented is recorded in the blank box, usually by a number. The space is used to record as much of the additional words needed for the lecture, including particularly cogent phrases describing the material, as well as subpoints of discussion. This may save just enough work on the part of the team that all the additional ("delta") salient points brought up in lecture can be covered.

If the lecture is actually more of a free-flowing discussion with many people introducing new points, it might be good to have each two-person team operate so that one transcribes the points the other person presents. Perhaps a brief discussion as how to best transcribe the point might be in order, so that there is a "secretarial consensus" of sorts.


At the University I am a student at (Undergraduate in Liberal Arts), we have mandated discussion sections for courses over a certain number of students. In one of these discussion sections, we have two people each week (who signed themselves up ahead of time) give a short presentation/summary of all of the readings for the week and a recap of the lectures.

I think that (depending on the format of your class) having your students themselves present to the class their findings might motivate them to do a better job at taking notes.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .