Soon, I'll likely teach a stats methods course at the masters level as part of my policy PHD program. If it were up to me, the class would only have one or two assignments determining your grade (putting attendance and other mandatory things that I can't change aside). This class would have lectures/readings, but 0 homework, no quizzes, no discussion posts, no nonsense.

The only assignments that determine your grade in this case are the final paper, and the first draft of that same paper. The paper would be a real paper, where students must collect, clean, and analyze a dataset, applying one of the methods we'd discuss that semester. I view there being several advantages to this.

The first advantage is pedagogical: to me, the only way for me to REALLY know if you understand something is by you writing it out to me and proving to me that you do. People can be nervous test takers. In-class tests have arbitrary times as determined by your class. The professor can sometimes choose questions that're worded unclearly or unfairly. A written paper? That you have many months to prepare about a topic that you choose? No, there's less of an excuse here. Why?

Well, you'll have books and freeware to explain the statistical material. You'll also have me for this purpose. You can also call me, come to office hours, email me for feedback on topics, data sources, whatever. You also have no quizzes or homework to worry about, so the burden my class has can't be used as an excuse for why the draft and final aren't done. In other words, you'll have all the time in the world and all the resources you could ever dream of to ensure the assignment goes well.

I believe it also prepares masters/PHD students for their careers one day where, if they go into academia, they'll be evaluated by their publications. A class that demands a quality paper they conceptualize and write is more representative of what they'll actually be doing in the real world as researchers, generally speaking. I'm not a professor yet, but I've always figured that this would be a good model for instruction (at least for stats methods in the social sciences). I was wondering if this model made sense or not from a pedagogical perspective.

EDIT: I should've specified more the first time, but yes, the degree to which I'd pursue this specific model will depend a lot on if it's a PHD class and if it's elective. If it's elective and PHD, naturally I'll presume everyone has certain working background knowledge (the masters level stats courses including causal inference) and wants to dive further. Now that I think about it more, this model would only be tenable for PHD students interested in methods. For a masters course, the way I've outlined it would be much too cruel of an expectation.

If I did anything like this at the masters level, I would likely only ask them to design a study, likely using synthetic data (that's what we did in my masters course), and implementation would be completely optional. I would also likely have additional assignments since I would presume most people have little to no exposure to the material and would need that to learn better

  • 2
    I think this is way too open ended for a class. Further, what sort of data are you anticipating them collecting? Remember, anything including human subjects has to go through your IRB.
    – Jon Custer
    May 12, 2023 at 14:39
  • 1
    Administrative data. I would want to teach the causal methods/advanced econometrics courses. So, it would be about Difference-in-Differences, Interrupted Time Series, Instrumental Variables, Synthetic Controls, Matching, Regression Discontinuity, and other common causal designs we see in policy/economics. I would also want to teach a class just on synthetic controls, but I'd still desire the same basic setup with modifications where needed. This does usually exclude situations where you'd need an IRB, naturally. May 12, 2023 at 14:44
  • 1
    A great idea in principle, but hard to do right. You will have to provide a lot of scaffolding and hands-on assistance toward the final term paper. May 12, 2023 at 17:02
  • 1
    > if they go into academia -- what if they don't? May 12, 2023 at 17:07
  • Well, it really depends on the level. If it's the PHD elective methods course, I'll presume you want to go into research. If it's masters/phd required, I won't teach it as if everyone "wants" to take the course/has requisite background knowledge May 13, 2023 at 12:47

3 Answers 3


This looks and feels to me like a typical capstone project with students who have had a chance to digest the material already.

As your students would work as individuals, you would really work them either very hard or the results will be disappointing.

Even if it involves group work, it assumes that your students can pick up on how to do statistics very quickly and on their own. A certain part of your class will not know how to start or even whether their project is feasible. You would need to get involved in their thinking process continually and for long periods of time. Another part of your class will feel that they pay you (or rather the university) to explain them statistics to you. Expect to have a conversation with the dean or your chair. It is hard to come up with a project that gives just weights to the complete process of statistical analysis as you envision it. Your colleagues might complain that your students will never have seen X or Y when they come out of class. If something goes wrong without the fault of a or the students, you have no fallback. You are running the risk of getting the worst teaching evaluations in your department ever and students will then hate you and your guts.

Now, here is what I can advise you to do based on my 30+ years of teaching. Do not take this as an absolute, others had different experiences and the personality of the instructor is very important in what works and what does not work.

  • First, I agree with you that an ideal setting of project based learning would show better results than what we typically end up doing. In fact, your proposal shares many elements with a flipped classroom.

  • Second, in Business Administration master programs, case based learning is prominent. You might pick up some ideas by looking at what they do. You will find that colleagues across campus are surprisingly accommodating if you say that you want to learn from them.

  • Third, since this is your first try and you do not yet know all the things that can go wrong, look into smaller projects, at least two per quarter or three per semester. This way, if without fault of the student something goes wrong, you can adjust to the failure.

  • Fourth, even Master students need motivation. If you flip the class-room, then each class-room session should start with a quiz on the material to be learned for this class. The quizzes allow students to learn time management, something that your proposal presumes.

  • Fifth, you need to specify what material to learn and how to learn it. Just because you find the material now easy does not mean that others do.

  • Sixth, students are in your class not just to learn statistics, but also to learn the skills to pull off a longer project. If they work with other students, they need to learn some team management methodology. They will probably need more guidance.

TLTR: Learning by doing is ideal, but you should see your vision of this class as a goal to achieved over many iterations of this class. Organizing project based learning is difficult and you might need to learn by trial and error. Limit the size of the errors so that they are not career ending. Include the experience of others, there is no need to reinvent a catastrophe. Remember that you are also tasked with teaching your students time management, planning, project management and so much more that I cannot think of at this moment.

Immediate advice: Doing something unusual only pays off if it is a success for all. You might not remember, but there was indeed a time when nobody ever was fired for buying the system from IBM.

  • You're quite right, this was quite helpful. I never considered the MBA programs which rely a lot on case studies, I'll look into those. I should also say that the way I would teach any course will depend a lot on if it's required and if it's a phd course. If it's an elective phd course I'll presume everyone has certain basic background knowledge. If it's masters, well then that assumption would be very unreasonable, and I'll tailor the class to reflect that. I likely wouldn't do this model for masters students, unless the candidates show unusual aptitude. May 13, 2023 at 12:52

On paper this sounds like an amazing class, one that I would have liked to take in graduate school. But I could easily see this going sideways very quickly. I can only speak from a practical perspective, not necessarily a theoretical, but there may be problems.

Whether or not this is well received is heavily dependant on what type of masters-level students you are teaching. Are they interested in a PhD or research work using these techniques? Or are they just caught up in this class as a requirement of their program? This could affect how willing they are to produce what you consider a "real" paper.

Another consideration is the level these students will be on day one of the class. A paper of the scale and quality that you imply almost demands that they already have enough experience with these methods to pick a topic and start developing it early in the course. But if this is a more introductory course how could they?

Following this train of thought, what happens if a student picks a topic and data that doesn't work out -just like real life research sometimes hits a dead end? Will they be penalized for this? How will you ensure students pick an appropriate topic with a usable data set and answerable question?

Finally, and maybe most importantly, does this format achieve the goal of the class, to teach statistics methodology?

Overall, these questions are just meant to help you think of this class from the students perspective. I don't mean to trash this idea, I love it. But I think to put it into practice you need adjust your expectations somewhat. I would suggest, at a minimum, more than just a single draft of the final paper. You may consider including a graded assignment at every step of the development process i.e. a proposal, a short lit review, a methods section (which is important given that this is a stats methods class), a first draft, perhaps another draft, and the final paper. Although following this format you risk turning this into a writing class.

Alternatively, perhaps you could require students submit a hypothetical methods section for each topic you cover, with the goal of picking one and turning it into their final project. This would ensure that they have mastered the content at each step of the way, while still working towards your end goal of a quality, "real" paper.

  • 1
    Much of what you describe is how I was taught methods (one paper, but say 6 grades, the proposal, literature, data, etc). I've edited my original question to discuss the level/experience of the course, since that'll matter a lot to me, and the students. I think one paper with multiple grades would be a good model to begin with. May 13, 2023 at 13:09

I think you need more structure, here. I think you are also designing a class which rewards only learning material in depth, but does not reward any learning in breadth.

To the first comment: It's all well and good to prepare students for the day when they will be judged solely on the quality of their publications. It is not all well and good to do that by just judging them solely on the quality of their publications, full stop. As others have said, this is a project-class, and without exception, every successful project-class I have ever taken breaks the projects down into smaller steps and assigns grades to those steps, often with presentations in class.

This has a lot of benefits:

  • It shows, it instructs, the students in the art of managing larger projects, where in this case the project is a statistical analysis and written document. This is a skill that can be instructed, and mastered. It doesn't just happen.

  • It ensures (not 'encourages', 'ensures') that you have up to date knowledge on the status of the various projects through the term, and it tries to ensure that the students are not drifting 35 degrees off the direction you want them to go. No one wants to work hard for three months, then find out there was no meeting of the minds with their professor, and now they get a C+ and their advisor wants to know what the heck just happened.

  • In-class presentations (for all of the various project stages) although many students hate them, can be incredibly useful. There is great virtue in one student seeing another student use a given technique in ways they would not have thought of. And although it's not pleasant, there's also great use in seeing mistakes caught and pruned early. Student A might not be making that mistake today, but seeing Student B make it and be corrected may prevent later errors.

But this of course all needs to be shaped, structured, managed by you, and communicated to the students. It may result in some loss of the total freedom they would have in their own future publications, but that is the price of learning.

To the second comment: If the only thing you truly intend to teach is how to apply already known techniques toward a larger project then these comments probably don't apply. But if you're also teach a toolkit or a spectrum of techniques, what you're doing here is telling the students very clearly that the only techniques they need to learn are the ones that apply to their projects. Students being students, some will take that to heart and ignore everything out of the scope of their project.

You must log in to answer this question.

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