A brief personal guide to doing research

Research is really hard. Try to be hard on writing and ideas, not on yourself.

This post is a collection of ideas about doing research that I often cover when presenting to PhD students. As a profession, we are really good at research, but we are not that good at teaching how to do research. This is a problem because many people struggle with this exact point. Doing research is really hard, and it would be nice if someone could teach us how to do it. Fortunately, lots of people have written guides for different parts of the research process (I have a lot of links collected on my website). But there are some things that I think are missing, and this post will cover some thoughts I have. I hope it is useful for someone.

Be hard on ideas, not on yourself

We start with the most important part. You should be hard on research ideas, not on yourself.¹ You have to be critical of your ideas, and you have to let others be critical of your ideas. This is a key part of doing better research. But realize that self-worth is not at stake, even though it may sometimes feel that way. If you associate your research ideas with self-worth, then you will probably not seek out feedback.

Thanks for reading Claes’s Substack! Subscribe for free to receive new posts and support my work.

I used to be really worried about my writing, and I did not want to let anyone read it. In fact, I dreaded going to my supervisor because I did not know how to handle feedback. This changed once I started working with a coauthor on a paper, and I learned that bad writing could be fixed. It wasn’t that I was a bad person, it was just that I didn’t write that specific paragraph well. That’s ok, because writing can be changed! Try to go from “OMG it’s all terrible” to “OMG it’s all terrible, now let’s make it better”.

So be hard on ideas, not on yourself. Seek out feedback, but try to not take it personally.

You should learn how to ignore (some) feedback

A common scenario: a PhD student is presenting a vague idea to a room full of researchers and gets stuck on a slide. Time passes. Everyone has time to think about the topic, and because it’s vague from the beginning, they will start to interpret it in their own way. Everyone in the room is smart, wants to help, and has ideas that they think are really interesting. The outcome is a lot of discussion around lots of different topics, with lots of suggestions for how to take the vague idea into six different directions. People want to help.

This is usually a disaster for the PhD student. Incorporating all the feedback will make it impossible to write a good paper, because there are just too many different things to do. Everyone is trying to help, but they will do so by talking about what they find interesting. What you need to do is find the feedback that relates to the best idea that you have, and just incorporate that part.

How can you do that? I think the best strategy is to ensure that you have a solid understanding of what your idea is about before starting the presentation. What is the one contribution of your paper? What is the one idea that you want to investigate? This is usually a painful process, where you will have to throw away many interesting ideas and thoughts to ensure you focus on one central contribution. This requires honesty with yourself about whether this is a good idea or not, and will require some trial and error. Do not expect to come up with the right idea directly.

But while it is painful to throw away many interesting things, it is also very useful because it allows you to throw away many interesting things. Not everything that is interesting should be included in your project. It will also allow you to tell all the well-meaning people in your presentation that, thank you, that is very interesting, but I am focusing on something else.

How to come up with ideas?

Good ideas often come from someone making connections between literatures, between theory and real-world details, between what’s interesting and what’s measurable. A very valuable thing is to sign up for email alerts for working papers or journals. That always helps me get a sense of what people are doing, which I think makes it easier to find connections. Beyond that, you have standard advice such as read newspapers, look at the world, etc. All those are probably useful, but what works is all a bit idiosyncratic. Find your own way!

But try to fail ideas fast. For a PhD student, there are many many ideas worth spending 1-2 weeks on. There are few ideas worth spending six months on. Be critical of the idea, not of yourself.

Another thought to keep in mind: Is there “alpha” in reading the same literature as everyone else? In non-finance parlance, can you get ahead by reading the same papers as everyone else? I think this is an interesting thought raised on Twitter by Alz. It might be better to try find your own niche. But I don’t know!

Some practical tips

Write like a newspaper, not like a mystery novel

You don’t reveal the murderer at the end of the paper. Academic writing is not a mystery novel. Instead, you want to write like a newspaper article: start with a summary of the most important bits, then go into details. Your readers are busy and want to know the research question, the method and the outcome very quickly. Make it easy for them.

Start anywhere. The first paragraph is always the hardest

I’m sorry to tell you, but if you are a researcher, you are now really just a writer. Lots of people struggle with writing, so try to make it easier for yourself. You don’t have to start at the start. The first paragraph and sentence are the hardest to write. Instead, if you have a nice idea that you want to write about but you are not feeling inspired, you can start with the easy parts! Write the data section or the background, or whatever you want at the moment.

Good academic writing is clear and to the point

Editing is where papers are won

Your first draft doesn’t have to be good. The first draft is just to get something down on paper. The second draft is to make you sound sane (not my quote, but I don’t remember who said it). Nobody writes cleanly on the first try. Clear writing is mostly rewriting. Again, this makes it really important not to assign self-worth to your writing. You have to be hard on the writing, because that’s how it gets better.

How can you become a better writer and learn how to edit? Take a look at this page and find a writing guide that you like. I like the concept of a “shitty first draft”. I think tools like ChatGPT are also very helpful, but you have to learn how to use them. I tried writing this post with the help of ChatGPT, but it just gave me a bunch of bullet points. That’s not useful or fun to read.

Summing up

Being a researcher is fundamentally different from being a student, and most students are not taught how to do research. They are taught how to understand research. There is a fundamental difference here: finding new ideas is different from solving problem sets, and writing a new paper is not the same as understanding an existing one. This often creates somewhat of a problem for people who have proven themselves to be very good at solving problem sets. Paul Niehaus refers to “a dramatic loss of structure”, creating problems for PhD students, and I completely understand why.

In part, it is difficult to teach how to do research because there is no one way to do it. Everyone has a different style, there is a lot of implicit knowledge, and you have to decide for yourself what you shine a light on. Try to be hard on ideas, not yourself, and expose yourself to feedback. I’m sure you will do great.

1. Credit for this phrasing belongs to my colleague Peter Andre.