Advice for PhD Economics Research

Transition to Research

Preamble

Your transition from taking classes to doing research is abrupt. You have spent most of your life becoming very skilled at listening, reading, and repeating (taking tests). Now, you are told to create. Your dissertation is due in a few years and should be filled with new research.

Most humans react similarly to this situation. You feel daunted. You feel anxious, uncertain. You are not sure how to start. You assume nobody else feels this way and that everybody else knows what to do. Consequently, you do not ask for help.

Your feelings are perfectly normal and fine. It is daunting; you're allowed to feel daunted, anxious, uncertain.

But, you should ask for help. Ask for help getting started with research from a professor. This does not mean you have to keep them as your advisor until you graduate; if it is not a good fit for whatever reason, the professor will be just as happy as you for you to find somebody else to be your advisor. Ask the professor for something to work on, to help you practice doing research (not just reading and learning about research).

If you have problems a professor can't help with, then find somebody else who can help. A large percentage of PhD students struggle with anxiety and depression; if you're among them, there are resources through Mizzou where you can ask for help, like the Counseling Center.

Specific Suggestions

The following are synthesized from various sources and my own experience.

1. Create milestones for yourself. Don't have a to-do list that just says “write dissertation.” Break down that big goal (dissertation) into smaller pieces, like each chapter. Then break down your first chapter into pieces. Keep breaking the pieces into smaller pieces until you have something specific you can accomplish today. Keep some medium-size pieces as “milestones” that you can celebrate. Passing your dissertation is an obvious proposal that is well worth celebrating, but hopefully you can find something to celebrate before then, too.
2. Structure your schedule, including times to relax. Without classes, you have a lot of flexibility. This is great: you can structure your own schedule to maximize your productivity. But if you don't create an explicit structure, you might both feel very unproductive and feel like you're working all the time. Figure out when your productivity is highest, medium, and lowest. Schedule your “difficult” research for your high-productivity times; e.g., if you're at your best when you wake up, use that time to try to write proofs or write a difficult piece of code or make presentation slides (or whatever you find most difficult). If possible, schedule your TA work in lower-productivity times; it's still important to do well, but (presumably) it requires less mental effort. In your lowest-productivity times, don't try to work; try to relax and enjoy another activity that restores you.
3. Just keep doing research. At first, it won't make sense; you have no perspective or deeper understanding or broad knowledge of the field. And you won't think your work is good; and you may well be correct. But keep going. You will fail. You will fail many times. You will feel frustrated. But keep going. For research (and running marathons and most other serious endeavors), learning-by-doing cannot be replaced by simply reading and watching and listening. Eventually things will make sense; eventually your research will be good. (But you will always keep failing and keep feeling frustrated sometimes because that's the nature of research; if you don't want to keep failing, then you will be happier to start looking for another job now.)
4. Avoid "productive procrastination." For example, in the short-term you may feel good/productive about taking an extra (not required) elective in Year 3, because you are much more skilled at taking classes than doing research, but you will almost certainly (like me and other students) regret this later when you realize how much potential dissertation research time you lost (and realize the class did not make you better at doing research).
5. Set up a regular (e.g., weekly) meeting with a professor (later your advisor) to ask for help. Most PhD students are bad at asking for help, so it's good to make a commitment mechanism like this. (I once got “lost” for a couple very unproductive months without meeting with my advisor; this happened more frequently to my roommate who then ended up dropping out, so please learn from his experience and don't skip meetings just because you have “nothing” to show your professor.) If you meet regularly with a professor (advisor), and you listen to their advice (and they are a decent advisor), then you will be able to graduate.
6. Take care of yourself. Take time to have fun, exercise, sleep, talk to a counselor if you're feeling really depressed and/or stressed, etc. The PhD program is an investment in your human capital, so it would be counter-productive to let it depreciate for any reason.

Writing

My advice is in Chapter 1 here

I also like John Cochrane's writing and presentation advice (although I do not agree with 100% of it)

Presentations

Links: Presentation Advice

Some of the writing advice in Chapter 1 here also applies to presenting: thinking about structure, simplicity, transitions, etc.

Presentation advice from Piazzesi, Schneider, and Tonetti (focusing on simple, clear, efficient communication)

Section 4 in John Cochrane's writing and presentation advice

Responding to Questions

For many students, the most stressful part of a presentation is being asked questions. This is partly intrinsic: you can practice your talk precisely, but you don't know which questions you'll be asked ahead of time. (Your advisor may have some educated guesses, though, and questions you're asked in the Research Workshop or conferences may well get asked again during your job talk.) But, it can reduce anxiety to have a procedure to follow for every question you're asked.

1. Reflective listening: repeat back the question in your own words. This has many benefits, including:
   • help catch any misunderstanding before it gets awkward
   • signal to the questioner that you care about understanding their question (especially important during a job talk)
   • gives you extra time to think of a response
   • gives you an opportunity to emphasize the part of the question that you think is most valuable (or that you can best respond to)
2. Classify/triage the question: decide which of the following categories the question is in, and respond accordingly.
• Clarification: somebody wants to make sure they're understanding properly, or wants clarification for something ambiguous. This is a good sign (they're trying to understand you!), and usually easy to answer, so you should always answer these fully.
• Critique: somebody asks about a limitation of your results, or about an alternative that is better in some way. Hopefully you are already well aware of the related literature (alternatives) and your research's limitations, so these should be relatively straightforward to answer. If it's a limitation, be honest about it, but feel free to emphasize the corresponding advantage; usually there is a tradeoff, and you accept certain weaknesses in return for other strengths. If it's another study on the same topic, again be honest about the positives/advantages of the other study, but feel free to briefly describe the advantages of your study; presumably yours is not worse in every dimension, or else you should not be presenting it. If somebody asks about a limitation or alternative you hadn't thought of, then (after some panic) do your best to first assess whether it is indeed a limitation or appropriate alternative, and then ask a quick follow-up question so you understand their point well enough to look into it further on your own. Make sure to write a note to yourself on a piece of paper, ideally with the person's name, and say thanks, you're making a note to look into it more later.
• Speculation: somebody asks you "what if...?" beyond the scope of your paper. This is common: people want to check if you have developed enough expertise to make an educated guess about something you have not actually worked on and practiced presenting. These questions are more difficult because you need to come up with a new idea in the moment, but also easier because there is no correct answer. Take your best guess, but also be very clear that it's only a guess (like "I'd conjecture that..." or "I haven't looked at it formally/empirically, but I'd guess that...").
• Trying to sound smart: somebody just wants to sound smart, not actually contribute anything meaningful. Especially in a job talk, you want everybody to feel good about your talk, so if somebody is insecure enough to waste everyone's time trying to sound smart, then (briefly) go ahead and compliment how great a question/comment they made. If you need to practice with your friends how to sound sincerely enthusiastic when you're not, then do so; it will backfire if your praise of the questioner sounds sarcastic!

Shawn's Suggestions

These suggestions are courtesy of Shawn Ni. He has advised countless PhD students at Mizzou and led the PhD Research Workshop for many years. I added some related details.

• Do not simply read your slides.
  • To help avoid this (and for other reasons): do not write in full sentences on slides.
  • Plan beforehand what to say for each slide; you can even write out what you plan to say in full sentences (and even bring these as notes in case you forget), but do not put the full sentences on the slides.
• Ensure everything on each slide is visible/legible. For example, don't copy a huge results table and scale it down to 10% where each number is only 6 pixels tall.
  • Don't use more than two levels within bullet point lists (like this). Not only does this help visibility, but it also alerts you to a poor structure. If you think you need a third level, either a) you are including too much detail for a talk or b) you need to split the slide into multiple slides.
• Practice
  • Including timing: I suggest recording how long each slide takes (rounded to the nearest half minute, say), and then thinking how you want to change the timing. E.g., you may realize a single slide is taking 5 minutes even though the topic is not primary. You can either try to write down times as you practice, or just record yourself (e.g., in Zoom) and then watch the recorded video to see time stamps.
• Practice…
• Practice!

Links: Slides in Beamer (LaTeX)

• Real example (in file _talk.tex)
• Older real example
• Dave's templates
• Overleaf's Beamer help page