Create your research system

We now begin the heart of this online guide for research and strategy. This chapter is akin to figuring out what you need to set up your research workshop. This is similar to setting up a woodworking shop in your garage. Below is a picture of mine. It could be better; I’m still developing my system for woodworking.

My research workbench or system is much more developed. I have been holding it over the last decade and a half, starting as a Ph.D. student, and every year I try to refine it little by little to make my work. Specifically figuring out ways to reduce the cost of producing high-quality research components.

The first element of the research system is your analytics toolkit. Excellent technical skills are the bedrock of a successful research career. Today, publishing requires both the understanding of theory and the ability to tease out meaningful insights from complex data sets. Even before you start your Ph.D., tool up. Here are four resources to get you started

R: This is perhaps the only statistical programming language you need to know. It is free and comprehensive (you can do visualization, machine learning, traditional econometrics, or write your own custom algorithms.) Even if you don’t use R that often, it is one language that all social scientists must learn. For Ph.D. students who may not have a lot of money or resources, I highly recommend learning R. I use vanilla R. Still, there are other packages like R-studio that may be very useful if you want something more user-friendly.

R Studio: this will make it easier to use R, and R studio is substantially faster for some use cases. R studio is free for academics, so it is worth a shot.

The other technology that I use for research is Stata. This is much more expensive, but I have found it indispensable since I learned how to use it as a junior faculty member in my first job. Stata is spectacular for traditional research analysis. If you want to produce publication-quality tables, use the latest clustering method for your standard errors, or just for some new bias in panel data, Stata is the right tool. There are different versions of Stata, some less expensive than others. Building up good Stata skills is worthwhile, as many of your co-authors across the field will likely have experience with it. This is true, especially if they are economists.

Stata SE or MP: If you are an economist or generally estimate Y = B0 + B1X type equations with worries about clustering standard errors or endogeneity. R may sometimes feel like overkill. Stata is my go-to software for most of my analysis. I learned Stata by working with a collaborator. With great Stata code, you can go from your raw data to publication quality tables with a press of the “do” button.

The final thing is that many young scholars are beginning to use Python. I have not learned Python, primarily because I’m of a different generation where it was not the dominant platform for empirical analysis and social sciences. However, as data sets have gotten larger and machine learning more critical, Python is a must-know tool for any minor social scientist.

Develop a writing workflow.

Building a process for clear and well-produced writing is paramount for success. Here are some tools and resources that can help you improve the quality of your written work.

Latex: Robert Hall, the Stanford economist, in an article about becoming a professional economist, said: “Pay close attention to the appearance and dissemination of your work. I hold the following controversial view that my economist wife thinks betrays a lack of spiritual development: There is a separating equilibrium between researchers who put out nicely typeset papers in Latex and those who struggle with the infirmities of Microsoft Word.” Learn Latex, and your readers will appreciate it.

Overleaf: Once you learn Latex, start using Overleaf. It is like Google Docs for Latex and helps you write beautiful latex manuscripts in a collaborative environment. Version 2 is even nicer, with the ability to add comments.

Grammarly: I am a fast typer, and sometimes I forget to include words in my writing. Grammarly is a bit pricey, but I use it all the time, so it has been worth it.

esttab: I can’t believe I used to create regression tables manually. If you are still creating tables by cutting and pasting your regression output into MS Word tables, please join us in the 21st century. Esttab will help you make beautiful Latex tables. The best part is you can link these directly to your latex files using the \include command, so every time you update your regression, your manuscript updates automatically too!

GoogleScholar BibTex export: I write in Latex. Google Scholar has a BibTex output feature that lets you cut and paste a BibTeX bib from Google Scholar. Beware, though, sometimes things missing or journal titles are awkwardly capitalized. But if you develop a good process, you only need to fix the errors once.

Dragon Dictation: Carpal tunnel is real. Having suffered from severe wrist and hand pain due to years of typing, perhaps lousy posture, and a host of other pathologies that came from working for the computer all day, I realize that having a good process around writing is really important. I now have a system where the first draft of many of my writings, or dictated and not typed. I’ve tried many voice dictation software. Initially, I used DragonDictate for the Macintosh. Unfortunately, they discontinued it, and I was left without dictation for nearly 2 years. I tried many alternatives, including the dictation built into the Macintosh, Google documents, and Dragon anywhere mobile with the smartphone. I found the solutions needed more important dimensions. This year I figured out a solution. I bought a cheap laptop – Hewlett-Packardwhere I installed Dragon individual professional, and I dictate within sublime. This has been a game-changer.

Sublime: Need a simple text editor with lots of power? Sublime is fantastic, and I find myself using it every day for the little bits of text work that I have to do.

Leverage the cloud.

Improve the processes around your work. Get the technology that reduces duplication, idle time, and inefficiency.

Dropbox: 10 years ago, I used to email my in-progress manuscripts to myself at the end of the day, so I could work with them on my home computer. Today, Dropbox is the one technology that has reduced the digital shipping waste I create. I usually keep one copy of a file—whether it is my data, notes, or code. I can work on these assets from anywhere. How amazing.

Google Docs: Writing a collaborative proposal? Responding to a reviewer letter? GoogleDocs has helped speed up these collaborative tasks.

Master your email and calendar: A few years ago, I realized that I sometimes checked 4 email addresses a day: an old Yahoo address, my work email, my new Gmail account, and a Gmail address for subscriptions or junk mail. Now, I’ve got one email address. I probably got 30 minutes back.

Learn to delegate

The best professors know how to break up their work into modular chunks and delegate it to others. This frees up time to do more important things. Sure, it might be instructive the first few times to clean your data, do a preliminary search of the literature, or develop a website for your research project. But it’s worth learning how to delegate these tasks so you can put your energies to more creative uses.

Hire someone on Upwork: You can hire people to do almost any task on Upwork. I’ve gotten people to copy-edit my articles, build a citation database on a topic I wanted to learn about, and scrape data from a website. Start with small projects and build your delegation skills on a platform like this.

Hire someone: Learn to break up your work into modular pieces and delegate the stuff that is probably not worth your time. Delegating is perhaps the master skill of the productive academic. If you want to learn how to delegate better work with someone who does this well. Start by hiring an undergraduate research assistant and get them to work on a small project. Remember to manage the work effectively; ask yourself these four questions.


MacBook Pro: I stopped using PCs about a decade ago, and my go-to computer is a MacBook Pro. The MacbookAir is a good entry point for a social scientist looking for a computer that can handle most of the software you need for statistical analysis and academic writing.

RAM is a barrier: If you can afford it, go for a computer with a great CPU and lots of RAM. But buying more RAM for your computer is worthwhile as your data sets increase in size.

SSC is critical. I bought a cheap desktop without an SSC. It was painfully slow.

Since I can’t run Dragon dictate on a Mac anymore, I had to buy a PC. I bought a desktop, but it was incredibly slow. I then bought a 350 dollar small laptop with an SSC drive. I only use it for dictation on Sublime and directly save my files (.txt) to Dropbox. This has been such an improvement. I can write at least 1k words daily because I can dictate rather than type.

Good monitor: Monitors are cheap. You can find large and high-quality monitors anywhere, really.

Get another monitor.

Keyboard: Get a comfortable keyboard. I use the Microsoft Sculpt Ergonomic Wireless keyboard at work.

Keep learning

Here are two hacks to help you get up-to-date on academic literature and find interesting ideas in popular literature.

How to read an academic paper: You should spend more time writing than reading. But having a good process for reading academic papers is vital. Check out these tips from Science magazine about reading academic papers. You could also have a computer read the papers by using software like NaturalReaders.

Audible: If you go to the gym, have a commute, or want to learn something new at the end of the day, start listening to audiobooks. A few years ago, I got a subscription to Audible. I’ve learned many new things about topics I didn’t know much about on my drive home. The selection of audiobooks available today is remarkable. You will find books that relate to your existing research area or a new area you would like to explore.

Some more advice.

Experiment: Try little things and see where they take you.