Sharique Hasan

Create your research system

This chapter is like setting up a workshop, whether it be for research or woodworking.

I have been developing my research system for the past decade and a half, starting from my years as a Ph.D. student. Every year, I work on refining it to make producing high-quality research components more cost-effective.

The first component of the research system is the analytics toolkit. Having excellent technical skills is critical to a successful research career. In today’s publishing world, it is essential to understand both theory and the ability to extract significant insights from complex data sets. Therefore, it is crucial to equip yourself even before starting your Ph.D. 

Here are four resources to help you get started:

  • R: This is perhaps the only statistical programming language one needs to know. It is free and comprehensive, and it can perform visualization, machine learning, traditional econometrics, or even write custom algorithms. Even if R is not used often, it is a language that all social scientists must learn. For Ph.D. students with limited resources, learning R is highly recommended. I use vanilla R, but other packages like R-studio can be more user-friendly.
  •  R Studio is a tool that simplifies the use of R and significantly speeds up some use cases. It’s free for academics, so it’s worth trying.
  • Stata is a pricier option, but I’ve found it invaluable since my days as a junior faculty member. It excels in traditional research analysis, producing high-quality tables, utilizing the latest clustering method for standard errors, and identifying bias in panel data. Various versions of Stata are available, with some being less expensive than others. Developing good Stata skills is worthwhile since many co-authors in the field probably have experience with it, particularly economists. Stata SE or MP: This is my preferred software for most of my analysis, especially for economists or those estimating Y = B0 + B1X type equations with concerns about clustering standard errors or endogeneity. Sometimes, R may feel like overkill. I learned to use Stata by working with a collaborator, and with well-written Stata code, you can quickly generate publication-quality tables with just a press of the “do” button.
  • Finally, many young scholars use Python, a must-know tool for any modern social scientist, as datasets have grown larger and machine learning has become more critical. Although I have yet to learn Python, being from a different generation where it was not the dominant platform for empirical analysis and social sciences, it is worth considering its use.

Building a writing system:

Developing a straightforward and rigorous writing process is vital to success. Here are some helpful tools and resources to improve the quality of your writing.

  • LaTeX: Robert Hall, a Stanford economist, wrote in an article about becoming a professional economist, “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.” Learning Latex can make a significant difference in the presentation of your work and how readers receive it. For me, LaTeX makes it easy to break the research article up into discrete components, each of which I can improve and still have a coherent manuscript (e.g., changing tables, adding figures, references, etc., without going through and updating the numbering in the text.) 
  • OverLeaf: If you have learned LaTeX, try using Overleaf, which is like Google Docs for LaTeX and allows you to create visually appealing manuscripts collaboratively. Version 2 is even better as it offers the option to add comments.
  • Grammarly: I tend to type quickly, and occasionally overlook words in my writing. While Grammarly comes with a cost, it is helpful and a worthwhile investment. (Update: ChatGPT might overtake Grammarly as the go-to software for grammar editing. These two technologies excel in different ways, but for now I find the edits from Grammarly a bit more to my taste).) 
  • esttab: I cannot 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, kindly join us in the 21st century. Esttab will help you create beautiful LaTeX tables. The best part is that you can link these tables directly to your latex files using the \include command, so your manuscript will update automatically every time you update your regression. What is even more incredible is that if you link your Overleaf to a Dropbox folder, you do not have to upload your .tex table files. They will automatically update once you produce them in Stata.
  • Google Scholar BibTex export: The GoogleScholar BibTex export feature is quite helpful for my research writing in Latex. It allows me to easily copy and paste a BibTeX bib from Google Scholar, saving me time manually entering the bibliographic information. However, it’s important to be cautious as sometimes specific details may need to be included, or the journal titles may need to be appropriately capitalized. But with a good process, errors can be fixed once, and the tool can be a significant time-saver in the long run.
  • Dragon Dictation: Carpal tunnel syndrome is a real issue. After suffering from severe wrist and hand pain caused by years of typing, poor posture, and various other computer-related problems, I realized the importance of having a good writing process. Now, I have a system where I dictate the first draft of many of my writings instead of typing them. I’ve tried many voice dictation software options. Initially, I used DragonDictate for Mac, but it was discontinued, leaving me without dictation for nearly two years. I tried several alternatives, including the dictation feature built into Mac and Google Docs and Dragon Anywhere on my smartphone, but I still needed to meet my needs entirely. This year, I found a solution. I bought a cheap Hewlett-Packard laptop, installed Dragon Individual Professional, and now I dictate within Sublime, which has been a game-changer.
  • Sublime: Sublime is an excellent text editor with a lot of power, and I use it every day for the small pieces of text work that I have to do.

Leverage the cloud.

  • Dropbox: Ten years ago, I emailed my in-progress manuscripts to myself at the end of each day, so I could continue working on them using my home computer. Today, Dropbox has been a game-changer, reducing the digital waste I create. I now keep a single copy of each file, whether it’s my data, notes, or code, and I can access and work on these files from anywhere. It’s truly amazing.
  • Google Docs: Collaborating on a proposal or responding to a reviewer letter? Google Docs has made these tasks much faster and easier. 
  • Email: It’s also essential to master your email and calendar. A few years back, I checked four email addresses daily: an old Yahoo address, my work email, a new Gmail account, and a Gmail address for subscriptions or junk mail. Consolidating to one email address has given me back around 30 minutes a day.

Learn to delegate:

The most effective professors understand the value of breaking down their work into manageable components and assigning them to others. This strategy not only saves time but also enables them to focus on more critical aspects of their work. While performing tasks such as data cleaning, conducting literature reviews, or developing research project websites might be helpful initially, it’s essential to learn how to delegate these tasks to devote your energy to more innovative uses.

  • Hire someone on Upwork: You can delegate almost any task on Upwork by hiring people. I have used it to get my articles copy-edited, build a citation database on a topic I wanted to learn about and scrape data from a website. You can start with small projects and develop your delegation skills on a platform like this. (Note: I think AI can do a lot of this well now, both custom software you develop and off-the-shelf AI like ChatGPT).
  • Hire someone: Learning to break up your work into modular pieces and delegating tasks that are not worth your time is perhaps the master skill of being a productive academic. Work with someone who excels in delegation to learn how to delegate better. Start by hiring an undergraduate research assistant and assigning them a small project. To manage the work effectively, ask yourself the four questions: what needs to be done, who will do it, how will the work be monitored, and what support will be provided?

Hardware

  • MacBook Pro: Ten years ago, I stopped using PCs and switched to a MacBook Pro as my go-to computer. If you’re a social scientist looking for a reliable computer that can handle most of the software required for statistical analysis and academic writing, the MacBook Air is an excellent entry-level option.
  • RAM is a barrier: Consider investing in a computer with a powerful CPU and ample RAM if it fits your budget. As your data sets grow, it’s worth purchasing additional RAM to enhance your computer’s performance.
  • SSC is critical: I had to purchase a PC since I can no longer run Dragon Dictate on my Mac. Initially, I bought a cheap desktop without an SSD drive, but it was painfully slow. Subsequently, I purchased a small laptop for $350, which had an SSD drive. I only use it for dictation on Sublime and directly save my files in .txt format to Dropbox. This has been a significant improvement as I can dictate instead of typing.
  • Good monitor: You can easily find large and high-quality monitors at affordable prices almost anywhere. Buy two.
  • Keyboard: Get a comfortable keyboard. I use the Microsoft Sculpt Ergonomic Wireless keyboard at work.

Keep learning

The following are two hacks that can help you keep up-to-date with academic literature and find intriguing concepts in popular literature.

Having a good process for reading academic papers is crucial, even though you should spend more time writing than reading. To help you with this, Science magazine [add link] has shared some tips on how to read academic papers effectively. Additionally, you can use software such as NaturalReaders to have your computer read the papers for you.

Audible: Listening to audiobooks is a great way to learn something new while doing other activities, such as going to the gym or commuting. I subscribed to Audible a few years ago, and since then, I have discovered many interesting topics during my drive home. The selection of audiobooks available today is impressive, and you can easily find books related to your research area or a new field you want to explore.

One more piece of advice:

Experiment:

Try little things and see where they take you.