Midd Predoc Guide

Created on October 01, 2023

2023

NB: minor revisions on language (e.g. typos) since first published.

1. Introduction

This guide is for Middlebury College students interested in “top” predoctoral research positions in the social sciences. Like other highly competitive jobs, these positions are affected by substantial information asymmetry. My goal is to reduce that barrier for Middlebury students relative to applicants from top research universities.

Everything here reflects my personal views. Depending on your background and goals, some suggestions may not be optimal for you. This guide is written to maximize your chances of landing the most competitive predoc roles, so parts of it may feel excessive if you are already a standout candidate or if you are targeting less competitive programs. Before acting on any advice here, I strongly recommend discussing it with faculty mentors who know your circumstances.

“Top” Predoc Programs?

Definition: “Top” is subjective; the right choice depends on your objective function and constraints. Here, I focus on pipeline predoc programs to top-7 economics departments. Some of these programs have median placements in top-7 Economics PhD programs. They are highly competitive and require substantial preparation, strong skills, and compelling observable signals.

Examples of top predoc programs include, but are not limited to:

  • MIT Blueprint Lab
  • MIT: Amy Finkelstein
  • MIT J-PAL
  • Dartmouth: Heidi Williams
  • Harvard: Melissa Dell, Claudia Goldin, Jesse Shapiro, David Yang
  • Harvard Opportunity Insights (OI)
  • Federal Reserve New York City
  • Microsoft Research
  • Stanford SIEPR: Matthew Gentzkow
  • Stanford GSB: Susan Athey, Guido Imbens
  • Princeton Industrial Relations Section
  • Chicago Booth: Pascal Noel
  • Chicago BFI: Magne Mogstad, Alexander Torgovitsky

Many other, less selective predocs also place researchers into top economics PhD programs, though often less frequently.

Why a “top” predoc program may NOT be for you

  • If you are not yet committed to pursuing an Economics PhD, you may prefer more general-purpose programs.
    • Examples include research roles at think tanks and law schools, economic consulting, quantitative roles at banks or hedge funds, and data science roles in tech or retail.
    • In these paths, people who later choose not to pursue an Economics PhD can still move into strong graduate programs, think tanks, law school, data science, consulting, or other industry jobs.
    • By contrast, some top predoc programs are designed primarily to place candidates into Economics PhDs and may be weaker springboards to other careers.
  • A “top” predoc does not necessarily maximize your learning or your individual treatment effect, given both selection bias and treatment-effect heterogeneity.
  • A “top” Economics program may matter less to your long-run life outcomes than you think.
    • Many do great work without going to a top school
      • John List went to U Wisconsin for undergrad and U of Wyoming for PhD
      • My mentor at Middlebury, Akhil Rao, one of the smartest people I know, went to CU Boulder for PhD
    • By contrast, many graduates of MIT/Harvard PhD programs leave academia or even leave the PhD.
    • Ultimately, it depends on your goals.

I find the readings below very helpful. Starting in Section 3, I rarely repeat points already covered there, so please read them first.

2.1 General-purpose Advice

  • Econ RA Guide
  • Susan Athey, Advice for Applying to Grad School in Economics
  • Jesse Shapiro, Notes on Applying for a PhD in Economics

2.2 On Writing Thesis / Research Paper

  • Plamen Nikolov, Writing Tips for Economics Research Papers
    • This is the thesis guide for Harvard Economics thesis students
  • Jesse Shapiro, Four Steps to an Applied Micro Paper
  • Jesse Shapiro, How to Give an Applied Micro Talk: Unauthoritative Notes
  • Deirdre McCloskey, Economical Writing. I do not think reading the full text is necessary; a good summary should suffice.
  • Claudia Goldin and Lawrence Katz, The Ten Most Important Rules of Writing Your Job Market Paper

2.3 On Coding / Data Work

  • Matthew Gentzkow and Jesse Shapiro, Code and Data for the Social Sciences: A Practitioner’s Guide
  • Matthew Gentzkow and Jesse Shapiro, Internal RA Manual
  • Jonathan Schwabish, An Economist’s Guide to Visualizing Data

3. Advice - Before Applying

3.1 Courses

What to take: The best course list depends on your field interests. Start with this list, then consult faculty mentors.

  • Empirical Research Methods
  • Applied Econometrics
  • Regression Analysis / Regression
  • Advanced Real Analysis Seminar / Real Analysis
  • Advanced Macro Theory / Macro Theory
  • Micro Theory
  • Advanced Probability Seminar / Measure Theory
  • Probability
  • Statistical Inference
  • Statistical Learning / Machine Learning
  • Intro to Computing
  • Data Structures
  • Elementary Topology
  • Linear Algebra
  • Multi-variable Calculus
  • Differential Equations
  • An Economics CW course
  • Any other coursework in Math, Economics, Statistics, and Computer Science

What about field courses in economics? They can be valuable for learning and undergraduate research, but they are usually less valued by PIs at top predoc programs, except when directly relevant to a specific lab (e.g., a health economics course for Amy Finkelstein’s lab). Field courses with more math and more exposure to academic papers are often more helpful than those with less.

Sequence:

  • Take the most challenging and “useful” or “practical” coursework as early as possible to improve your access to research opportunities.
    • Personal experience: In my sophomore year, a PI at Brookings explicitly noted in my interview that I was already taking real analysis, and that may have influenced the offer. (For context, many Harvard students take real analysis in their first math course; that is how those classes are structured.)
  • Some courses are offered only once every few years. Try to take them as early as possible by completing prerequisites early or speaking with the instructor.

3.2 Gain Research Experience

To:

  • build your CV
  • learn and practice how (best) to do research
  • develop relationships with potential future letter-writers
  • gain (potentially long-term) mentorship
  • learn about different fields of social science
  • develop ideas for your own thesis / independent research

As in paths like IBD and consulting, a common ladder is part-time RA → full-time summer RA → predoc, with project scale, your contributions, and PI name recognition typically increasing at each step.

3.2.1 Part-time / Semester RA @Midd

  • Part-time semester RA positions at Middlebury usually require little or no prior research experience.
    • Core economics courses and coding skills in Stata and R make you competitive
  • Because of search costs, not all semester RA positions are posted publicly.
    • So reach out to faculty.
  • At this stage, prioritize projects with the strongest learning opportunities and good PIs.
    • Your interest in the field / topic should be secondary
  • You may also consider RA positions in the statistics or political science department

3.2.2 Part-time RA Outside Middlebury

  • Some PIs run large labs with many part-time RAs handling manual or repetitive work.
    • These PIs are often prominent researchers with many projects and high labor demand.
    • Examples: Melissa Dell and David Yang.
  • These positions usually require stronger skills than an average part-time RA role at Middlebury.
  • You should expect no pay
  • How much you gain depends on your contribution.
    • Part-time RAs usually report directly to predocs, and many make limited contributions.
    • Personal experience: I know someone who became a PI’s part-time RA through cold email and eventually co-authored a paper with the PI as sole first author.
    • Work for someone whose research genuinely excites you.

How to find opportunities:

  • Visit the PIs’ personal website
    • Some PIs explicitly state on their websites that they hire part-time RAs
  • Check PI Twitter accounts; some post opportunities there
  • Check centralized postings such as predoc.org and Econ RA Twitter
  • Cold email PIs whose work interests you
    • Write a sincere, concise one-paragraph email explaining your interest and highlighting your skills.
      • Treat this as a cover letter where you have about 10 seconds to make the reader want to learn more.
    • Attach your CV, transcripts, writing sample, and code samples
    • Expect most PIs not to respond. Thank those who decline. Keep cold-emailing, but do not overdo it; it can reflect poorly if you accept or hold too many parallel opportunities and then turn most down.
  • Use Middlebury’s alumni network
    • A few alumni in the past two decades are now professors at top research institutions, and more alumni are current PhD students
    • Reach out to them and ask whether they have, or can refer you to, opportunities.

3.2.3 Summer RA @Midd

MiddEcon hires summer RAs around March or April.

Apply as soon as possible with your CV, transcripts, writing sample, and code samples.

3.2.4 Summer RA outside Midd

  • There are many summer RA opportunities / research fellowships / bootcamps outside Middlebury
  • These can be quite competitive, with slight variation across programs
    • A good rule of thumb: if any past Middlebury students have done it, then perhaps you can too
  • Formal summer positions are mostly paid
  • Applications open around late-autumn till spring

How can you find these opportunities? You should search online, but two easier channels are talking with economics faculty (who know both their own students and peers’ students) and speaking with alumni. A few conversations can quickly produce a long list of positions that Middlebury students can realistically target.

There are also informal summer research positions available through cold-emailing PIs. These are often unpaid, but Middlebury’s CCI may provide funding if you frame the role as a research internship.

  • You may get a summer 0555 credit for external summer research that results in concrete deliverables (a paper or at least a slide deck)
    • I may be the only person in recent history to have used this option, and Jeffrey Carpenter helped me. Contact him if you are interested.

3.3 Read (Lots of) Papers

A good researcher is creative, but where does creativity come from? You may be a genius and simply create from the void, but that is hard for me. Instead, I gain creativity from reading lots of papers, from which I learnt about new data sources, methodology, common arguments, theory concepts, ways to measure / quantify conceptual variables, and questions that were / are being asked by economists.

Reading is often hard at first, but it becomes easier with repetition as familiar concepts accumulate and new ideas become easier to map onto known ones. This is a long-term effort.

Here are some opportunities:

  • Follow NBER, IZA, and economics journals on LinkedIn, Twitter, and Instagram; shape your feed around interesting papers
  • Sign up for research summary emails from institutions like J-PAL or HBS
  • MiddEcon invites external speakers to present papers and talk with students. Sign up for a lunch if possible and attend their presentations
  • Join MiddEcon’s student advisory committee (SAC) for free lunch chat with candidates visiting to interview at MiddEcon
  • Ask MiddEcon faculty for lunch, offered for free at MiddDining
  • External opportunities for talks. For example, Harvard has a seminar series, most of which is streamed on Zoom
  • Econimate on YouTube

If you are new to economics research, summary books may be more useful than jumping directly into frontier papers. Books are usually more structured and systematic, so they provide a fuller view of a field, while individual papers often focus narrowly. Books are also typically better suited to less technical audiences. Examples include:

  • S. Dubner and S. Levitt, Freakonomics
  • D. Acemoglu and J. Robinson, Why Nations Fail
  • D. Acemoglu and J. Robinson, The Narrow Corridor
  • D. Acemoglu and S. Johnson, Power and Progress
  • D. Acemoglu and D. Autor, Lectures in Labor Economics
  • C. Goldin, The Race between Education and Technology
  • A. Banerjee and E. Duflo, Poor Economics
  • A. Banerjee and E. Duflo, Good Economics for Hard Times
  • A. Case and A. Deaton, Deaths of Despair and the Future of Capitalism
  • J. Angrist and S. Pishcke, Mastering Metrics
  • D. Kahneman, Thinking, Fast and Slow

3.4 Networking

This is one of the most important skills in your life. You learn about the job, the application processes (including how to write good CV / cover letters), and get some insider information (e.g. which PIs to avoid, what’s in the job interview, what are good opportunities, past placement, etc.)

Learn networking, cold-calling, and cold-emailing. Connect with alumni and others on LinkedIn, and ask about their jobs or PhD programs. Follow people on Twitter. Learn to leverage existing connections, including professors, to build new ones.

If this is difficult to do alone, ask CCI for help or join MiddCORE to practice.

3.5 Transfer / Study Away

If circumstances allow, consider transferring to, studying at, or exchanging with institutions that have top economics departments or rigorous training in Math, Economics, Statistics, and Computer Science, including but not limited to:

  • Any US institutions with a top Economics department
  • LSE / Oxbridge
  • Stockholm School of Economics
  • Swarthmore College (Middlebury’s official study away program)

These institutions often offer more advanced coursework (e.g., graduate classes, advanced real analysis, proof-heavy math) and have grading signals that may be more recognizable.

3.6 Feb Yourself

Most predocs start in early summer, but some PIs prefer candidates who can start earlier. Being a Feb does not hurt and can sometimes be an advantage, so do it if possible, all else equal.

3.7 Take the GRE

Many PIs value both competitiveness in grad school applications and qualitative & quantitative skills. A pretty GRE score helps with your application in both dimensions - and you will have to take it anyways before applying to grad school. (I have, however, encoutered no predocs that require GRE scores to apply - this is totally optional.)

3.8 Have Your Thesis / Writing Sample Ready

Most top predoc programs require, or at least accept (and value), a writing sample. Officially, it helps evaluate writing quality, but in practice it can also signal creativity, research potential, data reasoning, quantitative skill, judgment, organization, graphing ability, LaTeX proficiency, research interests, and depth of reading. With this in mind, have your thesis ready: written, reviewed by faculty and peers, revised, and proofread at least three times on different days.

4. Advice - Applying

4.1 Apply Early

Top positions open early, and many recruit on a rolling basis. For rolling searches, applying early offers a meaningful advantage because reviewing data tasks and conducting interviews is costly. Have the following ready by late August (written, reviewed by faculty/alumni/peers, revised, and proofread at least three times on different days):

  • CV
  • Annotated transcripts
  • Generic cover letter
  • Writing sample / thesis
  • A coding sample for each language you know
  • 2~3 recommendation letters
    • Some recommenders will wait until the last minute to submit
      • Personal experience: I know someone who reached (or passed) a final interview but lost the job because a letter arrived too late
    • Around mid-late summer, send over to your recommenders:
      • all application materials
      • a list of positions you are applying to, the JDs, how and by when to submit letters
      • a list of your highlights, most significant research contributions, and examples of your strongest qualities as a researcher
    • After that, check in regularly (e.g., weekly) to answer questions and make sure everything stays on track

4.2 Have 0 Typos

This is common advice for good reason. Typos can signal weak attention to detail, which many PIs care about. Poor attention to detail can also lead to avoidable research and coding mistakes that take days or weeks to fix, while reducing credibility. In your written materials (cover letter, CV, and writing sample), reduce typos as much as possible, but do not obsess.

Personal experience: I still received an offer from a PI who read my thesis, even though it had a major typo in the abstract.

4.3 Make Slides for Your Thesis

  • You will almost always be asked about your thesis or independent research during interviews. A slide deck helps you present clearly and professionally. Consider using Beamer.
  • If you have zero independent research projects, make slides for the RA project to which you contributed most

4.4 Craft HR-friendly Cover Letters

Depending on the position, your cover letter may be first screened by an HR with little quantitative background. The HR may have so many applications to review that they look for excuses to throw away yours. Don’t give them any excuse to. Don’t throw economics jargon everywhere - nobody will eventually read your jargons if the HR trashes your application. Be sincere and show your interest in the position. Ask your roommate majoring in studio arts to read it, and re-write until they feel compelled to hire / interview you.

Personal experience: In my junior year, I applied to many summer RA positions with an insincere, poorly written, jargon-heavy cover letter. I received one offer and zero other interviews, despite advanced classes at Harvard, strong technical skills, and substantial RA experience.

Bottom line: even if you could tell the counterfactual of every factual in the world, nobody cares unless the HR gives you a pass.

4.5 Annotate Your Transcripts

Middlebury’s transcript is not always easy to parse, and some top predocs (e.g., Amy Finkelstein’s) look for course details not shown on the standard transcript. Add annotations such as class year by semester, textbooks for key courses, and alternative course titles.

4.6 Connect To Other Applicants

Strong candidates often receive multiple offers. When they turn one down, some refer the opportunity to people in their network as a courtesy to the PI.

Personal experience: I know someone who was referred by a friend who declined a top predoc offer and then received that offer.

4.7 Networking

Before applying to a predoc position, talk with someone who works or has worked there. A short conversation can reveal a lot about the role and may save you from applying to a poor fit.