What does a typical pre-doc application entail?

1. A Common Application: You may be required to fill out an online application form, or send a CV, Cover letter, and writing and coding samples via email. For example, The Tobin Centre for Economic Policy at Yale University requires the following through a common application form for predoc fellowships: personal and educational background information, a ranking of prospective faculty supervisors, and a single PDF containing a cover letter that includes the contact information for 2-3 references, a resume/CV, a diversity statement (optional), unofficial transcript(s), a writing sample, and a coding sample.
2. Data Exercises or Skill Tests: At this stage, you will be sent a dataset along with instructions and questions and will be asked to complete the exercise in 48 hours (usually). You might have the option to solve the exercise in any statistical software, or be required to only use a particular one, depending on the application requirements.
3. Interviews: The interview may or may not be technical in nature. Depending on the positon, there might also be two interviews - a general interview about your background and interests, and a more technical interview that tests your knowledge about statistical or economic concepts relevant to the position.

Advice from Predocs and RAs

1) Is prior research experience required while applying for predocs? If yes, what advice would you give to those currently in their undergrad programs for building the skills and experience required?

I had prior research experience (both as an RA and doing my own research), but I don't know if it's necessary. For me, it was relatively easy to find short-term research assistant positions through my university at undergrad/Masters level. Sometimes these are only a couple of days doing something like digitalising pdfs, but it's really useful to have this kind of experience. The other thing you can do is to produce an optional dissertation as part of your undergraduate, if you don't have a compulsory one. I swapped out one of my final year modules for writing an Economics dissertation. This was really useful as I got to know more about the research experience and got to know faculty better.

2) Besides looking at postings on @econ_ra, how did you look for predoc and RA positions? Were there particular keywords on twitter, or websites and resources you found helpful?

I found predoc.org quite useful and I also checked the J-PAL job listings. Early on, I also made a list of the people that I would ideally like to be a predoc with and checked their websites regularly. In general though, there was very little that I found that wasn't listed on @econ_ra. Often third party websites don't take down job postings, so it's worth checking how old a job announcement is, and emailing to check if it's still open.

3) What advice would you give for clearing the data round? What should applicants focus on while answering relatively vague questions or questions that rely on their interpretation of results?

I did two of these and they both took quite a long time, so my impression was that a large part of this stage was basically signalling - as in, potentially a lot of marginal candidates do not think it is worth doing the problem set, so they screen themselves out at this point. I'm not sure if that is true, but it was my impression. For vague questions, I would recommend thinking about the two or three possible interpretations and answering them separately.

4) What types of questions were you asked in your interview? Are there books or resources applicants can refer to in order to prepare for the interview (in case technical questions are asked).

I only got asked one question in my interview, which was "tell me about you". This really threw me as I had prepared for shorter more specific questions, and I didn't have a succinct spiel about myself planned.

1) Is prior research experience required while applying for predocs? If yes, what advice would you give to those currently in their undergrad programs for building the skills and experience required?

It is implicitly required. Most of the postings will not list it as a requirement, but most competitive application require you to have done some sort of research in undergrad. This can be in the form of summer work. However, the best is with a professor since you have already done the kind of work you are applying for (even if it is in a different field). Advise would be to either do there own independent research work or find a professor at college who is looking for an RA. If both of these are not possible, find an internship that has some sort of research - you have to show that you understand what it takes to do at least basic research. 

2) Besides looking at postings on @econ_ra, how did you look for predoc and RA positions? Were there particular keywords on twitter, or websites and resources you found helpful?

I used a lot of google jobs. The words I used are: Research Assistant, Research Associate, RA, Economics Research Associate, Policy Research Associate. I also used Linkedin extensively - the job profiles are generally more detailed and you can get access to the people in the organisation so helps get in touch in case you want to use that approach. I looked directly at organisations I knew of (JPAL, IPA, etc) and looked at jobs on their websites. I also reached out directly to professors to ask if they wanted to hire an RA (this method is really a shot in the dark so probably the last resort or a longer way to get a job)

3) What advice would you give for clearing the data round? What should applicants focus on while answering relatively vague questions or questions that rely on their interpretation of results?

I would take this very seriously. The writing style is very important. Most organisations are starting to look for good writers so pay attention to how you are answering the questions. 

Honestly - I would focus on both. Clearing this round is key...most people get stuck here 

That choice also depends on the kind of RA job you are applying for. If you are doing to be mostly looking at data after it has been collected than focus on the latter, but if that's not the case then you also have to pay attention to the former.

4) What types of questions were you asked in your interview? Are there books or resources applicants can refer to in order to prepare for the interview (in case technical questions are asked).

There are no books as such that I looked at. It is a newer field so things are not that formal. 

Question topics: 

1. Previous Research work (focus is on this)
2. Independent research if done
3. How I spent my summers (generally important because they want to see if you have a history of being interested in research or their area of expertise)
4. Why am I applying for this job
5. Future plans  

You can also guide the conversation and tell them about things you want to highlight. Find opportunities to move the conversation in the way that you get to talk about experiences that are relevant...don't let the resume alone dictate the conversation. 

Talk in depth about actual data work done - as in if you have dealt with big data, regression experience etc

5) Do you have any other advice not covered in the questions above?

It's a slow process so have patience. It may seem like there aren't a lot of opportunities, but that's not true. There are a lot of ways to get research experience in this field so you just have to stick at it for some time. It took me 3 years to get the kind of RA role I was looking for :) 

• While applying, the CV should highlight that particular skill set which is most aligned to the job description mentioned in the advertisement
• I was given a dataset and codebook of the data. I was asked to formulate a research question, analyse the data to address the research question and write a short report on it. I was given 48 hours to complete this. I had to send the do file along with the write up. This was followed by an interview, where I was asked questions related to what methodology I used."

• Keep all code files well organised and in loops whenever possible for the coding tests. Personally, I like creating sections in the code but basically just ensure that it's super neat and runs smoothly.
• Starting applications early goes a long way. It's highly likely that you will not hear back from most places you apply to.
• Speak to current or past RAs about their experience before taking on the position.
• In my experience, follow up emails were very helpful. I always got a second round interview from places I sent follow up emails to.
• ALL data tests require at least one publication-ready table or figure to be created. This is an easy thing to prepare for in advance so you don't waste time on formatting options with the code during the timed test.

• Creating a CV using LaTeX: You can refer to Michael's cls file as well as a resume and CV created using it here
• 75% of Michael's predoc (at the Federal Reserve) involves coding. Here, and here, are examples of the coding he does on his job
• Other advice, " It’ll all be alright. You can only plan so much. Consider if the pre-doc is worth the time and salary loss in opportunity cost. If you can afford the flexibility, that’s great! If not, consider a regular job."

Advice from the Hiring-Side

In my experience, some prior data-intensive research exposure is essential. This could be in the form of a thesis and/or through experience working with a professor on a project. Specifically, I would say it is essential for candidates to have experience working with STATA; have taken coursework in econometrics; and have some RA/thesis experience. Some professors are also particular about grades in undergraduate/graduate coursework.

My advice for those currently in school who are certain they would like to do a PhD -- focus on your grades and on taking courses in math and econometrics. During one of your summers, try to work with a professor as an RA (in the field or in a lab). If you're unsure about a PhD, speak to people who are considering or have done a PhD in the field you are considering studying. This could be professors, current PhD students, or full-time RAs preparing their PhD applications.

2) What advice would you give for clearing the data round? What are some of the mistakes applicants commonly make that should be avoided? What should applicants focus on while answering relatively vague questions or questions that rely on their interpretation of results?

Be curious about your data and spend time understanding it. Is each row a person, a school, a child? Is it unique? What information do we have about these people/schools/children? Is there a lot of missing data? Is there anything else about the data that looks suspicious to you?

Real-world data is messy, and we look for people that have clean, correct, organized, efficient, and replicable code with lots of helpful comments to explain decisions that were made. We also want people to have a deep understanding of the data they are given and its limitations. My advice would be to take the extra step of going over and above what is asked for in the prompt.

As for questions that are vague or require interpretation, be logical and use this as an opportunity to shine. If you're running out of time, talk about what you might have done if you had more time and/or what additional information you would need to answer some of the questions you have about the data/topic. My general advice would be to focus on my advice in the previous two paragraphs and on understanding the data. As long as you're logical and clear and understand your data, you should be able to clear this round.

3) What types of questions are typically asked in the interview? Are there books or resources applicants can refer to in order to prepare for the interview (in case technical questions are asked).

Interviews typically include the usual questions on the candidates' background and interest in the position. You can also expect a question on your familiarity with the research the lab/professor does.

Technical interviews for pre-doc positions vary a lot in terms of their difficulty and the types of questions asked. My advice would be to have a good understanding of evaluation methods. My favorite resources on evaluation methods are:

1. The Mystery of Vanishing Benefits: an introduction to impact evaluation (link to paper)

2. Running Randomized Evaluations (link to the e-book) 

3. Table summarizing different evaluation methods (link to case studies -- See Module 2.2, Why Randomize, Page 8)

I would also encourage revising basic statistics (measures of central tendency, standard deviations, confidence intervals, etc) for the technical round.

4) Do you have any other advice from the hiring side not covered in the questions above?

Pre-doc positions can be extremely competitive. Apply widely and don't get disheartened if you don't make it for a role. To find pre-doc opportunities, keep an eye-out for the JPAL annual recruitment drive, NBER pre-doc listings, and the @econ_ra Twitter handle. Best of luck!

• Make it well commented but don't make the comments too long
• Make sure everything is labeled clearly
• If you can't solve something but have been trying, make sure you lay out the steps and where you think you're going wrong/what your limitations are
• Make the files as easy to read and download as possible
• Avoid using excel if possible. Some people at my organization frowned upon over reliance on excel.

Resources for various components of the application