Writing a Research Statement
Writing a research statement was one of the more challenging things that I’ve had to do in my career, because it reflected a transition point between being a PhD student, to thinking more like a PI. As a student, I become used to thinking of research in terms of separate projects: I would motivate an aim, run some experiments, output a research paper, and then move onto the next project (obviously, the process is far, far less straight-forward than that trite summary, but the point is that I had become used to thinking of research as atomized outputs). But as I was looking down the tunnel to the next stage in my career, the research statement was the main document that made me realize that kind of thinking would no longer suffice: instead of simply describing a bunch of separate projects, I now had to convince hiring committees that a) I had a larger research vision, and b) this larger research vision would serve as a wellspring for sustained research, as opposed to a scattering of one-off projects.
I got a lot of mentorship in this process - postdocs, faculty mentors, and my supervisor were all incredibly supportive and helped review multiple iterations of my drafts, providing detailed feedback. However, one thing that struck me about the process was how inequitable it was: I was fortunate to have a lot of intensive one-on-one attention in developing my vision, but not everyone has access to this. Job adverts are remarkably low in detail about what a research statement actually entails, and if you’re unlucky with mentorship, and you don’t know where to look, you might not actually know what a research statement looks like or is supposed to accomplish. By sharing my own research statements and a few high-level tips, I’m hoping to (minorly) level the playing field a little.
Disclaimer: Everything in this article is from my own perspective, and other people have have different experiences with the process. Not everything I write here will work for everyone! I encourage you to seek out multiple perspectives.
Download the first draft and the final version of my research statements here.
Some contextual notes: I’m a computational biologist. Different fields have different standards for research statements. Some fields only require 1-2 pages; other fields expect up to 10 pages. I would check in with mentors in your area to understand field-specific nuances (tip: ask recently hired assistant professors, since they were just through the process and are most “up to date”). I also mostly targeted my application towards research-intensive institutions. If you’re targeting an institution that emphasizes a blend of teaching/research (e.g. R2s, PUAs, SLACs), your research statement should look different.
At the core, the research statement should a) summarize your research interests, b) explain your future plans for research in the position you’re applying for, c) convince the hiring committee that your research is impactful and fundable, and d) you are capable of carrying out independent and sustained research program.
That’s a lot to achieve in one document. As a result, the research statement is a tight balancing act between many elements:
- Establish a research vision. As I explained in the introduction, this a transition point that many PhD students struggle with, since it’s not a muscle that not everyone exercises at the PhD level. What are the larger problems that motivate you? Why are these problems exciting and impactful? At the same time, even if you can pitch an incredibly exciting, impactful, and ambitious research vision, it’s no good if you can’t convince the hiring committee that you can actually do it. So the research statement needs to include just enough details to show that you have some specific plans and the know-how to execute these plans (but not enough that your statement gets lost in the trees and loses conciseness). One common format is to propose three research projects unified by a central vision, as having some specific projects gives a sense of direction and an opportunity to impress technical knowledge.
- The research statement should be a forward facing document. The CV describes that you did in the past; the research statement should be more focused on what you will do in the future. At the same time, convincing research builds upon past experience and doesn’t just materialize out of nowhere. Thus, you should also emphasize your skillset and past successes in a way that doesn’t make them the focal point of your research statement, but supports what you have written about your goals.
- Demonstrate independence and demonstrate that your proposed research is unique. One common pitfall is when candidates propose research that is basically what their advisor(s) do. For example, a research statement that is simply a straight-forward extension of the PhD or postdoc work can give this impression. At the core, the committee wants to know if you can propose new ideas even without the support of an official mentor. But there is also a pragmatic element to this: it’s usually desirable to show that you’re not in direct competition with your previous lab. This is a larger point that should be impressed on by the research statement. Ideally, you’re proposing an impactful research vision: but if the research vision has impact, why aren’t other scientists already working on it? Thus, you should demonstrate that even though you’re working on an important problem, you have some unique angle or skills that differentiate you from other people working on similar problems - including your advisor, but also the larger scientific community at hand.
- Finally, even as you show that your research is ambitious, impactful, and independent, you still need to convince the hiring committee that you can actually execute it. Usually, this is a point that has to executed from multiple fronts. I alluded to some of these above: referencing specific projects and past experience. Another point that (some) researchers choose to incorporate is pointing to collaborators. Not only does this help in proposing a more ambitious research agenda by pointing to skills or resources that you alone may not be able to cover, but it shows that you are connected to a larger community as an academic (and presumably, this community cares about your research.) As with the other points, this needs to be executed in moderation: too much, and you’ll raise questions about how independent your research program will be.
My research statement as an example
The advice I gave above is rather abstract - it can be hard to see how to actually execute some of these points. To make this more concrete, I encourage you to look at how my own research statement evolved from the first draft to the final draft, with again, ample help from my supervisor and other mentors.
Here are some things to look out for:
- I improved the scope and impact of my research vision: This is especially apparent in how I introduce my research statement. In the first draft, I am narrowly focused on machine learning for microscopy images. This was mostly what I did in my PhD, so it felt “safe” when I was first drafting the statement. In my final draft, I propose a more ambitious vision of applying deep learning in biology in general. Note that none of the research projects that I am proposing have drastically changed. Instead, I’ve just framed these projects to have more widespread impact (than just the limited subset of biologists who work with high-throughput image screens alone.)
- I inverted the chronological order of each project section into a “supporting experience” section: My first draft tries to tell my story in a chronological order, but I ultimately decided this detracted from the vision of the research statement. In the final draft, I separate my past work, and instead put it at the bottom of each section as evidence that I can carry out the research. Hopefully you can see that again, much of the content has not changed. Instead, what this new structure does is alter the motivation underlying each project: instead of being the rationale for past work, the motivation for each section now comes off as more arching subgoals (e.g. “make computational analysis the first option biologist use to understand their datasets”, “enhance the depth of insight across many experiments in many labs, not just big dataset generated by a few centralized groups”) that I will address in my research program, and the past work is shown as evidence that I’ve made strides towards these motivations.
- I consolidated the background into a single section: This helps centralize what I view my core strength and differentiating angle is. I go to pains to emphasize that this is a recent angle (i.e. “I pioneered one of the first self-supervised methods for biology”), and to emphasize that even as the strategy becomes popular, I still take a different perspective than the field at large ("[Other work] repurpose methods from mainstream data domains. I design methods athat incorporate biological assumptions.") Emphasizing that these are newer ideas that I established helps answer questions of independence and differentiation. It also helps set up the rest of the proposal, which is proposing a significant amount of breadth: typically, it would be challenging for a researcher to work on so many different modalities and domains, but I suggest that this is feasible through naming specific collaborations, and emphasizing that I still have a “core” background uniting work in all of these different domains, which is my methodological expertise.
One final tip is that everyone’s research statement will look different. Some of the strategies I executed in my research statement make sense for me, but would not make sense for everyone: this includes not just content, but even the format and structure of the statement. I suspect this is why so little “official” guidance or templates exist for research statements - everyone has a different path through their academic training, so everyone’s story will look different. In other words, the flexibility of the research statement makes it daunting, but is also a strength you can leverage: what would be the best way to tell your story?