How to Write a Successful Level I DHAG Proposal
Let me let you in on a secret: We don’t receive as many high-quality Level I proposals as we’d like to see. We get plenty of terrific Level II and Level III’s, but the Level I’s tend to receive lower scores from our peer reviewers.
Of course, there’s a potential upside to this: Write a compelling Level I proposal and you are in an excellent position to get an award. At the Level II and Level III, it is much harder -- even really terrific proposals often don’t get funded because there simply isn’t enough money. But we try to set aside some funds for each level, so a well-articulated Level I application could be your ticket to an NEH award.
So let’s talk about the Level I proposal and what makes for a good one.
The idea behind a Level I project is that it can be “high risk/high reward.” Put another way, we are looking for interesting, innovative, experimental, new ideas, even if they have a high potential to fail. It’s an opportunity to figure things out so you are better prepared to tackle a big project. Because of the relatively low dollar amount (no more than $50K), we are willing to take on more risk for an idea with lots of potential. By contrast, at the Level II and especially at the Level III, there is a much lower risk tolerance; the peer reviewers expect that you’ve already completed an earlier start-up or prototyping phase and will want you to convince them your project is ready to succeed.
One reason the Level I’s get lower grades, I suspect, can be attributed to the peer review process. Let’s face it: peer review was designed, in part, to mitigate risk. Peer reviewers are looking for flaws to criticize. An early experimental project will always have pitfalls and risk. So it can be hard for peer reviewers to give a Level I the highest grade when they can identify flaw A, flaw B and risk Z.
So how do we address this problem? One way is on our end: in ODH we plan to do a better job explaining to the peer reviewers that they need to place more emphasis on the positive potential of the project and try not to penalize applicants because they haven’t answered every essential question or risk. The very point of the Level I is to give the project team some time and money to experiment, learn, refine, and figure out what those pitfalls are and how to best move their project forward. As with any early-stage research, you try, you fail, you succeed, you document what you did, and the field learns from it.
In addition, we’ve made another change on our end: Level I proposals are now reviewed on a special panel where they only compete against other Level I’s. That way, the peer reviewers aren’t comparing them to the much more mature Level II and III proposals.
What can you do, as an applicant, to prepare a better Level I proposal? For one thing, if your project is truly at the early stage apply for a Level I. We get a fair number of such projects who puff up their budget and add in additional, but less well described, activities to try to fit into the Level II category. Panelists may read these and say “If this was a Level I, I would have given it a higher grade. But they haven’t done the preliminary work to warrant a Level II.” Heed that advice.
When writing up your Level I, here are things to think about:
- talk about the potential of your project.
- talk about what you know.
- talk about what you don’t know.
- be completely upfront about the risks involved.
This is hard to do. I’ve read hundreds and hundreds of applications over the years. Scholars can have trouble acknowledging what they don’t know. But if you acknowledge an issue up front, panelists are less likely to penalize you for it. At the Level I stage, this is key. In a Level I application, you are supposed to have a lot of unanswered questions. You are supposed to encounter risks. But if you don’t acknowledge them, the panelists may pounce.
This might be best expressed in an example:
Hypothetical Level I Grant proposal
Below is an example of a cool project idea, but one that will likely receive a middling grade:
“A team of digital historians plans to address the important topic of how President Franklin Roosevelt’s personal relationship with British Prime Minister Winston Churchill greatly influenced the outcome of World War II. We have access to a remarkable trove of recently digitized letters, reports, and memoranda from both the US and UK governments and we plan to use topic modeling to help us identify key patterns, themes, actors, and places. This Level I grant will help to launch this important project.”
The panelists might say things like “exciting idea” and “this has potential.” But then the negative comments will emerge, ultimately sinking the project:
- The computational linguist on the panel will say “what kind of topic modelling? I’m not sure that will work? What about word vectors instead? Why didn’t they mention that? Have they consulted with any NLP people? This is going to be harder than they think.”
- The librarian on the panel will say “what about dirty OCR? What about handwritten marginalia? I don’t think they understand the condition of that corpus. They didn’t talk about it! Do they have a librarian or archivist on their team? This is going to be harder than they think.”
- The historian on the panel will say “I’m not so confident they will learn anything new. This is well-worn territory and they haven’t even read the materials yet! This is going to be harder than they think.”
Now your project ends up with a mediocre grade and won’t be funded. The panelists are convinced that you are in over your head.
The Same Proposal -- Improved
So how might you have written this Level I proposal to prevent these issues from dragging your grade down? Here is a new version of the proposal. This time, the project director is more upfront about what they don’t know:
“Many historians have written about the personal relationship between President Franklin Roosevelt and Prime Minister Winston Churchill and how it might have influenced strategy in the Pacific theater [cite some examples]. We have recently uncovered a large number of potentially important documents that we think might help us gain some new insight into this relationship [talk about these documents and note the cool stuff you found and how it might give you insight into the Roosevelt/Churchill relationship. Also mention what condition/format they are in and perhaps provide samples of the documents as part of the 10 pages of the appendix that you can use to bolster your application]. These documents are voluminous so we feel it might be useful to use text analysis techniques to help us get a good grasp on the collection. We are requesting support at the Level I stage to do some experiments to see if we might get closer to answering this important question. Here are some things we plan to try:
- We will be working with an archivist, Fred Jones, whose institution has given us permission to use this archive for our work (see his letter of support). It has already been digitized, and Jones thinks the OCR is pretty clean, but we will need time to review it with an eye toward data analysis. This could be a risk -- it might take valuable time to clean it. But we are going to work closely with Jones and if it turns out the data is poor, we still plan to publish documentation on the state of the corpus so that others can learn from us.
- We have a computational linguist on our team, Jane Doe. Doe has experience working with historical materials similar to this document collection [cite examples] and we plan to try a few different text analysis techniques to see if we can zero in on the important historical actors, places, and themes in the corpus (see appendix for our technical work plan) and we will document the success or failure of each. Hopefully, we will gain some insight toward our historical question. Even if we are ultimately unsuccessful, we will have some robust documentation on the pros and cons of these textual analysis techniques for this kind of corpus. We will publish our results, whether they be positive or negative, which may help other researchers doing similar work.
- There is certainly a lot of risk in our project -- risk in the OCR quality, risk in the ability to find a useful data analysis technique, and, of course, risk that we won’t be able to provide insight about our humanities research question. This is why we are requesting a Level I grant. We feel that regardless of what impediments we may encounter, we will make good progress toward our research goals, help to document the condition of an important historical corpus, and document the pros and cons of different data analysis techniques. At the end of the project, we will be in a position, if appropriate, to seek additional funding once we have a better handle on next steps. If all goes well, it might lead to an important new research avenue for military historians, presidential historians and others.”
In the revised example above, the project director is making a good case for their experimental project and its potential. At the same time, they are addressing and acknowledging, in advance, the kinds of things the peer reviewers are likely to bring up.
One last bit of advice: Don’t be afraid to use the full range of Level I funding. Level I projects can be anywhere from $10,000 to $50,000. Most people ask for funds in the top of that range (and that’s fine for many projects). But there may be times when your project only needs a very small amount. At the recent #ODHatTen 10th anniversary meeting, ODH grantee Matt Kirschenbaum noted that he holds the record for the smallest ever ODH grant of $11,708. His project, Approaches to Managing and Collecting Born-Digital Literary Materials for Scholarly Use, requested a small amount of money for some initial planning meetings to help shape a large and important project around the preservation of born-digital documents. These early meetings put the project in a good place to later get other, larger grants.
Good luck with your project! Oh, and two more things we always like to emphasize:
- Send us a draft! (It helps. Really.)
- Want to become an expert at grant writing? Volunteer to be a panelist and you can read a whole bunch of them.