When we’re supporting grant proposal development, we commonly field questions about which preliminary data to use and how much constitutes “enough”. While there’s no hard-and-fast rule for incorporating preliminary data, there are some helpful strategies to approach it. It boils down to understanding a few key points: what exactly are “preliminary” data, what is their purpose, and when and how do you use them. 

Perhaps most importantly, it’s about thinking about your preliminary data as a powerful part of your research story. How you choose it, frame it, and weave it into your proposal narrative can meaningfully shift how reviewers experience your entire application. Here, we discuss approaching the incorporation of preliminary data as a strategic asset from the earliest stages of proposal development.

What Are Preliminary Data, and What Are They For?

In the strictest sense, preliminary data are unpublished results. In practice, there is often some wiggle room in what can be used as “preliminary”: data from pilot studies; development of a tool, method, or other resource supporting the proposed work; replication or extension of findings from other groups; or even published findings from indirectly related work that supports, for example, prior successes of the approach or team. 

Their purpose in a grant proposal is threefold. First, they establish credibility: they signal to reviewers that you and your group are equipped to do the proposed work. Second, they reduce perceived risk/establish feasibility: they demonstrate that the central premise of your proposal has at least some grounding in real observations, not just theoretical reasoning, and that what you want to do is possible. Third, and perhaps most important, they build the logical case for your aims or objectives—each piece of data should move the reviewer closer to the conclusion that your proposed research is the natural next step.

Understanding these purposes is what makes it possible to be strategic about which data to include and how to frame them. With that foundation in place, let’s look at how to put it into practice.

Lead with Function, Not Volume

One of the most common mistakes is treating the preliminary data section as a summary of everything the research has produced that’s even tangentially related to the proposal. More data is not more persuasive. A reviewer who has to mentally dig up the connection between five disparate datasets and your central hypothesis may not reach the conclusion you want them to reach.

To approach this more intentionally, ask yourself, “What is the minimum set of findings that makes the most compelling case for this specific project?” Then, consider whether you already have that set of findings or need to go back to data collection. Next, consider what each piece of this “minimum effective data set” is actually doing. In most strong proposals, preliminary data perform one or more of these functions:

• Establishing that the phenomenon you’re studying is real and meaningful
• Demonstrating that your team has the technical capacity to execute the proposed plans
• Showing that your central hypothesis is supported by initial evidence, not proven, but pointed in a compelling direction
• Supporting that your planned approach can produce the intended outcomes
• Closing a gap that a reviewer might otherwise flag as a weakness

Map your available data against those functions. If a piece doesn’t perform any of them, it probably doesn’t belong, even if it represents significant work.

Frame Data to Build Forward Momentum

Preliminary data, whether compiled in a stand-alone section or woven through, are not meant to read like the results section of a research paper. They also aren’t necessarily meant to capture the natural, chronological history of how you got to this point (although sometimes that is a powerful story). That distinction matters more than it might seem.

In a manuscript, results serve a completed story; sometimes authors tell this as a series of chronological steps. In a grant proposal, preliminary data serve a story that hasn’t happened yet. That means the framing needs to be explicitly forward-looking: each data point should make the next step feel necessary and/or possible. Additionally, because these are typically unpublished works, the chronology leading to the current proposal may include a number of steps that aren’t necessary to reviewers’ understanding of where you’re headed.

A useful test: after presenting each part of your preliminary data, can you write a sentence that begins “These findings raise the question of/demonstrate that…” and have the answer point directly to one of your proposed aims or hypotheses? If so, the data are doing their job. If the transition feels forced or requires several inferential steps, revisit the structure or reconsider whether that data belongs.

Order matters too. Think about the logical sequence a reviewer needs to follow to arrive at your central hypothesis:

• Data that establishes the phenomenon should typically come before data that probes the mechanism
• Data that validate your technical approach can often be most effective when placed near the step they directly support, rather than grouped with all other preliminary findings (unless the application instructions require full separation of preliminary data). 

When You Don’t Have the Right Amount

This is a real situation, and it’s worth addressing directly. Early-career investigators, researchers pivoting to a new area, and those proposing genuinely novel methodologies may feel they don’t have enough preliminary data for the proposal they want to write. Other times, PIs might feel that they have to hide some of their preliminary data so they don’t look like they’re too close to getting the outcomes they propose (e.g., “all but published”).

Here are a few ways to deal:

Match the mechanism to your evidence. Some funding mechanisms are explicitly designed for proposals with less established preliminary data: early-stage investigator designations, exploratory or developmental mechanisms, and some internal or center-based grants. Understanding which mechanism fits your stage of evidence matters as much as writing what you have well. 

Let your framing do the heavy lifting. Reviewers generally understand the difference between limited preliminary data because the investigator hasn’t done the work, and limited preliminary data because the central question is genuinely new. Your significance and innovation sections can help bridge that gap. A novel question with a coherent conceptual rationale and a feasible technical approach can make a strong case even without a deep bench of supporting findings. Alternatively, if you feel the need to hide some of your data because you’ve done too much, reconsider your goal(s). For example, perhaps you’ve already nearly answered the question in Aim 1. Rather than keep your data under wraps, consider where that data is actually pointing you to go next, and make that the goal of Aim 1.

Be strategic about what you generate before you submit. Think about which key data would most dramatically lower perceived risk for a reviewer. Focus your pre-submission time there, rather than spreading effort across multiple partial experiments or analyses.

Get feedback. Discuss with collaborators, and/or ask a few peers to review your research objectives and preliminary data. Ask them what they view as the weak points, and determine if/how you can fill those gaps.

Use Data to Get Ahead of Reviewer Concerns

Reviewers who find a concern the applicant already anticipated and addressed respond differently than reviewers who feel they’ve discovered something the applicant missed. The former can improve your score. The latter becomes a weakness in the critique, and some weaknesses can be hard even for favorable reviewers to counter in a panel discussion.

Think through your proposal from the perspective of a skeptical reviewer. Where would a reasonable investigator push back? For each point of potential skepticism, ask whether you have data that speaks to it.

This is also worth keeping in mind during the planning phase. If you know a grant application submission is coming in 18 months, you can run specific pilot studies to strengthen that future application.

Start Planning the Next Application Now

The most important shift for researchers who want to use preliminary data more effectively is a planning one: think of the data you’re generating now as the potential foundation for future applications.

When you’re deciding which direction to push a project in the coming months, add this to the decision tree: which path generates findings that would be most compelling as the foundation for the next proposal?

Grants are infrastructure for the bigger vision, not the vision itself (If that framing is new or worth revisiting, this post on keeping your research vision bigger than any one grant offers a useful companion perspective). The time to think about the data that will make your next application strong is long before you sit down to write it.

What single piece of evidence would most strengthen the application you’re planning to submit next?