How to Write Project Proposal from A Collection Of

What a Research Proposal Actually Has to Prove

A research proposal has to satisfy two requirements that sit in tension with each other. It has to prove you know a field well enough to say, with authority, that a specific piece of work in it hasn't been done — and it has to propose doing work that, by definition, doesn't exist yet. Mastery of what's known, and a credible claim about what's missing, in the same document.

Your source collection is what satisfies both halves at once. The papers, books, technical reports, datasets, and grey literature you've gathered demonstrate the mastery; the gap in what those sources collectively address, taken together, is the argument for why your proposed work matters.

This guide is for academic researchers, PhD candidates, and graduate students writing:

• Thesis proposals
• Grant proposals (NSF, NIH, foundation grants)
• Research project proposals for lab groups
• Postdoctoral research proposals

The process — literature mapping, gap identification, contribution argument — applies to all of these, with variations in length and emphasis.

---

The Three Tasks Your Source Collection Accomplishes

When you've assembled a source collection for a research area, that collection needs to accomplish three things for a proposal:

1. Demonstrate that the problem exists and matters Your collection should include sources that establish the significance of the research area: evidence that the problem is real, documented, and has measurable consequences. For scientific research, this is usually empirical literature; for humanistic research, it may be primary sources or contextual historical/theoretical work.

2. Map what has already been done Your collection documents the state of the field — who has worked on this question, with what approaches, in what contexts, with what results. This is the literature review's function: showing you know the field so thoroughly that you can identify what's missing.

3. Make the gap argument The gap is the heart of any research proposal: why is there something that your collection of sources, taken together, doesn't address? The gap must be:

• Specific (not "no one has studied this" but "no study has examined X in context Y using method Z")
• Genuine (a real absence in the literature, not a configuration you prefer)
• Significant (if it were addressed, it would matter — to the field, to practice, to policy)
• Addressable (your proposed approach is suited to filling it)

---

Step 1: Map Your Source Collection Systematically

Before writing, map your sources systematically:

SOURCE COLLECTION MAP

Research area: [Topic]
Total sources: [N papers/books/reports]

By scope of coverage:
  Foundational work (established the field/question):
    1. [Author, Year, Title, Journal/Publisher]
       What it established: [...]
  
  Empirical studies (what has been measured/tested):
    1. [Author, Year] — Population/context studied: [...]
       Method: [...]
       Key finding: [...]
  
  Theoretical/conceptual work (frameworks for understanding):
    1. [Author, Year] — Framework proposed: [...]
       How I'll use this: [...]
  
  Applied/practice work (how concepts have been applied):
    1. [Author, Year] — Application context: [...]

By methodology:
  Quantitative studies: [List]
  Qualitative studies: [List]
  Mixed methods: [List]
  Theoretical: [List]
  Review papers: [List]

By context/population studied:
  [Context A]: [Sources studying this context]
  [Context B]: [Sources studying this context]
  [Gap context — studied rarely or not at all]: [...]

By time period:
  [Historical coverage of the literature]
  [Most recent work: Year]
  [Whether the field has changed significantly recently]

---

Step 2: Identify the Gaps

With the source collection mapped, identify where the gaps are. There are five types of research gaps, each of which can justify a proposal:

Population gap: The question has been studied, but not with [specific population]. Studies on X consistently focus on adult populations; the question in adolescent populations is unaddressed.

Context gap: The question has been studied in [context A] but not [context B]. Extensive research in high-income countries; the question in low-resource settings is understudied.

Methodological gap: Existing studies have used [method A] to study this question; [method B] would provide a different form of evidence that could confirm, refute, or extend what [method A] has found.

Conceptual/theoretical gap: The field uses [framework X] to understand this phenomenon; there's a plausible argument that [framework Y] would reveal something different or important that [framework X] misses.

Temporal gap: Research on this question is dated; significant changes in the phenomenon since [year] mean existing findings may not apply to current conditions.

GAP IDENTIFICATION

Gap type: [Population / Context / Methodological / Conceptual / Temporal]
Specific gap: "[Precise description — what hasn't been studied, in what context, with what method]"
Evidence the gap is real (from your source collection):
  "The [X] literature, as represented by [Author A, Year], [Author B, Year], and [Author C, Year], 
  consistently [studies population / uses method / applies framework] but does not [the gap]."
Why this gap matters:
  "Filling this gap would advance [the field / practice / policy] by [specific contribution]"
Your proposed approach addresses this gap by: [Briefly]

---

Step 3: Build the Evidence for Significance

Even a well-identified gap needs a significance argument: why does filling this gap matter? Your source collection should include material that establishes:

The practical significance: What happens in the world because this question is unanswered? Whose practice, policy, or well-being is affected?

The theoretical significance: What would answering this question contribute to the field's conceptual development? Does it test a theoretical prediction? Extend a framework to a new domain? Challenge an assumption?

The methodological significance: If you're proposing a new method or the application of an established method in a new context, why does that matter? What will the method reveal that existing approaches can't?

---

Step 4: Draft the Proposal

RESEARCH PROPOSAL STRUCTURE

Title: [Proposed project title]
Submitted by: [Name] | Date: | Funding agency / committee: [...]

ABSTRACT / EXECUTIVE SUMMARY (200-300 words)
[The problem, the gap, the proposed approach, the expected contribution — compressed]

SIGNIFICANCE AND INNOVATION
Why this research matters:
  "[Evidence from source collection that establishes the significance of the problem]"
  [Author, Year] — what this source establishes
  
The gap:
  "[Specific description of what the literature hasn't addressed]"
  Evidence of the gap: "[Citations showing what has been done — the absence of 
  research on the specific gap follows from what's present]"
  
BACKGROUND / LITERATURE REVIEW (varies by proposal length)
[Organized by theme, NOT by source — each paragraph addresses a theme with relevant sources]
  Theme 1: [...]
    "[Author A (Year)] found that [X]... [Author B (Year)] extended this to show [Y]... 
    The combined implication is [Z]."
  Theme 2: [...]

THE GAP AND THE PROPOSED CONTRIBUTION
[Specific, precise gap statement derived from the literature map]
"The literature addresses [what it covers], but [specific gap] remains unaddressed.
This gap matters because [significance]. This proposal addresses it by [approach]."

PROPOSED RESEARCH QUESTIONS
1. [Primary question — the one the proposal is built around]
2. [Secondary questions — supporting or exploratory]

APPROACH AND METHODS
[What you will do and why this approach is suited to the gap]
[Justified by citations: "This approach has been validated in [comparable context] by [Author, Year]"]

EXPECTED CONTRIBUTIONS
[What the project will add to the field — theoretical, empirical, practical]

TIMELINE
[Realistic phases with milestones]

REFERENCES
[Complete citation list — formatted per appropriate style guide]

---

Before/After Worked Example

Context: A computational social science PhD candidate is writing a grant proposal for an NSF Doctoral Dissertation Research Improvement Grant. Her dissertation focuses on how algorithmic recommendation systems shape political information exposure. Her source collection includes 34 papers.

Source collection map (excerpt):

• Foundational: Eli Pariser (2011) on "filter bubbles"; Sunstein (2017) on group polarization online
• Empirical (large-scale audit studies): Ribeiro et al. (2020) on YouTube recommendation pathways; Guess et al. (2019) on Facebook political content; Huszár et al. (2022) on Twitter political amplification
• Empirical (survey/experimental): Prior (2007) on media choice and polarization; Guess et al. (2021) on misinformation sharing
• Methodological: Bandy (2021) on auditing algorithmic systems; Freelon (2018) on computational social science methods
• Context gaps identified: Most audit studies focus on the US or Western European platforms; few examine the same algorithms' behavior in different regulatory or cultural contexts

Before (vague gap argument, literature as list):

Many researchers have studied recommendation systems and political content. Pariser wrote about filter bubbles. Ribeiro et al. studied YouTube. Guess et al. studied Facebook. This research proposes to extend this work by studying cross-national differences in algorithmic recommendation.

No argument for why the gap matters; the literature is listed, not analyzed; no connection between what the existing literature shows and why the gap is significant.

After (gap argument from systematic source mapping):

---

NSF DDRIG Proposal: Algorithmic Amplification of Political Content Across Regulatory Contexts

SIGNIFICANCE AND INNOVATION

The amplification of political content by recommendation algorithms has become a central concern for democratic theory and information governance. Audit studies have documented that recommendation pathways on YouTube (Ribeiro et al., 2020), political content amplification on Twitter (Huszár et al., 2022), and news feed composition on Facebook (Guess et al., 2019) systematically favor content with higher engagement signals, often at the cost of authoritative or mainstream sources. These findings have driven regulatory responses — most significantly the EU's Digital Services Act (2022), which mandates transparency and accountability from designated platforms.

THE GAP

Existing audit studies share a critical limitation: they are conducted almost entirely in US or Western European contexts using accounts with English-language settings, operating platforms in a single regulatory environment. The same algorithms are deployed globally; the question of whether regulatory context, cultural context, or language modifies algorithmic amplification behavior is almost entirely unstudied in the computational social science literature.

Bandy's (2021) methodological review of algorithmic audit studies notes that "geographic and linguistic diversity remains one of the most significant gaps in the audit literature." This proposal is designed to address that specific gap.

WHY THIS GAP MATTERS

If regulatory interventions (such as the DSA) modify algorithmic behavior — and if that modification is detectable through audit methodology — then comparative cross-national audits before and after regulatory enforcement are essential to evaluating whether regulation works. Without this evidence, regulatory policy is being made without empirical grounding on whether platforms' compliance actually changes what content they amplify. This is a gap at the intersection of regulatory policy and information governance with direct implications for both.

PROPOSED RESEARCH QUESTIONS

1. Does the same recommendation algorithm produce different political content amplification patterns across regulatory jurisdictions (US, EU-subject, non-EU)?

2. Is platform regulatory compliance (post-DSA) associated with measurable changes in political content amplification in EU-regulated contexts versus non-regulated contexts?

APPROACH [Methods: cross-national audit with matched sockpuppet accounts, pre/post DSA enforcement comparison; justified by: Freelon (2018) computational methods; Bandy (2021) audit methodology; prior cross-national work in adjacent areas]

REFERENCES

• Bandy, J. (2021). Problematic Machine Behavior. ACM CSCW.
• Freelon, D. (2018). Computational Research in the Post-API Age. Political Communication, 35(4).
• Guess, A.M., et al. (2019). Less Than You Think. Science Advances.
• Huszár, F., et al. (2022). Algorithmic amplification of politics on Twitter. PNAS.
• Pariser, E. (2011). The Filter Bubble. Penguin Press.
• Ribeiro, M.H., et al. (2020). Auditing Radicalization Pathways on YouTube. ACM FAT.
• Sunstein, C. (2017). #Republic. Princeton University Press.

---

The gap is specific (cross-national comparison, regulatory context); the gap argument connects directly to why it matters (regulatory policy without empirical grounding); the evidence map makes clear that the gap is real (the prior audit literature doesn't do this).

---

Prompts to Reuse

Project Proposal From Collection of Sources

I'm writing a [thesis / grant / research] proposal for [Topic].
Source collection: [N sources across [fields/subfields/methodologies]]

Source collection map:
Foundational/theoretical:
  1. [Author, Year] — [What framework or question this established]
  
Empirical studies (what has been studied):
  1. [Author, Year] — Population: [...] Method: [...] Finding: [...]
  
Context/population coverage:
  Studied: [Contexts/populations well-represented in collection]
  Under-represented: [Contexts/populations with fewer or no studies]

Gap identified:
  Type: [Population / Context / Methodological / Conceptual / Temporal]
  Specific gap: "[Precise description of what the literature hasn't addressed]"
  Evidence from collection that the gap is real: [Which sources collectively show the absence]
  Why the gap matters: [Theoretical/practical/policy significance]

Draft a research proposal that:
1. Opens with significance (why this research area matters — from collection)
2. Maps the literature thematically (NOT source-by-source — by theme, each paragraph covering a theme across sources)
3. Makes the gap argument explicitly and precisely
4. Proposes specific research questions targeted at the gap
5. Outlines an approach suited to filling the gap
6. Argues for the expected contribution (what answering the question adds to the field)

Literature review rules:
- Organize by theme, NOT by source ("Author A found X; Author B found Y" is a list, not a review)
- Each paragraph should synthesize across sources: "Studies consistently show X ([A, Year]; [B, Year]; [C, Year])"
- The gap should follow naturally from the literature map — what the collection covers and what it leaves out
- Citation style: [APA / MLA / Chicago / ACS / etc.]

Do NOT: write a literature review that catalogues sources without synthesizing them; present the gap as a configuration preference rather than a genuine absence in the literature

---

Key Takeaways

1. Map your source collection before writing: systematic mapping by scope, method, and context reveals the gaps that make the proposal's contribution argument — without the map, the gap is at best vague.
2. Five types of research gaps, all valid: population, context, methodological, conceptual, temporal — each can anchor a research proposal; identify which type your gap represents.
3. The gap must be specific and genuine: "no one has studied this in my city" is usually not a genuine research gap; "no studies have examined [specific question] in [specific context] using [specific method]" may be.
4. Organize the literature review by theme, not by source: synthesizing across sources in each paragraph demonstrates mastery; listing sources sequentially does not.
5. The significance argument connects the gap to what would be different if the gap were filled: why would answering this question matter? — field advancement, practice improvement, policy grounding — this is the proposal's persuasive core.

---

Conclusion

A research proposal from a collection of sources is only as strong as the gap argument it makes — and a gap argument is only as strong as the source collection that identifies it. When you've systematically mapped what your collection covers and what it leaves unaddressed, the gap is not something you have to invent; it emerges from the map. The proposal's work is then to argue that the gap is specific, genuine, significant, and addressable — and that your proposed approach is suited to filling it. Your source collection provides both the evidence of mastery and the evidence of absence. The proposal uses both.

See also: Building a Personal Knowledge Base.