A research project roadmap: from question to manuscript

Stetleaf Academy

Reviewed by Alex Juusela, MD, MPH, FACOG, Maternal-Fetal Medicine Physician · Last reviewed 2026-11-05

A research project is easier to report when its question, methods and decisions remain traceable from the beginning. This roadmap offers an educational sequence from an initial question to a submission-ready manuscript. It is a planning aid, not a universal protocol, an ethics determination or a promise of publication. Different designs need different expertise and approval pathways; adapt the sequence with the people responsible for your project.

Why it matters

Reporting guidance helps identify the information that a reader will later need. The EQUATOR Network promotes the use of reporting guidelines and provides resources for health research reporting [1]. For randomized-trial protocols, SPIRIT 2025 supplies a specific framework for documenting the planned work [2]. Other designs require their own planning and reporting approach.

Incomplete or unusable reporting can waste the effort invested in research [3]. Writing is therefore not simply a final formatting step. Preserving definitions, dated plans and reasons for changes makes later reporting more straightforward. That does not turn a reporting checklist into a design manual or ensure a project is feasible. It helps keep the eventual manuscript connected to what was planned and done.

Key ideas

Stage 1: turn an interest into a question

Write down the population, comparison or phenomenon, outcome and setting as applicable. Identify what existing evidence does not answer and why the question matters. Separate a broad topic from a question that the available methods and resources can address. Keep a short decision record when the question changes, so the team understands which version is current.

A preliminary literature search can reveal overlapping work and help refine terminology. Preserve the search notes and the sources you actually read. Do not build the rationale from citation titles alone or treat a convenient dataset as a sufficient explanation of the question.

Stage 2: match design and reporting guidance

Choose the design with appropriate methodological advice, then identify its reporting framework. The guideline follows the design, not the clinical topic. Use the following starting points, checking extensions and versions for the actual project.

Project output	Reporting starting point
Randomized-trial protocol	SPIRIT 2025 [2]
Cohort, case-control or cross-sectional report	STROBE [4]
Systematic review	PRISMA 2020 [5]
Randomized-trial results report	CONSORT 2025 [6]

This is a short navigation table, not an exhaustive catalogue. Diagnostic, prediction, qualitative and case-report work have other guidance. See How to choose the right reporting guideline rather than forcing every project into one of these four rows.

Stage 3: document methods and responsibilities

Define eligibility, measurements, outcomes and the planned analysis at a level that supports the work. Assign responsibility for data collection, data management, analysis and documentation. Get suitable methodological input before committing to choices that cannot be repaired later. Record how missing information and deviations will be handled, while recognizing that the plan may require justified amendments.

For a trial, connect the protocol to the statistical analysis plan and registration record as applicable [2]. For any design, preserve dated versions. A folder of files named final, final-new and final-really does not show which plan governed the work.

Stage 4: resolve oversight and access before starting

Determine the relevant ethics, governance, data-access and registration requirements with the responsible bodies. This roadmap does not decide whether approval, consent or an exemption is required. Record the determination and its scope, and do not assume that access to records is permission to use or share them for research.

Keep private material in the approved environment with access limited to the authorized team. Plan what may later be shared and under which restrictions. A general intention to share data should not override consent, governance requirements or protections for individuals.

Stage 5: collect and manage a traceable record

Use defined variable names, a data dictionary and a record of transformations. Preserve the original data according to the applicable governance plan; conduct cleaning and analysis through documented steps. Log amendments, unexpected issues and decisions that affect interpretation. Resolve unclear entries with an explicit process rather than silently changing them.

The practical aim is that another authorized team member can understand how the analysis dataset was produced. It is not necessary to expose private data in order to describe that process clearly in the manuscript.

Stage 6: analyze and separate plans from discoveries

Work from the dated analysis plan and document departures. Distinguish prespecified analyses from exploratory work. Preserve outputs and code with enough context to identify the dataset and method used. Ask for appropriate statistical review of estimates, uncertainty, missing-data handling and interpretation rather than expecting a writing checklist to settle those questions.

Unexpected results may generate useful new questions. Label those questions honestly instead of rewriting the original objective as though the new analysis had always been intended.

Stage 7: write, check and prepare for submission

Draft the methods from the study record and the results from checked outputs. Keep tables, figures, abstract and main text consistent. Use the reporting checklist to point to locations in the manuscript, explaining genuine non-applicability rather than merely ticking boxes. The discussion should distinguish findings from interpretation and address the limitations of the work.

Read the target journal's current instructions and resolve required statements with the team. Confirm contributor roles and approvals before submission. The choice to submit, post or share outside the project is a separate authorized action, not an automatic consequence of completing this roadmap.

Worked example

Illustrative example (invented). Not a real study.

An invented team wants to examine whether longer travel distance is associated with attendance at a follow-up clinic.

Question. The team defines attendance and its assessment window before extracting records.
Design. Because nobody assigns travel distance, it plans an observational cohort and identifies STROBE as the reporting starting point.
Oversight. The team seeks the required governance determination and data-access authorization before using records. This example does not assume the outcome of that review.
Documentation. The plan names the exposure definition, potential confounders, missing-data approach and people responsible for each task.
Reporting. Later, the team explains a measurement limitation and a documented change rather than hiding them. Its checklist points to the corresponding manuscript passages.

No patient data or results are supplied. The example shows a sequence of decisions, not proof that the proposed design answers the question adequately.

Common pitfalls

Starting with data rather than a question. Explain why the project matters and what it can address.
Choosing a checklist at submission only. Read it early enough to preserve relevant information.
Leaving responsibilities implicit. Assign owners for methods, data and documentation.
Treating access as permission. Confirm the relevant governance requirements.
Losing the original plan. Preserve dated versions and reasons for changes.
Calling exploratory work prespecified. Keep the distinction visible in the report.

Quick self-check, in our own words

I can state the question and identify its design.
I selected a reporting framework and checked for extensions.
I documented methods, roles and oversight requirements.
I can trace the analysis dataset and outputs to their source records.
I preserved amendments and distinguished exploratory analyses.
I reconciled the manuscript, tables, figures and checklist.
The team has a separate process for approving external submission or sharing.

References

Simera I, Moher D, Hirst A, Hoey J, Schulz KF, Altman DG. Transparent and accurate reporting increases reliability, utility, and impact of your research: reporting guidelines and the EQUATOR Network. BMC Med. 2010;8:24. doi:10.1186/1741-7015-8-24. PMID: 20420659.
Chan AW, Boutron I, Hopewell S, Moher D, Schulz KF, Collins GS, et al. SPIRIT 2025 statement: updated guideline for protocols of randomised trials. BMJ. 2025;389:e081477. doi:10.1136/bmj-2024-081477. PMID: 40294953.
Glasziou P, Altman DG, Bossuyt P, Boutron I, Clarke M, Julious S, et al. Reducing waste from incomplete or unusable reports of biomedical research. Lancet. 2014;383(9913):267-76. doi:10.1016/S0140-6736(13)62228-X. PMID: 24411647.
von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. PLoS Med. 2007;4(10):e296. doi:10.1371/journal.pmed.0040296. PMID: 17941714.
Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71. PMID: 33782057.
Hopewell S, Chan AW, Collins GS, Hróbjartsson A, Moher D, Schulz KF, et al. CONSORT 2025 statement: updated guideline for reporting randomised trials. BMJ. 2025;389:e081123. doi:10.1136/bmj-2024-081123. PMID: 40228833.