For nearly a century, the scientific paper has worn the same reliable jacket: Introduction, Methods, Results, and Discussion, better known as IMRAD. It is tidy, familiar, journal-friendly, and deeply useful. It became widely adopted in medical journals during the twentieth century, and the ICMJE still describes it as the usual structure for original research articles because it reflects the broad process of scientific discovery. (PMC)
But science has changed.
Today’s research articles often contain multi-omics datasets, biomaterials, machine learning models, imaging pipelines, clinical cohorts, computational simulations, preprints, code repositories, supplementary files, reporting checklists, and claims that stretch from molecular mechanism to clinical translation. IMRAD still works, but sometimes it behaves like a four-room house trying to host a whole research city. 🏙️🔬
So let us imagine a better format.
Not a replacement born from rebellion for rebellion’s sake. A format designed for the modern reader, reviewer, editor, student, machine reader, and future researcher trying to reproduce the work five years later.
I call it the CREDENCE format.
It is built around a simple principle:
A research article should not be organized only by what the authors did. It should be organized by what the authors claim, what evidence supports each claim, how strong that evidence is, and how others can reuse or challenge it.
Why IMRAD is brilliant, and where it struggles
IMRAD is elegant because it gives science a familiar storyline:
| IMRAD section | Core question |
|---|---|
| Introduction | Why was the study needed? |
| Methods | How was it done? |
| Results | What was found? |
| Discussion | What does it mean? |
This is still valuable. The problem is not that IMRAD is wrong. The problem is that IMRAD often hides the most important thing in a paper: the claim-evidence relationship.
In many articles, the reader must hunt across sections like a detective with a tired magnifying glass:
The claim is in the Abstract.
The rationale is in the Introduction.
The method is buried in Methods.
The result is Figure 3C.
The control is in Supplementary Figure 8.
The limitation is in the final paragraph.
The data are somewhere else entirely.
The code is “available upon reasonable request,” which often means “vanished into the mist.”
This is not reader-friendly. Gopen and Swan’s classic essay on scientific writing argued that readers interpret scientific prose according to structural expectations, not merely grammar. When structure misdirects attention, clarity suffers. (Gatsby)
IMRAD gives readers a map of the article. CREDENCE gives readers a map of the evidence.
The CREDENCE format
CREDENCE stands for:
| Letter | Section | Purpose |
|---|---|---|
| C | Context and Consequence | What problem matters, and why now? |
| R | Research Question and Claim Ledger | What exactly is being claimed? |
| E | Experimental or Analytical Design | How was the question tested? |
| D | Data and Evidence Modules | What evidence supports each claim? |
| E | Error, Uncertainty, and Robustness | How fragile or reliable are the findings? |
| N | Novelty in Relation to Prior Work | What is truly new, and what is not? |
| C | Caveats, Ethics, and Boundaries | What should readers not conclude? |
| E | Extension, Reuse, and Next Experiments | How can others build on this work? |
The acronym is not just decorative. The promise of the format is credibility. The paper should earn the reader’s confidence step by step.
Let us walk through it.
1. Context and Consequence
The first section replaces the traditional broad Introduction with a sharper opening.
Instead of starting with a grand tour of the field, this section answers:
What is the unresolved problem?
Who or what is affected by it?
Why does the problem remain unsolved?
What would change if the gap were addressed?
A traditional IMRAD Introduction often sprawls. It begins with global burden, then history, then mechanisms, then a pile of prior studies, then finally the actual research question appears at the end like a shy animal emerging from shrubbery.
CREDENCE asks for the question earlier.
Example opening in CREDENCE style:
“Current cardiac patches for myocardial infarction face three linked barriers: poor wet adhesion, limited electrical integration, and uncontrolled local drug release. These barriers reduce their translational potential because a patch must remain attached to a moving epicardial surface, support electromechanical compatibility, and deliver therapy without burst-related toxicity. This study tests whether an integrated biomaterial design can address these barriers at the prototype stage.”
This is not a literature dump. It is a launchpad.
Why this is better than IMRAD
IMRAD often encourages the writer to begin with the universe and slowly zoom in. CREDENCE begins with the problem and consequence. The reader knows within seconds why the work exists.
2. Research Question and Claim Ledger
This is the heart of the new format.
Every paper should include a Claim Ledger, ideally as a table near the beginning.
A Claim Ledger lists each major claim, the evidence used to support it, and the strength of that evidence.
| Claim no. | Claim | Evidence type | Strength | Not yet shown |
|---|---|---|---|---|
| 1 | The hydrogel can be fabricated reproducibly | Composition, gelation, imaging | Strong if replicated | Scale-up |
| 2 | The patch releases hesperidin over 72 h | UV-Vis release assay | Moderate | Release in serum/in vivo |
| 3 | rGO may support conductivity | Material composition only | Weak unless measured | Conductivity, cardiomyocyte coupling |
| 4 | Sucker geometry may improve adhesion | FEA model | Preliminary | Wet myocardial adhesion test |
| 5 | Patch supports cytocompatibility | 3T3-L1 MTT assay | Preliminary | Cardiac cells, immune response |
This single table would prevent many common scientific-writing sins: overclaiming, unclear novelty, missing controls, and inflated conclusions.
The Claim Ledger also helps reviewers. Instead of asking, “Where did they prove this?” the reviewer can immediately see whether the authors are making a measured claim, an inferred claim, or a proposed claim.
Why this is better than IMRAD
IMRAD lets claims scatter across the paper. CREDENCE pins them to the wall, politely but firmly. 🧷
This improves honesty. It also improves writing because the author must decide: Did we measure this, infer this, or merely propose it?
3. Experimental or Analytical Design
This section replaces the traditional Methods opening, but with an important difference.
Instead of listing procedures immediately, it first explains the logic of the study design.
It asks:
What were the experimental units?
What were the controls?
What comparisons matter?
What variables were measured?
What would falsify or weaken the hypothesis?
What was planned before analysis?
What was exploratory?
The ICMJE recommends that methods provide sufficient details about procedures, materials, and statistical methods for readers to understand and evaluate the study. (ICMJE) CREDENCE keeps that requirement but adds a design logic layer before the protocol details.
Example:
“The study was designed to test material feasibility in three stages: fabrication, functional characterization, and computational mechanical assessment. The primary material comparator was the alginate-silk control hydrogel. The main prespecified readouts were swelling ratio, cumulative hesperidin release, cell viability, pore morphology, FTIR spectral shifts, XRD phase behavior, and simulated von Mises stress.”
This tells the reader not only what was done, but why each piece exists.
Why this is better than IMRAD
Traditional Methods can become a recipe book. CREDENCE turns it into a strategy document plus recipe book. The reader sees the experimental chessboard before the moves begin. ♟️
4. Data and Evidence Modules
This section is the biggest departure from IMRAD.
Instead of separating all Methods from all Results, CREDENCE organizes the paper into Evidence Modules. Each module is built around one claim.
Each Evidence Module has the same structure:
Claim being tested
Method used
Result
Control/comparator
Interpretation
Confidence level
What remains unresolved
For example:
Evidence Module 2: Hesperidin release from the hydrogel
Claim: The hydrogel provides controlled hesperidin release.
Method: UV-Vis quantification of hesperidin release in PBS at 37 °C.
Result: 67.5% release occurred within 30 min, followed by slower release up to 72 h.
Interpretation: The result supports burst-dominant biphasic release, not fully sustained release.
Confidence: Moderate for in vitro PBS release; low for in vivo pharmacokinetics.
Unresolved: Release under serum-rich, enzymatic, mechanically dynamic conditions.
This is cleaner than forcing readers to jump between Methods, Results, and Discussion. It also makes the writing more resistant to hype because each claim must sit beside its evidence.
Why this is better than IMRAD
IMRAD is chronology-centered: first we explain all methods, then all results, then all meaning.
CREDENCE is evidence-centered: each claim travels with its method, result, interpretation, and limitation.
This is especially useful for interdisciplinary papers. A biomaterials paper may contain chemistry, microscopy, mechanics, cell biology, drug release, and simulation. In IMRAD, these threads can tangle into a glorious spaghetti monster. In CREDENCE, each thread gets its own spool. 🍝
5. Error, Uncertainty, and Robustness
Every paper should have a section that does not merely say “limitations,” but actively examines fragility.
This section asks:
How variable are the data?
Were results replicated?
Were alternative explanations tested?
How sensitive are conclusions to assumptions?
What controls failed or were absent?
What negative or null results occurred?
What analysis choices could change the conclusion?
This is where statistical reporting, sensitivity analysis, model assumptions, missing data, batch effects, and negative results live.
The EQUATOR Network maintains a searchable library of reporting guidelines because different study types need specific reporting standards for completeness and transparency. (EQUATOR Network) CREDENCE is not a replacement for CONSORT, PRISMA, STROBE, ARRIVE, or other guidelines. It is a narrative architecture that makes room for them.
Example robustness paragraph:
“The FEA results are sensitive to assumed Young’s modulus, boundary conditions, and suction pressure. Because the myocardium was modeled using simplified material properties, the simulation should be interpreted as a design-screening tool rather than proof of adhesion safety. Experimental pull-off and cyclic wet-adhesion testing are required.”
That paragraph is not weakness. It is intellectual hygiene.
Why this is better than IMRAD
In many IMRAD papers, uncertainty is exiled to the final paragraph, where it appears as a ritual apology. CREDENCE brings uncertainty into the central architecture of the paper.
A good paper should not hide its caveats in the attic. It should give them chairs at the table.
6. Novelty in Relation to Prior Work
This section forces authors to answer one of the most abused questions in research writing:
What is actually new?
Not what sounds new. Not what the authors wish were new. Not what becomes new after ignoring inconvenient literature.
This section should include a Novelty Matrix.
| Prior work already showed | This study adds | Still not shown |
|---|---|---|
| Drug-loaded cardiac patches exist | Hesperidin in this specific hydrogel system | In vivo myocardial delivery |
| Conductive cardiac patches exist | rGO included in SF/alginate/gelatin/dECM system | Functional conductivity |
| Wet adhesive patches exist | Octopus-sucker-inspired geometry proposed | Ex vivo wet adhesion |
| dECM hydrogels exist | Cardiac dECM integrated into this composite | Retained bioactivity quantified |
This would dramatically improve thesis and manuscript writing. It prevents lazy firstness claims, such as “this is the first study,” unless the authors can defend the claim.
Why this is better than IMRAD
IMRAD usually places novelty in the final paragraph of the Introduction and maybe again in the Discussion. CREDENCE gives novelty its own courtroom.
The paper must show not only what it did, but what it did relative to the closest prior art.
That is better scholarship and fewer reviewer thunderbolts. ⚡
7. Caveats, Ethics, and Boundaries
This section is different from the robustness section.
Robustness asks: How reliable are the findings?
Boundaries ask: Where should these findings not be applied?
This section includes:
Biological boundaries
Clinical boundaries
Species/model boundaries
Ethical considerations
Safety concerns
Generalizability
Misuse potential
Regulatory relevance
Conflicts of interest
For biomedical studies, this is crucial. A hydrogel that is cytocompatible in fibroblasts is not automatically safe for the heart. A machine learning model that works in one hospital is not automatically fair across populations. A drug that works in mice is not automatically a therapy.
Example:
“The present findings should not be interpreted as evidence of myocardial regeneration. The study does not include cardiomyocyte functional assays, ex vivo myocardial adhesion testing, immune compatibility, or in vivo MI repair. Therefore, the patch should be considered a prototype platform rather than a therapeutic candidate.”
This kind of writing protects the reader, the author, and the science.
Why this is better than IMRAD
IMRAD often lets boundaries blur into the Discussion. CREDENCE makes them explicit. This reduces translational exaggeration, one of the great glitter-traps of modern biomedical writing.
8. Extension, Reuse, and Next Experiments
The final section asks:
What can others reuse?
Where are the data?
Where is the code?
What exact experiment should be done next?
What would move the work from prototype to validation?
What would disprove or revise the central claim?
This section should include a Reuse Box.
| Reusable item | Location |
|---|---|
| Raw data | Repository link |
| Analysis code | GitHub/Zenodo |
| Protocol | Supplementary protocol |
| CAD/STL files | Repository |
| Statistical script | Repository |
| Reporting checklist | Supplementary file |
It should also include a Next Experiment Ladder:
| Stage | Next experiment | Purpose |
|---|---|---|
| Material validation | Conductivity under hydrated conditions | Test rGO function |
| Adhesion validation | Pull-off on wet porcine myocardium | Test sucker geometry |
| Cell validation | hiPSC-cardiomyocyte beating assay | Test electrical/biological compatibility |
| Mechanistic validation | ROS, VEGF, Cx43 assays | Test drug and conductive mechanisms |
| Translational validation | Small-animal MI model | Test repair potential |
Why this is better than IMRAD
IMRAD usually ends with a broad future-scope paragraph. CREDENCE ends with a usable research roadmap. It helps the next researcher know exactly where to place the next brick.
Demonstration: IMRAD versus CREDENCE
Let us imagine a paper on a biomimetic cardiac patch.
In IMRAD
The paper might look like this:
Introduction: MI causes cardiac damage. Patches are promising. Adhesion and conductivity are problems. We developed a new patch.
Methods: We prepared silk fibroin, alginate, gelatin, dECM, rGO, tannic acid, and hesperidin hydrogel. We performed SEM, FTIR, XRD, release assay, MTT, and FEA.
Results: The patch was porous. FTIR showed interactions. XRD showed amorphous distribution. Hesperidin released over 72 hours. Cells survived. FEA showed lower stress.
Discussion: The patch may help regenerate myocardium. Future studies are needed.
This is acceptable. But it leaves dangerous ambiguity. Which result supports which claim? How strong is the evidence? What was not tested? Is the release really sustained? Was conductivity measured? Was adhesion tested?
In CREDENCE
The same study becomes:
Context and Consequence: Post-MI patches need wet adhesion, local delivery, and electrical compatibility.
Research Question and Claim Ledger: Five claims are listed, with evidence strength.
Design: Prototype material study with defined controls and readouts.
Evidence Module 1: Fabrication and morphology.
Evidence Module 2: Hesperidin release, interpreted as burst-dominant biphasic release.
Evidence Module 3: Cytocompatibility in 3T3-L1 cells, framed as preliminary.
Evidence Module 4: rGO-containing design, clearly marked as not yet conductivity-validated.
Evidence Module 5: Sucker geometry FEA, clearly marked as simulation-only.
Robustness: Missing direct conductivity, wet adhesion, cardiomyocyte function, in vivo testing.
Novelty Matrix: Specific combination is novel; broad category is not.
Boundaries: Not evidence of myocardial regeneration.
Extension: Next experiments and reusable CAD/protocol/data.
The second version is harder to exaggerate and easier to trust.
That is the central advantage.
The CREDENCE abstract
Even the abstract changes.
Traditional abstracts often follow Background, Methods, Results, Conclusion. CREDENCE abstracts would include a mini claim ledger:
Context: Cardiac patches require wet adhesion, local drug delivery, and electrical compatibility.
Question: We tested whether a hesperidin-loaded, rGO-containing SF/alginate/gelatin/cardiac-dECM hydrogel with octopus-sucker-inspired geometry could serve as a prototype cardiac patch.
Evidence: The hydrogel was fabricated and characterized by SEM, FTIR, XRD, swelling, release assay, MTT, and FEA.
Findings: Hesperidin release was burst-dominant and biphasic over 72 h. 3T3-L1 viability supported preliminary cytocompatibility. FEA predicted lower stress concentration in the sucker-inspired geometry.
Boundaries: Conductivity, myocardial adhesion, cardiomyocyte function, and in vivo repair were not demonstrated.
Next step: Hydrated conductivity, cyclic wet-adhesion, cardiac-cell, and MI-model validation are required.
This abstract would be bracingly honest. A reviewer may still criticize the work, but they cannot accuse the authors of selling a telescope as a spaceship. 🚀
Why CREDENCE is better for readers
Readers do not read papers linearly anymore. They scan abstracts, inspect figures, jump to methods, search keywords, open supplementary files, and compare claims with data.
CREDENCE supports this behavior.
| Reader need | IMRAD weakness | CREDENCE solution |
|---|---|---|
| Understand main claims quickly | Claims scattered | Claim Ledger |
| Judge evidence strength | Methods/results separated | Evidence Modules |
| Assess overclaiming | Limitations buried | Caveats and Boundaries |
| Compare with prior work | Novelty often vague | Novelty Matrix |
| Reproduce study | Protocols scattered | Reuse Box |
| Plan next work | Future scope vague | Next Experiment Ladder |
| Review efficiently | Reviewer must reconstruct logic | Claim-evidence map visible |
It is not merely a new order of sections. It is a new moral posture: show the claim, show the evidence, show the uncertainty.
What would journals gain?
Journals would gain more transparent submissions.
Reviewers could evaluate manuscripts faster because the authors have already mapped claims to evidence. Editors could detect overclaiming earlier. Reporting guidelines could be placed where they belong, inside the design and robustness logic rather than as afterthought checklists. The EQUATOR Network’s role as a reporting-guideline library would become even more powerful when manuscripts had a dedicated place for study-specific reporting requirements. (EQUATOR Network)
CREDENCE also helps prevent “supplementary exile,” where critical controls are banished to supplementary materials. If a claim depends on a control, the control belongs in the Evidence Module.
What would authors gain?
Authors often discover weaknesses only after reviewer comments arrive with tiny knives. CREDENCE reveals those weaknesses before submission.
When authors prepare the Claim Ledger, they may realize:
“We claimed conductivity but never measured conductivity.”
“We claimed sustained release but most drug came out in 30 minutes.”
“We claimed biocompatibility but tested only one non-cardiac cell line.”
“We claimed novelty but forgot three close prior papers.”
“We claimed mechanism but showed correlation.”
This is painful, but useful. The format turns the manuscript into a pre-review machine.
What would students gain?
Students often learn IMRAD as a container. They ask, “What goes in the Introduction?” or “How long should the Discussion be?”
CREDENCE teaches a deeper habit:
Every scientific claim must carry its evidence passport.
That habit is more valuable than any section heading.
A student using CREDENCE would learn to write with calibrated certainty. Not timidly. Not theatrically. Precisely.
Could CREDENCE replace IMRAD tomorrow?
No, and it does not need to.
A realistic path would be hybrid adoption. Journals could keep IMRAD but require CREDENCE elements:
Claim Ledger after the Introduction
Evidence Modules within Results
Robustness section before Discussion
Novelty Matrix in Discussion
Reuse Box before References
In other words, CREDENCE can be used as a paper architecture or as an overlay on IMRAD.
For example:
| IMRAD section | CREDENCE upgrade |
|---|---|
| Introduction | Add Context, Consequence, Claim Ledger |
| Methods | Add Design Logic and reporting checklist |
| Results | Convert to Evidence Modules |
| Discussion | Add Novelty Matrix and Boundaries |
| Data availability | Expand into Reuse and Next Experiment section |
This makes the proposal practical. Science does not need another format war. It needs better thinking made visible.
A possible CREDENCE article template
Here is how a full research article could look:
Title
Precise, claim-calibrated, no therapeutic overreach.
Structured Abstract
Context, Question, Evidence, Findings, Boundaries, Reuse.
1. Context and Consequence
Problem, gap, stakes.
2. Research Question and Claim Ledger
Primary question, secondary questions, claims, evidence strength.
3. Experimental or Analytical Design
Study logic, controls, variables, prespecified outcomes, exploratory outcomes.
4. Data and Evidence Modules
Each module pairs method, result, interpretation, and unresolved issue.
5. Error, Uncertainty, and Robustness
Statistics, sensitivity, negative results, missing controls, alternative explanations.
6. Novelty in Relation to Prior Work
Closest literature, novelty matrix, what is new and not new.
7. Caveats, Ethics, and Boundaries
Generalizability, safety, clinical limits, ethical issues.
8. Extension, Reuse, and Next Experiments
Data/code/protocol availability, next experiment ladder.
References
Complete and relevant.
Supplementary Evidence Archive
Raw images, full protocols, code, data, checklists.
Final thought: IMRAD tells the story of discovery. CREDENCE tells the story of evidence.
IMRAD was built for a world in which the research article was mostly a narrative of scientific work. It still does that well. But modern science needs papers that are not only readable but inspectable, reusable, and claim-calibrated.
CREDENCE does not ask authors to be less ambitious. It asks them to be more exact.
It does not flatten scientific storytelling. It gives the story a skeleton strong enough to carry data, doubt, novelty, and future use.
The future research article should not be a polished tunnel through which conclusions rush past the reader. It should be a transparent machine: every gear visible, every claim traceable, every uncertainty labeled, every next experiment waiting like a lit doorway.
That is the article format science deserves next. 🧬📜✨
References
Sollaci LB, Pereira MG. The introduction, methods, results, and discussion structure: a fifty-year survey. Journal of the Medical Library Association. 2004;92(3):364-367. PMID: 15243643. The study describes the historical adoption of IMRAD in leading medical journals. (PMC)
International Committee of Medical Journal Editors. Preparing a Manuscript for Submission to a Medical Journal. The ICMJE states that original research articles are usually divided into Introduction, Methods, Results, and Discussion, and that IMRAD reflects the process of scientific discovery. (ICMJE)
Gopen GD, Swan JA. The Science of Scientific Writing. American Scientist. 1990;78(6):550-558. This essay explains how reader expectations shape clarity in scientific prose. (Gatsby)
EQUATOR Network. Search for reporting guidelines. The EQUATOR library provides a searchable collection of reporting guidelines for different health research designs. (EQUATOR Network)
No comments:
Post a Comment