Reverse-engineering I — IMRaD architecture
Week 4 · 2105620 Research Communication Skills for Chemical Engineers
27 August 2026
This week has no classroom
How Week 4 works
- Watch three recordings, about fifteen minutes each. Watch them in order
- Do the reverse outline of the paper you already chose. One page
- Submit before Week 5. It is the front half of Assignment 1, not extra work
There is no session on 27 and 30 August. The work is not cancelled — it moved onto your desk.
What the three recordings do
Segment map
One
What each section is for
The four jobs of a paper, and the promises the sections make to each other
Two
The reverse outline
How to see a structure you cannot see while reading forward
Three
Where arguments break
Five failures that are invisible in the text and obvious in the outline
Recording one is conceptual. Two is a procedure you will copy. Three is the diagnostic work, and it is what you submit.
What each section is for
Segment one · about 15 minutes
Four sections, four jobs — and four promises the sections make to each other.
A paper is not a report of what
you did, in the order you did it.
It is an argument, assembled so that a stranger can check it.
Four sections, four jobs
IMRaD
| Introduction | Establishes that a specific question is worth answering, and that it is not yet answered |
| Methods | Makes the answer checkable. Someone else could do this and find out whether you are right |
| Results | States what was observed. Nothing about what it means |
| Discussion | Converts observation into claim, and states the boundary of that claim |
Read the right-hand column as job descriptions. A section that does a different job is a structural fault, however well written it is.
The sections make promises to each other
Dependencies
- Every aim stated in the Introduction is answered in the Results
- Every number in the Results is produced by something in the Methods
- Every claim in the Discussion traces back to a specific Result
- Nothing new appears in the Discussion — no new data, no new method
These four are checkable by a reader who does not understand your chemistry. That is exactly why reviewers check them first.
Where sections do each other’s jobs
The three common trespasses
- Results that interpret — “the improvement is attributed to a more uniform interphase”. Nothing was observed. That belongs in the Discussion
- A Discussion that reports — three paragraphs restating the figures. The reader has already read the figures
- Methods that argue — “this technique was chosen because it is superior to X”. Justification, not procedure
All three read perfectly well sentence by sentence. That is why they survive to submission.
The abstract is the same argument, compressed
One paragraph, four moves
| Sentence 1–2 | The context, and the specific problem — the Introduction |
| Sentence 3 | What was done — the Methods |
| Sentence 4–5 | What was found, with numbers — the Results |
| Sentence 6 | What it means, and how far — the Discussion |
If you cannot compress a paper into these six sentences, the problem is usually not the abstract.
You cannot see a structure
while you are reading forward.
Which is why the next fifteen minutes exist.
The reverse outline
Segment two · about 15 minutes
One line per paragraph, recording what the paragraph does. Five steps, and you will use it for the rest of your career.
An outline written backwards, from a finished text
Definition
One line per paragraph, recording what the paragraph does — not what it says.
- A summary compresses content. It tells you the paragraph is about zinc dendrites
- A reverse outline records function. It tells you the paragraph narrows a general problem to a specific one
- Only the second one lets you see whether the argument holds together
Five steps
The procedure
- Number the paragraphs — in the margin, or as a list. You will refer to them by number from here on
- Write the function in ten words or fewer — start with a verb. If it will not fit, the paragraph has more than one job
- Tag each line with the job it serves — Introduction, Methods, Results or Discussion, regardless of the section it is printed in
- Collapse to section level — read the lines for one section together and write a single line for the section
- Read only the outline — close the paper. Everything wrong with the architecture is now on one page
Step five is the one people skip, and it is the step where the work pays off.
Verbs that describe a job
Vocabulary
establishes situates |
Puts the work in a context the reader already accepts |
narrows specifies |
Moves from a general area to a particular question |
reports quantifies |
States an observation, with or without numbers |
justifies motivates |
Gives a reason for a choice |
contrasts reconciles |
Sets two findings against each other, or resolves them |
concedes bounds |
Admits a limit on what can be claimed |
claims attributes |
Asserts a mechanism or an explanation |
Worked example — an Introduction, outlined
Constructed for this exercise
| P1 | Establishes that grid storage demand is growing |
| P2 | Narrows to aqueous zinc as a low-cost option |
| P3 | Reports that dendrites limit zinc anode lifetime |
| P4 | Reviews electrolyte additives as a mitigation |
| P5 | Narrows to aqueous zinc as a low-cost option |
| P6 | States the aim: screen four additives at high current density |
Six lines. Nothing about chemistry. One of them is a problem, and you found it in four seconds.
What a healthy Introduction looks like in outline
The shape
- Establishes a context the reader already accepts
- Narrows, once or twice, each time tighter
- Reports what is known, and then what is not
- States the gap — as a cost, not as an absence
- States the aim, in terms that the Results can answer
Five moves, in that order, and the gap is one of them. In most drafts the gap is missing, buried, or written as “little work has been done”.
Worked example — a Discussion, outlined
The section you also have to submit
| P1 | Restates the main finding |
| P2 | Attributes the effect to a more uniform interphase |
| P3 | Compares the result with two earlier studies |
| P4 | Restates the main finding, with different wording |
| P5 | Concedes that only symmetric cells were tested |
| P6 | Proposes future work on full cells |
Two findings, from six lines: P4 doubles P1, and P2 claims a mechanism that no Result measures.
When a paragraph will not fit in ten words
The most useful thing the limit does
- Two verbs in your line — the paragraph is doing two jobs. It should usually be two paragraphs
- No verb you are happy with — the paragraph may be doing no job at all. Ask what would be lost by deleting it
- The line is longer than the paragraph deserves — a common sign of a literature-review paragraph that wandered in
Do not fight the limit. When a line will not fit, that is the outline telling you something about the text, not about your writing.
Ten words. A verb.
Function, not content.
Everything in segment three depends on you doing it that way.
Where arguments break
Segment three · about 15 minutes
Five failures that are invisible in the text and obvious in the outline. This segment is where your submission comes from.
Five breaks, all invisible in the text
The diagnostic list
| The orphan aim | An aim in the Introduction with no matching Result |
| The unsupported claim | A Discussion claim with no Result behind it |
| The invisible method | A number in the Results that no Method produces |
| The doubled move | Two paragraphs with the same function |
| The buried gap | The gap arrives late, or never arrives at all |
Each one is a broken promise from the earlier list. That is not a coincidence — it is where the list came from.
The orphan aim
Break one
| Introduction, P6 | States three aims: screen additives, identify the mechanism, assess full-cell viability |
| Results | Reports screening. Reports post-mortem imaging. |
| Discussion | Claims a mechanism. Says nothing about full cells. |
Aim three is an orphan. It was promised on page two and abandoned by page nine — and the paper never admits it.
The unsupported claim
Break two
| Discussion, P3 | Attributes the improvement to a more uniform interphase |
| Results | Reports cycle life, voltage hysteresis, and surface images |
| The problem | No measurement of the interphase exists anywhere in the paper |
The claim may well be true. It is still not supported by this paper, and that is a different thing.
The invisible method
Break three
| Results | Reports coulombic efficiency of 99.4 % over 400 cycles |
| Methods | Describes cell assembly, electrolyte preparation, and cycling protocol |
| The problem | No definition of how efficiency was calculated, and no statement of what counts as failure |
Two papers can report the same number and mean different things by it. That is not a small problem; it is why half the literature does not compare.
How to write a finding
The form your submission must use
Location · Function · What is missing
- ✅ Introduction P6 states an aim on full-cell viability. No Result addresses it and the Discussion does not acknowledge the omission.
- ❌ The introduction is not very well connected to the results.
The first sentence can be checked in one minute. The second cannot be checked at all — and it is what most appraisals contain.
Outline your own draft
before a reviewer does.
The five breaks are cheap to fix at outline stage and expensive after submission.
What you submit
Due before Week 5 · one page
- The outline — numbered lines for the Introduction and the Discussion of your chosen paper. Ten words, verb first
- The collapse — one line for each of those two sections
- Two findings — written as location, function, what is missing
- The AI declaration — attached, as for every assignment from Week 3 onwards
Same paper you chose for Assignment 1. This page becomes the front half of that appraisal — it is not additional work.
Before Week 5
Three things
- Submit the outline. Before the Week 5 session, not during it
- Keep the paper. Week 5 takes an Introduction apart move by move, using yours
- Bring the matrix too. Week 5 asks you to write the gap statement your matrix implies
Week 5 is a working session. It assumes you arrive with an outline and a matrix. Without them you will watch other people work.
Reverse-engineering II — the Introduction and the gap
Next · Week 5 · 3 and 6 September
Bring the outline you submitted and the evidence matrix from Week 3. Week 5 is a working session.
Before you close the video · 45 seconds
How was this week?
Five questions. Anonymous — no name, no email, no login.
The fourth question is the one I actually use.
An asynchronous week is the one I can see least. This form is the only signal I get.
oxidized-challenge-ed9.notion.site
Scan it before you close the tab