Results is "what happened," Discussion is "what it means." If the Introduction is a "declaration of what is lacking," the Discussion is a "declaration of what is updated." If they are not mirrors, the structure is broken.
The Lab Forum lecture "Writing Papers" covered the whole picture of submission, and "Writing Papers — Figures" covered the design of figures. This page covers the stretch from the figures being frozen to the manuscript being complete — Draft-A to Draft-I of the Paper Card.
There are two prerequisites. The figures have been frozen in Figures-C. The Core Question and Key Contributions have been fixed in Framing-A and B. With these two, the Draft is not a task of writing from a blank page, but a task of unfolding decided materials into text.
B4: Learn what each chapter of your thesis is responsible for. M & D: Check whether the Introduction and Discussion of your own manuscript are mirrors. Postdocs: Apply the five-paragraph structure of the Discussion to a junior's manuscript.
Methods → Results → Discussion → Introduction → Conclusion → Abstract → Polish
In a manuscript, the order of reading and the order of writing differ. The reader reads Abstract → Introduction → …, but the writer writes from the other end.
The reason is simple: the Introduction cannot be written without knowing "what this paper has shown." At the end of the Introduction comes "Here we show that…". That must match the first sentence of the Discussion. If you write the Introduction before the Discussion, you either announce something you did not show, or fail to announce what you did show.
We write from Methods because it is closest to "facts" and the place where you get lost the least. When you cannot write, write from where you can. "Against the fear of the blank page, start from the caption of Figure 1," stated in Writing Papers, is the same principle.
The Introduction and the Discussion are mirrors
The Introduction and the Discussion are mirrors. The Introduction declares "the field knows this much, and this is what is lacking (the Gap)," and the Discussion declares "with this result the Gap is filled in this way, and the field's assumption is updated in this way." If what the Introduction declared to be lacking is not recovered in the Discussion, the structure of the paper is broken.
The division of roles between Results and Discussion is just as important. Results is only "what happened." Discussion is only "what it means." When the two mix, the reader can no longer distinguish fact from interpretation, and the reviewer writes "Overclaiming."
The purpose of Methods is to write the foundation of facts that others can reproduce. It is where the "reproducibility" of Lab OS Chapter 06 is implemented as text.
Four blocks
"Details omitted" is an abandonment of reproducibility. Conversely, reusing the Methods of your master's thesis in a journal paper is accepted if you cite the source (see the self-plagiarism table in Writing Papers). Methods is the only place in a paper where "writing to a template" is allowed.
The purpose of Results is to write only the facts that correspond to the figures.
Template: "Figure 1 shows … X increased as … The difference between A and B was …" Include numbers, confidence intervals, and the direction of the comparison.
If you feel like writing "why it happens" in Results, move it to the Direct Answer or Assumption Refinement of the Discussion. Once even a single interpretation slips into Results, the reader reads all subsequent facts as interpretation-laden.
The purpose of the Discussion is to answer the Core Question and give the results meaning (an update of the field's beliefs). Write five paragraphs, in this order. The order traces the outputs of Framing-A (Core Question, Gap, Hidden assumption, Why it matters) in reverse.
Copy the Core Question and answer it directly in one or two sentences. Summarize the results concisely; numbers may be included. Do not widen the interpretation here. "This study shows that … Our results indicate that …"
The first sentence of this paragraph becomes the "Here we show" of the Introduction and the first sentence of the Conclusion. Say the same thing in three places. If they diverge, the claim of the paper is not yet settled.
Reread the Gap sentence of the Introduction, make explicit "what was lacking," and write how the present results fill it. Do not create a new Gap here. "Our findings address the limitation that … This resolves the uncertainty regarding …"
Write one assumption of existing research and how it changes. Write it as an "update," not a "denial." An aggressive style gains nothing, considering that a reviewer may hold that assumption. "This challenges the assumption that … Instead, our results suggest that …"
The Hidden assumption and Our update in the Comparison Table of Framing-A are the direct material for this paragraph.
Reread the opening of the Introduction (Broad Context) and write how that context changes. Make explicit where it affects theory, application, or design. Do not save the world. "These findings change how we understand … This has implications for …"
The commitment of Lab OS Chapter 02 matters here. A paper committed to explanation has implications for understanding; a paper committed to control has implications for design. Do not write implications your evidence does not reach.
Write two or three constraints of this study, how they affect the results, and indicate the next research direction. Do not be defensive. "This study assumes … One limitation is … Future work should …"
A paper that writes its limitations honestly withstands review better. If you can separate discrepancy (reducible uncertainty) from unknowability (irreducible uncertainty), the Limitations become not a confession of weakness but a bridge to the next inquiry.
The purpose of the Introduction is to place the provisional belief settled in the Discussion correctly in the context of the field. The Narrative Skeleton of Lab OS Chapter 04 was a blueprint written before the data came in. Here, with the Discussion settled, you rewrite it.
Five elements
New claims grow in the Introduction because, while writing, you think "I can say this too." If you can say it, go back to the Discussion and add a paragraph. Do not announce in the Introduction what is not in the Discussion.
Check: Is it not a repetition of Results? / Is it not a copy of the Discussion? / Is it not overclaiming?
Check: Does it match the Introduction? / Does it match the Discussion? / Does it contain no new information? / Is it within 200–250 words?
The passing condition is that the claim is understood from the Abstract alone. Most readers read only the Abstract. Reviewers, too, first fix the paper's claim from the Abstract, then check whether the body supports it.
Once all sections are in place, audit the structure.
The comments reviewers write are surprisingly similar.
All four can be eliminated at the Framing and Figures stages of the Paper Card. Before submission, write these four yourself and be able to point to where in the manuscript the answer to each lies.
Produce a manuscript in which the claim can be read through. Do not write from the top, do not aim for perfection, do not worry too much about the fine points of English, do not overthink the reviewer response. This is the stage where "separate generation from editing," stated in Writing Papers, is carried out across the whole manuscript. Mark the points where you want comments, and hand it to Team Meeting or your supervisor.
Aim for the minimum completeness that withstands review. Cut what can be cut. Write the answers to the anticipated reviewer comments. Decide what to move to the Appendix.
Do not keep polishing forever. Prioritize clarity of logic over perfect English expression. For sentence-by-sentence checking, use the five items of Writing Paper Skills (strength of claim, reviewer's challenge, correspondence with evidence, validity of comparison, safer rewrite).
Implement the commitments of Lab OS Chapter 02 as verbs.
| Commitment | Verbs you can use |
|---|---|
| Description | observed, measured, found |
| Explanation | indicates, is consistent with, is controlled by (after eliminating alternative explanations) |
| Prediction | predicts, generalizes to |
| Control | enables, improves, reduces |
Use "suggests" only in the Discussion, as a weak commitment to explanation. Not in Results. "demonstrates" and "proves" only when the evidence reaches that level.