Research and Academic Writing for Graduate Students · Part 5
When students hear that a draft “lacks logic” or “does not flow well,” they often start editing sentences. They add transitions, rearrange paragraphs, and polish their English. These revisions can help, but the underlying problem often lies in the connections between ideas.
The draft may not adequately connect why a problem deserves attention, what existing knowledge leaves unresolved, how the proposed method addresses that problem, and what conclusions the results support.
The logic of a paper concerns whether these connections are valid. Its flow concerns whether the explanation follows an order that allows readers to understand those connections. Individual statements can be correct while the links between them are weak. A study can also have a sound argument that readers struggle to follow because of the way it is presented.
Both deserve attention when writing a paper.
1. From what you did to what you learned
Researchers know their own work well. They remember collecting data, developing a method, adjusting experimental conditions, and analyzing results. It is therefore natural for a first draft to follow the order in which the work was done.
But the order in which researchers carry out a study may differ from the order in which readers need to understand it.
Readers need to understand why a method is necessary before learning the details of its development. Before reading results, they need to know what the experiment was intended to establish. By the time they reach a conclusion, they should already have encountered the evidence supporting it.
When organizing a paper, ask both questions:
“What did I do next?”
“What does the reader need to know next?”
Not every task deserves equal space simply because it required effort. Give priority to material that helps readers understand and assess the central claim.
2. Organizing your thinking through claims, reasons, and evidence
A useful way to examine an argument is to distinguish its claims, reasons, and evidence.
| Element | Role | Question to ask |
|---|---|---|
| Claim | A judgment or conclusion you want readers to accept | What am I arguing? |
| Reason | A basis for accepting that claim | Why do I make this judgment? |
| Evidence | Material supporting the reason | What establishes it? |
A claim may express the main contribution of the paper or a smaller judgment within a section or paragraph. Evidence may include measurements, comparative experiments, mathematical derivations, simulation results, observations, and prior research.
Reasons and evidence often overlap in everyday language, but separating them can help when reviewing a paper. A reason may itself be another judgment. Asking how that judgment was established takes you to the evidence behind it.
The Craft of Research treats these relationships as the basic structure of a research argument. It also emphasizes explaining what justifies the connections and responding to anticipated objections.[1]
There is no need to apply this structure mechanically to every paragraph. You might state a claim before presenting evidence, or describe observations before drawing a conclusion. The roles of the elements and the connections between them matter more than a fixed order.
3. Evidence still needs explanation
Students sometimes assume that a table or figure provides sufficient support for a claim. Presenting data and explaining what those data mean, however, are separate tasks.
Readers need to know more than the size of a number.
- What is it being compared with?
- Under what conditions was the result obtained?
- Does the chosen metric adequately represent the property being claimed?
- How far can the observed difference support a conclusion?
Without these explanations, readers must guess the author's reasoning. Something that seems too obvious for the author to mention may be exactly what a first-time reader needs.
In Toulmin's model of argument, a warrant justifies the inference from data to a claim. For a researcher writing a paper, it is useful to think of it through this question:[2]
“Why does this evidence support this claim?”
After presenting results, explain how they answer the research question. Repeating the numbers in a table as sentences does not, by itself, accomplish that task.
4. The question in the introduction must carry through to the conclusion
The overall flow becomes easier to assess when you consider the role of each part of the paper.
| Part of the paper | Role in the argument |
|---|---|
| Introduction | Establish the problem and the gap in existing knowledge, leading to the research question and objectives. |
| Methods | Explain the approach used to answer the question and why it is appropriate. |
| Results | Present evidence that allows the research question to be assessed. |
| Discussion | Explain what the results mean, how they relate to existing knowledge, and the limits of their interpretation. |
| Conclusion | State the answers and contributions supported by the preceding argument. |
Section structures vary across journals and types of research. Even so, readers should be able to trace the connections among the problem, the approach to testing it, the results, their interpretation, and the conclusion.
The flow breaks when an issue presented as important in the introduction is not addressed in the experiments, or when the conclusion emphasizes properties different from those actually measured. This can happen when the research question is too broad or when the introduction and conclusion have not been revised to reflect the study that was actually completed.
After finishing a draft, read the research objectives in the introduction alongside the conclusion. Simply checking whether the paper answers the question it initially posed can reveal substantial problems.
5. Every paragraph has a job
Even when the overall structure is clear, the reasoning within a paragraph can become scattered. A paragraph may contain several competing claims or collect related information without making a central point.
After reading a paragraph, you should be able to answer, “What is this paragraph saying?” The remaining sentences should help explain, support, or qualify that point.
Connections between paragraphs matter too. What readers learn in one paragraph should prepare them for the next. When a new problem appears without preparation, or an explanation assumes a concept that has not been defined, readers have to go back and reconstruct the argument.
Transitions signal relationships that already exist. Adding “furthermore,” “however,” or “therefore” between statements does not create a logical connection where none is clear.
Pay particular attention to sentences containing “therefore.” If readers can accept everything before it and still reasonably reject the conclusion that follows, a necessary premise or piece of evidence may be missing.
6. Engineering implications also need logical support
Engineering papers often need to explain practical implications alongside contributions to knowledge. Here, too, it is necessary to examine the distance between the findings and the benefits being anticipated.
An improvement in a metric, a functioning system, practical applicability, and economic benefits may be related, but they are different judgments. A result at one level does not automatically establish an outcome at the next.
Distinguish directly demonstrated results from potential applications inferred from them. When discussing an application, explain the conditions that would need to be met.
This distinction helps readers understand the contribution accurately and make informed decisions about further research or implementation.
Precision matters more than forceful wording. Claims with clearly stated conditions and scope are easier to assess and easier for later studies to build upon.
7. Limitations determine the scope of a claim
Limitations should not become a conventional paragraph added at the end of the discussion. If the data or study design have limitations, those limitations should also shape the scope of the claims made earlier.
Stating a broad conclusion with certainty and then adding “further validation is needed” does not repair the earlier reasoning. If something requires further validation, the main text and conclusion should distinguish what has already been established from what remains open.
This does not mean that every sentence must sound tentative. State well-supported findings clearly. Keep three categories distinct:
- What has been established
- How those findings are interpreted
- What remains an untested possibility
This distinction also helps during revision. When a claim lacks support, consider whether it requires additional analysis, a clearer explanation of existing evidence, or a narrower formulation.
8. Questions to ask after completing a draft
Before polishing individual sentences, consider the following:
- Can I state the central claim in one or two sentences?
- Do the results and conclusion answer the research question posed in the introduction?
- Can I point to the evidence supporting each important claim?
- Have I explained why that evidence supports the claim?
- Have I extended the conclusion beyond what the study directly established?
- Can I explain the role of each paragraph and its relationship to the paragraphs around it?
If these questions are difficult to answer, revise the structure before concentrating on smoother prose. Sometimes an explanation needs to be added. Sometimes material unrelated to the main argument needs to be removed.
Readers have not shared every step of the researcher's experience. They must understand the need for the study, assess the method, and follow the results to a conclusion using the paper alone. Improving the logic and flow of a paper means making that reasoning possible without leaving gaps for readers to fill.
References and further reading
[1] Booth, W. C., Colomb, G. G., Williams, J. M., Bizup, J., & FitzGerald, W. T. (2024). The Craft of Research (5th ed.). University of Chicago Press.
A useful guide to organizing claims, reasons, and evidence, explaining their connections, and responding to objections. Chapters 5–9 are particularly relevant.
Publisher's page
[2] Toulmin, S. E. (2003). The Uses of Argument (Updated ed.; original work published 1958). Cambridge University Press.
Useful for examining the assumptions and justifications that connect data to claims.
Publisher's page
Research and Academic Writing for Graduate Students