Research and Academic Writing for Graduate Students, Part 7
A colleague recently told me about a graduate student who was hesitating to begin writing a paper because the student could not judge whether the research was publishable. The supervisor had suggested that it was time to start writing, but the student remained doubtful about the prospects for publication. I have encountered similar situations with my own students. That conversation prompted me to put together some points that may help students assess whether their research is ready for publication.
“Professor, are these results enough for a paper?”
This is an understandable question. A student has spent months collecting data, developing a method, and running experiments. There are results. Yet, when it is time to write, the student is unsure whether the work is sufficient for publication. Is a small improvement over an existing method enough? Are more experiments needed? Should another methodological component be added?
Judging publishability is difficult when research experience is limited. Students may not yet be familiar with the contributions and validation expected by journals in their field. Even experienced researchers cannot predict an editorial decision with certainty.
However, difficulty predicting a review outcome should be distinguished from difficulty explaining what one's own research has established.
The former reflects uncertainty in scholarly evaluation. The latter may signal a need to revisit the research question, study design, and meaning of the results. If the work began with a development objective but no research question, or if the findings have not been positioned within the literature, judging publishability becomes difficult even for the researcher.
When a student asks this question, I therefore ask what the study has revealed before asking what has been developed.
The question to ask is not simply, “Have I done enough work for a paper?” It is, “What has this study added to knowledge, and is the evidence sufficient to support it?”
1. Why should research results be published?
For graduate students, papers may be graduation requirements or credentials for employment and further study. Those purposes, however, do not fully explain why research should be published.
Research seeks answers to questions that are not yet adequately understood. If an answer remains on a researcher's computer or in an internal laboratory report, others have little opportunity to examine or use it. Making findings public allows someone else to formulate the next question, test the findings under different conditions, or apply them to a practical problem.
Publishing a paper turns an individual researcher's results into an account of knowledge that others can examine and use. Published knowledge may subsequently be revised or challenged. Scholarship advances through that continuing process of scrutiny and revision.
Writing also matters to the researcher. Something that seemed clear during an experiment often becomes less clear when it must be explained in words. Connecting the motivation, choice of method, and interpretation of the results can reveal gaps in the study design.
Writing therefore need not wait until every aspect of a study is complete. Once the question and main results have taken shape, drafting can help identify missing evidence and guide additional analysis or experiments. A supervisor's suggestion to start writing may be an invitation to undertake precisely this assessment together.
Rather than waiting for certainty about publication, it is often more useful to write and identify the contribution and remaining weaknesses concretely.
2. Scholarly contribution is central to publishability
Students sometimes present the work they have performed as the scholarly contribution.
“I developed a new model.”
“I integrated several types of data.”
“I built an automated system.”
Such work may be necessary for a study. These statements alone, however, do not explain its scholarly contribution. Readers want to understand how the work changes existing knowledge, as well as what it took to develop the method or system.
Consider the following two descriptions.
“A model was developed to detect infrastructure damage by combining images and sensor data.”
“Under low-light conditions in which images alone were insufficient to distinguish damage, sensor information reduced detection errors; the study identified the conditions under which this benefit appeared and those under which it was limited.”
The first describes what was developed. The second describes what was learned. For the second statement to constitute a contribution, the finding must not already be adequately established in the literature, and it must be supported by appropriate experiments.
Scholarly contributions are not limited to new theories or algorithms. A study may identify conditions under which an existing explanation fails, explain why earlier findings conflict, or provide reliable empirical evidence about an important phenomenon. A new dataset or measurement method can also be valuable if it enables questions that were previously difficult to address.
Performance improvements can be contributions too. But an explanation that ends at “accuracy improved” leaves their significance underdeveloped. The researcher should be able to explain which limitation was addressed, whether the improvement is consistent, when it occurs, and what additional costs or constraints accompany it.
A useful exercise is to complete the following sentence.
“Previous research has not adequately established ______. Through ______, this study found ______.”
If it is difficult to make that statement specific, first consider whether the question and contribution are clear, before assuming that more experiments are needed.
3. Why is it difficult to judge whether one's research is publishable?
There are several possible reasons. A student may have limited experience reading and writing papers or may apply an excessively demanding standard to personal results. When examining the research itself, however, I pay particular attention to two questions.
First, was the study conducted after a clear research question had been formulated?
Second, does the researcher understand where the study stands within existing knowledge?
Setting only a development objective
“Developing an AI-based prediction system” is a development objective. By itself, it does not identify the scholarly question the study is intended to answer.
For example, “Does incorporating physical constraints into learning improve predictive generalization when observations are scarce?” can be a research question. Addressing it requires varying the amount of data, comparing models with and without physical constraints, and evaluating performance under conditions not used for training.
If system development is the only objective, a student might instead compare accuracy against an existing model once, using the largest available training dataset, and stop there. This produces results, but it does not establish the effect under limited-data conditions.
The number of experiments matters less than whether they were designed to answer the research question.
If a hypothesis is proposed, the researcher should consider in advance which outcomes would support it and which would count against it. Testing only conditions that produce the expected result is not an adequate examination of the hypothesis.
The language of “accepting” and “rejecting” hypotheses also requires care. Empirical studies often establish support under particular conditions and data rather than prove that a hypothesis is universally true. In statistical testing, failure to reject a null hypothesis does not establish that it is true.
Not every study requires formal hypothesis testing. Exploratory and methodological studies are also legitimate. Regardless of study type, however, it should be clear what the researcher intended to learn and what evidence would justify an answer.
Failing to connect findings to existing knowledge
Even with experimental results, a researcher who does not sufficiently understand the literature will struggle to interpret their significance. Something new to the student may already be well known in the field. Conversely, a numerically small difference may expose an important limitation of an established assumption.
A literature review is more than collecting papers to cite in the introduction. It distinguishes what is known from what remains uncertain and establishes where the new findings belong.
This is the positioning of the research. Using a recent technology does not, by itself, explain that position. The researcher must be able to state what knowledge is added relative to previous studies.
Some students respond to uncertainty about publishability by adding another module or experiment. Before doing so, check the following connection.
Gap in existing knowledge → Research question → Study design → Results → New knowledge
If that connection is broken, doing more work may not solve the problem.
4. What does peer review assess?
Students preparing their first paper should understand the process between submission and publication.
Papers commonly reach publication after revision and further assessment rather than acceptance exactly as submitted. Authors may receive revision requests after the first review, submit a revised manuscript and response letter, and then receive further requests. Multiple rounds are common, but there is no fixed number: the process depends on the journal and the manuscript. 1
A typical sequence is as follows.
Submission and initial editorial assessment
The editor assesses the fit with the journal, the clarity of the contribution, and whether the manuscript is ready for external review. A paper may be rejected at this stage without external review. This is called a desk rejection.Review by subject specialists
Researchers in the field examine the importance of the question, originality of the contribution, validity of the methods, and connection between the results and conclusions.Editorial decision and author revision
The editor considers the reviews and decides whether to reject the manuscript, invite revision, or accept it. Authors invited to revise modify the manuscript and submit a response letter explaining how each comment has been addressed.Reassessment of the revised manuscript
The editor or reviewers examine the changes and responses. Unresolved issues may lead to further revision. Not every revised manuscript returns to external reviewers; the editor may make the assessment directly. 2Acceptance and preparation for publication
After acceptance, the paper proceeds through production steps such as editing and proof correction.
A request for revision should therefore not be treated as evidence that the study has failed. Equally, an invitation to revise does not guarantee acceptance. What matters most is whether the substantive issues have been resolved, rather than which round of review the paper has reached.
How did peer review begin?
Peer review is a system in which researchers assess manuscripts written by others in their field. It is familiar today, but it did not begin in its present form.
Accounts of its history often refer to Philosophical Transactions, established in Britain in 1665. It would be inaccurate, however, to assume that modern external peer review was already established at its launch. Early publication decisions were primarily made by the editor. When the Royal Society formally took over the journal in 1752, a committee-based assessment system was introduced. 3
In 1832, the Royal Society introduced a more formal review process drawing on the expertise of its Fellows. This history illustrates a gradual development from editorial judgment and assessment within a learned society toward specialist evaluation. It is therefore more accurate to understand peer review as a system that developed in stages than to assign the entire modern institution a single founding date. 4
The rationale is to expose authors' claims to assessment by other knowledgeable researchers, rather than leave them solely to the judgment of the authors or editor.
What do reviewers look for?
Peer review does not primarily assess how long the authors worked or how much they produced. It asks whether the resulting claims are academically meaningful and supported by the evidence presented.
Reviewers commonly consider questions such as these:
- Is the research question important and clear?
- Does the paper contribute knowledge or methodology beyond existing studies?
- Are the study design and comparisons appropriate for the question?
- Do the results support the conclusions?
- Have plausible alternative explanations and important limitations been examined?
- Is the method explained sufficiently for others to understand and scrutinize it?
Reviewers provide comments and recommendations; the editor weighs them and makes a decision. This is not simply a majority vote among reviewers.
Suppose a reviewer says that comparative experiments are insufficient. Before deciding that one additional experiment will resolve the issue, identify what the reviewer could not establish. The concern may be the effect of the proposed method, the fairness of the comparison, or whether the result holds only for a particular dataset. Each requires a different response.
Insufficient explanation can be addressed through clearer writing. Insufficient evidence cannot be resolved by changing the wording alone.
Peer review is not perfect. Reviewers differ in expertise and perspective, and they may misunderstand a manuscript or overlook something important. Authors do not have to accept every suggestion uncritically. When they disagree, they can explain their position respectfully with evidence and reasoning.
Before concluding that a reviewer has misunderstood, however, consider whether the manuscript invited that reading. Logic that is familiar to the author may not be clear to the reader.
What students should learn from peer review goes beyond wording that satisfies reviewers. It includes exposing their claims to other researchers' questions, strengthening the evidence, and defining precisely how far the claims can extend.
5. When and how should a target journal be selected?
It is useful to consider possible journals early in a study rather than first thinking about them after the research is complete. Doing so helps clarify which readers need the knowledge the study is intended to produce.
Start with the journal's scope and audience. Even within AI research, a journal focused on algorithmic advances may expect different explanations and validation from one focused on contributions to construction and infrastructure.
Next, read recent related papers in the journal. Go beyond titles and abstracts: examine the questions they address and the evidence used to support their conclusions.
| Consideration | Question to ask |
|---|---|
| Topic and audience | Why should this journal's readers know about my findings? |
| Type of contribution | Is the emphasis on methodology, explanation of a phenomenon, or empirical analysis? |
| Validation | Are the comparisons and tests needed to support my claims in place? |
| Scope of claims | Am I generalizing too far from a single case? |
| Submission conditions | Do the article type, length requirements, and costs fit the study? |
Choosing a target journal is about finding a community that can evaluate and use the contribution, rather than simply following a journal's current trends. The target can also change as the findings develop.
However, selecting a less competitive journal does not resolve an unclear contribution or invalid evaluation. Journals may differ in the scope of contribution they expect, but sound evidence is necessary wherever the work is submitted.
6. How do international and domestic journals differ?
It is common to assume that journals published abroad are stronger and domestic journals weaker. The country of publication, however, is not a reliable measure of scholarly quality. Korean societies publish international journals, and publication overseas does not automatically make a journal trustworthy.
More useful distinctions concern the main readership and the context of the problems addressed.
For a journal serving international readers, authors need to explain why a case from a particular country or site matters to researchers elsewhere. A study of Korean infrastructure or institutions can have international significance when its findings address a broader scholarly question.
A journal primarily serving domestic readers can discuss problems closely connected to national institutions, design standards, and industrial conditions in detail. Communicating usable knowledge to Korean practitioners in Korean is also an important role.
| Aspect | Journals primarily serving international readers | Journals primarily serving domestic readers |
|---|---|---|
| Main readership | Researchers and practitioners across countries | Researchers and practitioners within the country |
| Context to explain | Provide enough background for readers unfamiliar with the setting | Discuss national institutions and field conditions in detail |
| Presentation of contribution | Emphasize principles, methods, and findings relevant beyond the immediate setting | Emphasize knowledge and applications relevant to domestic problems |
| Shared requirements | Clear questions, valid methods, sufficient evidence, and research integrity | Clear questions, valid methods, sufficient evidence, and research integrity |
These are not rigid boundaries. A journal serving domestic readers may also publish broadly applicable theories and methods. The more useful question is who should read the study and what knowledge those readers should gain.
7. How should JCR Impact Factors, quartiles, and percentiles be interpreted?
When looking for a target journal, students encounter questions such as “What is its Impact Factor?”, “Is it a Q1 journal?”, and “Where does it rank within the field?” These indicators describe citation influence and relative position within a category. They serve as supplementary indicators when assessing a journal's standing. To interpret them properly, it is necessary to understand what each measures.
What does the Impact Factor indicate?
Journal Citation Reports (JCR) is Clarivate's journal citation analysis resource. One of its main indicators is the Journal Impact Factor (JIF), often shortened to IF.
For a given year, the JIF divides citations received in that year to the journal's publications from the preceding two years by the number of eligible scholarly items, such as articles and reviews, published in those two years. The publication types whose citations enter the numerator do not correspond perfectly to those counted in the denominator, so the JIF is not strictly the simple average citation count of every item published. 7
For example, suppose a journal published 100 eligible papers in 2023 and 2024, and its publications from those years received 600 citations in 2025. Its 2025 JIF would be 6.0. The numerator counts citations occurring in 2025 alone, not citations accumulated over the preceding two years.
A higher IF indicates that recent publications in the journal are being cited relatively actively. It does not mean that every paper in the journal receives a similarly high number of citations. Citations may be concentrated in a small number of papers.
Fields also differ in the number of researchers, publication volume, reference-list lengths, and citation speed. Direct comparisons between journals in different fields based only on IF therefore have limitations.
What are Q1, Q2, Q3, and Q4?
Quartiles help assess a journal's relative position within a field. The commonly cited JCR quartiles, Q1 through Q4, group journals according to JIF rank within a subject category. A category identifies the subject field; a quartile identifies a position within its ranking. 5
| Quartile | Approximate position in the category's JIF ranking |
|---|---|
| Q1 | Top 25% |
| Q2 | Below the top 25%, through the top 50% |
| Q3 | Below the top 50%, through the top 75% |
| Q4 | Remaining 25% |
This is a conceptual description. Tied ranks can mean that the four groups do not contain exactly equal numbers of journals. A journal assigned to multiple categories may also have a different quartile in each. When describing a journal as Q1, check both the year and the category. 6
How does a percentile differ from “top X percent”?
Quartiles divide journals into four groups. Percentiles describe relative position more finely. A higher JIF percentile indicates a higher position within the category. 5
Do not confuse a percentile with “top X percent.” A percentile of approximately 90 means roughly the top 10%, not the top 90%.
| Expression | Approximate interpretation |
|---|---|
| 90th percentile | Top 10% |
| 75th percentile | Top 25% |
| 50th percentile | Around the middle of the category |
Exact values depend on the ranking method and treatment of ties. Use the values reported in JCR when checking a particular journal.
How much do these indicators tell us about journal quality?
The IF, quartile, and percentile are useful reference points when assessing a journal's standing. They help show how actively a journal is cited and where it sits relative to other journals in its field. Consulting them when comparing potential target journals is reasonable.
Journal quality, however, is not fully captured by a single number. Subject expertise, editorial and review practices, research integrity, readership, and influence on scholarship and practice also matter. Citation indicators principally describe the influence reflected in citations.
Journal-level indicators must also be distinguished from the value of an individual paper. Publication in a Q1 journal does not by itself establish the validity of a particular paper's methods and conclusions. A paper in a Q3 or Q4 journal may still provide a reliable answer to an important question. Individual papers must ultimately be assessed by reading their questions, methods, evidence, and findings.
It is therefore inappropriate to assume that a certain number of experiments qualifies a paper for Q1 and slightly fewer for Q2. Journals differ in audience, scope, editorial policy, and expected contribution.
My advice is to first clarify the knowledge the study contributes, then identify journals whose readers need that knowledge, and finally examine both their recent papers and citation indicators. Metrics can inform journal selection, but they cannot explain the contribution of the study on the researcher's behalf.
8. What should you be able to explain before submission?
When assessing publishability, try answering the following six questions.
What is not yet adequately known?
Can you identify a gap in knowledge with reference to the literature?What question does this study answer?
Can you describe what you sought to learn, rather than merely what you developed?Can the study design answer that question?
Are the comparisons and evaluation methods appropriate, and have alternative explanations been considered?What has been learned from the results?
Can you explain the meaning of the findings beyond listing performance values?How far can the claims extend?
Can you distinguish what has been established from what remains untested?Why should this journal's readers know about the findings?
Is there a clear connection between the contribution and the target readership?
Clear answers do not guarantee acceptance. They do, however, establish a basis for judging readiness. Even if additional experiments are needed, the researcher can explain what those experiments are intended to establish.
If the answers are unclear, increasing complexity or the number of experiments is not automatically the next step. Return to the literature and clarify the question and the position of the findings. If the existing experiments cannot answer the question, improve the design.
An interesting question may also emerge after the results have been examined. That is legitimate, but a question discovered after the fact should not be presented as though it had been specified from the outset. Explain that the finding emerged through exploration and, when necessary, examine it using independent data or additional experiments.
Nor should a study be dismissed as unpublishable merely because an expected hypothesis was not supported. If an important question was investigated with a valid method, an unexpected result may help reveal the limits of existing knowledge. Conversely, even an improvement in the expected direction is difficult to defend if the comparison was unfair or the interpretation insufficiently supported.
Researchers cannot guarantee an editorial outcome in advance. They should nevertheless be able to explain the question, the evidence obtained, and what the results add to existing knowledge. Discussions with a supervisor become more useful when these points are laid out concretely, rather than reduced to “Will it be accepted or not?”
When a student asks whether the work is enough for a paper, the first thing I want to hear is not how much work has been done.
“What was previously unknown, what has this study established, and what evidence supports that conclusion?”
Being able to answer in your own words can help you move beyond vague anxiety about publication. Even if the answer is not yet complete, it clarifies what needs to be organized or strengthened. Writing a paper is itself part of making that answer clear.
References
Research and Academic Writing for Graduate Students