What Is Plagiarism and How to Avoid It in a Thesis
What Is Plagiarism in a Thesis
Plagiarism means using another person's ideas, arguments, data, wording, images, or structure while presenting them as your own without properly acknowledging the source. In a quantitative study, this mistake can appear in the literature review, theoretical foundation, scale, methodology, interpretation of results, or reference list.
You do not need to copy an entire long paragraph to commit plagiarism. A sentence translated from an English source, a research model taken from an article, or a scale table reproduced without citing its source also needs to be acknowledged. Changing a few words, reversing the order of two clauses, and leaving out the citation does not create an independent argument.
If you are working on a graduation thesis, your goal is not to force the software to report the lowest possible number. Your goal is to make it clear to the reader which ideas come from the literature, which data you collected, which analysis you performed, and which interpretations are your own.
A Standard Structure for Avoiding Plagiarism
You need to address plagiarism throughout the research process, from choosing the topic to exporting the final version. The table below helps you divide the work into manageable parts instead of waiting until the night before submission to upload the entire file to a checking tool.
| Section to check | What you need to do | When to do it |
|---|---|---|
| Ideas and model | Clearly record the articles, books, or studies that provide the basis for the model, hypotheses, and relationships between variables | When developing the proposal |
| Literature review | Summarize the sources in your own understanding and cite them immediately after the main idea | While reading the literature |
| Scale and questionnaire | State the source of the indicators, the adaptation context, and the translation method | Before collecting data |
| Data and results | Distinguish your own data from illustrative output or data reported in previous studies | When running SPSS or SmartPLS |
| Reference list | Compare every in-text citation with the final reference list | Before checking Turnitin |
| Final draft | Read the similarity report paragraph by paragraph rather than looking only at the total percentage | After completing the content |
You can also review these Graduation thesis topics to check whether your topic has a sufficiently clear scope. A topic that is too broad often leads you to collect paragraphs from many different sources and combine them without keeping track of where they came from.
Steps for Writing the Research Without Plagiarism
Step 1: Finalize the Research Question and Literature Scope
Before searching for articles, write a short paragraph describing the topic, survey population, dependent variable, independent variables, and research context. For example, you might study students' intention to use e-wallets in Ho Chi Minh City. This scope helps you search for literature on usage intention, financial technology, students, and suitable models instead of downloading dozens of unrelated papers.
Once your topic selection method is clear, create a tracking table with the source title, year, objective, research sample, main variables, findings, and the section where you expect to use the source. Do not save only the URL. After several weeks, a URL will not tell you which idea belongs to which author.
Step 2: Read the Literature and Take Notes in Your Own Language
When you reach an important passage, close the article or switch to your notes window before rewriting it. Record the main idea in the way you understand it, then reopen the source to verify the terminology, numbers, and author names. This reduces the chance of unconsciously copying the original sentence structure.
Each note should have one of three labels: direct quotation, paraphrase, or my idea. A direct quotation must be placed in quotation marks and include the page number if your department requires it. A paraphrase still needs a citation because changing the wording does not make the original idea yours.
Step 3: Organize the Literature by Theme Instead of Combining Articles One by One
Your literature review needs to show how the studies connect to one another. You can organize it by concept, variable group, context, or the development of the model over time. Each paragraph should include a topic sentence, evidence from several sources, similarities and differences, and a transition to the research gap.
For example, instead of writing four paragraphs that begin with the names of four authors, write: “Studies of perceived usefulness often treat perceived usefulness as a factor explaining technology usage intention. However, findings among younger users may differ depending on the service context and their familiarity with the platform.” Then place the relevant sources after the claims they actually support.
You need to check whether the source really says what you are writing. A study that demonstrates a relationship in banking does not automatically demonstrate the same relationship in higher education.
Step 4: Add the Citation as Soon as You Introduce the Source
Do not leave citations until the final step. When you write a claim based on the literature, insert the source in the sentence or at the end of the paragraph. For example: “Cronbach's Alpha is used to assess the internal consistency of indicators within the same scale (Cronbach, 1951).”
If you use the Alpha criterion of 0.7, you can state the source clearly (Nunnally, 1978). For an exploratory scale, some studies accept Alpha from 0.6 and require an explanation based on the context, supported by (Hair et al., 2010). This type of citation shows the committee where your number comes from instead of presenting it as a rule you invented.
Step 5: Record the Sources of the Scale, Data, and Model
In quantitative research, plagiarism is often overlooked in the scale table. If you use indicators from a previous study, record the author, year, original construct, adaptation method, and source for each group of indicators. Do not present the questionnaire as if you created all of it yourself when its foundation comes from a published scale.
The same applies when your model is based on TAM, TPB, UTAUT, or SERVQUAL. State the appropriate theoretical sources. For example, TAM is associated with (Davis, 1989), TPB with (Ajzen, 1991), and UTAUT with (Venkatesh et al., 2003). You can adapt the model to the Vietnamese context, but you must identify which parts are inherited from earlier research and which parts are your proposal.
Step 6: Check the Match Between Citations and References
Create two lists and compare them manually. The first list contains every author and year that appears in the text. The second contains every source in the reference section. Every in-text citation must have a corresponding reference, and every source in the reference list should be used somewhere, unless your department allows you to list sources you have read but not cited.
Check the spelling of author names, publication years, author order, and titles. A source that appears in the reference list but is never cited does not make an unsupported paragraph acceptable. Conversely, adding a source at the end of a paragraph when that source does not support the claim is also an academic integrity problem.
How to write this in your thesis
You can replace the bracketed sections with information from your actual topic. The sentences below are adaptable wording frameworks, not substitutes for reading and citing the original sources.
| Situation | Adaptable sentence |
|---|---|
| Introducing a concept | “[Concept] is understood as [brief definition], according to the approach of [author, year].” |
| Synthesizing studies | “Previous studies show that [variable X] is related to [variable Y], but the findings differ depending on [context or population].” |
| Stating a gap | “Within the literature reviewed, evidence concerning [relationship or context] in [location] remains limited.” |
| Adapting a scale | “The [construct] scale was adapted from [author, year], with the wording adjusted to suit [survey population].” |
| Describing the data | “The analysis included [number] valid observations collected from [population] during [period].” |
| Reporting results | “The results show that [variable X] has a [positive or negative] effect on [variable Y], with [coefficient or p-value] equal to [value].” |
| Explaining a source | “The [criterion name] criterion was selected based on [source], in line with the [exploratory or confirmatory] objective of the study.” |
A results sentence must use the actual figures from your .sav or .csv file. Do not enter a number that looks attractive simply to complete the paragraph. If you are running a regression, for example, the Durbin-Watson condition between 1 and 3 is supported by (Field, 2013), while your own result must come from the Model Summary table in SPSS.
When reporting PLS-SEM, distinguish measurement indicators from structural model indicators. Outer loading from 0.7 upward can be supported by (Chin, 1998), while HTMT below 0.85 or 0.90, depending on how conceptually close the constructs are, can be cited to (Henseler et al., 2015). Do not add CFI, TLI, or RMSEA to your thesis simply because you saw them in a CB-SEM paper.
A Short Plagiarism Example
Suppose you are writing a thesis about the factors affecting students' intention to use online learning applications. A weak paragraph might read: “According to many studies, perceived usefulness affects usage intention. Perceived ease of use also affects usage intention. These two factors are very important in technology acceptance.” The paragraph is too general, does not identify the studies used, and makes the phrase “very important” without evidence.
A stronger version would be: “In technology acceptance research, perceived usefulness describes the extent to which users believe that a system improves their task performance, while perceived ease of use reflects the extent to which users perceive the system as easy to use. These two concepts are foundational components of Davis's Technology Acceptance Model (Davis, 1989). This study tests whether the two concepts explain students' intention to use online learning applications.”
The second paragraph still requires you to check the definitions against the original source. The phrase “this study” identifies your scope and choice, while the TAM foundation is clearly acknowledged. If you add hypotheses, separate the argument from previous research from the hypothesis you propose.
For quantitative results, a shortened paragraph could read: “The regression results show that perceived usefulness has a positive coefficient for usage intention. This conclusion is based on the sign of the coefficient and the statistical significance level in the Coefficients table, not on the writer's impression. The specific value should be reported from the analysis of a survey sample consisting of [number of valid observations] observations.”
If this is a draft, replace every bracketed section with the actual figures, state the coefficient and significance level, and report the decision concerning the hypothesis. Do not use figures from the reference article as a substitute for your own data.
Common Problems That Lead to Follow-Up Questions from the Committee
The first mistake is assuming that replacing words with synonyms is enough to avoid plagiarism. The committee can recognize the same argument structure, the same sequence of ideas, and the same example even after the wording has changed. You need to read, understand, synthesize, and cite the source instead of relying only on a paraphrasing tool.
The second mistake is placing one citation at the end of an overly long paragraph. The reader cannot tell whether the source supports the entire paragraph or only the final sentence. Divide the paragraph according to each idea and place the source close to the claim it supports.
The third mistake is using a secondary source as if it were the original source. You read a Vietnamese thesis, see that its author mentions a model from an international article, and then add the international article to your reference list without reading the original. The safer approach is to find and read the original source. If you cannot access it, follow your institution's rules for secondary citations instead of implying that you read the original.
The fourth mistake is deleting every passage flagged by the software. A similarity report may highlight model names, technical terms, source titles, method descriptions, or required wording. Read each matching source and determine whether the match is a valid quotation or an uncited copied passage.
The fifth mistake is confusing the similarity percentage with a plagiarism decision. Turnitin and similar tools help identify overlap, but the final decision also depends on the citation method, department rules, and the supervisor's assessment. You can review the Turnitin check guide to prepare the draft and read the report source by source.
The sixth mistake is paying someone to write the thesis or accepting an analysis file without the original data. The risk is not limited to plagiarism. When the committee asks why you removed an indicator, why you selected a model, or how a result was generated, you may not be able to explain it. For a thesis and a research project, every results table should be traceable to the data file and analysis procedure.
Frequently asked questions
What is plagiarism if I have changed the wording?
If you are still using another person's idea, argument, data, or structure, you still need a citation even when you have rewritten it in your own words. Good paraphrasing makes the text clearer and shows that you understand the source, but it does not remove the original author's right to acknowledgment.
What Turnitin percentage is considered safe?
There is no single number that applies to every university, department, and type of assignment. A report with a low percentage can still contain an important copied passage if the source is not cited. Review each matching group, exclude sections that your institution permits you to exclude, and ask your supervisor about the internal threshold.
Does a direct quotation count as plagiarism?
A direct quotation may be acceptable when it is marked correctly, fully cited, and used only when necessary. If you copy many long passages even with citations, the thesis may still be judged excessively dependent on the literature. In a quantitative thesis, prioritize synthesis and paraphrasing, keeping the original wording only when a definition or statement requires a high degree of precision.
Is using ChatGPT to rewrite a source considered plagiarism?
A rewriting tool cannot replace source verification. If you put another author's paragraph into a tool and use the result without citing it, the underlying use of the idea has not changed. You also need to verify every author name, year, definition, and number because a tool may generate nonexistent sources or assign the wrong content to a real source.
At what stage should I check for plagiarism?
Check at least twice. Run the first check after completing the draft so you can fix missing sources, unmatched citations, and copying that occurred while synthesizing the literature. Run the final check after locking the content, formatting, tables, and reference list because late edits can create new matching passages.
Open your thesis file, make a list of the passages that use sources without citations, and then check each source before uploading the final version to the checking tool. If you need help turning this process into a draft with citations, results tables, and consistent formatting, you can review the M5 chapter-writing module from DoThesis.