Paraphrasing in a thesis means expressing an idea from a source independently while preserving its meaning and acknowledging its author. Replacing words with synonyms is not enough. You also need to check whether your revision has changed the participants, figures, conditions or certainty of the claim.
The difficult cases are often sentences that sound smoother than the original but no longer say the same thing. “Participants reported” becomes “people do”; a difference between groups becomes an improvement following an intervention. This guide gives you academic paraphrasing practice using a short example from a genuine publication and a complete, explicitly fictional miniabstract. It also shows when quotation or summary is a better choice.
Paraphrase, quotation and summary: choose the purpose
First decide why you are introducing the publication. Do you need the exact wording of a definition, a detailed account of one finding, or a short overview of a longer argument? That purpose determines the form. You do not have to turn every quotation into a paraphrase or expand every summary.
- A quotation preserves the exact wording. Mark it according to your chosen style and provide the source and required locator. It is particularly useful when you are analysing the author's language itself.
- A paraphrase presents a selected idea from another writer using your own construction. It retains important conditions and still requires a citation.
- A summary selects the main points of a longer passage. It may omit details that do not serve the current purpose, but that must not change the meaning. A summary also identifies its source.
A theoretical chapter can use all three forms. Your own commentary does a different job: comparison, evaluation of relevance or identification of a limitation. Readers should be able to tell where your account of the source ends and your reasoning begins. For the mechanics of referencing, see the guide to bibliographies and footnotes.
How to paraphrase academic sources step by step
Read the passage in context
Do not start by rewriting its first sentence. Establish the question the author is answering, the meaning of key terms and whether the passage reports a finding, proposes a hypothesis or discusses another publication. Read the preceding and following sentences. They may contain a condition that changes the meaning of your entire note.
Record what you must preserve
In your working note, separate the subject, the finding and its limits. For research, add the design, recruitment method, denominator for a percentage or mean, and timing of measurement. Retain expressions such as “may”, “in this sample” and “self-reported” when they qualify the claim. They are not decorative words to remove for brevity.
Put the source aside and draft your version
Explain the idea to someone familiar with the subject but not with this particular article. Build a sentence that serves your paragraph's question. If you can remember only individual words without looking back, return to the reading. For a difficult passage, first explain its meaning simply, then refine the academic wording.
Compare both meaning and wording
Check whether you have added a cause, certainty or generalisation. Then assess whether the expression is independent. Retain the correct disciplinary terms rather than replacing them with approximate alternatives merely to look different. If you need a distinctive phrase exactly as the author wrote it, treat it as a quotation and mark it appropriately.
Add the citation and connect it to your argument
The reference should allow readers to identify the publication and understand which material comes from it. Next check what your paraphrase contributes to the paragraph. A citation alone does not explain why you discuss this particular study. Show its relevance to your chapter's question while keeping the source's finding separate from your own commentary.
Barbara Braid describes this sequence in the chapter “Pisanie artykułów naukowych” in the Strefa Nauki handbook. Its purpose is understanding followed by verification, rather than mechanically counting replaced words. No single number of changes can by itself establish that a paraphrase is acceptable.
A genuine publication: what must not disappear?
A sampling limitation changes the scope of the sentence
Genuine source, original commentary
Anna Dolot's 2020 article examines remote work: 327 purposively selected participants during the pandemic. The methods state: “Dobór próby badawczej miał charakter celowy”.
Dolot (2020) describes the experiences of 327 purposively selected people working remotely during the pandemic. The findings concern the participants, not automatically all workers.
“Dolot proved how all employees view remote work” goes beyond the source. The revised paraphrase retains attribution and scope. You can use it to build your argument, but cannot add representativeness because a shorter sentence sounds more confident. Full publication details appear in the sources below.
Notice that names, sample sizes and design labels do not need decorative synonyms. Your task is to explain their significance independently. Comparing several publications requires an additional step: establishing whether they concern comparable phenomena. The guide to synthesising a literature review covers that task.
A full exercise: the numbers remain but the conclusion changes
The miniabstract below and all its figures were invented solely for paraphrasing practice. They describe neither a completed study nor an existing publication. Do not attach a fictional author, year or DOI, or include them in the reference list of your thesis.
Read the fictional miniabstract
Practice material
A one-time questionnaire study aimed to describe the association between self-reported weekly planning and perceived study overload. Participation was voluntary; 120 students from two programmes at one university responded. Regular planning was reported by 50 participants; the remaining 70 reported irregular planning or none. The overload question, rated from 1 to 5, received 48 and 66 responses respectively, with means of 2.8 and 3.4. In this sample, regular planning was accompanied by a lower self-reported overload rating. No test of the difference or intervention was conducted, so the comparison does not establish whether planning reduces overload. Volunteer recruitment limits generalisation.
Before reading the revision, identify who responded, how many participants contributed to the means, and whether the study compared two groups or the same people at two times. These are the details most easily lost when shortening a passage.
Compare two paraphrases
A version that changes the meaning
A study of 120 students proved that weekly planning significantly reduces study overload. Participants who started planning reduced their overload from 3.4 to 2.8 points.
A paraphrase faithful to the fictional source
A one-time questionnaire involved 120 volunteers from two programmes at one university. Of the 50 regular planners, 48 rated their overload, giving a mean of 2.8 on a scale from 1 to 5. In the second group of 70, 66 answered this question, with a mean of 3.4. This describes a difference between two groups' self-reports. Without an intervention or a test of the difference, an effect of planning was not established; volunteer recruitment also limits generalisation.
The revised version refers only to the fictional miniabstract above. In this exercise, that acknowledgement is not replaced with a fabricated citation to a nonexistent publication.
- “Proved” and “reduces” turn a description of an association into a causal claim.
- “Significantly” in this context suggests a statistical test that was not conducted.
- “From 3.4 to 2.8” suggests a change after participants began planning. The source instead compared two groups at one time.
- 120 participants is not the denominator for both means. There were 114 overload ratings: 48 and 66. Six participants did not answer this question.
- One university and volunteer recruitment limit the observation's scope. Removing them can imply a broader conclusion than the material supports.
The mean difference is 0.6 scale points. It is neither an improvement in the same people nor a difference of 0.6 percentage points. The means alone do not establish statistical or practical significance. When revising an account of a genuine publication, also check how the authors calculated the result and which analyses they actually performed.
How to shorten the passage and separate your own commentary
If your paragraph only needs to introduce a cautious observation, you do not have to reproduce every figure. You can summarise the fictional miniabstract and keep the details in your notes. If your argument depends on the size of the difference or missing responses, however, the relevant information belongs in the text.
In a volunteer questionnaire at one university, regular weekly planners reported lower average study overload than the other participants. A one-time comparison without an intervention or a test of the difference does not establish that planning caused the observed difference.
This summarises one selected idea from the exercise. You could separately add your own comment: “A subsequent study should distinguish planning from other differences between the groups more clearly.” That proposal belongs to the writer discussing the source. It must not be presented as a questionnaire finding or attributed to the source's authors if they did not make it.
Where should the paraphrase citation go?
Place it where readers can identify the idea it supports. Beginning with “Dolot (2020) describes…” immediately attributes the finding. If you then move to your own commentary or another publication, signal that change. A citation only at the end of a whole paragraph can be ambiguous if you have already mixed several sources.
There is no single format for every thesis. Your chosen style determines the reference structure and locator requirements. A page number is particularly useful when readers need to find a specific idea in a long publication; whether it is compulsory for a paraphrase depends on the style and local instructions. Give the page in the actual edition, not the preview screen number. Never invent an unchecked page number.
A reference-list entry alone does not identify which sentences came from that publication. The same applies when you paraphrase material from another language. Translation does not turn someone else's concept into your finding. If you know a source only through another publication, do not imply that you read the original: use the appropriate secondary-citation format or obtain the primary text.
What if your paraphrase is still too close?
Stop editing word by word. Write a one-sentence answer to: “What did the author establish here, and what does it not demonstrate?” Then build a new passage around your argument before comparing it with the original. You do not need to force changes to specialist concepts, scale names or established terminology. An unsuitable synonym can reduce precision.
If you still need the author's exact wording, use a properly marked quotation. Not every difficult passage needs a longer paraphrase. A good chapter includes your explanations and comparisons rather than a chain of rewritten accounts. For paragraph construction, see the examples of developing a theoretical chapter.
An automated tool may suggest fluent wording, but that does not verify fidelity to the source. Check each changed condition, figure and citation, alongside the rules for tool use in your work. Do not judge correctness solely by a similarity percentage. If you have a report, see the guide to JSA and similarity reports.
Paraphrasing checklist before sending your chapter
- I understand the passage and have checked its context in the publication.
- My wording preserves the concepts, scope, figures and important qualifications.
- I have not turned an opinion into a fact, association into causation, or a group difference into change over time.
- Distinctive expressions taken word for word are marked as quotations.
- The citation identifies a genuine, checked source; its locator is not a guess.
- Readers can distinguish the source's finding, a summary and my own commentary.
- The passage answers my paragraph's question rather than merely making the chapter longer.
Sources and their use
- Natalia Kohtamäki, “Praktyczny przewodnik antyplagiatowy”, Faculty of Law and Administration, UKSW, publication date unverified, pp. 9 to 11. Used for the distinction between paraphrase and quotation and the need for attribution.
- Barbara Braid, “Pisanie artykułów naukowych”, in “Podręcznik powarsztatowy Strefy Nauki”, University of Szczecin, 2021, pp. 50 and 51. Used for paraphrasing and note-taking methods.
- University of Wrocław Anti-Plagiarism Support Team, “Czy parafrazowanie zwalnia z zaznaczania cytatu?”, 2024. Used to distinguish independent composition from surface-level word substitution.
- Anna Dolot, “Wpływ pandemii COVID-19 na pracę zdalną - perspektywa pracownika”, “e-mentor” 2020, 1(83), pp. 35 to 43, DOI: 10.15219/em83.1456. The genuine example draws on the methods and sample section.
- University of Cambridge Subject Libraries, “Good academic practice and avoiding plagiarism: Paraphrasing”, publication date unverified. Supplementary guidance on independent wording and acknowledgement.
Sources checked on 25 October 2026. The editorial examples were written for this guide. The miniabstract about planning and overload is entirely fictional and must not be cited as a research finding.
Frequently asked questions
Does a paraphrase require a source citation?
Yes, when it presents someone else's idea or finding. I check source attribution for both paraphrases and direct quotes. Verifying that a quote matches the original requires access to the cited publication.
Putting a sentence in your own words does not remove the need to acknowledge the source. However, this does not mean you must add a footnote to every sentence that sets out your own argument.
Do you check whether the sources in the bibliography exist?
Within the agreed scope, I check whether bibliographic entries can be identified and flag entries that raise doubts. Not finding a publication in a search does not prove that it does not exist. Checking a quotation or whether a reference accurately represents a publication requires access to the source.
What is the difference between proofreading and editing?
Proofreading focuses on language accuracy, spelling and punctuation. Editing also addresses sentence structure, style, paragraph organisation, flow and the logic of your argument. I tailor the scope to the condition of your text and the changes it needs.
I offer both proofreading and editing for bachelor's and master's theses, coursework, reports, other academic papers and presentations.
Does a high percentage of similarity mean plagiarism?
A percentage alone is not enough for such an assessment. I must check the indicated fragments, their sources, the citation style, and the context within the thesis. My analysis helps determine the necessary editorial corrections; the evaluation of the report belongs to the supervisor.
Have the sources, but your draft still follows them too closely?
Send your chapter extract, the publications you used and your guidelines. I can help refine sentences and paragraphs, preserve the meaning of the findings you discuss and organise citations. We will agree on the scope of editing and source checks after reviewing your materials.
