Research Planning, Academic Integrity and AI

[TURKIYE]

Research Planning and Academic Integrity: References, Quotation and AI​

Your university, course and receiving organisation determine permitted tool use and disclosure. Responsibility for the research remains with the researcher.

A research plan is not a promise that the findings will prove your opinion. It explains what you will ask, how you will answer and the limits of what you can establish. A broad idea and scattered reading may reveal, during writing, that the question exceeds available data or that the method does not measure the intended concept. A good plan helps identify problems early rather than simply meeting an admission or thesis-approval requirement.

Academic integrity begins at this stage, before any result. How you select reading, distinguish other people's text from your notes, record methodological decisions and protect participant data affects the final work. AI tools make responsibility, verification and disclosure especially important. This guide offers a practical way to organise research while keeping academic decisions with the researcher.

Organising documents and an online application​

From a broad subject to a researchable question​

Begin with a subject, identify a problem you can describe, then decide which part is feasible within available time and resources. Technology in education is a broad subject; a question about a particular practice in a defined group is closer to a study. Narrowing a question can strengthen its ability to produce a reliable answer without reducing its importance. State the participants or material, context, period and limits of the claim.

Separate the question from your expected finding. You may anticipate a relationship, but design the study so you can recognise and explain a different result. Presenting the plan as proof of a settled idea encourages selective reading of evidence. Explain the problem's significance, existing knowledge, what needs examining and the possible contribution. Do not claim a topic has never been studied before adequately searching the literature.

Connected components of a plan​

A plan commonly includes a problem, question, aims, academic background, method, resources, timetable and ethical considerations. The template, length and detail follow institute or supervisor requirements and may differ by discipline. Ask how every part connects to what precedes it. An aim the data cannot address, a tool that does not measure the concept or a timetable omitting permissions indicates a break in the reasoning.

Describe methods sufficiently to make the steps understandable, rather than listing impressive terms. Explain selection, collection, analysis and what supports confidence in the conclusion. Text selection may be central in documentary research, measurement in an experiment, and recruitment, analysis and confidentiality in interviews. Adapt the approach to the question instead of borrowing another discipline's template unchanged.

Read literature to build an argument​

Keep a reading record of each work's details, question, method, principal findings, limits and relevance. Go beyond separate summaries to identify agreement, disagreement and methodological or contextual gaps. Two works may differ because they study different groups or define a concept differently. Explaining the difference is more useful than collecting many titles without analysis.

Distinguish complete reading from knowledge gained through an abstract or another work's citation. If you cannot access the original, do not imply you examined its details. Record search terms and exclusion criteria to understand how the reading list developed. Consult newer work in fast-changing fields while retaining relevant foundational research. Quality and relevance matter more than references that do not contribute to the argument.

Using other people's material​

Use	What to do	Risk avoided
Direct quotation	Identify the quoted text and attribute it under the guide	Presenting another person's words as your own
Paraphrase	Reconstruct the meaning and acknowledge its origin	Replacing words while retaining the source's structure
Summary	Condense accurately with attribution	Removing qualifications and widening conclusions
Data or a figure	Identify the source and check permissions	Treating attribution as a substitute for all rights
An idea or method	Explain its origin and use	Reducing integrity to similar wording alone

Integrity covers ideas, methods and data as well as verbatim text. Some material requires permission, so naming its creator does not resolve every restriction. Follow the programme's writing guide for in-text references and the list of works used. Formatting matters, but accurate attribution and an accurate account of what a work says matter beyond appearance.

Paraphrasing without disguised copying​

Read a passage until you understand the argument, then express its meaning in your own wording and your paragraph's logic, preserving its limits. Compare it with the original for distortion or wording that still requires quotation. Synonyms or translation into another language do not make the idea yours. Acknowledgement remains necessary when relying on a particular work.

In an illustrative example, a study reports an association in a limited sample. An inaccurate paraphrase turns this into a claim that the factor causes the outcome for everyone. An accurate account preserves the association and study context, then discusses its relevance to your research. No invented paper or result is needed to explain the principle: retain certainty, sample limits and methodological boundaries when summarising.

Notes that prevent confusion​

Separate verbatim extracts, your summaries and your own ideas in the working record. Mark quotations when you copy them and record their location and citation details immediately. Delaying the step can blur the boundary between your words and someone else's. Do not rely on memory when writing a chapter months after reading.

A reference manager helps organise material but does not verify the author, title, year or pages for you. Check those details and retain material only as permitted. When changing citation styles or merging files, inspect broken or duplicate references. Do not fill the bibliography with unused works to create an impression of more reading than you actually did.

Data and permissions before collection​

Establish early whether the study involves human participants, sensitive data or institutional permissions. Ask the supervisor or ethics office about required approvals before implementation; requirements vary by design and institution. Do not start surveys or interviews assuming retrospective approval. Include permission time in the plan and address possible refusal or delay.

Limit personal information to what the question needs, decide where it is stored, who can access it and how findings avoid unnecessary identification. Participants need appropriate information about purpose and use under the study's procedures. Removing names alone may not anonymise someone if remaining details can be combined to identify them. Treat workplace or placement data carefully even when another role gives you access.

Preparing an application for study​

Deciding whether to use AI​

Start with the course, institute, journal and, where applicable, funder's policy. Language assistance may be permitted with disclosure, while a particular assignment may restrict or prohibit tools. Failure to find a convenient policy paragraph does not mean every use is allowed. Ask about the specific task: organising ideas, improving wording, writing code or analysing data.

Before use, establish what you will do, what the tool will do and how its output will be verified. Do not enter participant data, confidential manuscripts or protected material into an external service without checking permission and institutional rules. Where personal data may be used, anonymise appropriately outside the tool before permitted entry, remembering that details may still identify people. Service settings and content-processing policies are part of the decision. If confidentiality or verification cannot be controlled, choose another way to complete the task.

Verification, responsibility and disclosure​

A tool may produce convincing but incorrect text, a nonexistent reference or a summary removing essential conditions. Open every work you intend to rely on and confirm its existence, content and relevance. A well-formatted reference is not evidence. For tool-assisted analysis or code, understand the steps and test cases you can evaluate. Do not insert results whose reasoning you do not understand.

Keep a suitable record of the task, tool, actual use and verification, in accordance with policy. Disclose use where and how the organisation requires, whether in methods or a designated statement. Disclosure does not authorise every use. An illustrative statement to adapt might say: “A tool assisted with language revision; the researcher reviewed the changes and verified the content.” Do not use that description if the tool also produced substantial analysis or methodological decisions.

Similarity is not a complete integrity judgement​

A similarity report is an inspection tool with institution-specific limits and procedures. No single percentage provides a national guarantee of acceptable work. Matches may concern names, terminology or properly quoted material. Integrity problems may also involve ideas or data the report does not detect. Read highlighted passages and their reasons rather than focusing solely on the number.

Do not try to defeat a report by deleting letters, turning text into images or using automatic rewriting to conceal quotation. Correct attribution, quotation and actual writing, and reconsider whether the passage is necessary. Follow the required report and final-version review procedure. Discuss problems before submission rather than making a file technically different while leaving the work inaccurately attributed.

Integrity in findings and authorship​

Do not invent missing responses or discard findings inconsistent with expectations without a defensible methodological reason. Record exclusions, amendments and data limitations, distinguishing advance plans from decisions made after seeing results. Label practice material as practice rather than study data. An unexpected result may be valuable when presented honestly within the design's limits.

Discuss authorship and contributions in collaborative work and follow applicable contribution and responsibility rules. A person's name on a project does not replace an accurate description of their role. Do not submit previous assignments or research extracts as new work without following reuse and attribution rules. Explain relationships between works to the supervisor early so boundaries are clear before publication or assessment.

A hypothetical plan open to review​

Huda wants to understand delays in student project submission. Starting from a broad question about education quality, she narrows it to one group's perceptions of work organisation. She sets participant selection criteria, interview and analysis methods, and what the study can and cannot establish. Before inviting participation, she asks about ethics approval and permission, adjusting the timetable for review and possible delay.

She uses a permitted tool to review the clarity of some questions without entering participant responses, then reviews its suggestions with the supervisor. Her reading notes distinguish quotations from personal observations, and she opens every work before citing it. When interviews do not support an initial expectation, she revises the interpretation instead of adding data. The plan and decision record explain what she actually did, rather than constructing a successful-looking story after results are complete.

Test your understanding before adopting the text​

Choose an important draft paragraph and explain it aloud without looking. Can you identify whose idea it is, the supporting evidence, limits and why you used it? If not, reread the original or rebuild the argument. Apply the same check to tool-assisted analysis: understand inputs, steps, results and possible failure, rather than only how to copy the output.

Repeat the review for the question, a methods section and a central conclusion. Ask a suitable reader to identify unclear reasoning, then address its cause instead of adding terminology. This is personal practice in understanding your work; it does not replace supervision or institutional requirements. Its value is detecting text that appears coherent but whose reasoning you cannot defend.

Recording research steps and author responsibility​

Step	What to retain	What to check
Question	Scope and revisions	Feasibility with available resources
References	Claim, reference and location	Actual reading and accurate attribution
Data	Permission and processing steps	Privacy and record integrity
Digital tool	Use and limitations where disclosure is required	Course or institutional policy

Completing paperwork for study​

Before submitting the plan or research​

• The question is defined and answerable with available material.
• Aims, methods and analysis connect to the question.
• I understand the institute's template and approval requirements.
• My notes distinguish copied text, summaries and my own ideas.
• I have checked each work's details and what the reference attributes to it.
• I have acknowledged ideas, data and figures.
• Data storage and access follow the applicable rules.
• I understand tool and disclosure policies and have verified output.
• I have reviewed similarity passages and their causes.
• Findings, limits and contributions are presented without fabrication or concealment.

Frequently asked questions​

Does translating a text make it my own wording?
No. Changing language does not remove the need to acknowledge the origin of the idea and expression. A translated quotation may need identifying under the writing guide.

Is inclusion in the final reference list sufficient?
The reader must also know where you relied on the work in the text, using the required style. A separate title list does not identify the origin of each idea or item.

Does disclosing a tool permit every use?
No. Check permission for the task first, verify output and disclose under the policy. Accuracy, confidentiality and integrity do not become the tool's responsibility.

Can an approved plan change?
A revision may need a procedure or approval, especially when participants, tools or data change. Discuss it before implementation and record its reason and effects.

What if I find an error after submission?
Identify the error, its effect and necessary correction, then contact the supervising body under its procedure. Clear, early correction is preferable to continued circulation of a result you know needs revision.

Related topics​

• Grading System & GPA in Türkiye 2026: Pass/Fail, Retakes,
• Course Registration in Türkiye 2026: Add/Drop, Sections, and GPA
• Internships in Türkiye 2026: How to Land a Strong Internship W
• Language Tests for Studying in Türkiye 2026: TÖMER v
• Buying a Laptop for Study in Türkiye 2026: Best Specs by Major
• Student Services on e-Devlet in Türkiye 2026: Student C
• Studying at Turkish Universities
• University and Programme Recognition

​