Module 10 · 3 lessons

Research ethics and integrity

Two different obligations: to the people in your research, and to everyone who will read it. Almost nobody sets out to do this badly.

Free — no account needed to read

Two different obligations: to the people in your research, and to everyone who will read it.

Why it matters

Almost nobody sets out to do this badly. The failures are small, defensible-feeling decisions made under pressure, which is exactly why knowing their names in advance is what prevents them.

What it teaches

• Informed consent, privacy, confidentiality and anonymisation
• Vulnerable participants and where extra care is owed
• Fabrication, falsification and plagiarism
• p-hacking, HARKing, selective reporting and cherry-picking
• Authorship and citation ethics
• Responsible use of AI in research

When you reach for it

• You are about to collect data from people
• A result is nearly significant and you are considering one more analysis
• You used an AI tool somewhere in the work and are not sure what to say about it

The common misreading

That ethics is a form you complete before the real work. It is a set of decisions that recur throughout, most of them long after any approval was granted.

What this module will not tell you

This module will not tell you what your institution or jurisdiction requires. Rules for consent, data protection and approval differ by country, field and organisation, and you must check yours — nothing here substitutes for that.

Think first

You promised participants anonymity. Your sample includes the only woman in a 40-person engineering team, and her quote is the most important in your findings. What now?

Write down what you would want to be told, and what you would want control over, if you were a participant in your own study.

The lessons

• What you owe the people in your research — Consent is about comprehension and freedom, not a signature. Promise only the confidentiality you can keep. (9 min)
• What you owe everybody who reads it — Decide the analysis before you see the result. If you decide after, say so — exploratory is a respectable label. (9 min)
• Using AI in research, responsibly — Use AI for finding, drafting and structuring. Verify everything factual, never let it be your only analyst, and disclose what you used. (9 min)

Practise it: Write the ethics section for your own work

Work through what your project owes participants and readers, in writing.

1. Write what you would tell a participant, in plain language, under 150 words.
2. Name who could be identified from your data even without names, and what you will do about it.
3. Write your analysis plan before you see any results, and date it.
4. List every AI tool you have used and what each was used for.
5. Name the one pressure most likely to tempt you into a selective decision later, and the rule you are setting now to handle it.

What a considered answer tends to contain (not a mark scheme): A considered answer identifies a real deductive disclosure risk specific to the setting, and names a pressure honestly rather than claiming none exists.

Use it on real work: Put it into the project

Record the ethical position of a real project where it will be seen.

1. Write the consent language into the project brief or your materials.
2. Record your dated analysis plan before collecting or analysing.
3. Add an AI-use note listing tools and tasks.
4. Check every AI-supplied citation in your library against its source, and remove any that do not resolve.

Reflect

Which small, defensible-sounding decision would you personally be most tempted by under deadline pressure?