26134 · Responsible Evidence-based Decisions
Ethical Use of Data
Week 11 is Module 5, the 'responsible' half of the subject's name, and it is fully examinable short-answer territory. You learn rule- versus principle-based ethics, the four ethical principles, consequentialist versus deontological frameworks, anonymisation and data security, and the Five Safes framework. Expect an exam vignette asking you to reason through a data-use scenario against these frameworks.
What this chapter covers
- 01Rule-based (clarity, consistency, compliance) vs principle-based (flexibility, judgement) approaches
- 02The four ethical principles: respect for persons, beneficence, justice, respect for law and public interest
- 03Respect for persons and informed consent; extra protection for those with diminished autonomy
- 04Consequentialist (judge by outcomes) vs deontological (judge by duties/means) frameworks
- 05Anonymisation / de-identification, secure data handling, and re-identification risk
- 06The Five Safes framework: safe projects, people, settings, data, outputs
- 07Why ethics matters for evidence-based decisions — data can serve or harm
Applying ethics frameworks to a data-reuse scenario
- +1Respect for persons. The data were collected for fulfilment, not marketing, so this is a secondary use: was consent obtained or is it within reasonable expectations? Respect for persons calls for informed consent and honouring people's autonomy over their data.
- +1Beneficence and justice. Weigh benefits (better-targeted offers) against risks (re-identification, unwanted profiling), and minimise the risks. Justice asks whether those benefits and risks are distributed fairly, and that no group is unfairly targeted or excluded.
- +1Respect for law and public interest. Check privacy-law compliance for both the internal reuse and the external share, and act transparently and accountably. Contrast a consequentialist view (net benefit may justify it) with a deontological view (duties such as consent and confidentiality may forbid it regardless of benefit).
- +1Five Safes on the vendor share. Safe projects (is this use appropriate?), safe people (is the vendor trained and authorised?), safe settings (is the environment secure?), safe data (is the extract sufficiently de-identified?), and safe outputs (do released results avoid disclosure risk?). A responsible decision satisfies all five.
Key terms
- Rule-based vs principle-based ethics
- A rule-based approach follows specific stated rules, giving clarity, consistency and compliance; a principle-based approach reasons from broad principles, giving flexibility for novel situations where rules do not fit — important for new digital-age data uses.
- The four ethical principles
- Respect for persons (autonomy and informed consent), beneficence (minimise risks, maximise benefits, do no harm), justice (fair distribution of risks and benefits) and respect for law and public interest (legal compliance, transparency, accountability).
- Consequentialist vs deontological frameworks
- Consequentialism judges an action by its outcomes (e.g. a utilitarian cost–benefit weighing); deontology judges it by duties and the means themselves, regardless of outcome. The two can disagree on the same data-use decision.
- Anonymisation / de-identification
- Removing or masking identifying information so individuals cannot readily be recognised in a dataset. It reduces but does not fully remove re-identification risk, so it is paired with secure handling and access controls.
- Five Safes framework
- A structured way to manage safe data access across five dimensions: safe projects (appropriate use), safe people (trained/authorised users), safe data (sufficiently de-identified), safe settings (secure environment) and safe outputs (results without disclosure risk).
- Informed consent
- Agreement to the use of one's data given with adequate understanding of what is collected and why. It operationalises respect for persons, and secondary uses beyond the original purpose usually require fresh consent or a clear lawful basis.
Ethical Use of Data FAQ
Is data ethics really examinable in a statistics subject?
Yes. Week 11 is Module 5 and the final exam explicitly spans all five modules, including data communication and data ethics, so a short-answer ethics vignette is fair game. Because it is qualitative, prepare it as a structured framework you can apply, not as content to compute — students who skip it lose easy marks.
What is the difference between consequentialist and deontological reasoning?
Consequentialist reasoning judges an action purely by its outcomes — if the net benefit is positive, it is justified. Deontological reasoning judges by duties and the nature of the act itself — some actions (breaking consent, violating confidentiality) are wrong regardless of benefit. A good answer applies both lenses to the same scenario.
What are the Five Safes and how do I use them?
They are five dimensions for safe data access: safe projects (is the use appropriate?), safe people (are users trained and authorised?), safe data (is it sufficiently de-identified?), safe settings (is the environment secure?) and safe outputs (do released results avoid disclosure risk?). Apply each one explicitly to the scenario in the question.
Can AI help me with data ethics in 26134?
Yes, as a study aid. Sia can help you structure an ethics answer — mapping a scenario to the four principles, contrasting consequentialist and deontological views, and running the Five Safes — and check that your reasoning is complete. Use it to rehearse; it does not write your graded assessment, and the UTS academic-integrity policy applies.
Exam move
Do not skip ethics — it is examinable and quick to secure if you prepare a structure. Memorise three frameworks you can deploy on any vignette: the four principles (respect for persons, beneficence, justice, respect for law and public interest), the consequentialist-versus-deontological contrast, and the Five Safes (projects, people, settings, data, outputs). Practise applying all three to a short data-use scenario, writing one or two sentences per element, because the marks reward explicit mapping, not a general opinion or a yes/no verdict. Keep a one-page ethics framework summary in your printed exam notes. When you are unsure how to structure a response, ask Sia to model the mapping on a fresh scenario and critique yours; confirm assessment details on Canvas.
Working through Ethical Use of Data in 26134? Sia is AskSia’s AI Statistics tutor — ask any 26134 Ethical Use of Data question and get a clear, step-by-step explanation grounded in how 26134 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
