Chapter 20 — Privacy and data protection: foundational cases
Lesson map
Lesson 20 published. Map, example, EX058.
flowchart TD
N0["Learning outcomes"]
N1["20.1 Re-identification"]
N2["20.2 Sparse data"]
N3["20.3 Behaviour and choices"]
N4["20.4 Applying lessons"]
N5["Synthesis"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5
N5 --> EX["Worked example"]
EX --> Q["Exercise EX058"]
Logical drawing
flowchart LR
N0["Learning outcomes"]
N1["20.1 Re-identification"]
N2["20.2 Sparse data"]
N3["20.3 Behaviour and choices"]
N4["20.4 Applying lessons"]
N5["Synthesis"]
N0 --> N1 --> N2 --> N3 --> N4 --> N5 --> Q["EX058"]
[Learning outcomes]
|
v
[20.1 Re-identification]
|
v
[20.2 Sparse data]
|
v
[20.3 Behaviour and choices]
|
v
[20.4 Applying lessons]
|
v
[Synthesis]
|
v
[Worked example] --> [Exercise EX058]
Learning outcomes
You will (1) learn from published privacy research cases; (2) identify the lesson each case teaches; (3) apply lessons to schools.
20.1 Re-identification
Elena studies Sweeney's (2002) demonstration that a few attributes such as postcode, birth date and sex could re-identify people in supposedly anonymous health records. The lesson is that removing names does not make data anonymous. For the trust, this means that releasing pupil-level data to researchers without names but with birth date and school could still identify children, and must be treated as personal data with proper safeguards. Sweeney showed how few attributes identify people. Birth date and school can be enough for small schools.
20.2 Sparse data
Narayanan and Shmatikov (2008) showed that a dataset of film ratings could be linked to public reviews using a few known ratings. The lesson is that detailed behavioural data are highly identifying. Elena applies it to learning apps, which record detailed activity such as answers and timing. Such data could reveal individual children even without names, which strengthens her concern about apps sharing activity logs with third parties. App logs record each click and its time. A few timestamps can single out one child.
20.3 Behaviour and choices
Acquisti, Brandimarte and Loewenstein (2015) review evidence that people's privacy choices are uncertain, context-dependent and malleable. The lesson for the trust is that parents' consent depends heavily on how choices are presented. Elena recommends neutral consent forms that do not pre-tick options or use pressure, consistent with the EDPB consent guidelines. She notes that research on behaviour helps design fairer choices. Acquisti and colleagues show that people's privacy choices are easily shaped by design and context, so the trust's forms present choices neutrally.
20.4 Applying lessons
Elena writes a short guide for the trust: never treat nameless data as anonymous without analysis; treat learning activity data as identifying; and design consent choices neutrally. Each lesson is linked to a published case. Dr. Asante notes that grounding guidance in evidence makes it more persuasive to busy head teachers than general warnings, because the evidence shows that the risks are real rather than theoretical. Her slides cite each study. Each slide pairs one study with one school example, so head teachers can see why the research matters locally.
Worked example
Elena draws three lessons from Sweeney (2002), Narayanan and Shmatikov (2008) and Acquisti et al. (2015) and turns them into a guide for the trust.
Synthetic case. The guide is synthetic; the cited works and EDPB Guidelines 05/2020 are real.
Errors and limits
Lessons from other domains need care when transferred to children's data.
Exercises
- EX058 — Sweeney lesson. What lesson does Sweeney's work teach? Rubric: 3 = removing names does not make data anonymous;; 2 = close;; 1 = names enough;; 0 = none.
- EX059 — Activity logs. Why are learning app activity logs identifying? Rubric: 3 = detailed behavioural data can link to individuals like sparse ratings;; 2 = mentions detail;; 1 = not identifying;; 0 = none.
- EX060 — Neutral consent. What consent design follows from behavioural research? Rubric: 3 = neutral forms without pre-ticked boxes or pressure;; 2 = mentions design;; 1 = pre-tick;; 0 = none.
Synthesis
Chapter 20 learns from cases. Chapter 21 builds system models.
Continue studying — chapter 20 of 40
← Chapter 19 (draft) · Course chapter index · Chapter 21 (draft) →