Key Idea: Data is the content topic everything else leans on. Learn the ladder from data to wisdom, the seven life-cycle stages and the four characteristics of big data — then learn a real case for each dilemma, because that is what the markbands pay for.
Paper 1
- Part a: define data, distinguish it from information, name types
- Part b: explain a security measure or a collection method
- Part c: evaluate a system that turns personal data into decisions
Paper 2
- Q1: read a figure off a data source
- Q2: explain what a source’s evidence does and does not show
- Q4: synthesise sources on who a dataset serves
Everywhere else
- Algorithms and AI both run on the data ideas taught here
- The dilemmas recur in every other content topic
- The HL extension’s interventions are mostly data interventions
The ladder, and where machines stop
Four words that are not the same, and the exam does not treat them as the same. The useful part is the top: a machine climbs three rungs and cannot climb the fourth.
| Step | What it adds | Can a machine do it? |
|---|---|---|
| Data | nothing yet — raw facts | Yes, better than we can |
| Information | context: what, when, about whom | Yes, automatically |
| Knowledge | the pattern, so it can be acted on | Yes, often better than we can |
| Wisdom | judgement about whether it should be | No — that is a choice about values |
A system can be accurate about a group and still be the wrong basis for a decision about one person in it.
Seven stages, and the one that causes arguments
The life cycle in order
- Create, collect or extract — the data comes into existence.
- Store — somewhere it can be found again.
- Process — cleaned, sorted, combined.
- Analyse — patterns are looked for.
- Access — people or systems read it.
- Preserve — kept usable as formats and discs age.
- Reuse — used again, for something it was not collected for.
Important: Consent was given once, for one purpose. Reuse puts the same data behind a decision the person never had a chance to refuse — which is why The EU General Data Protection Regulation (GDPR) (European Union, in force since May 2018) attaches a lawful basis to each purpose rather than to the data.
Big data, and the characteristic most answers forget
| What it means | What it makes hard | |
|---|---|---|
| Volume | more than a person could read | must be spread over many machines |
| Variety | text, images, video, sensors | does not fit rows and columns |
| Velocity | arriving continuously | decisions before it is all in |
| Veracity | some of it is wrong or missing | volume HIDES errors, it does not fix them |
Three of the four are engineering problems that have largely been solved. Veracity is not, and naming it opens the counter-argument a part c answer needs.
Exam-style questions
Outline what is meant by metadata.
🔒 Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.
Explain two reasons why an organization might choose to mask data rather than encrypt it.
🔒 Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.
Evaluate the use of large-scale data collection by a national health service.
🔒 Model answer plan
See the mark-by-mark plan — for / against / judgement, with marking guidance — in study mode.
Quick check
Define data and information Data is raw facts with no context. Information is data given context, so that it means something.
Name the seven life-cycle stages Create/collect/extract, store, process, analyse, access, preserve, reuse.
Name the four characteristics of big data Volume, variety, velocity, veracity.
The three data-security methods Encryption (lock it with a key), masking (realistic fakes), erasure (overwrite so it cannot be recovered).
Why is “anonymous” data often not anonymous? A few details together — postcode, date of birth, sex — identify most individuals, and joining datasets makes the link easy.
What makes something a dilemma? The harm and the benefit come from the same cause, so removing one removes the other.
Exam tips
- Every long answer in this topic should name a real case with a date and a place. Twenty of them, refreshed each term, is enough.
- Keep “link” and “cause” apart. A pattern is enough to predict with and never enough to blame with.
- Say which life-cycle stage you are talking about. It is the fastest way to turn a vague answer into a specific one.
- In a part c on data, put the benefit first — starting with harms makes an answer read as one-sided even when it is not.