C14 · Topic 2.3 · 40 minutes · DAT-2.A, DAT-2.B, DAT-2.C
Patterns and limits of inference
By the end you can:
- Calculate and qualify a descriptive claim.
- Distinguish association from causation.
- Explain risks of linking datasets.
Before this lesson: topics 2.1.
- Metadata
- Data that describes other data, such as units, collection time or meaning of a field.
- Association
- A pattern of related values that does not alone establish causation.
Understand
A descriptive result should name the analysed records and processing choices. A difference between groups in a small convenience sample does not prove a universal difference or identify its cause. Consider who was recorded, who was missed and alternative explanations for a pattern.
Combining records can reveal information not obvious in either source alone, including sensitive information. Removing names does not guarantee anonymity if locations, times or other indirect identifiers can be linked. Evaluate whether the new information is necessary and whether the use is appropriate before linking.
Worked example
After documented cleaning, the four fictional trips average 25.5 minutes; three bus trips average 30 minutes and one walk takes 12.
- The group averages describe these records under the stated cleaning policy.
- Bus trips may cover greater distances, so the result does not establish a causal effect of transport mode.
- A precise location/time table linked with another source could identify routines even if names were removed.
Explore each concept
Choose a concept, change its inputs and follow the explanation. Each example models the stated rule or scenario; completing it is practice rather than evidence of mastery.
- Metadata versus primary data
- Descriptive claims versus causes
- Combining records and re-identification
Change it · step through · explain
Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.
Ready · Make a prediction
- QuestionmetadataWaiting
- Time unit metadataminutesWaiting
Ready to explore: Metadata, association and linked data
Why: Change an input, predict the first effect, then advance one step.
Try this: Change one input or choose another listed case. Predict the effect, reset and compare the states and governing rule.
Step through and experiment
Change it · step through · explain
Change an input, predict what will happen, then advance one step. Read why the state changes, reset and experiment. This fixed teaching model does not save progress or award XP.
Ready · Make a prediction
- Input 1deduplicateWaiting
- Input 2excludeWaiting
Ready to experiment
Why: Predict the first change, then advance one step.
Try this: Change only one cleaning policy. Which numerator or denominator changes?
Check a misconception
Claim: A correlation proves that one measured variable causes the other.
An association may have confounders, selection effects or other explanations.
Guided practice
Write a supported one-sentence claim about this sample.
Work with a partner. Make a prediction, compare your reasoning, then identify the rule that supports your answer.
Independent transfer
Explain a privacy risk from linking anonymous-looking trip timestamps with a public class timetable.
Exit check
What additional evidence could improve a transport comparison?
Write your reasoning before comparing with a classmate. If your teacher assigned this task, submit it through My classes to save your work and receive feedback.
Check yourself
Your practice so far
Check an answer to see its rationale.