Operational definitions, proxies and ethics
We often measure something easy to record as a clue about something harder to observe. Check how good that clue is and whether using it could harm people.
An operational definition is the explicit rule that turns a concept into recorded values. A proxy is a measurable stand-in for something not directly measured. Platform minutes may include an idle browser tab; low online activity may reflect offline study, work or caring responsibilities.
Reliability concerns consistency of measurement under comparable conditions. Measurement validity concerns whether it captures the intended concept. This is a different use of "validity" from a Week 3 range check. A consistently recorded login count can be reliable while being an incomplete measure of engagement.
Consider who benefits, who bears false alarms or missed cases, whether less intrusive data would suffice, and whether affected people can understand and challenge a decision. A sensitive attribute might be useful for auditing unequal outcomes without being an appropriate input for an individual decision. Predictive usefulness alone does not settle that choice.
Sources: W1 PDF pp. 3, 13-17, 22-30, 32-53, 57-60; L1 Parts 1-5.
Practice problems
Supplementary study questions on the lecture concepts — not verbatim exam questions. Each answer counts toward spaced review.
1. Conceptual understanding
Definitions and distinctions.
A login counter records consistently but misses offline study. Which distinction matters?
Why might a useful predictive feature still be inappropriate?
2. Application
Apply the concept to a case or calculation.
Why can platform minutes overstate active study?
A device gives inconsistent readings under the same conditions. What is in question?
3. Interpretation and limitations
Evaluate assumptions and justify conclusions.
Before using an early-warning model, which question matters?
Low online activity should be interpreted how?
Explain it in writing
Self-checked, not marked. Write a few sentences, then compare with the example answer. Your writing is saved as you type.
Show an example answer
A login counter may record each login consistently, yet logins do not fully represent learning. Consistency and measuring the right concept are different questions.
Check: did you answer the question, use evidence or working, and avoid claiming more than the evidence supports?
Show an example answer
It can help check whether outcomes differ unfairly across groups. That does not automatically justify using the attribute to make decisions about individual students.
Check: did you answer the question, use evidence or working, and avoid claiming more than the evidence supports?