Data 2 — What the Number Counts
Rung 2 · the proxy · one to two periods
The most important lesson in the strand, and the one the whole Academy already depends on without teaching directly.
Every student here will eventually be handed this sentence:
"This map shows where people complain about rats. Not necessarily where rats are. That distinction is going to matter."
That is the second paragraph of the New York investigation, and _planning/the-gis-stack.md calls the step it teaches "the single most important step on the ladder." Right now a student meets it for the first time inside an assignment, in a map, once.
This lesson teaches it as a subject, off the map, before the map.
Part A — Three claims that do not measure what they say
Give the claims without comment. Do not signpost that they are flawed.
| The claim, as it is stated | How it was actually measured |
|---|---|
| The average human attention span is eight seconds | average time spent on a webpage |
| These are the ten most regretted college majors | a survey of people who chose to answer a survey about regret |
| This percentage of people shower daily | people were asked, and answered |
For each, students fill in three lines. The frame is fixed and used word for word — the same discipline the NYC investigation applies to correlations, for the same reason:
This number measures ________. It is named as if it measures ________. The gap between those matters when ________.
The third line is the one that separates the lesson from a gotcha. A proxy is not a lie. Webpage dwell time is a real measurement of a real thing. The question is what you are allowed to conclude from it, and when the gap stops being harmless.
Land this explicitly: the eight-second attention span is one of the most repeated statistics of the last decade and it is not supported by the measurement behind it. It survives because it sounds right and nobody asked question 1.
Part B — Build a proxy, then break it
Hand each table one thing that cannot be measured directly:
- how safe a neighbourhood is
- how good a teacher is
- how hungry a city is
- how much school spirit a class has
- how well the Academy is doing
Three tasks, in order, and the order matters:
1. Name a proxy you could actually collect. Not a perfect one. A real one, that a real organisation could gather with real money.
2. Name what your proxy would miss. Every proxy misses something. Naming it is the work.
3. Name who your proxy would miss.
That third question is load-bearing and it is where the lesson stops being about statistics.
Crime reports measure who calls the police, which is not the same as who is harmed. A survey at an event measures who came to the event. Test scores measure who tests well, on that day, in that room.
Every proxy has a population it cannot see. Find yours.
Then the standing frame, which is the rat map's sentence generalised. Students write one for their own proxy, word for word:
This dataset shows where people ________, not where ________ is.
Part C — Turn it on the Academy
Ten minutes, and it is the connective tissue to the rest of the year:
The Academy will collect feedback at the Appley ceremony. Who is in that room? Who is not? What does feedback collected there measure, and what is it named as if it measures?
That is courses/is1/units/the-handoff/ Phase 1's central question, arriving before the data exists rather than after. When the class runs the Full Cycle on their own survey in DATA1, they will already own the objection.
Assessment
Critical Analysis rubric.
- SKEPTICAL — the whole lesson, and here it is earned rather than granted
- ACCESSIBLE — Part B's "who does it miss" invites the student's own position, and should be allowed to
- PROFOUND is available and should be graded only where the student names a proxy that misses them or someone they know, and says what that costs
- CONTRASTING — measured versus named, three times over
Where it runs
Year 10, IS1, as part of MAPS1 — before the New York rat map.
courses/is1/SCHEDULE.md lists MAPS1 as "not captured" and the operator confirmed "Maps1 is the lead in to the GIS work (rats in IS1)." This is that lead-in's first half; 04-the-map-is-a-chart.md is its second.
Running it before the investigation changes what the investigation is. The rat map's famous sentence stops being a warning the assignment hands the student and becomes a case of something they were taught a week earlier. The student gets to recognise it rather than be told it.
Open
- Confirm MAPS1's position against DATA1's "last month you mapped 311 complaints" reference. The strand assumes MAPS1 → NYC MAP → DATA1 in that order, close together.
- Part B's five subjects are a starting set. The best one in practice is likely "how well the Academy is doing" because it has a real answer with real stakes, but it should not be the only one on offer.