Data 1 — Read the Chart
Rung 1 · what it says · one period, repeatable
Runs the deck the Academy already owns: library/investigations/_protocol/evaluating-data.pptx, 33 slides. That record ends "the deck alone is half the lesson." This is the other half.
The opener, and it is the whole method
Put slide 2 up. Say nothing.
Olympic Gold Medals (by planet)
Ask one question: what is wrong with this chart?
Nothing is wrong with it. Every number on it is correct. Every medal in the history of the Olympics was won on Earth, so the chart is accurate, complete, and honest.
It is also worthless.
A chart can be completely correct and still tell you nothing.
That is the thing to install before the serious slides arrive. It is not "charts lie." Charts mostly do not lie. They answer a question, and the question is chosen by whoever made it.
Then hand over the three questions the students will now ask of every chart for four years:
| The question | What it catches | |
|---|---|---|
| 1 | What is being counted? | the planet slide, and every proxy in Data 2 |
| 2 | Compared to what? | a number with no denominator and no baseline |
| 3 | Who made this, and what did they want me to think? | framing, selection, the headline doing the arguing |
Write them on the board. They stay there for the unit.
The run — rounds of three
The deck is ~28 charts. Do not do 28. Do nine to twelve, in rounds of three.
Per chart: ninety seconds, silent, then each table writes one sentence answering one of the three questions — the table picks which, and must say which.
Then each chart gets a bucket, and this is the design decision that makes the lesson work:
| Bucket | Means |
|---|---|
| SOUND | the chart answers a real question and answers it fairly |
| SOUND BUT FRAMED | the numbers hold; the title, scale, or selection is doing the arguing |
| JUNK | the chart does not support what it appears to say |
A table may use JUNK at most twice in the whole period.
That cap is the point. Given permission to be suspicious, a fifteen-year-old will call everything a lie by chart four, and that is not skepticism — it is a different way of not thinking.
SKEPTICAL vs UNCRITICAL is the rubric's pair. Blanket suspicion is a third thing and it is not on the rubric. Being unable to tell a good chart from a bad one is the same failure as believing all of them.
Every bucket must be defended in one sentence. A bucket without a reason is not a bucket.
The traps worth landing on
The deck's record already names them. Choose from these when picking the run — each one teaches a distinct failure:
- Median Wage and Happiness in US States — correlation presented as cause, and "happiness" is a number somebody constructed. Two failures in one chart.
- Deaths from Police Shootings: Gender Gap 9x Larger than Race Gap — the two gaps are computed on different bases. The headline does the arguing, and the chart is not wrong. This is the hardest one in the deck and the best.
- Average Human Attention Span, based on average webpage browsing time — the whole problem is in the subtitle. Save it; it is the bridge to Data 2.
- Top 10 Most Regretted College Majors — self-reported regret, from people who chose to answer a survey about regret.
- Six Figure Income, Race, and Birth Circumstance — the data is sound. What it appears to say depends entirely on which comparison is drawn. This one belongs in SOUND BUT FRAMED and students will fight about it, which is the desired outcome.
Exit ticket
One chart from today that you thought was junk and decided was sound. What changed your mind?
If nothing changed your mind today, say so — and say whether that is because the charts were clear or because you had already decided.
The second half of that question is borrowed from the Handoff vote, deliberately. It is the same honesty being asked for.
Assessment
Critical Analysis rubric. This lesson makes four reachable:
- CONCISE — one sentence per chart is the constraint
- STRUCTURED — the three questions are the structure, and the bucket forces a judgment
- CONTRASTING — the bucket is a comparison against the other two buckets
- SKEPTICAL — at entry level, and bounded by the JUNK cap
PROFOUND and WIDE ANGLE are not reachable here and should not be graded. This is the entry rung.
Where it runs
Year 9, ECS, before SF MAP. Two reasons:
_planning/the-progression.mdrecords that no ECS assignment asks a student to say what they think — every one is build-or-research, and CONCISE and ACCESSIBLE are the year's entry criteria with nothing deliberately developing them. This is a talking lesson that costs one period and needs no materials the Academy does not own.- It is the same doc's item 5, "run one seminar in ECS — cost: zero, protocol exists." Run this with the seminar table card and roles and item 5 is closed.
It also repeats well. Running the shorter data-visualizations.pptx deck the same way in a later year is a legitimate second pass — see 03-the-shape-of-the-picture.md, which uses that deck for a different purpose.
Open
- Seven slides in the deck carry no text layer. Their content is not recoverable from the file and needs eyes on the deck before the run list is fixed.
- Is
data-visualizations.pptxa subset of this deck or a distinct set? Unresolved in the inventory. If it is a subset, Data 3 needs its own charts. - The bucket cap is set at two on judgment, not evidence. Watch one run and adjust.