See how an axis changes the impression, while the values stay put.
The local lesson files could not load. Open the supplied complete source or reload the working preview.
A 10–15 MINUTE INVESTIGATION · APPROXIMATELY AGE 15
A bold headline.
A closer look.
At the fictional Northbank Flight Fair, a small change becomes a big claim. Find out what the evidence can actually tell you.
Compare a convenient selection with tickets from the whole population.
Separate groups, then make a careful claim of your own.
All names, headlines and data are fictional. No account or personal information is needed. Your notes stay in this tab; refreshing or resetting clears them.
Useful starting skills: reading a basic graph and finding an average. Hints and data tables are available throughout.
01 / GRAPH FRAMING
Did the change get bigger?
Both line charts use the same six fictional session times. Adjust the second chart and watch the scale, not just the slope.
All six values remain visible. Number inputs work with typing or arrow keys.
Data table for both charts: 6 synthetic sessions
GUIDED PRACTICE
Try a new range, then choose what changed.
Need a hint?
Read the first and last values in the table, then compare them with the metric cards. A zoomed line chart can reveal a small variation. It is not automatically deceptive; the labels, context and accompanying claim matter.
02 / SAMPLING
Who got a ticket?
The fair has 60 fictional glider trials. Its front table displays only the 20 long-wing designs. Does a sample from that table speak for the whole fair?
Select a method and draw. Sampling is without replacement; a seed lets you repeat a draw exactly.
PopulationSample
Open the comparison and sample tablesTHE NUMBERS BEHIND THE PICTURE
Random tickets can select any of the 60 trials. This seeded shuffle simulates equal-chance selection; it does not guarantee a balanced individual draw.
Accessible comparison and selected-trial tables
View the complete synthetic population: 60 trials
Draw at least one sample before recording your conclusion.
Need a hint?
Try all 20 front-table tickets. Which category is still missing? Then try random tickets with the same sample size. Separate a method’s coverage from the chance variation in one draw.
03 / CORRELATION
What else could explain the link?
Here are 24 different fictional trials. Longer practice and longer flights appear together. But people choosing different glider designs also practised for different lengths of time.
Circle: short wingsTriangle: long wingsDashed line: linear fit
Open the plotted data tablePATTERN CHECK
Association alone does not establish causation. Design is a plausible third-variable explanation in this constructed example; it is not proof of a particular causal mechanism.
Optional: what does r mean?
The correlation coefficient r describes the direction and strength of a linear association. Near 0 means little linear association, not proof that two things are independent. These small datasets are deliberately constructed, not statistical estimates from a real study.
Accessible table of the currently plotted trials
Compare at least one separate design group, then choose a conclusion.
Need a hint?
Compare “all designs” with each separate design. Could design affect flight time and also be linked to how long someone practised? A follow-up could hold design and test conditions constant while randomly assigning practice durations.
04 / FRESH APPLICATION
Write a claim you can defend.
THE NORTHBANK NOTICE · FICTIONAL HEADLINE
“Practice Pack Doubles Flight Time!”
Ten volunteers from the long-wing display tried a new pack. The fair has 10 long-wing and 40 short-wing registrations. None of the short-wing trials were measured here. There was no control group or random assignment.
LOOK AT THE EVIDENCE
The shortened axis doubles the visible bar height, not the flight time. The source values, the sampling frame and the study design each deserve a look.
Need a starting point?
Compare the means, then ask who was included and what would distinguish a practice effect from other changes. This is a new case: use all three tools from your notebook.
Data table: 10 synthetic paired trials and both means
Build an evidence-based conclusion
Your selected claim and evidence will appear here.
Name the change, support it with numbers, and keep the limitations visible.
Prefer to write your own nuanced conclusion?
This text stays in this tab. The app does not grade or understand your writing.
Compare with one possible conclusion
In these ten long-wing volunteer trials, the mean rose from 12.0 to 12.6 seconds, a 5% increase. The cropped bar chart makes that increase look much larger. The selected sample and lack of a fair control mean we cannot infer the effect for all gliders or say the pack caused it. Other cautious wording that fits those facts is reasonable.
YOUR EVIDENCE NOTEBOOK
Keep the observation.
Keep the caveat.
A good conclusion says what you saw and where the evidence stops.
No automatic prose grading. Notes remain in memory in this tab only. Refresh or reset clears them.
Try this next, offline
Sketch one fictional dataset twice using different scales. Give both drawings to a friend or educator. Ask them to write one conclusion that stays true for both, then name one thing the data cannot establish.