"Lab day," Comet says, placing two cones twenty steps apart on the grass by the bench.
"Shuttle walks," Wren says. "Cone to cone and back is one walk. Twelve of them, timed in whole seconds."
"And on one walk, stop for a sip of water at the far cone," Comet says. "On purpose. We want an outlier."
Comet walks, Wren clicks, Nova logs. "21, 22, 22, 23, 23, 23, 24, 24, 25, 25, 26, 38 seconds. The water stop was the 38."
"Fences," Wren says, working fast. "Q1 22.5, Q3 25, IQR 2.5. Upper fence 28.75. The 38 is an outlier."
Nova shows both means. "With it, 24.7. Without it, 23.5. Would you like a hint? Check the median too."
"23.5 with, 23 without," Comet says. "It moved by half a second. Now you run yours."
The crew's twelve shuttle walks are made up for the Run Log. Yours will differ, and that is the point.
| Walk | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 |
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Seconds | 21 | 22 | 22 | 23 | 23 | 23 | 24 | 24 | 25 | 25 | 26 | 38 |
| Measure | With the water stop | Without it |
|---|---|---|
| Mean | 24.7 | 23.5 |
| Median | 23.5 | 23 |
| Fences | 18.75 and 28.75 | 17.5 and 29.5 |
| Statement | True or false? |
|---|---|
| 25 + 1.5 × 2.5 = 28.75 | ? |
| The water stop should be reported as an outlier, not hidden in the mean. | ? |
| Removing an outlier always changes the median by a large amount. | ? |
| A normal model would fit the shuttle walks well, outlier included. | ? |
Good lab work. Tomorrow you prove why the median resists, and you learn when the normal model should not be used.