Rocket plunges his hand into the bean jar. "I bet I grab exactly 20 beans every time."
"Every time?" Raven asks. "That sounds like a question with many answers."
Rocket counts his handful onto the tray. "Eighteen. Hmm. One more try." He grabs again. "Twenty-one."
Nova hovers above the tray. "How many handfuls would make a fair picture?"
"Ten," Rocket decides. "Ten grabs, ten counts, one dot plot."
Raven writes the counts as he goes. "What do you notice so far?"
"They are all close to 20," Rocket says. "But almost never exactly 20."
"Close but varied," Nova says. "That is a distribution. Let us see its shape."
Today you collect your own data set and draw its shape. The question is statistical: how many beans fit in one of your handfuls?
Each grab gives a different count. Ten grabs give you a distribution.
Before you start, make a prediction. Write one number you think will come up most often. You will check it at the end.
Rocket predicted 20 every time. Here is what really happened. Read the table, then the plot.
His counts reach from 18 to 23. The tallest stack is at 20, so his prediction was close to the peak.
But only three grabs hit 20 exactly. The other seven missed by one, two or three beans. That spread is normal.
| Grab | Beans |
|---|---|
| 1 | 18 |
| 2 | 21 |
| 3 | 19 |
| 4 | 22 |
| 5 | 20 |
| 6 | 19 |
| 7 | 23 |
| 8 | 20 |
| 9 | 20 |
| 10 | 21 |
Rocket's prediction was a single number. His data was ten numbers with a shape. Both are useful, but they are not the same thing.
A prediction is what you expect before you look. Data is what you find when you look. Good scientists write both down.
Rocket's peak landed on his prediction, 20. But his counts spread 5 beans wide. A handful is never exactly the same twice.
When your own plot is done, compare it with your prediction the same way. Was your number the peak, near the peak, or far off?
| Statement | True or false? |
|---|---|
| Ten handfuls of beans will almost always give ten different counts. | ? |
| Rocket's counts spread from 18 to 23. | ? |
| A prediction is what you expect before you collect the data. | ? |
| If your prediction misses the peak, your data is wrong. | ? |
Your first home-made distribution. Tomorrow the crew opens the real Field Day survey.