The relationship that holds in a sample may not hold in the population.
To see this for yourself, imagine the following situation:
You randomly select strangers to flip a coin.
Each stranger flips the coin once.
You record the stranger's height and whether the stranger's flip was "heads" or "tails."
After drawing a sample of 3, you calculate the correlation between height and the result of the coin flip.
Pressing the button below simulates your doing this.
Press the button several times.
You will see that the correlation between height and coin flips, which is 0, in the population, is not always 0 in the sample.
| Height (inches) | Coin Flip Result |
|---|