 |
 |
 |
 |
 |
 |
 |
| l |
If we
want greater confidence that an interval
|
|
|
|
calculated
from our data will contain the true value,
|
|
|
|
we
have to use a wider interval.
|
|
|
| l |
To
double the precision of a 95% confidence interval
|
|
|
(i.e.halve
the width of the confidence interval), we
|
|
|
|
need
to take 4 times as many observations.
|
|