|
|
Sampling
is the process of taking a smaller group of subjects/scores
from a larger population. In statistical inference, we test hypotheses
about the population using samples. |
|
|
Samples do not replicate the population.
The discrepancy between sample values and population scores is known as
sampling error. |
|
|
The sampling distribution is what we use
to help us decide whether a particular sample came from a given (or known)
population. It tells us what to expect when we draw one sample from a
larger population. |
|
|
A sampling distribution is the frequency
distribution of the value of a statistic calculated on an infinite number
of samples of a given size drawn from a known population. |
|
|
To test hypotheses, we place the value of
a statistic for a single sample on the sampling distribution. If the sampling
distribution shows a low probability (unexpected), then we conclude that
the sample did not come from the population. If the sampling distribution
shows a high proability (expected), then we conclude that the sample did
come from the population. |
|
|
Every statistic has a sampling distribution. |
|
|
Sampling distributions vary in shape depending
on what statistic is used and either the sample size or degrees of freedom
for that statistic. |
