Inferential statistics are based on the concept of making decisions using distributions of sample statistics. Such a distribution is called a sampling distribution. The reason to use samples is that populations are usually too large or impossible to test.

If you set up a sampling distribution, you can judge the relative position of your sample statistic as compared to the population statistic. With this information, you can make a judgment called a hypothesis test. Hypothesis testing is discussed in our next workshop.

In any case - there are different shapes of sampling distributions.
Some distributions are unimodal, symmetric and similar to the bell shape of the normal distribution (t-distribution). Some distributions are asymmetrical such as the chi-square and the F distributions.