ABSTRACT
A sample is good if it yields a statistic that is close enough to the parameter of interest. The concept of the sampling distribution of a statistic is the most important idea in statistical inference. Virtually every inferential procedure involves knowing, approximating, or at least assuming knowledge of the sampling distribution of some statistic. It turns out that many statistics have a normal sampling distribution. In the case of a sampling distribution of some statistic, this means that most of the possible samples give a statistic that is fairly close to the parameter, with relatively few samples giving a statistic that is far from the parameter. A histogram is a graph that shows the proportion of individuals that have each possible value on some variable or statistic. The larger the sample, the closer a statistic will tend to be to the parameter of interest. So the sampling distribution will be more concentrated toward the center.
