Question 124:

Asked on March 20, 2008

Tags: t-distribution , Small Sample , Z-Statistic

1

Answer:

No answer provided yet.While the answer depends on the context (medical experiments versus political surveys for example), many see anything less than 30 observations in each sample as being considered small. This stems from the fact that at around 30, the t-distribution starts to have values closer to the Z-distribution. The latter is used on a test of the population and the former on the test of the sample distribution. As the sample sizes increase the two approach each other until at infinity, they are identical.