I know that the larger population, of size m, follows a normal distribution. I have some sample that is size n, and follows its own particular distribution, which is not normal. Given my sample sizes, what are the chances that I would get the distribution that I did in my sample?

Additionally, if I know that the larger population does not follow a normal distribution, but I still know what its distribution is, is there a way to do this same calculation?

  • $\begingroup$ The question is difficult to answer, since it is somewhat imprecise. A Kolmogorov-Smirnov test (see Wikipedia for a start) may be useful to you. More old-school traditional, particularly for the normal, are Chi-squared tests. $\endgroup$ – André Nicolas Aug 28 '14 at 16:05

For your first question: if a sample is from a continuous distribution, then the probability that you would get your exact sample is zero. However, the p-value from a normality test will give you the probability of getting a sample at least as non-normal as yours.

For the second question: you cannot pick out the "true" distribution using a sample, but there are test analogous to the normality tests for other distributions (e.g., weibull, gamma, etc)


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.