P.Mean: What distribution does this data come from? (created 2008-07-23).

This page is moving to a new website.

I'm very interested in assessing distributional fits for empirical data and I've found tidbits of information here and there but no real good source. Could you recommend a few good sources?

Can I ask why you want to test distributional fits? In many settings, the actual distribution of the random variable is less important than other considerations, such as how drop-outs are handled. I can think of a few applications where the actual distribution is critically important (most notably when you are trying to make extrapolations to the extremes of the distribution. Most of the time, however, it doesn't matter much whether your data comes from a Weibull distribution or a Gamma distribution. Both are moderately skewed and a bounded below by zero.

Of course, I would end up writing this last statement before noting that the author of the inquiry is in reinsurance. Clearly this is an area where the extremes are important. You may find that several distributions model your data well in the middle, but they diverge strongly at the extremes. So you may have to live with the fact that there is only an ambiguous solution to your problem. I'll see what I can come up with, though.

This work is licensed under a Creative Commons Attribution 3.0 United States License. This page was written by Steve Simon and was last modified on 2010-04-01. Need more information? I have a page with general help resources. You can also browse for pages similar to this one at Category: Modeling issues.