Fitting a distribution to data using MLE

I have a dataset to which I'd like to fit a distribution using MLE. In the old days, I would have taken a guess and then used a KS test to see if the fit were good. I've been led to believe that modern (i.e. written in the last 10 years) can do this for me.

It's been a while since I've actually worked for a living, so please be patient with me.

The data is more-or-less symmetrical and is leptokurtic. My prior is that a t-distribution with n < 20 might work.

What package do I want to use and which algorithm within that package? Is there a website into which I can just plug the data to get a guess? What's the best (basic) citation on this topic?


Me: 55, MS ms stat (1994), senior manager who likes to get his hands dirty.


TS Contributor
I generally use Minitab which has a distribution identification menupoint with a good choice of the most frequently seen distributions. Just searching around I found this shareware:

If t seems to be a good choice maybe you could try the Weibull distribution - the universal workhorse of distribution fits