@ Copyright 2005, Jeff Miller.
This program and its documentation may be duplicated and used without charge for any educational or noncommercial purposes. For commercial use, please contact the author.
Suppose you have data consisting of numerical measurements from two conditions--call them "experimental" and "control" conditions--with a significantly higher mean in the experimental condition. This program computes a likelihood ratio test to see whether the difference between the two conditions is a "uniform effect" or a "mixture effect". Intuitively, the idea of a uniform effect is that all of the scores in the experimental condition are increased relative to what they would have been in the control condition. With a mixture effect, however, only some of the scores in the experimental condition are affected; the rest of the scores in this condition are the same as they would have been without the manipulation (i.e., the same as they would have been in the control condition). In addition to analyzing real data to test for uniform versus mixture effects, the program can also be used to generate simulated data to explore the statistical properties of the likelihood ratio test.
MixTest can test for mixture effects involving more than 30 statistical distributions, including the normal, uniform, gamma, exponential, lognormal, logistic, Wald, Weibull, binomial, geometric, etc. It can also carry out the test for distributions derived from these by convolution, mixturing, censoring, order statistics, etc. The program runs in a cmd window, with input from text files and output to text files.