Here we review some basic statistical principles for defining differences between treatment groups.  We extend that to address the statistical challenge that comes from doing thousands of comparisons in a microarray experiment.