Features of statistics education reform such as active learning, analysis of real data, use of technology, and authentic assessment are claimed to enhance student learning.  In presentations and publications innovators in statistical education often assert that instruction has been improved without providing convincing evidence.  To demonstrate design and analysis of experiments in statistics education we employed a three period, two-treatment crossover experiment to evaluate use of multimedia educational software (ActivStats™) in an introductory statistical methods course.  We used within semester exam scores and scores on selected exam questions to compare performance of lab sections assigned to ActivStats or Minitab.   Student attitudes and beliefs associated with  these packages were also investigated.