Abstract:
|
This presentation will address several unrelated but interesting consulting problems concerning the design, analysis, and interpretation of experiments. When is it appropriate to omit from an analysis some entirely valid but specifically selected data values in order to achieve a p-value less than the magical 0.05? When is the obvious conclusion . Wrong? Can it be beneficial to discount an investigator's assurances when designing an experiment? What does statistics have to do with Twinkies? Examples answering these and other questions of general interest provide some engaging teaching examples.
|