Math Forum Discussions

Richard Scheaffer

Posts: 440
Registered: 12/6/04

Experiments
Posted: Sep 6, 1996 9:52 AM

Plain Text

With regard to the recent discussion on Planning and Conducting Experiments,
Bob Hayden was right on target, as usual. The changes were intended as a
clarification, not an expansion. These are concepts one should be familiar
with when planning a study or critiquing someone else's plan.

These can be discussed without actually completing the inferential analysis
of data from designed experiments. Often, a graphical display is good enough
for a preliminary analysis. On the other hand, all of these concepts can be
illustrated with two treatments, where the comparisons (paired or unpaired) can
be made with inferential trechniques covered in the course.

We do not intend that you teach analysis of variance; we do intend that students
appreciate the fact that study design is one of the very important aspects of
statistical reasoning, some would say THE most important. It has been said that
the randomized, comparative experiment is one of the most important
contributions to science in the twentieth century. Should not students being
introduced to the subject know this?

Dick Scheaffer