- Emphasize that uncertainty is often unavoidable; we
can best deal with it by seeking
to know where it may occur and trying to estimate how large it is.
- Be willing to say, "I
don't know" when appropriate.
- Point out the differences
between ordinary and technical uses of words.
- Be sure to include some
discussion of skewed distributions.
- Emphasize that every
frequentist statistical inference technique depends on model assumptions.
- Form the habit of checking if the model
assumptions are reasonable before applying a
procedure.
- Expect your students to
do the same.
- Give assessment questions
that ask the student to decide which techniques are appropriate.
- Discuss robustness.
- When a test fails to
reject the null hypothesis, do not accept the null hypothesis unless a power calculation has shown that the test
will detect a practically
significant difference, or unless there is some other carefully thought
out decision criterion that has been met.
- Expect your students to
do the same.
- Remember, and emphasize,
that one study does not prove anything.
- In particular, do not use
strong language such as "We conclude that ...", "This proves that ...",
'This shows that ... is ...."
- Instead, use more honest
language such as, "These data support the claim that ..." or "This
experiment suggests that ..."
- Expect your students to
do the same.
- In introductory courses,
try to caution your students about the problems with multiple
inference (and the file drawer problem),
even if you can't go into detail.
- In advanced courses, be sure to discuss the problems of multiple inference and the file drawer problem.
- Try to use a textbook that includes the points above. The introductory textbooks by De Veaux, Velleman and Bock generally are good choices for this at their level. I occasionally post brief reviews of statistics textbooks on my blog at http://www.ma.utexas.edu/blogs/mks/.
- Make use of web demos that can help students understand concepts and cautions. The following sites include links to some such demos:
- http://www.ma.utexas.edu/users/mks/CommonMistakes2013/commonmistakeshome2013.html
- http://www.ma.utexAS.edu/users/mks/M358KInstructorMaterials.htm
(scroll down to near the bottom of the page; these are keyed to one of
the textbooks by De Veaux,Velleman, and Bock.)

- For more suggestions for
introductory courses, see American Statistical Association (2005,
reformatted 2010),
*Guidelines for Assessment and Instruction in Statistics Education (GAISE) College Report*(can be downloaded from http://www.amstat.org/education/gaise/index.cfm) - Make use of other resources for professional development for statistics teachers:
- The Consortium for the Advancement of Undergraduate Statistics Education (CAUSE) offers workshops, webinars, and links to resources for statistics teachers.
- The Journal of Statistics Education is available on line to anyone for no charge. It publishes three issues per year containing peer-reviewed articles related to statistics education. Their home page has information on how to receive e-mail announcements when new issues are published, as well as links to an Interactive computing Archive, a Data Archive, and other resources.
- Join
the American Statistical Association
and its Statistical Education Section. The latter will give you access
to a list serve that can be useful for asking questions about teaching
statistics.