Well, David, I would argue that a point null hypothesis involving continous
variables is NEVER absolutely true -- but, at the same time, I would argue
that when it is nearly true, given typical power in behavioral research, we
get "significant" results most often when, by luck, we get a sample in which
the apparent effect is larger than it really is in the population. For a
better understanding of this opinion, please read the materials cited in the
readings I recommend to my students when they are reading Chapter 8 of David
Howell's marvelous text book, most especially Frank Schmidt's recent
divisional address to the APA, which is written in relatively simple
language, something most statisticians cannot do. The readings are listed
at: http://core.ecu.edu/psyc/wuenschk/read30.htm
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
++++++ Karl L. Wuensch, Department of Psychology, East Carolina University,
Greenville NC 27858-4353 Voice: 252-328-4102 Fax: 252-328-6283
[EMAIL PROTECTED] http://core.ecu.edu/psyc/wuenschk/klw.htm
