One of my students has just finished a thesis on Equivalence Testing in
which she showed an inconsistency in the SAS TOST (Two T-Test). Her
formulation I gave her, and she worked out power and sample size
calculations to go with it. We want to publish it before handing out her
good work. If you are interested to follow this up, I will make contact
with her, but please do not broadcast this.
Wanzer Drane
>>> [email protected] 01/09/04 05:58PM >>>
Statalisters,
Does anyone have Stata code for equivalence testing? I may have a need
to
do some analyses using these concepts but have discovered that the
publicly
available Stata code, -equim- and -equip- (sg21, STB-17 & STB-24), has
problems.
When I inquired about the problems I found, the author of the code
stated
that "all the equivalence test programs have a problem when the
observed
difference is less than delta" and that "...when using the equim
versions,
the lower tail of the t CI is wrong." He goes on to say that "I have
no
interest in these programs and will not be updating then under any
conditions that I can now foresee." He also states that "If someone
else
wanted to modify these programs, they have my full backing (but
probably
should not expect any help from me in doing so)."
I am aware of two (non-Stata) packages that perform these analyses,
EquivTest ($675) and ZumaStat ($79), but have no experience with
either
package. Has anyone tried either of these and would you be willing to
provide comments about your experience?
Any other leads would also be appreciated.
Tom Steichen
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
