Long-term feature: stop based on desired power

[rbcavanaugh]

From GF:

• app could stop based on a researcher's desired power to detect some effect (like change between two administrations)
• to discuss with WH

[gfergadiotis]

This would be a cool feature...

[william-hula]

Yeah, cool idea, but makes me nervous. Seems too much like "we'll just keep collecting data until we get the result we want."

[gfergadiotis]

Sure, even though they might conclude that there is no real difference or that the detected difference is so un-important that the cost of treatment does not worth it. In my mind, I was thinking something like a “power surface” similar to power curves but with an extra dimension such as “average SEM as a function of length”.

Having said that, given that this is a whole new project, I suspect now might not be the time to explore it. So, let’s put this in the back burner.

G

Gerasimos Fergadiotis, PhD., CCC-SLP
Associate Professor
Speech & Hearing Sciences
527 SW Hall Street, RM430G
University Center Building
Portland, OR 97201
Phone: 503-725-2217
Web page: http://aaldresearch.wix.com/aald

On Jul 20, 2021, at 10:47 PM, William Hula @.***> wrote:

Yeah, cool idea, but makes me nervous. Seems too much like "we'll just keep collecting data until we get the result we want."

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rbcavanaugh/pnt/issues/11#issuecomment-883649573, or unsubscribe https://github.com/notifications/unsubscribe-auth/AL6NBQFKTQWSEU442NISQ2TTYXHEZANCNFSM5ACQWUWQ.

[rbcavanaugh]

back burner it is!