tag:blogger.com,1999:blog-659101895904690224.post46966776203870799..comments2023-12-05T12:22:10.177+01:00Comments on Klimablogg: Statistical power of temperature trendsHans Petter Jacobsenhttp://www.blogger.com/profile/09403991111027540923noreply@blogger.comBlogger1125tag:blogger.com,1999:blog-659101895904690224.post-15923850406079990002018-12-13T01:01:00.612+01:002018-12-13T01:01:00.612+01:00Hi
Thanks for the nice blog :)
Correct we should...Hi<br /><br />Thanks for the nice blog :)<br /><br />Correct we should focus on the priori power, not the post-hoc power.<br />The post-hoc power is directly correlated with the p-value.<br /> (when calculating base on the observed effect)<br /><br />The null assumption distribution should be non-central t, which is similar to a "shifted t", but not symmetric<br /><br />You can see an accurate dynamic chart with h1 uses "non-central t" distribution in the following calculator:<br />http://www.statskingdom.com/30_test_power.html<br />just use "t" distributionAnonymoushttps://www.blogger.com/profile/13346073376553938948noreply@blogger.com