Hi Anderson - first, this is tremendous, thanks for your replies!
however, sadly, I believe I may still be doing something wrong....
I made a featquery result available (where I don't think the 95%-CI should reveal a significant mean cope): https://drive.google.com/file/d/0B3ThVM_Z6djNa3hwVTgtMTVXMWc/view?usp=sharing
(and in excel format: https://drive.google.com/file/d/0B3ThVM_Z6djNNllhUkZUUXNvQlk/view?usp=sharing)
Here, you can see:
cope mean: 0.07969
varcope mean: 0.007069
The way I understood your reply was that instead of using the standard error of the mean cope, we should use the mean varcope to compute 95%-CIs...
95%-CI= mean varcope x critical t-value = 0.007069 x 2.7764 = 0.019627
But, alas, I must be doing something wrong here, eh?
(and, as an aside, I tried to reproduce the mean t-value provided in the table using the forumula you provided: cope/sqrt(varcope) --- however, this did not reproduce the result in the table: i.e. 0.9478 was computed vs. 0.893, from the table. here too, I must be doing something wrong.)
Thanks again for your efforts!
HUGE help.
Ken
|