Message-ID: <52C636BF.4070507@gmail.com>
Date: 2014-01-03T04:04:15Z
From: Ben Bolker
Subject: Sample size calculation
In-Reply-To: <loom.20140102T133356-952@post.gmane.org>
On 14-01-02 07:42 AM, Bonitta gianluca wrote:
> Thank You for soon reply fisrt of all
>
> i have the dataset and r script maby i can send and you can help to
> prototipe the script in r ...for example
>
>
>
> for professor Ben :
>
> i had seen long power pack but i think this need repeated measure for the
> use i have not repeated measure ...do You think is possible to suite long
> power pack for my case?
I didn't necessarily mean that you should use longpower, I was just
giving a list of packages geared toward mixed model power analysis.
>
> I have also try simulative way for post hoc power calculation.... but this
> yeld pap power analysis paradox... when p-value is "big" power is low ,
> when p-value is "small" ower is hight so i prefer to determinate a priori
> power ....
I didn't mean that you should do a post hoc power analysis (I
completely agree with you that post hoc power analysis is a bad idea!)
>
> thank you so much
>
I just pushed an example of what I mean by simulation-based power
analysis to Rpubs: http://rpubs.com/bbolker/11703
cheers
Ben Bolker