PDA

View Full Version : limits of agreement

Matt Taylor
07-25-2004, 07:14 PM
Are there any statistically minded people out there as I have a basic knowledge of stats.

I'm using a limits of agreement - LA (Bland and Altman) for comparing two different methods of turning bias.Turning bias, the preferred direction of turn, is measured as a % of total left turns, thus 100% = all left turns, 0% = all right turns and 50% = equal bias. The problem is when I run LA between the two methods, my results suggest that method two will be 64% below or 55% above method 1 (mean 63.7% sd 39.3%) , thus resulting in a negative %. Firstly, is this correct can I have a negative %, secondly can LA be run when dealing with % which cant go below 0 or above 100% with regards to turning bias. any suggestions
hope that made sense - will post responses
thanks to all
Matt

-----------------------------------------------------------------
To unsubscribe send SIGNOFF BIOMCH-L to LISTSERV@nic.surfnet.nl
For information and archives: http://isb.ri.ccf.org/biomch-l