tag:blogger.com,1999:blog-8499895524521663926.post5968260949596767870..comments2024-10-08T05:57:43.045-04:00Comments on Phylogenetic Tools for Comparative Biology: Phylogenetic regression when estimated Pagel's λ > 1Liam Revellhttp://www.blogger.com/profile/04314686830842384151noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-8499895524521663926.post-21889930171182515892019-03-20T18:11:15.101-04:002019-03-20T18:11:15.101-04:00one slight update - estimated gls lambda is 1.108 ...one slight update - estimated gls lambda is 1.108 (not 1.18)Anonymoushttps://www.blogger.com/profile/11550933579634319893noreply@blogger.comtag:blogger.com,1999:blog-8499895524521663926.post-90235433483131919052019-03-20T18:01:54.905-04:002019-03-20T18:01:54.905-04:00Hi Liam,
I am late to this post but I am wondering...Hi Liam,<br />I am late to this post but I am wondering if you have any updated information regarding how to best approach cases where your estimated lambda is greater than 1 (when estimated using ML in nlme's gls function). <br />I am dealing with a case where phyl.resid (using phytools) estimates lambda as .95 but the gls estimates lambda as 1.18. The results from using a 1.18 lambda in the gls function seem pretty outrageous and are quite different from assuming a brownian motion model (lambda=1). Using the lambda above 1 completely alters the post-hoc estimates of some of my independent categorical variables in ways that seem to make no sense. <br /><br />Given the odd post-hoc estimates, I am inclined to limit lambda to 1 but was wondering if you had any further suggestions. <br />Thanks! Anonymoushttps://www.blogger.com/profile/11550933579634319893noreply@blogger.com