[Mesquitelist] Presentation of multiple likelihood-ratio tests, extraction of multiple P values from Mesquite
n.hobbhahn at gmail.com
Wed Mar 28 04:36:21 PDT 2012
Dear fellow list users,
I am using the following likelihood-ratio tests in Mesquite to examine the evolution of a binary trait over a phylogeny:
1. Asymmetry likelihood-ratio test to compare asymmetrical and one-parameter models of trait evolution
2. BiSSE ln likelihood difference for a) speciation rate, b) extinction rate, and c) diversification rate.
I am running these analyses over a sample of 1000 trees to account for phylogenetic uncertainty. To do this, I open a New Bar & Line Chart for Trees under "Analysis" and set up the analyses. The output then contains mean, median, and 95% confidence intervals for the likelihood-ratio statistic D.
My questions are:
1. Is it statistically correct to present average D values with their 95% confidence intervals? If not, could you please advise how the results from multiple LH-ratio tests should be presented?
2. IF I present average D values, is it correct to use the chi-square value with the corresponding number of degrees of freedom for assessment of statistical significance?
3. Alternatively, one could follow the approach by DeWitt Smith et al (2008, in Evolution 62(4): THE ROLE OF POLLINATOR SHIFTS IN THE FLORAL DIVERSIFICATION OF IOCHROMA (SOLANACEAE)), who report the average P value from multiple comparisons. IF this approach is preferable to that suggested in 2., then
4. How can I extract not only the LH-ratio values for each tree, but also the associated P values from Mesquite?
Thank you very much for your interest and time,
Dr. Nina Hobbhahn
Lab of Prof. S. D. Johnson
School of Life Sciences
University of KwaZulu-Natal
Private Bag X01
Scottsville, Pietermaritzburg, 3201
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Mesquitelist