Hello!
I am working on a manuscript on sexual dimorphism in an aquatic
invertebrate, where we have estimated sexual dimorphism (SD) for 7 different
traits in four populations (a total of 28 SD-estimates). We have used the
following formula for estimating SD: 100 * (mean male trait value - mean
female trait value)/overall trait mean).
Then, we have used these SD-estimates to perform a GLM against other
interesting variables, such as the intersexual genetic correlations for each
of the traits.
Here are my questions:
1. Is there any procedure in "R" you would recommend that takes in to
account the sampling variance of the SD-estimates, rather than using the
mean value of each (which is supposed to reduce error and increase Type
I-error rates?
2. Is there a procedure to estimate SE for ratios such as this SD-estimate?
3. The data in these GLM:s might not be entirely statistically
non-independent (i. e. intersexual genetic correlations). Can you recommend
any R-procedure (package) that can deal with this problem (e. g.
bootstrapping or resampling)?
Many thanks in advance for input!
Erik Svensson
--
View this message in context:
http://r.789695.n4.nabble.com/Standard-errors-of-sexual-dimorphism-tp3785770p3785770.html
Sent from the R help mailing list archive at Nabble.com.
