Re: CIs from average beta estimates

[Shawn Craik]

Here is a question from a friend who is working with nest survival data.

I've model-averaged the nest survival models in my candidate set to properly characterize the variables' parameters. I'm pretty sure I calculated everything correctly, working my way through the formulas in Ch 6. My problem is this: ALL confidence intervals, including the intercept, cross zero, and I have no idea why since there are 4 variables, including age and 3 nest-site variables, that really do seem to influence nest success - their CIs never cross zero in any of the individual models. Does this indicate overall poor model fit or just a huge variation in my data set?

[darryl]

First off, your friend may want to look at some of the issues covered here:
http://www.phidot.org/forum/viewtopic.php?t=996