CI from product variance

[cootney]

I have probably a straight forward problem that's been giving me a lot of grief (I'm fairly new to Mark). I'm estimating overall survival (over 4 intervals to 40-days post-hatch) using the CJS model. I model averaged my real parameters and obtained what I hope is the correct overall SE using the Delta method (pg.11 - variance of product of survival rates). My next question is how to estimate CI's from that estimate of variance...the CI side note in the Delta Method chapter delves into back-transforming logit estimates, however I'm not sure how to do that with a product SE? It can't be as simple as mean +/- SE*1.96. I've been through the Mark book and Delta Method Appendix several times and can't seem to find exactly what I need. Consulting a statistician resulted in a recommendation to estimate variance by bootstrap simulations and calculate from there...so any advice?

[cooch]

I have probably a straight forward problem that's been giving me a lot of grief (I'm fairly new to Mark). I'm estimating overall survival (over 4 intervals to 40-days post-hatch) using the CJS model. I model averaged my real parameters and obtained what I hope is the correct overall SE using the Delta method (pg.11 - variance of product of survival rates). My next question is how to estimate CI's from that estimate of variance...the CI side note in the Delta Method chapter delves into back-transforming logit estimates, however I'm not sure how to do that with a product SE? It can't be as simple as mean +/- SE*1.96.

Sure - simply calculate the 95% CI on the logit scale (using +/- 1.96SE), then backtransform onto the normal probability scale.