glmmADMB definition of levels for random component
Did anyone ever answer you? It seems as though this may have slipped through the cracks.
Yes, M. Brooks answered me 15 days ago (thank you so much for your fast reply) giving me an approach similar to (a): mod.admb.1<-glmmadmb(ndiscarded~ [fixef]+(1|year_trip)+(1|year_trip_haul),...) being the ranefs recoded as mentioned in my option (a)
As it turns out you're right to be concerned in this case, the model specification you have (~year + (1|trip/haul), leavin out all the other stuff) does assume that trips are consistent across years. Your option (a) [recode to an "implicitly nested" rather than "explicitly nested" coding] would work. I don't like option (b) because it would require fitting a separate parameter for all of your other predictor variables in each year [i.e. it would be equivalent to introducing a (quarter+deep+duration):year interaction], which you may not want to do. Option (c) doesn't work so well either, because it introduces year as a random effect, which (unless you have coded year as numeric in order to look for a log-linear trend in by-catch with year) will be completely redundant with the fixed effect of year in your model. If you don't want to use option (a), then I would recommend the following
The trouble I have regarding the option (a) is that I'm assuming that sampled trips come from the overall trip population (T), and from my point of view this is not true; Following the sampling scheme, trip "t" is randomly drawn from the trip population "T_i" from the year "i" . Many factors (such as shift in fishing behaviour and techniques, changes in fish populations, yearly quotas...),makes me think that trip population can not be assumed to be the same across years. I expect that using the (a) approach produce an over-inflation of the random effect sigmas.
option (d):
ndiscarded~year+ [other fixed effects] + (1|year:trip)+(1|year:trip:haul)
which expands the trip/haul nesting (which would otherwise expand to trip+trip:haul) to include an interaction with year. If this gives you trouble you could also do this even more explicitly by adding a 'yeartrip' variable to your data set which was defined as interaction(year,trip) and similarly for 'yeartriphaul' ...
I guess this (d) model could be a more practical approach to my dataset, let's see what happen! thanks Ben and Mollie for your invaluable help! #------------------------------------------ Juan Santos juan.santos at vi.ieo.es CO VIGO, Instituto Espa?ol de Oceanograf?a #------------------------------------------