Hello simone77, thank you for your reply.

simone77 wrote:have you checked how many individuals you have from each plot?

I think the number of individuals in each plot is reasonable for a random effect:

plot = p10 p11 p12 p13 p14 p15 p16 p17 p18 p19 p20 p21

n = 84 73 84 84 62 63 96 64 56 108 81 87

simone77 wrote:It could help knowing which error messages you get when you run the third phrase.Did you try to run simpler model (e.g. [i+r(plot)]) to see if you get the same error messages?

For some reason I am now able to call GEMACO on all three sentences I wrote before. However, I still get the following error when I try to run them or the model you suggested

[i+r(Plot)]:

"Not enough input arguments.

Error in modme/F1_age_evt_al (line 32)

Error in modme/f_age_al (line 20)

Error in menumuse>Run_Callback (line 1775)

Error in gui_mainfcn (line 95)

Error in menumuse (line 46)

Error in matlab.graphics.internal.figfile.FigFile/read>@(hObject,eventdata)menumuse('Run_Callback',hObject,eventdata,guidata(hObject))

Error while evaluating UIControl Callback"

Nothing changes if I write an equal sentence for survival and transition or if I keep one of them constant with the random effect only in the other.

The beta parameters that appear in IVFV seem ok. There is one(or two) beta with "al." after the betas for the fixed effects.

simone77 wrote:On the other hand, I wonder whether the model you are trying to run is very far from the "truth", sometimes I have had problems of convergence when I used a certain structure for a parameter that seemingly was not plausible for the data (what about the AIC and parameter identifiability of [t.plot] and [t+plot]?).

I had run

[t.Plot] for survival and transition just as a way to try to find the problem. I know the amount of data I have is not enough to estimate well the huge number of parameters. This is actually the reason why I wanted Plot as a random effect in the first place! The model

[t.Plot] converges, but with 13 possibly redundant parameters. The estimated rank equals the number of mathematical parameters (176).

Any ideas?

___

Uschi