Barker Robust Design

[claudiapenaloza]

Guillaume,

We received uncorrected proofs a few weeks back, then corrected and sent them back to the journal... so it should be any day now (I'm sorry I don't have the latest version). I have you on my radar as soon as it comes out.

About the Barker RD in RMark, I vaguely remember running a Barker RD in Rmark some time ago... but I don't remember what gamma.prime and gamma.doubleprime correspond to. Jeff would know, but is currently not available(?).
However, you should be able to deduce it if your are running analyses on simulated data... fix S and p to 1 and start playing around with the gamma's... how do the encounter histories change as you change gamma' (and gamma")? That should give you your answer (I think ).

Again... Jeff would definitely know (and I'm sorry but I can't find the script I ran way back when...).

[bacollier]

@claudiapenaloza Thanks for your reply. I also asked directly to Gary White during EURING. After I had to manage other analyses, so, I take time now to thank you.

Do you know when this paper about the Barker Robust Design will be available, and do you (or Jeff) know if a RMARK example will be provided ?

After succeeding to use the Barker RD with MARK, I analysed several (~100) simulated dataset with RMARK and I was wondering if gamma" and gamma' used in RMARK referred to a" and a' (the notation in the Barker RD) or referred to the regular notation in RD ?
Interpretation would not be the same depending on the answer as a" = 1 - gamma" and a' = 1- gamma' (regular notation in RD)

Thanks for your help !

Guillaume

I am not in front of my work laptop right now, but if you look in the RMark folder where you installed the package, there is a file called MarkModels.pdf and one called parameters.txt that define all the models and their resulting parameters. I suspect (guessing because I can't check right now) that Jeff has it set up to mirror how Gary set it up in MARK and you should be able to check it against other models in how it is used from the above files, so you might be able to track the definitions down that way as well.

As for an example, if there is not one in there it likely does not exist, but feel free to contribute one and I am sure it will be included.

\bret

[claudiapenaloza]

... there is a file called MarkModels.pdf and one called parameters.txt that define all the models and their resulting parameters. I suspect (guessing because I can't check right now) that Jeff has it set up to mirror how Gary set it up in MARK and you should be able to check it against other models in how it is used from the above files, so you might be able to track the definitions down that way as well.
\bret

That's the first place I checked... but Jeff just has gamma' and gamma"... I'm guessing they are defined as in the RD, because I don't think Bill came up with "availability" (a' and a") until he wrote the Barker RD ms... which happened way after the model was implemented in MARK and RMark. I'm not sure though .

[bacollier]

That's the first place I checked... but Jeff just has gamma' and gamma"... I'm guessing they are defined as in the RD, because I don't think Bill came up with "availability" (a' and a") until he wrote the Barker RD ms... which happened way after the model was implemented in MARK and RMark. I'm not sure though .

Yeah, I think your above is correct (at least best I can tell digging round in the source) I don't see it defined using a' or a" terminology, but I could have missed it. Will let you know if I stumble upon it, but this one may have to wait for Jeff to respond.

\bret

[Guillaume Souchay]

Thanks Claudia and Brett for your suggestions.

I looked at the "parameter.txt" file also. I guess gamma' and gamma" are the same as in the regular RD. However, in MARK, the model included parameters a' and a" and are defined as availability in the MARK help. So, depending on how Jeff implemented the model in Rmark (based on the notation of the RD or of the Barker RD) ... I will wait for Jeff to be sure about that !!

Maybe to add information in the RMARK file "parameter.txt" or to change the name in RMARK may avoid misunderstanding

Guillaume

[jlaake]

Guillaume-

If you'll send me a dbf/fpt file with a single model run of Barker RD from MARK I'll make the fixes necessary in RMark. It sounds as if some of the parameters have changed names since I added it to RMark or I had them wrong in the first place.

--jeff

[jlaake]

This was simply a matter of renaming GammaPrime to aPrime and GammaDoublePrime to aDoublePrime in the parameters.txt file that is installed in R/.../Library/RMark. Guillaume tested it and reported that it worked correctly with the change.

I'll post the new version to CRAN in awhile but in the meantime you can either make the changes yourself with an editor for those models or send me an email and I'll send the new file to you.

--jeff

[claudiapenaloza]

Thanks for the change in RMark Jeff... and no, you didn't have them wrong... you were just a little out of the loop.

As I mentioned before, the Barker RD had been implemented in both MARK and RMark way before Bill had the time/right data-set to complete and write the manuscript. "Availability" didn't come to be until he wrote the manuscript... and there was some back and forth with Gary about the parameter name change (or rather, definition of different parameters).

I guess Gary put a' and a" into MARK once they had settled on using "availability" for the ms, and none of us (i.e., me) thought to give you the heads-up to update RMark accordingly (though I confess I haven't used the model since before the ms was written and didn't know Gary had put "availability" into MARK either). Its all settled now , thank you.

p.s.: many of you may have noticed I just sent you the print proofs of the Barker RD ms. If you didn't get the .pdf and want it, please post your email and I'll send it to you.