jolly

questions concerning analysis/theory using software from the Patuxent Software Archive

jolly

Postby ruth » Mon Apr 21, 2008 11:06 am

Hello,
I'm a new-user of JOLLY and having a problem getting started... please could someone explain to me what the FORTRAN format of recapture records in the input file is, and how do I use it to describe my data?
Thank you!
Ruth
ruth
 
Posts: 3
Joined: Mon Apr 21, 2008 9:32 am

Re: jolly

Postby cooch » Mon Apr 21, 2008 11:13 am

ruth wrote:Hello,
I'm a new-user of JOLLY and having a problem getting started... please could someone explain to me what the FORTRAN format of recapture records in the input file is, and how do I use it to describe my data?
Thank you!
Ruth


First, don't use program JOLLY. While it *was* a seminal piece of software when it was first released, it is no longer a robust analysis platform. You really should consider using MARK, or M-SURGE, or the equivalent. Program MARK might be the most accessible. Simply point your browser at

http://www.phidot.org/software/mark/


and go from there. Second, I'd then suggest the following sequence:

1. put your own data in a locked box, and don't open it until you're ready. Resist the temptation to attack your own data until you're ready...

2. ...which means, some level of mastery of the basics. This can be done most efficiently (but not without some work), but working through chapters 1 -> 8 of the MARK book (at minimum). The book can be found at


http://www.phidot.org/software/mark/docs/book/
cooch
 
Posts: 1461
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

Re: jolly

Postby cooch » Mon Apr 21, 2008 11:17 am

ruth wrote:Hello,
I'm a new-user of JOLLY and having a problem getting started... please could someone explain to me what the FORTRAN format of recapture records in the input file is, and how do I use it to describe my data?
Thank you!
Ruth


Just saw that you posted this in the Patuxent Archive part of the forum (as you should have - given your query was about JOLLY). As such, I was probably incorrect in assuming you'd stumbled onto JOLLY by mistake, rather than choice - your posting in the 'correct' part of the forum suggests otherwise.

However, my general suggestions in my initial reply still hold. I think you'd get more at this stage by working with MARK (or M-SURGE) than JOLLY, which is very limited in capability relative to newer software applications.
cooch
 
Posts: 1461
Joined: Thu May 15, 2003 4:11 pm
Location: Cornell University

jolly

Postby jlaake » Mon Apr 21, 2008 11:32 am

I agree with Evan and my advice would be to use MARK and avoid Jolly altogether. Personally I think the program should be removed from the Patuxent website unless the optimization code can be improved. Many folks are misled because of its apparent simplicity. The most useful model (constant Phi, and time-varying p) will often not converge for adequate data sets. As noted in MARK, the formulation of the Jolly model is difficult to optimize; however, you can use the SIMANNEAL option in MARK or you can use alternative formulations for the likelihood like POPAN or the Pradel models.

-jeff
jlaake
 
Posts: 1177
Joined: Fri May 12, 2006 12:50 pm
Location: Escondido, CA

Postby ruth » Mon Apr 21, 2008 3:52 pm

Thank you very much for your replies and advice. I appreciate that MARK is a far superior program and it would certainly be my first choice if it weren't for a couple of things.. I do not want to do anything particularly complicated with my recapture data - I just need rough/crude estimates of population sizes (from which I can calculate the proportion of the population sampled, to be entered into a paternity assignment analysis). I would ideally still have liked to use MARK for this but I have a very tight time constraint (PhD thesis deadline just around the corner) and it was suggested to me that I simply do not have the time to master MARK. So I'm basically looking for a relatively quick and easy way of very roughly estimating population size...?
ruth
 
Posts: 3
Joined: Mon Apr 21, 2008 9:32 am

POPAN instead?

Postby dhewitt » Tue Apr 22, 2008 4:45 pm

I'm basically looking for a relatively quick and easy way of very roughly estimating population size...?


Aren't we all? (Sorry, I couldn't resist.)

What about POPAN? I've never used it, but it looks a little less troublesome than MARK for what you want. You might even be able to run it from the online interface:
http://www.math.sfu.ca/~cschwarz/Carlan.online/
dhewitt
 
Posts: 150
Joined: Tue Nov 06, 2007 12:35 pm
Location: Klamath Falls, OR, USA

POPAN vs Mark

Postby cschwarz@stat.sfu.ca » Tue Apr 22, 2008 4:57 pm

Most of the features of POPAN are already in MARK.

I will take you just as long to read the mark manual as to read the POPAN manual. I suggest you read Chapter 14 of the GIM book as it as a worked example in great detail on how to use the POPAN module in MARK.
cschwarz@stat.sfu.ca
 
Posts: 43
Joined: Mon Jun 09, 2003 1:59 pm
Location: Simon Fraser University

Oops

Postby dhewitt » Tue Apr 22, 2008 5:10 pm

Apologies all around. Looks like you're in for it Ruth.
dhewitt
 
Posts: 150
Joined: Tue Nov 06, 2007 12:35 pm
Location: Klamath Falls, OR, USA

Postby abreton » Wed Apr 23, 2008 3:52 pm

I'll throw another suggestion out there, however I agree that Chapters 1-8 are probably the best option. Consider picking up a book by Amstrup, McDonald and Manly, "Handbook of Capture-Recapture Analysis", 2005, Princeton University Press. From this text, the following may be enough to get you through a basic analysis,

Chapter 1-2 (35 pages ) Intro to CMR, early closed models
Chapter 3 (20 pages) primarily covers the Jolly-Seber model (the default model in POPAN)
Chapter 9, all sections except 9.6 and 9.8 (~59 pages)
Chapter 10 (7 pages)

That's a total of 121 pages, some of which you'll need to read a few times but still not a major commitment. Note that you'll be introduced to the Peterson-Lincoln abundance and variance estimators in Chapter 2. If you want "rough" estimates for almost no effort, you could consider using these "primitive" estimators. But better yet, see recent developments in this area from Carl Schwarz: http://www.stat.sfu.ca/~cschwarz/Peters ... /paper.pdf. From Carl and a previous phidot post, "This is a work in progress and could change in the future."

In Chapter 9, you'll find [basic] instructions on how to import data and build, run and interpret model results in MARK. You may discover that the Jolly-Seber (JS) model is more than you need, perhaps you only want/need to estimate abundance and your data meet the assumptions of closed models outlined in Chapter 2? If so, then you may be better off fitting the Huggins-type closed population model (see Chapter 4 and the example in Chapter 9).

I suspect that if you kept the analysis very simple, you could acquire robust estimates of abundance (and other parameters) after a commitment of about 7 days; maybe less depending on your previous modeling experience. After making this effort, perhaps someone close to you can critique your analysis leading to refinements.
abreton
 
Posts: 111
Joined: Tue Apr 25, 2006 8:18 pm
Location: Insight Database Design and Consulting

Postby ruth » Thu May 08, 2008 6:22 pm

Thank you all for your suggestions and taking the time to give me this useful advice - REALLY helpful and much appreciated!!
ruth
 
Posts: 3
Joined: Mon Apr 21, 2008 9:32 am


Return to analysis help

Who is online

Users browsing this forum: No registered users and 1 guest