How to enter more than 2 groups in mark Resight

[shikha bisht]

i have mark-resight data from 20 different groups (villages) on stray dogs.there is just one primary session and three secondary occasions in that, wanted to know how to incorporate it in my encounter history file...don't know much about mark resight apart form the basics that i read in the MARK book, trying to help a friend with his analysis who needs it real quick!
i was thinking could it be like this
111 21 0000000000000000000;
101 3 0000000000000000000;
010 5 0000000000000000000;
000 6 0000000000000000000;
111 0 4 0000000000000000000;
110 0 5 0000000000000000000;
000 0 7 0000000000000000000;
and so on , like we remove one zero from the right hand and add to the left hand?

[sbonner]

Hi Shikha,

The answer to your question depends on whether or not dogs can move between villages. If the dogs can move between villages then you need to use a multi-state model. I'm going to assume that this is not the case, and the villages form separate groups. In that case, you can use the formatting discussed on page 2-3 of the MARK book extending it from 2 groups for sex to 20 groups for the villages.

There are only 7 possible observable histories, so your inp file should contain only 7 lines. Given the data that you have provided so far the first line should look something like:

111 21 4 . . . . . . . . . . . . . . . . . .;

where you fill in the dots with the counts for the remaining 18 villages. You then need to repeat this for the remaining histories (110,101,011,001,010,100).

Hope that helps,

Simon

[birdman]

I'm not sure I understand your input file, OR the exact structure of the study. You'll need to provide more detail. Some questions to help:
If you just have one primary session, with three secondaries, do you mean you sampled each village one time, with three individual trap sessions over several days or what?
Your title asks about entering more than 2 groups. Is each village a group? If so you just tell MARK at the input interface how many groups you want to use. In your input file you'll have a column for each group, and (assuming there are no individual covariates for the animals) you put the number of individuals with each encounter history in each appropriate group column, separated by spaces; if none for that encounter history, make sure you put a zero. This is all over-viewed pretty well in the Gentle Introduction.
Lastly, I don't understand the example input file you've provided. You should have your encounter histories as you've already listed them, followed by frequency columns for each group. What are all of the zeros on the end? As long as you aren't including individual covariates, I think it would look something like this (Assuming only 3 groups and all possible capture histories occurring at least once):

/*Encounter Hist, Grp1 Grp2 Grp3*/
111 21 4 12;
110 12 7 2;
101 16 3 12;
011 2 15 6;
010 5 17 9;
001 6 4 2;

Note that you can't have a 000 encounter history, because by definition you haven't caught any of these individuals. I hope this serves to show, if nothing else, that you've some reading to do before you're going to be able to tackle this analysis. Folks here will help, but you should spend some time working through some examples in the Gentle Intro before diving into a real dataset. Good luck.

[bmcclintock]

Hi Shikha,

Just clarifying previous posts -- if you're using the logit-normal mark resight model, then '000' encounter histories must be included. Building on Birdman's example with 3 groups, here's an example input file:

/* Begin Input File */

/*Encounter Hist, Grp1 Grp2 Grp3*/
111 21 4 12;
110 12 7 2;
101 16 3 12;
011 2 15 6;
010 5 17 9;
001 6 4 2;
000 6 7 4;

/* Unmarked sightings Grp1*/
Unmarked Seen Group=1;
96;

/* Unmarked sightings Grp2*/
Unmarked Seen Group=2;
56;

/* Unmarked sightings Grp3*/
Unmarked Seen Group=3;
34;

/* Marked but unidentified sightings Grp1*/
Marked Unidentified Group=1;
0 0 0;

/* Marked but unidentified sightings Grp2*/
Marked Unidentified Group=2;
1 0 1;

/* Marked but unidentified sightings Grp3*/
Marked Unidentified Group=3;
2 1 0;

/* End Input File */

Cheers,
Brett