Format for MARK input file?

[JonL]

If one reads page 2-4 in "Program MARK A gentle introduction" there is one version of an input file for two categories. The two last columns give number of males and females (e.g.) respectively. For one history per individual this means "0 1" or "1 0" (for males and females respectively). Page 2-8 seems to present an alternative (also for) input for one individual per line of history. In this case there is always a "1" in the first column after the history proper. The second column is either 1 or 0. With this format, males and females would be coded "1 1" or "1 0". Is either format correct or have I misunderstood something?

How does these notations relate to the input of "attribute groups". Where do I enter males and females respectively (first or second entry) in the "Enter specifications for MARK analysis" panel?

Jon

[cooch]

If one reads page 2-4 in "Program MARK A gentle introduction" there is one version of an input file for two categories. The two last columns give number of males and females (e.g.) respectively. For one history per individual this means "0 1" or "1 0" (for males and females respectively).

It makes no difference if you use '1 0' for males (and thus '0 1' for females) or vice versa. The only thing you need to do is remember which column corresponds to which sex.

Page 2-8 seems to present an alternative (also for) input for one individual per line of history. In this case there is always a "1" in the first column after the history proper. The second column is either 1 or 0. With this format, males and females would be coded "1 1" or "1 0". Is either format correct or have I misunderstood something?

The latter - in that example, the use of a covariate to indicate sex is being demonstrated. The first 1 is simply the frequency, while the 0 or 1 in the next field indicates males or females. This is simply one way of coding for a classification variable that is convenient if the classification variable is binary (i.e., has two states, like sex). In general, it is preferable not to do this (for reasons discussed elsewhere), but there are occasions where it is convenient to do so.

How does these notations relate to the input of "attribute groups". Where do I enter males and females respectively (first or second entry) in the "Enter specifications for MARK analysis" panel?

In the case where gender is coded as an individual covariate, you don't specify them as separate attribute groups - you would specify the individual covariate as 'sex' (or some such). The key is to remember that if you use that approach, that the classification variable is now being treated as an individual covariate, and that is where you specify it it in the numerical estimation window.

[fhensley]

I guess I'm having similar problems. I can't get any model to run because I get the error "not enough groups counts specified on encounter history"

Is MARK unhappy with how I have designated my 4 categories (hatchling, juvenile, male, female) or is it unhappy with my sample sizes? Here's how my .inp data lines look:

/*11 encounter intervals, 59 individuals*/
/* 25 hatchlings, 15 juvies, 12 males, 7 females*/

/*1-3*/ 000100000000010;
/*1-8*/ 000110000001000;

Why won't MARK read the first 11 as encounters and then recognize that animal 1-3 is a male, and animal 1-8 is a hatchling??

Thanks for any help for the newbie!

Frank Hensley

[jlaake]

You need spaces between ch and each frequency

/*1-3*/ 00010000000 0 0 1 0;
/*1-8*/ 00011000000 1 0 0 0;

[cooch]

/*11 encounter intervals, 59 individuals*/
/* 25 hatchlings, 15 juvies, 12 males, 7 females*/

/*1-3*/ 000100000000010;
/*1-8*/ 000110000001000;

Why won't MARK read the first 11 as encounters and then recognize that animal 1-3 is a male, and animal 1-8 is a hatchling??

Thanks for any help for the newbie!

As per Jeff's answer, you need spaces between the encounter history, and other elements of interest. As per *all* of the examples in Chapter 2, the *only* contiguous string in each line of the .INP file is the encounter history itself. All other elements (dummy variables, covariates,...) are separated by *at least* one blank space.