Group coding in inp file and assigning Group labels in Mark

[njstricker]

Here's the short form to my question: How does Mark actually assign the group names to groups in your input file?

Here's the long version. I have an input file with an encounter history for each individual, followed by the frequency column (n=1 for each record) and 2 columns for the dummy coding of my groups--birds banded as hatch year (HY= 1 0) and after hatch year (AHY= 0 1). So, a recapture-only record might be: 00010011001 1 0 1; for an AHY bird. In setting up the analysis, I inform Mark that I have 2 groups, and their labels, HY and AHY.

How does Mark assign the group labels to the data? Does it read the columns (i.e. 1 in the first column = HY, 1 in second column = AHY)? Or does it read rows (i.e. coding encountered in the first row--e.g. first record-- is assigned HY, the next different coding encountered is assigned AHY)?

Thanks!
Nathan

[cooch]

Here's the short form to my question: How does Mark actually assign the group names to groups in your input file?

Here's the long version. I have an input file with an encounter history for each individual, followed by the frequency column (n=1 for each record) and 2 columns for the dummy coding of my groups--birds banded as hatch year (HY= 1 0) and after hatch year (AHY= 0 1). So, a recapture-only record might be: 00010011001 1 0 1; for an AHY bird. In setting up the analysis, I inform Mark that I have 2 groups, and their labels, HY and AHY.

You've not set up your INP file correctly. You have 2 groups - HY and AHY. So, you need 2 columns after the encounter history. Let '1 0' indicate HY, and '0 1' indicate AHY. So

1011010 1 0; <- would be a single individual banded as HY
1011101 0 1; <- would be a single individual banded as AHY

You don't need the extra column to indicate frequency. This is indicated explicitly by the coding. The following might make this clearer:

101011 45 62; <- 45 individuals banded as HY, and 62 individuals banded as AHY, each with the same encounter history.

This is laid out fairly well in Chapter 2.

As for how MARK knows which groups are which, the columns are read left -> right, and as such, you have to remember which columns correspond to what, and then label them accordingly in the data sepcification window.

Have another read of Chapter 2, and then Chapter 4.

[njstricker]

Got it! I misunderstood the association between the frequency and the dummy coding. The 1 IS the frequency, not just a group indicator.

My follow-up question then is, once I make the correction to my input file, do I need to create a new analysis file in Mark, or can I rerun my previous analyses from the old file from the Results Browser/PIM?

[cooch]

Got it! I misunderstood the association between the frequency and the dummy coding. The 1 IS the frequency, not just a group indicator.

Correct - read Chapter 2, carefully. That will help. Basically, the columns in the INP file are generally not indicator variables in the sense they are in design matrices (which is covered in considerable detail in Chapter 7).

My follow-up question then is, once I make the correction to my input file, do I need to create a new analysis file in Mark, or can I rerun my previous analyses from the old file from the Results Browser/PIM?

You need to start over...

[njstricker]

Thanks! The use of dummy variables is a little different here than in some other applications I've used.

Back to the drawing board, er, PIM. . .