Program RELEASE error problem

[birdman]

I am working on a 10 year dataset of CJS data on a small passerine. Data are somewhat sparse, but on the whole it is a pretty good dataset. I’ve run the models to test the hypotheses we are most interested in concerning survival, and found that for any model incorporating temporal variation, the models just don’t produce meaningful parameter estimates or confidence intervals. Age models with a constant age specific structure (>1yr versus <1yr) produce good estimates and reasonable SEs. This makes sense to me as recapture rates were pretty low overall.

I want to assess GOF using RELEASE as a standalone as outlined in Chapter 5. I modified my input file with the following (including a couple of encounter histories):

/*EH Adult Juv Number*/
PROC CHMATRIX OCCASIONS=10 GROUPS=2 SUMMARY FULLM;
0100000000 1 0; /*66-24525*/
0100000000 1 0; /*66-24526*/

I then found my DOS prompt (haven’t been there in a while!) and entered:
REL_32 I=C:\Users\me\Documents\SALS.inp O=RelOut.LST
And hit enter.

I get the following message:
“Length of command exceeded MAXCHR/2 characters. RELEAS ERROR TERMINATION"

Can anyone give me any hints as to where my problem might lie? I’ve searched the forum and can’t get a hit that seems immediately relevant.

Thanks,
Pru

[birdman]

Following up with a bit more information. I have tried reducing the input file to a single group, and have tried using the minimal allowed input that should run the analysis according the GI.

i.e. PROC CHMATRIX OCCASIONS=10 GROUPS=1;

Regardless of what I do I get the error. I also found that it is producing the specified output file, and that it aborts the run 2 lines from the end of the input file. The readout at the end of the file, including the last few lines of accepted data:

"
The following capture history is a duplicate of number 49 above.
INPUT --- /*940-09833*/
INPUT --- 0000000100 1 0;
The following capture history is a duplicate of number 49 above.
INPUT --- /*940-09834*/
INPUT --- 0000000100 1 0;
INPUT --- /*940-09834*/
INPUT --- 0000000100 1 0;

Length of command exceeded MAXCHR/2 characters.
"
So it appears to read recordID 940-09834 two times, then it gives the error. I checked the input file... There are two records after this one, and there are no irregularities in the input file at this point. I tried deleting the last two records and running the file again. I get the same response, just 2 records higher, at 940-09832.

That's all I can add at the moment.

[cooch]

i.e. PROC CHMATRIX OCCASIONS=10 GROUPS=1;

OK, so you have 1 group, which is what causes the problem since...

Code: Select all

  INPUT ---       0000000100 1 0;
  INPUT ---       0000000100 1 0;
  INPUT ---       0000000100 1 0;


...these encounter histories have 2 groups (i.e., to frequency columns).

[birdman]

i.e. PROC CHMATRIX OCCASIONS=10 GROUPS=1;

OK, so you have 1 group, which is what causes the problem since...

Code: Select all

  INPUT ---       0000000100 1 0;
  INPUT ---       0000000100 1 0;
  INPUT ---       0000000100 1 0;


...these encounter histories have 2 groups (i.e., to frequency columns).

Evan, thanks for the response. As you note, in this example, I incorrectly mixed the single group input with the 2 group output. However, that is a mistake in transcription on my part. As noted in the original message, I started with two groups, but then compressed to one to see if that cleared the problem. In the second analysis I actually did reduce to a single grouping column.

In any case, running it with the original file (2 grps) or the newly produced one (1 grp) I continue to generate the error. I wish it were simply misgrouping. Any other thoughts?