Error message--group counts

[njstricker]

Hi!

I'm attempting to run a multi-state recaptures only analysis, looking at transition probabilities among 5 different groups. When attempting to run the analysis, I keep getting the error message: "Not enough group counts specified on the encounter history." I have gone through the input file several times--the encounter history and dummy coding for the groups appears to be fine.

I've read elsewhere on this forum that MARK counts the number of semi-colons to determine if there are enough records for the analysis. Does MARK do this count by group, and if so, could that be the source of the error message? I have 2 states, 5 groups, 698 records, but only a few records for a couple of the groups. If the low number of records for a couple of groups is the cause of the error, then I can go back and reorganize my groups, merging some together. But, I'd like to find out first if there might be some other reason for this error message before spending time reworking my data set. Any ideas?

Thanks,

Nathan

[ganghis]

It's almost certainly an error with how the encounter histories have been entered. With questions of this type, please include a small segment of your encounter history file.

[njstricker]

Here are a few lines of the encounter history:

PPAAP 01000;
AAAAA 01000;
PAAAA 01000;
AAAPA 01000;
AAAPA 01000;
AAAAA 10000;
PPAAA 10000;
AAPPA 10000;
AAAAA 10000;
PAPPP 10000;

Hope this helps. 5 periods, 2 states (P,A) and 5 groups. The smallest number of encounter history counts for any one group is 24. There are 698 total records.

[cooch]

Here are a few lines of the encounter history:

PPAAP 01000;
AAAAA 01000;
PAAAA 01000;
AAAPA 01000;
AAAPA 01000;
AAAAA 10000;
PPAAA 10000;
AAPPA 10000;
AAAAA 10000;
PAPPP 10000;

Hope this helps. 5 periods, 2 states (P,A) and 5 groups. The smallest number of encounter history counts for any one group is 24. There are 698 total records.

Simple - groups must be separated by at least a single space (as is noted in several places in Chapter 2). So, for example

PAPPP 10000; INCORRECT

PAPPP 1 0 0 0 0; CORRECT

So, all you need to do is separate the group coding variable columns with at least a space.

[njstricker]

Thanks! I'll give that a try.