Pre-Defined Models

Contents – Index

Pre-Defined Models

From the Results Browser Window menu, Run | Pre-Defined Model(s) allows you to have MARK build a model for you.  A list of the possible models for group*time (g*t) are provided for PIM coding, and for Design Matrix coding.  For PIM coding, possible models for each parameter are the dot (.), group (g), time (t), and group*time (g*t) models, with all permutations across all the parameters in the model.  The same models are also available for the Design Matrxi coding, but this coding also allows additive models, such as group + time (g+t).  For Design Matrix coding, the PIMs used are the g*t coding for each parameter.  Thus, this option will not allow you to build age-structured models, because age-structured models require a different set of PIMs than the g*t setup.  There is an option under the File | Preferences dialog window (Set Preferences) to change the design matrix coding to use the first row of a time effect as the refernce row, rather than the last row.

Another limitation of the Pre-Defined Models option is that only analyses with less than 512 possible models are considered.  For example, a live recaptures analysis with groups and time would generate 16 possible PIM models.  However, a robust design data analysis with 5 sessions would generate 2^18 possible models because each of the 5 sessions would have 3 parameters (giving 15 different combinations), plus the survival, immigration, and emigration parameters.  Therefore, the Pre-Defined Models option will not work for this complex analysis.
 
Closed Captures Data Types

For the closed captures data types, including the Huggins data type and the Huggins random effects data type, the set of 8 models from Otis et al. (1978) is available to chose from.