Data Type Change

Under the PIM menu choice in the Results Browser window, you can choose to change the data type of the models you are constructing. Specifically, the PIMs are changed to match the new data type that you have chosen. Any PIM Chart, PIM Windows, or Design Matrix Window that are open at the time are automatically closed, because all these windows depend on the structure of the PIMs for the current data type.

Adding models for a different data type to the existing Results Browser window can be very confusing. For example, the models added may use a different likelihood function, and hence AIC values will not correspond to existing models. Further complications result with the model averaging procedure, because the parameters selected to average may not exist in all the data types present in the Results Browser, or the indexing of parameters will change, making the model averaging results invalid.

The main reason for including the Data Type Change option in MARK is to allow experienced users to explore other models for their data set, but they must be cognizant of other problems they may be creating for themselves.

When you select the PIM | Change Data Type menu choice, you are given a list of the data types with which the Encounter Histories are compatible, including the current data type. Select one of the data types, and the PIMs will be changed accordingly. As an example, if you have initially selected the Live Recaptures data type (Cormack-Jolly-Seber), the the following data types are compatible with this data format:

- Jolly-Seber Lambda -- Burnham's parameterization of the Jolly-Seber data type with parameters phi, p, lambda, and N

- Pradel Survival and Seniority data type with parameters phi, p, and gamma

- Pradel Survival and Lambda data type with parameters phi, p, and lambda

- Pradel Survival and Recruitment data type with parameters phi, p, and f

- POPAN data type with parameters phi, p, pent, and N

- Link-Barker Jolly-Seber data type with parameters phi, p, and f

You can select any of these models from the list presented, and now compute estimates for this data type.

As another example, the closed captures data type contains 12 models. You can select any of these 12 models from the Change Data Type menu.

Care must be taken when model averaging real parameters to be sure the same parameter is being averaged across all the models in the results browser. A check box is provided in the model averaging dialog window to select only models for the same data type.