DETECTING PARAMETER REDUNDANCY

Necessary and sufficient conditions are established for the parameter redundancy of a wide class of nonlinear models for data distributed ac cording to the exponential family. The likelihood surfaces for paramet er-redundant models possess completely hat ridges. Whether a model is parameter redundant can be established by checking the rank of a deriv ative matrix, using a symbolic algebra package. A feature of contingen cy table applications is the need to extend conclusions from particula r to general dimensions. We meet this via an extension theorem. Exampl es are given from the area of animal survival estimation using mark-re capture/recovery data.