Individual Covariates and Real Parameter Estimates
Individual Covariates and Real Parameter Estimates
When individual covariates are used in a model, the real parameter estimates and the derived parameter estimates are different for each set of individual covariate values in the model. In the Run Window, 3 options are available to select what individual covariate values are used to compute the real and derived parameter estimates displayed in the output.
The first option is to use the individual covariate values for the first encounter history in the encounter histories file. This option just provides a (possibly) reasonable set of real parameter estimates for an individual.
The second option is to use the mean value of each of the individual covariates in the model. This option normally makes a lot of sense, except in cases where dummy variables are included in the individual covariates. For example, suppose adults are coded as 1 and subadults coded as 0. What does it mean to compute real parameter values for the mean of this age variable?
The last option is to allow the user to specify values of the individual covariates to use to compute the real parameter estimates. This option is particularly useful when combined with the Standardize Individual Covariates option, because the mean of a standardized covariate is 0, and roughly 95% of the values will be <1.96, and roughly 95% of the values will be >-1.96. Thus a range of the real parameter estimates can be computed by using a range of the individual covariate values.
The value of the individual covariate used to compute the real and derived parameters can be changed in the model output using the ReGenerate Real Derived Model(s) menu choice under the Results Browser Run menu. This option is useful for model averaging when some of the models were run with different individual covariate values than other models.
You can graph the relationship between a real paramete and an individual covarariate with the Individual Covariate Plot capability.