Zelig4: Everyone's Statistical Software

Description Usage Arguments Value Note Author(s) Examples

The param method is used by developers to specify simulated and fixed ancillary parameters of the Zelig statistical model. That is, this method is used between the zelig2 function and the qi as a helper function that specifies all the necessary details needed to simulate quantities of interest, given the fitted statistical model produced by the zelig2 function.

1	param(obj, num, ...)

`obj`	a `zelig` object
`num`	an integer specifying the number of simulations to sample
`...`	optional parameters which will likely be ignored

The main purpose of the param function is to return a list of key-value pairs, specifuing information that should be shared between the qi function and the fitted statistical model (produced by the zelig2 function. This list can contain the following entries:

`simulations`	specifies a set of simulated parameters used to describe the statistical model's underlying distribution
`alpha`	specifies the fixed (non-simulated) ancillary parameters used by the statistical model's underlying distribution
`family`	specifies a family object used to implicitly define the `link` and `linkinv` functions. That is, this specifies the "link" and "inverse link" functions of generalized linear models
`link`	specifies the `link` function to be used. This parameter is largely unimportant compared to the "inverse link" function
`linkinv`	specifies the `linkinv` function to be used.

The 'param' function is a method meant to be overloaded by Zelig Developers

Matt Owen mowen@iq.harvard.edu

param.some.model <- function (obj, num, ...) {
  list(
       simulations = NULL,
       alpha = NULL,
       link = NULL,
       linkinv = NULL,
       fam = NULL
       )
}