Contents
- Index
Bootstrap Options
The default for the bootstrap goodness-of-fit procedure is to compute the deviance, degrees of freedom of the deviance, and c-hat. However, to do this, the rank of the variance covariance matrix must be computed to determine the number of parameters estimated. To get valid values for the number of parameters estimated, the variance estimation method used is the 2ndPart option (2nd partials). This procedure is computationally intensive. If the degrees of freedom of the deviance and c-hat are not required, then only the deviance is provided in the output, and the time to compute a valid variance-covariance matrix is not required.
Note that the bootstap goodness-of-fit procedure is different than the bootstrap data procedure.