Object Reference : Object View and Procedure Reference : Equation
Influence statistics.
Displays influence statistics to discover influential observations, or outliers.
equation_name.infstats(options) stats_list [@ save_names]
If no stats_list is provided all of the statistics will be displayed. save_names is an optional list of names for storing the statistics into series in the workfile. The save_names should match the order in which the keywords in stats_list are entered.
Show a table of the statistics (the default is to display a graph view of the specified statistics).
rows = key
The number of observations/rows to display in the table, where key can be either “50”, “100” (default), “150”, or “200”.
sort = key
Sort order for the table, where key can be “r” (Residual - default), “rs” (RStudent), “df” (DFFITS), “dr” (Dropped Residual), “cov” (COVRATIO), “h” (diagonal elements of the hat matrix).
Display the table by the sort order rather than by the observation order.
Force the dialog to appear from within a program.
The stats_list parameter is a list of keywords indicating which statistics to display. It may take on the values:
The studentized residual: the t-statistic on a dummy variable that is equal to 1 on that observation only.
The scaled difference in fitted values.
Dropped residual: the estimated residual for that observation had the equation been run without that observation.
The ratio of the covariance matrix of the coefficients with and without that observation.
Diagonal elements of the hat matrix:
eq1.infstats(t, rows=150, sort=rs) rstudent covratio dffits @ rstuds covs
will display a table showing the 150 largest RSTUDENT statistics, along with the corresponding COVRATIO and DFFITS statistics. It will save the RSTUDENT and COVRATIO statistics into the series in the workfile named RSTUDS and COVS, respectively.
See also “Influence Statistics” . See also Equation::infbetas.