help for rvlrplot

Plot residuals vs lagged residuals for time series data

rvlrplot [ , residualtype qualifier scatter_options plot(plot) ]


rvlrplot plots residuals from the last model versus lagged (i.e. lag 1) residuals. Data must have been tsset previously.

By default, residuals are whatever predict, res produces after a model. The plot is restricted to the estimation sample.


The correlation between residuals and lagged residuals is calculated quietly. This may be retrieved by return list and then (say) used in a second pass giving the correlation on the graph.

If the order of observations is meaningful, but data have not been tsset, this may be done by

gen t = _n tsset t


residualtype specifies a type of residual other than the default. The following types are currently supported: anscombe, deviance, likelihood, pearson, residuals, response, rstandard, rstudent, score, working.

qualifier specifies one of standardized, studentized, modified, adjusted.

scatter_options are options of twoway scatter.

plot(plot) provides a way to add other plots to the generated graph; see plot_option.


. tsset time . regress whatever time . rvlrplot


Nicholas J. Cox, University of Durham, U.K. n.j.cox@durham.ac.uk

Also see

On-line: help for predict, scatter, tsset, modeldiag