help devr2


devr2 -- Calculates a deviance based R-squared measure for models estimated using the [R] glm command. The measure is based on Cameron and Windmeijer, 1997.


devr2 calculates a deviance based R-squared measure for regression models in the exponential family with known scale parameter. This is equal to

1-[(deviance of the fitted model)/(deviance of the constant only model)]

It measures the proportionate reduction in recoverable information due to the inclusion of regressors, where information is measured by the estimated Kullback-Leibler divergence, and may be loosely interpreted as the fraction of uncertainty explained by the fitted model.

Further details are given in Cameron and Windmeijer, 1997.

Currently, the command can only be used following model estimation using the [R] glm command.

Saved results

devr2 saves the following in r():

Scalars r(dev_model) Deviance of the fitted model r(dev_null) Deviance of the constant only model r(devr2) Deviance based R-squared value


Brilleman, S. email:


Cameron, A. C. & Windmeijer, F. A. G. (1997). An R-squared measure of goodness of fit for some common nonlinear regression models. Journal of Econometrics 77:329-342.

Also see

Online: [R] glm