https://communities.sas.com/t5/SAS-Procedures/How-to-compute-marginal-effects-in-general-linear-model/m-p/67887#M19436 I guess the marginal effect of factor x_i is just (<I>beta_i*yhat</I>), where <I>beta_i</I> is the estimated coefficient for x_i, and <I>yhat</I> is the predicted value of y. It seems that the built-in distribution of y doesn't make a difference. Only the link function counts. Maybe it's always the partial derivative of the inverse link function with respect to x_i for all generalized linear models.<BR /> <BR /> I figured this out by playing it in STATA.<BR /> For dataset,<BR /> /*<BR /> x1 x2 y<BR /> 1 2 33.11<BR /> 2 1 2.5<BR /> 3 0 0.4<BR /> 4 1 0.1<BR /> 5 2 0.2 <BR /> */<BR /> Try commands<BR /> /*<BR /> glm y x1 x2, family(gamma) link(log)<BR /> mfx<BR /> */<BR /> or<BR /> /*<BR /> glm y x1 x2, family(poisson) link(log)<BR /> mfx<BR /> */<BR /> The marginal effects just follow the same rule. Tue, 06 Jan 2009 21:38:21 GMT deleted_user 2009-01-06T21:38:21Z https://communities.sas.com/t5/SAS-Procedures/How-to-compute-marginal-effects-in-general-linear-model/m-p/67886#M19435 Hi :<BR /> <BR /> How to get the marginal effects of each independent variable(x1,x2,x3) in general linear model with Y(dependent variable) continuous and link function is log and built-in distribution is gamma.<BR /> <BR /> <BR /> Thank you! Mon, 05 Jan 2009 17:48:41 GMT https://communities.sas.com/t5/SAS-Procedures/How-to-compute-marginal-effects-in-general-linear-model/m-p/67886#M19435 deleted_user 2009-01-05T17:48:41Z https://communities.sas.com/t5/SAS-Procedures/How-to-compute-marginal-effects-in-general-linear-model/m-p/67887#M19436 I guess the marginal effect of factor x_i is just (<I>beta_i*yhat</I>), where <I>beta_i</I> is the estimated coefficient for x_i, and <I>yhat</I> is the predicted value of y. It seems that the built-in distribution of y doesn't make a difference. Only the link function counts. Maybe it's always the partial derivative of the inverse link function with respect to x_i for all generalized linear models.<BR /> <BR /> I figured this out by playing it in STATA.<BR /> For dataset,<BR /> /*<BR /> x1 x2 y<BR /> 1 2 33.11<BR /> 2 1 2.5<BR /> 3 0 0.4<BR /> 4 1 0.1<BR /> 5 2 0.2 <BR /> */<BR /> Try commands<BR /> /*<BR /> glm y x1 x2, family(gamma) link(log)<BR /> mfx<BR /> */<BR /> or<BR /> /*<BR /> glm y x1 x2, family(poisson) link(log)<BR /> mfx<BR /> */<BR /> The marginal effects just follow the same rule. Tue, 06 Jan 2009 21:38:21 GMT https://communities.sas.com/t5/SAS-Procedures/How-to-compute-marginal-effects-in-general-linear-model/m-p/67887#M19436 deleted_user 2009-01-06T21:38:21Z