Standardized Beta Coefficient

[Paul Smith]

Having run my linear regression analysis that contains two independent variables and one dependent variable, the Standardized Beta Coefficient returned for both independent variables is .803 and .347 respectively.

Is is correct to assume that in terms of relative importance to the model, independent variable 1 (.803) has 2.31 times (803/.347) more importance than independent variable 2 (.347)?

Thanks

[Karel Novak]

What standardized beta coefficients do, is that they compare the strength of the average effect of each individual independent variable to the dependent variable. Comparability is achieved via standardization of each indenpendent variable so the mean is 0 and SD 1.

Therefore, I wouldn't really use the word "importance" since it may imply substantial importance, which is rather unquantifiable and strongly dependent on the context of research area. But you can say that you measured which variable strongly explains the variation of dependent variable

[Phil Bromiley]

Alternatively, for a one standard deviation change in both variables, the predicted value increases .8 of a standard deviation (of the dv) for one and .347 of a standard deviation for the other.