When we take the log, we change the units of measurement of the dependent variable. You can do it even the dependent variable if ratio. The spread of the residuals changes systematically with the values of the dependent variable. The purpose of the transformation is to remove that systematic change in spread. Sometime, taking the log can normalize the variable. However, the coefficients of the independent variables will be changed. It depends on whether you take the log on both dependent and independent variables or not. If you take the log on both sides, then the coefficient of x is the elasticity of y respect to x. If you take the log only on dependent variable, then the 100*coefficient of x is the semi-elasticity of y respect to x.
The spread of the residuals changes systematically with the values of the dependent variable. The purpose of the transformation is to remove that systematic change in spread. Sometime, taking the log can normalize the variable.
扫码加好友，拉您进群
全部版块
我的主页
收藏
