Interpretation of logarithms in a regression

 

If you do not see the menu on the left please click here

 

Taken from Introduction to Econometrics from Stock and Watson, 2003, p. 215:

 

Y=B0 + B1*ln(X) + u ~ A 1% change in X is associated with a change in Y of 0.01*B1

 

ln(Y)=B0 + B1*X + u ~ A change in X by one unit (∆X=1) is associated with a (exp(B1) - 1)*100 % change in Y

 

ln(Y)=B0 + B1*ln(X) + u ~ A 1% change in X is associated with a B1% change in Y, so B1 is the elasticity of Y with respect to X.

 

 

 

Out-of sample test

 

SOURCE: http://www.stata.com/help.cgi?predict

 

    predict can be used to make in-sample or out-of-sample predictions:

 

        6)  predict calculates the requested statistic for all possible

            observations, whether they were used in fitting the model or not.

            predict does this for the standard options (1) through (3) and

            generally does this for estimator-specific options (4).

 

        7)  predict newvar if e(sample), ...  restricts the prediction to the

            estimation subsample.

 

        8)  Some statistics make sense only with respect to the estimation

            subsample.  In such cases, the calculation is automatically

            restricted to the estimation subsample, and the documentation for

            the specific option states this.  Even so, you can still specify

            if e(sample) if you are uncertain.

 

        9)  predict can make out-of-sample predictions even using other

            datasets.  In particular, you can

 

            . use ds1

             (fit a model)

            . use two              /* another dataset */

            . predict yhat, ...    /* fill in the predictions */”