Explanatory power of a log-linear model

Can someone explain Question ID: 485704 Please?

Why cant we decide based on the Standard Error differences of the two models?

Ans given by question :

"To actually determine the explanatory power for sales itself, fitted values for the log-linear trend would have to be determined and then compared to the original data. The given information does not allow for such a comparison. "

you can find the full question here : http://mycfaspace.com/index.php/forum/32-cfa-level-2-quantitative-methods/8737-2013-cfa-level-2-quantitative-analysis-session-3-reading-13-practice-questions-sample-questions

MUCH APPRECIATED!!!

Double post.

Simply put, the dependent variables are different-- one model is predicting Sales and the other model is predicting the natural log of Sales. You cannot directly compare model fit for these models because the left-hand sides (dependent variables) are not the same.

To make the comparison, you would need to get the predicted Y-values from the Ln(Sales) model, giving you Ln(y-hat) values. Then, you would take the inverse of the natural log for each of these values. From here, you could compute an R-squared value for the natural log model, a pseudo R-squared. You could compare this pseudo R-squared to the R-squared of the model predicting Sales.

Hope this helps!

Edit: Also, to answer your original question-- you can’t compare the standard errors in the same way that you can’t compare r-squared or the f-tests for these models (or any models with different Dependent Variables)— each of these model-based statistics is a function of Y-values and predicted Y-values. The models don’t use the same Y-values and predicted Y-values (as I mentioned previously, it’d be like comparing oranges to apples).

Thanks a lot!! it makes perfect sense!

Glad to help!