+ Reply to Thread
Results 1 to 3 of 3

Thread: log-lin regression vs lin-lin

  1. #1
    Points: 69, Level: 1
    Level completed: 38%, Points required for next Level: 31

    Posts
    2
    Thanks
    1
    Thanked 0 Times in 0 Posts

    log-lin regression vs lin-lin




    Hi guys,
    I'm a new guest in this forum so first of all I'm Giovanni and nice to meet you.

    I'm solving a problem set of Econometrics course and it's the first time I face non linear reg. In fact, prof asked us to intrepret the result of the non linear reg.

    The dataset is the most popular of all, that is the dependence of income to year of education. The independent variable (year of education) is called educ and the dependent variable (income) is called y.

    After some descriptive analysis, we have to choose the best model (lin-lin, log-lin, log-log).

    I choose the first one, cause it's the higher adjusted R2.

    Now, he asked us to multiply the regressand * 1000 and analyze the results, in the special case of log-lin results.

    My question is: why the B1 remain the same and the constant modifies? (pic attached).

    Thank you guys!

    Giovanni
    Attached Thumbnails Attached Thumbnails Click image for larger version

Name:	Schermata 2015-03-15 alle 16.21.16.jpg‎
Views:	82
Size:	30.5 KB
ID:	5160   Click image for larger version

Name:	Schermata 2015-03-15 alle 16.21.09.jpg‎
Views:	48
Size:	26.2 KB
ID:	5161  

  2. #2
    TS Contributor
    Points: 5,246, Level: 46
    Level completed: 48%, Points required for next Level: 104
    maartenbuis's Avatar
    Location
    Konstanz
    Posts
    372
    Thanks
    3
    Thanked 146 Times in 123 Posts

    Re: log-lin regression vs lin-lin

    You cannot compare R^2s (adjusted or not) across models with when you applied a non-linear transformation to the response/dependent/explained/y-variable. The logic is that R^2 measures the proportion of explained variance in y, and by non-linearly transforming y you change that variance, i.e. the denominator in that proportion.

  3. The Following User Says Thank You to maartenbuis For This Useful Post:

    gionni25 (03-16-2015)

  4. #3
    Points: 69, Level: 1
    Level completed: 38%, Points required for next Level: 31

    Posts
    2
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Re: log-lin regression vs lin-lin


    Quote Originally Posted by maartenbuis View Post
    You cannot compare R^2s (adjusted or not) across models with when you applied a non-linear transformation to the response/dependent/explained/y-variable. The logic is that R^2 measures the proportion of explained variance in y, and by non-linearly transforming y you change that variance, i.e. the denominator in that proportion.
    Ah, ok!

    What criterium I have to follow in chasing type of model?

    Thanks

+ Reply to Thread

           




Tags for this Thread

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts






Advertise on Talk Stats