+ Reply to Thread
Results 1 to 4 of 4

Thread: Interpreting log-linear coefficients

  1. #1
    Points: 226, Level: 4
    Level completed: 52%, Points required for next Level: 24

    Posts
    2
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Interpreting log-linear coefficients




    Hi there,

    I am fairly new to regression and have been set the task of modelling a data set in the form ln(y) = bx1 +bx2 +bx3 +bx4 +bx5 + e.

    In the regression output the coefficient value of b range from -0.081 to 0.61. I understand, that for b1=0.61, that a 1 unit increase in x1 will result in a 0.61 increase in ln(y). Using exp(0.61)=1.84, I have interpreted this as a 84% increase in y. For negative coefficients, there seems to be some indication of a different process - can anyone shed any light on this for me?

    At the moment, for x3=-0.081, I have used exp(-0.081)=0.92, indicating a 8% decrease in y for every unit change in x. Is this correct?

    Am I approaching this regression output in the correct way? Any advise would really be helpful and much appreciated.

    Many Thanks,
    Adam

  2. #2
    Points: 651, Level: 13
    Level completed: 2%, Points required for next Level: 49

    Location
    Groningen, Netherlands
    Posts
    43
    Thanks
    2
    Thanked 2 Times in 2 Posts

    Re: Interpreting log-linear coefficients

    Uhm, tricky question of which i can't figure out the answer. I also looked at this source, where the same question was raised:
    http://www.mrexcel.com/forum/showthread.php?t=301293
    I think the answer is complicated. If I had such a task, it would be more easier if you make your model completely multiplicative, that is:
    Ln(y)= b1 Ln x1 + b2 Ln x2 etc. or y= x1^b1 * x2^b2 etc. However, that was not your task if i'm right. Hope you find the answer in the link.

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

    adam15 (04-19-2012)

  4. #3
    Points: 226, Level: 4
    Level completed: 52%, Points required for next Level: 24

    Posts
    2
    Thanks
    1
    Thanked 0 Times in 0 Posts

    Re: Interpreting log-linear coefficients

    Thanks Eyecatcher. The link was helpful. I think I will focus on the effect each independent has upon the ln(y).

  5. #4
    Points: 1,980, Level: 26
    Level completed: 80%, Points required for next Level: 20
    derksheng's Avatar
    Posts
    247
    Thanks
    49
    Thanked 35 Times in 28 Posts

    Re: Interpreting log-linear coefficients


    ln(y) = a + bx + e
    y = e^(a+bx+e)


    dy/dx = b*e^(a+bx+e) = b*y
    elasticity = (dy/y)/(dx/x) = b*y*(x/y) = b*x

    Same with multiple independent variables.

+ Reply to 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