+ Reply to Thread
Results 1 to 5 of 5

Thread: Comparison of Coefficients of two different Regression Models

  1. #1
    Points: 4, Level: 1
    Level completed: 7%, Points required for next Level: 46

    Posts
    1
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Comparison of Coefficients of two different Regression Models




    Hi everyone,

    I am analyzing the impact of price and promotion activities on sales in two separate 2-Way-Fixed Effects models. One model is for private labels, the other for national labels.

    I now want to compare the coefficients of the two models with each other. My goal is to be able to state which type of price and promotion activity has a stronger effect for which brand.

    Can anyone suggest an appropriate method? For linear models I found the Chow Test.

    Thank you very much in advance"

  2. #2
    Points: 2,462, Level: 30
    Level completed: 8%, Points required for next Level: 138

    Posts
    200
    Thanks
    20
    Thanked 48 Times in 43 Posts

    Re: Comparison of Coefficients of two different Regression Models

    So your goal is to say if two regresson coefficients are significantly different? If so: you have from both regression models the coefficient (beta) and the standard error of this coefficient (se). Assuming that the distribution of the betas is approximately normal (e.g., if both regression analyses rely on more than 30 samples) you can calculate it in R in the following way using a z-statistic:

    SE <- sqrt(se1^2+se2^2) #that is the approximate standard error of the difference
    z<-(beta1-beta2)/(SE) #that is the z-transformed difference
    p <- 2*pnorm(-abs(z)) #that is the p-value

  3. #3
    Points: 2,000, Level: 26
    Level completed: 40%, Points required for next Level: 150

    Location
    New Zealand
    Posts
    228
    Thanks
    3
    Thanked 48 Times in 47 Posts

    Re: Comparison of Coefficients of two different Regression Models

    Is the data suitable for combining into a 3 way model with private/national as a factor?

  4. #4
    TS Contributor
    Points: 12,501, Level: 73
    Level completed: 13%, Points required for next Level: 349
    rogojel's Avatar
    Location
    I work in Europe, live in Hungary
    Posts
    1,491
    Thanks
    162
    Thanked 334 Times in 314 Posts

    Re: Comparison of Coefficients of two different Regression Models

    hi,
    could you merge the two datasets, add a discrete variable for label type and anlyse it this way? It would provide all the info you need and some more if you check interactions as well.

    regards

  5. #5
    Fortran must die
    Points: 62,031, Level: 100
    Level completed: 0%, Points required for next Level: 0
    noetsi's Avatar
    Posts
    6,538
    Thanks
    693
    Thanked 916 Times in 875 Posts

    Re: Comparison of Coefficients of two different Regression Models


    I believe, although its been a long time, that there are problems in comparing across data sets with different distribution (which has been suggested above). Are you using the same data set to build both models, or different data sets.

    I think merging the data with a private/public dummy makes more sense. You might find an intervention effect so that the regression model is different for private and public units.
    "Very few theories have been abandoned because they were found to be invalid on the basis of empirical evidence...." Spanos, 1995

+ 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