+ Reply to Thread
Results 1 to 2 of 2

Thread: Unique vs. Shared Variance Multiple Regression

  1. #1
    Points: 495, Level: 9
    Level completed: 90%, Points required for next Level: 5

    Posts
    22
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Unique vs. Shared Variance Multiple Regression




    Hi all,

    I am having trouble with interpreting the explained variance of my multiple regression model. I ran a hierarchical model with nine predictors overall, with five in the first block and four in the second. The addition of the four new blocks explained an extra 24% of the variance of the DV after controlling for the other five variables in block one.


    Three of the four new predictors were significant predictors of the DV. How do I determine the unique variance explained by each of the variances? I should mention that some of my variables are dummy coded so interpreting the standardised beta is not possible. My sample size is large so the study is not underpowered.

    Do I run a new model with the eight predictors at step 1, and then put the variable I want to check the unique variance for at step 2? Then I would check the r2 change value. Or is there another method?

    The problem is, when I do the aforementioned R2 change method, one of the significant variables explains about 11% of the variance in the DV. The other two significant DVs contribute less than 1% to the prediction of the DV when I control for all other variables. Yet when the four variables are added to the model together, they explain an additional 24% of the variance.

    The variables were found to be significant predictors, I don't understand how they contribute so little to the variance individually and yet collectively, the four variables explain far more than their unique contributions suggest. Is there something I'm missing here?

  2. #2
    Points: 495, Level: 9
    Level completed: 90%, Points required for next Level: 5

    Posts
    22
    Thanks
    13
    Thanked 0 Times in 0 Posts

    Re: Unique vs. Shared Variance Multiple Regression


    Anyone? My supervisor is going to beat me up at this rate lol

+ 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