+ Reply to Thread
Results 1 to 4 of 4

Thread: Is this data acceptable for a multiple regression?

  1. #1
    Points: 258, Level: 5
    Level completed: 16%, Points required for next Level: 42

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Is this data acceptable for a multiple regression?




    Hi all,

    I have data from a community with MIN size (g) MAX size (g) MAX size without neighbours (g), Height (cm) and Area (cm^2). I have this for about 27 species.

    I want to see which of those variables is the best predictor of ramet counts, which I also measured in this community.

    All of the data are normally distributed post log-transformation. I have already performed multiple regression analyses, but when I attempt to calculate the partial correlation coefficients I seem to run into an issue where my partial coefficients are greater than the overall Rsq value. I am wondering if perhaps I violated some assumption that I didn't take into consideration?

    Thanks.

  2. #2
    TS Contributor
    Points: 17,749, Level: 84
    Level completed: 80%, Points required for next Level: 101
    Karabiner's Avatar
    Location
    FC Schalke 04, Germany
    Posts
    2,540
    Thanks
    56
    Thanked 640 Times in 602 Posts

    Re: Is this data acceptable for a multiple regression?

    my partial coefficients are greater than the overall Rsq value.
    Do you compare here R with (partial) r?

    Kind regards

    K.

  3. #3
    Points: 258, Level: 5
    Level completed: 16%, Points required for next Level: 42

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Is this data acceptable for a multiple regression?

    Quote Originally Posted by Karabiner View Post
    Do you compare here R with (partial) r?

    Kind regards

    K.
    Thanks for your reply,

    To clarify:

    I am comparing the R square for the best/top model with the partial coefficients of determination for each variable in the top model. So for example: the best model is ramet counts ~ height + MAX Size. The R square for that model is 0.45

    When I attempt to figure out the partial coefficients of determination I divide the residual SE of the reduced model by the residual SE of the best model and divide by the residual SE of the reduced model. So let's say I was trying to figure out how much variation in ramet counts was explained by height if I hold MAX size constant. I get a partial coefficient of determination of 0.46 for MAX size and 0.42 for height.

  4. #4
    Points: 258, Level: 5
    Level completed: 16%, Points required for next Level: 42

    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Is this data acceptable for a multiple regression?


    I meant (residual SE reduced model - residual SE top model) / residual SE reduced model. Sorry for the confusion!

+ 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