+ Reply to Thread
Results 1 to 4 of 4

Thread: Principal component analysis before regression anallysis

  1. #1
    Points: 261, Level: 5
    Level completed: 22%, Points required for next Level: 39

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Principal component analysis before regression anallysis




    Hi this is my first post and i've searched for this answer in vain. I have 50 participants in an experiment i conducted and i want to use a regression analysis but i have 9 independent variables. Since i know that many of these are somewhat related, i wanted to conduct a PCA before doing my regression analysis. Conducting the PCA, i get 1 factor + 3 other variables that can't be explained by factorization (spelling?). Anyway all independent variables in the factor fall under a theoretical concept that makes a lot of sense. My question is, can i input the 3 IV and the factor in my regression model? would that be a good idea?

    Any input is much appreciated

    Thanks in advance

  2. #2
    Points: 2,626, Level: 31
    Level completed: 18%, Points required for next Level: 124

    Location
    Dallas, TX
    Posts
    311
    Thanks
    12
    Thanked 90 Times in 88 Posts

    Re: Principal component analysis before regression anallysis

    I don't see any issue with including the factor along with 3 IVs. If three IVs can't be grouped into any factor then they are independent and should be treated like that.

  3. #3
    ggplot2orBust
    Points: 71,220, Level: 100
    Level completed: 0%, Points required for next Level: 0
    Awards:
    User with most referrers
    trinker's Avatar
    Location
    Buffalo, NY
    Posts
    4,417
    Thanks
    1,811
    Thanked 928 Times in 809 Posts

    Re: Principal component analysis before regression anallysis

    Conducting the PCA, i get 1 factor + 3 other variables that can't be explained by factorization (spelling?).
    I don't get this. The program will continue to parse out and make linear combinations for as many breaks as you tell it. How can you say what 3 variables aren't explained by factorization?

    Another approach is to use a hierarchical regression approach and add the correlated (look at the correlation matrix searching for higher correlated factors) factors into the model as a block.
    "If you torture the data long enough it will eventually confess."
    -Ronald Harry Coase -

  4. #4
    Points: 261, Level: 5
    Level completed: 22%, Points required for next Level: 39

    Posts
    2
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Principal component analysis before regression anallysis


    What i should have said is that i took out the 3 IV at different iterations because they were too "complex" and had less than 50% explained variance in communalities!
    Last edited by jeandaniel; 04-07-2012 at 10:16 PM.

+ 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