+ Reply to Thread
Results 1 to 4 of 4

Thread: Logistic regression analysis - Large beta coefficients

  1. #1
    Points: 218, Level: 4
    Level completed: 36%, Points required for next Level: 32

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Logistic regression analysis - Large beta coefficients




    Hi,

    I have just run a logistic regression analysis on 7 IVs and a dichotomous DV. The IVs are regression-weighted composite scores of Likert scale items.
    The output in SPSS gives me VERY large beta coefficients and p-values very close to 1. To check for multicollinearity, I ran linear regression analyses for every IV. The VIF never exceeded 3. What could be the problem, if it is not multicollinearity?

    Thanks for your help!

  2. #2
    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: Logistic regression analysis - Large beta coefficients

    How big is your data set and what do the standard errors look like?
    "If you torture the data long enough it will eventually confess."
    -Ronald Harry Coase -

  3. #3
    TS Contributor
    Points: 12,501, Level: 73
    Level completed: 13%, Points required for next Level: 349

    Posts
    951
    Thanks
    0
    Thanked 103 Times in 100 Posts

    Re: Logistic regression analysis - Large beta coefficients

    Quote Originally Posted by adrmck17 View Post
    VERY large beta coefficients and p-values very close to 1.
    The reference value for your IV has 0 non-events (or 0 events, if you have large negative beta). I'd look into that, # of non-events/events at certain levels of the IVs.

  4. #4
    Points: 218, Level: 4
    Level completed: 36%, Points required for next Level: 32

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Logistic regression analysis - Large beta coefficients


    Thanks for your replies!

    Quote Originally Posted by trinker View Post
    How big is your data set and what do the standard errors look like?
    I have just over 100 cases. What will the Standard Errors tell me?

    Quote Originally Posted by Mean Joe View Post
    The reference value for your IV has 0 non-events (or 0 events, if you have large negative beta). I'd look into that, # of non-events/events at certain levels of the IVs.
    The number of events ("Yes" = 1) and non-events ("No" = 0) is split 54% and 46%. In this case, would the levels of my IVs be the Likert scale items (Strongly Disagree (1) - Strongly Agree (5))?

+ 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