+ Reply to Thread
Results 1 to 5 of 5

Thread: Variation in significance Categorical Regression Analysis (CATREG)

  1. #1
    Points: 14, Level: 1
    Level completed: 27%, Points required for next Level: 36

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Variation in significance Categorical Regression Analysis (CATREG)




    Hello,
    I am doing Categorical Regression Analysis (CATREG) with SPSS. However, every time i run the analysis, the standard error and p-value vary slightly, although all the variables both DV and IV remain same. This makes some IV's significant sometimes (<0.05), and sometimes insignificant (>0.05). Example: for one variable the p-values are 0.044, 0.037, 0.083, 0.046, 0.054 in 5 different runs.
    The beta value remains same for all the different runs. I presume this happens because the std error is calculated by bootstrap method which re-samples randomly. If this is true then how should i produce the final results?

    My sample size contains 920 observations.
    IV is a categorical variable
    DV: 11 variables, 9 nominal and 2 ordinal

    Would be much appreciated if anyone could help on this.

    Thanks
    Faqhrul Islam

  2. #2
    Points: 20,006, Level: 89
    Level completed: 32%, Points required for next Level: 344

    Posts
    568
    Thanks
    50
    Thanked 20 Times in 19 Posts

    Re: Variation in significance Categorical Regression Analysis (CATREG)

    Haven't used SPSS in a long while, but is there an option to specify how many boot strap samples to select? I'd go with something really high (>10K) then try running that again and comparing them.

  3. #3
    Points: 14, Level: 1
    Level completed: 27%, Points required for next Level: 36

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Variation in significance Categorical Regression Analysis (CATREG)

    Thanks James.
    Apparently re-sampling isn't required for my analysis. I tried with no re-sampling. But still the results are different. Do not think boot strapping is causing that problem.

  4. #4
    Omega Contributor
    Points: 38,396, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    7,001
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: Variation in significance Categorical Regression Analysis (CATREG)

    So are the SEs coming from bootstrap or not? In other programs you would set the seed value, since some procedures use say atomic clock for selecting the random starting seed value.


    Could this be the issue?
    Stop cowardice, ban guns!

  5. #5
    Points: 14, Level: 1
    Level completed: 27%, Points required for next Level: 36

    Posts
    3
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Re: Variation in significance Categorical Regression Analysis (CATREG)


    Yes, the SE varies because of Bootstrap which by nature is Monte Carlo. Even if no re-sampling is selected, SPSS does bootstrap(1000) by default. However, .632 Bootstrap can be chosen by the user to enhance the bootstrapping. I am trying with 15000 samples of .632 bootstrap which seemingly minimizes the variations. Not sure yet how to select the final model or present results!!

+ Reply to Thread

           




Tags for this 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