+ Reply to Thread
Results 1 to 4 of 4

Thread: 1 sample K-S-Test output

  1. #1
    Points: 1,984, Level: 26
    Level completed: 84%, Points required for next Level: 16

    Location
    Germany/Scotland
    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Question 1 sample K-S-Test output




    Hi there,

    I got told that I would need to do a test to look at my data and whether it is normally distributed or not. So I did the 1 sample K-S- test and got the following output:

    One-Sample Kolmogorov-Smirnov Test

    Binge / no binge

    N 120
    Normal Parameters Mean 2,5083
    Std. Deviation 1,22300

    Most Extreme Differences Absolute ,200
    Positive ,200
    Negative -,197
    Kolmogorov-Smirnov Z 2,187
    Asymp. Sig. (2-tailed) ,000


    It seems a bit weird to me that the Asymp. Sig. (p) would be 0.000 without any other numbers behind it....does it mean, that my data is skewed and not normally distributed? If this is the case, can I still perform a multinominal logistic regression? Or would I have to do some non-parametric test (in which case I would be even more clueless than I am at the moment) ?

    Thanks for your help!

    PS for some reason I can't generate a proper table for my output...sorry...hope it is not too confusing

  2. #2
    Points: 2,101, Level: 27
    Level completed: 68%, Points required for next Level: 49

    Location
    Belgium
    Posts
    87
    Thanks
    0
    Thanked 0 Times in 0 Posts
    This test tests the following:

    H0 : population is normal
    HA : population is not normal

    Since your p-value is very small, it means that if the null hypothesis were true, the probability that we would have observed a KS test statistic of 2.187 is 0.000 (obviously never zero, but it's very very very very small). Since this is the case, we reject the null hypothesis.

    More info: http://faculty.chass.ncsu.edu/garson/PA765/kolmo.htm

  3. #3
    Points: 1,984, Level: 26
    Level completed: 84%, Points required for next Level: 16

    Location
    Germany/Scotland
    Posts
    4
    Thanks
    0
    Thanked 0 Times in 0 Posts
    I just realised that I actually do not have continous measures but categorical ones...so I can't actually use this test!

    Thank you so much for your reply anyway! stupid me!

  4. #4
    Points: 2,101, Level: 27
    Level completed: 68%, Points required for next Level: 49

    Location
    Belgium
    Posts
    87
    Thanks
    0
    Thanked 0 Times in 0 Posts

    You would model it with a normal distribution though... Everything that would fall between 5 and 6 would be category 5 (for example). It's just like modelling a discrete binomial or Poisson distribution with the normal distribution, as long as it's a good approximation.

+ Reply to Thread

           




Similar Threads

  1. Replies: 0
    Last Post: 03-05-2011, 04:40 AM
  2. Mulivariate output vs within-subjects output
    By regressedout in forum Statistics
    Replies: 0
    Last Post: 05-07-2009, 03:58 PM
  3. Mcnemar test output
    By yoda69 in forum Statistics
    Replies: 0
    Last Post: 03-13-2008, 08:23 AM
  4. Replies: 0
    Last Post: 10-26-2007, 01:25 PM
  5. median test or 2 sample kolmogorov test ?
    By sirhoppel in forum Statistics
    Replies: 0
    Last Post: 05-13-2006, 07:27 AM

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