+ Reply to Thread
Results 1 to 3 of 3

Thread: One more chi-square question

  1. #1
    Points: 52, Level: 1
    Level completed: 4%, Points required for next Level: 48

    Posts
    5
    Thanks
    2
    Thanked 0 Times in 0 Posts

    One more chi-square question




    On a table of p-values for a chi-square test, I'm looking at the significance values on the one axis. Can someone explain that in layman's terms? I want to know, specifically, if I can use that value to say "there is an X% chance that these two events are linked" or something similar.

    In looking around online, I found one example using the data below. The chi-square value ends up being 1.668 and the Df are 1. That puts me between 0.1 and 0.25 (10% and 25%) level of significance, leading me to accept my null hypothesis that the new book did not have anything to do with the change in scores. Am I able to say that "There is between a 10% and 25% chance that the book had something to do with the change in scores, and a 75%-90% chance that it was just due to random events" or is that not what those significance levels mean?

    If that's not what they mean, is there a way to figure out what that percentage would be? I keep seeing references to "the likelihood that an additional data point would be more extreme than the existing value" or something similar, but that doesn't make any sense to me at all in a context like this.
    Attached Images  

  2. #2
    Omega Contributor
    Points: 38,334, Level: 100
    Level completed: 0%, Points required for next Level: 0
    hlsmith's Avatar
    Location
    Not Ames, IA
    Posts
    6,998
    Thanks
    398
    Thanked 1,186 Times in 1,147 Posts

    Re: One more chi-square question

    Somebody may correct me but: given the null hypothesis is true, you would see similar results 10-25% of the time. So the difference between them are not that uncommon under the null hypothesis.
    Stop cowardice, ban guns!

  3. #3
    Points: 52, Level: 1
    Level completed: 4%, Points required for next Level: 48

    Posts
    5
    Thanks
    2
    Thanked 0 Times in 0 Posts

    Re: One more chi-square question


    Quote Originally Posted by hlsmith View Post
    Somebody may correct me but: given the null hypothesis is true, you would see similar results 10-25% of the time. So the difference between them are not that uncommon under the null hypothesis.
    With that logic though if you take it to the extreme, in this case down to a chi-square value of 0.005 or so, that gives you a percentage of 95% or so. I can't see that meaning that 95% of the time you'd get a value of 0.005.

    Also, just running up the chart, that would be saying that 95% of the time you'd be getting a value of 0.005, but also 90% of the time you'd be getting 0.02, 80% of the time you'd be getting a value of 0.06, etc. That is well over 100% obviously when you put it all together, so I don't see how that would be the meaning. Along the lines of your answer though, could my original example there be something like "10-25% of the time you would get that value or more? Does that make sense?

+ 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