+ Reply to Thread
Results 1 to 5 of 5

Thread: Acceptable significance level for Chi-Square statistic

  1. #1
    Points: 1,845, Level: 25
    Level completed: 45%, Points required for next Level: 55

    Posts
    13
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Smile Acceptable significance level for Chi-Square statistic




    Hi All,
    I need to know if the significance level calculated for a particular Chi-square statistic is less than or equal to 0.05, can we reject the respective null hypothesis?

    Cheers,
    Sach

  2. #2
    TS Contributor
    Points: 21,169, Level: 91
    Level completed: 64%, Points required for next Level: 181
    Awards:
    Downloads
    gianmarco's Avatar
    Location
    Italy
    Posts
    1,014
    Thanks
    153
    Thanked 199 Times in 147 Posts
    Hi!
    Yes, you can.

    Gm

  3. #3
    Points: 2,198, Level: 28
    Level completed: 32%, Points required for next Level: 102

    Posts
    275
    Thanks
    0
    Thanked 1 Time in 1 Post
    Generally, yes. .05 significance is what is usually accepted in the scientific literature, but do not treat .05 as some magic number. It's just the cutoff that they arbitrarily chose; in reality, a p-value of .051 and .049 are basically the same.

  4. #4
    Ninja say what!?!
    Points: 8,297, Level: 61
    Level completed: 49%, Points required for next Level: 153
    Link's Avatar
    Posts
    1,165
    Thanks
    37
    Thanked 84 Times in 76 Posts
    Quote Originally Posted by squareandrare View Post
    Generally, yes. .05 significance is what is usually accepted in the scientific literature, but do not treat .05 as some magic number. It's just the cutoff that they arbitrarily chose; in reality, a p-value of .051 and .049 are basically the same.
    You should definitely keep that in mind! So many people I come across put all their hopes into getting a value less than .05 without the acceptance of what was just stated!

  5. #5
    Points: 6,576, Level: 53
    Level completed: 13%, Points required for next Level: 174

    Location
    Atlanta
    Posts
    113
    Thanks
    0
    Thanked 11 Times in 3 Posts

    There is a very nice website
    "The Little Handbook of Statistical Practice" and it has really nice history of
    Why p = 0.05?

    Speaking of p-values for chi-square test. I was asked to create a measure to check how badly people create schedule. The more uniform schedule the better. You don't want to schedule 20 arrivals and 20 departures in one hour - this will overwhelm your workers.

    I decided to use chi-square test to test for even distribution of the event during the day. Split 24 hour clock into time buckets, use chi-square test p-value as measure for "uniform" distribution (equal probabilities for each bucket).

    It works great, however I would like to have reasonable cut point, I feel that usual p-value of 0.05 and 0.1 still leave lots of bad schedules. Plus we not exactly dealing with random numbers. We will more likely accept null because people do try create reasonable schedule.

    I I move my cut point up say to 0.25 or 0.3 I'll get really good results. Can I do this? Well, of course I can, but is it reasonable? Will my explanation above about non-randomness of events be reasonable justification for larger cut point?

    Thanks!
    Many people would sooner die than think. In fact they do. - B. Russell

+ Reply to Thread

           




Similar Threads

  1. level of significance
    By jenclark242 in forum Statistics
    Replies: 0
    Last Post: 10-02-2010, 09:36 PM
  2. significance level help
    By sunstone in forum Statistics
    Replies: 0
    Last Post: 11-12-2008, 08:47 AM
  3. level of significance
    By mustachman in forum Statistics
    Replies: 2
    Last Post: 07-01-2008, 12:10 AM
  4. level of significance- pls help
    By sweetx in forum Probability
    Replies: 2
    Last Post: 11-03-2006, 08:58 AM
  5. Significance Level
    By nat9 in forum R
    Replies: 1
    Last Post: 10-21-2006, 09:53 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