+ Reply to Thread
Results 1 to 2 of 2

Thread: Making sense of Chi-square and Cramer's V

  1. #1
    Points: 6,193, Level: 51
    Level completed: 22%, Points required for next Level: 157

    Posts
    43
    Thanks
    0
    Thanked 0 Times in 0 Posts

    Making sense of Chi-square and Cramer's V




    I read a lot about chi-square tests that do not ever talk about Cramer's V (or comparable stat), but if I'm understanding correctly, it seems like a chi-square test is not complete without something like the Cramer's V as a check.

    So would the following be an accurate summary of how to interpret Cramer's V?

    When doing a chi-square test, you can state that an association exists between the two categorical variables only if BOTH the p-value < .05 (or whatever your threshold is) AND the Cramer's V is "high".

    If p-value > .05 then the probability of this result happening again is low. Or if the Cramer's V is "low", then any apparent association is nullified since it's due only to a large sample size.

  2. #2
    Super Moderator
    Points: 19,674, Level: 88
    Level completed: 65%, Points required for next Level: 176
    bugman's Avatar
    Posts
    1,867
    Thanks
    197
    Thanked 229 Times in 183 Posts

    Cramers V is a kind of post hoc test that tests the strength of the association. So, you can have an existing association with a low p-value and a low Cramers V, but you can say that the association is stronger as the v-value increases - in a similar way that you can have a significantly different from zero slope in a regression with a low pearsons correlation coefficient.

    It is a bit more tricky to interpret because Cramers V will depend also on the number of levels of the categorical variables.

    Phil

+ Reply to Thread

           




Similar Threads

  1. Cramer's V & Phi or Fisher's exact test?
    By micdhack in forum SPSS
    Replies: 1
    Last Post: 02-08-2011, 02:24 PM
  2. Replies: 0
    Last Post: 07-12-2010, 03:55 AM
  3. HELP! Doesn't make sense
    By JRM1973 in forum Statistics
    Replies: 2
    Last Post: 07-24-2008, 04:27 PM
  4. Replies: 1
    Last Post: 02-27-2008, 03:34 PM
  5. Making sense of data transformation
    By omgitstiff in forum Statistics
    Replies: 2
    Last Post: 12-13-2007, 12:36 PM

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