CP2 - Chi-square test

Discussion in 'CP2' started by Bharti Singla, Mar 24, 2022.

  1. Bharti Singla

    Bharti Singla Senior Member

    IFoA, April 2020 - Chi Square test

    Method 1

    Paper A - They have used a chi-square test to validate the random numbers fall into uniform distribution. Here, they have used an excel formula =CHHINV(probability, DOF) and gives 16.9 and compares it with sum of (a-e)^2/e

    Method 2
    Paper B - Here, they have used an another type of chi-square test using an excel formula =CHISQ.TEST(actual range, expected range) and compared it with probability 95%

    But if I use the method 2 in Paper A by using the actual and expected ranges, I am getting a P-value of 71% which is less than 95% and test fails.
    How do we know which formula to use for chi-square test?
     
  2. DamienW

    DamienW Member

    I was confused about this earlier. I think the we should use the probability of 5% rather than 95% (which is extremely high) in paper 2 for testing. Hence, if you compare the P-value of 71% with 5%, the test should pass.
     
  3. Steve Hales

    Steve Hales ActEd Tutor Staff Member

    It would indeed be more appropriate for the value in 'Data checks'!F10 to be 5% rather than 95%.
    This is because the CHISQ.TEST function returns the right-tailed probability of the chi-squared distribution.
     
    Sarah Byrne likes this.

Share This Page