iSixSigma

Hypothesis Testing Clarification.

Six Sigma – iSixSigma Forums Old Forums General Hypothesis Testing Clarification.

Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #35507

    Brad M.
    Participant

    Can someone give me a laymen’s explanation as to why you calculate the “F Test” statistic before you calculate the “t Test” statistic?  I understand that you are looking for the variances to be “statistically” similar with the “F Test” but what happens if you find that they are not?  Does this mean that the “t Test” cannot be calculated or will be meaningless if calculated?  Also, two processes with statistically dissimilar variances but statistically similar means is significant isn’t it?  Wouldn’t the process with less variation in it be a better process even if the means are statistically similar?   Thanks for your feedback.

    0
    #100070

    Ken Feldman
    Participant

    The Minitab default for the 2-sample t test is that the variances are unequal.  There is an option to select equal variances.  It is useful to do the F test first to ascertain which is more representative of the data.

    0
    #101185

    Deanb
    Participant

    The purpose of running the F test first is to determine which of the 2 T-tests to use. If the F-test shows equal variances, then use the T-test for equal variances. If the F-test shows unequal variances, then use the T-test for unequal variances. The point is to improve the precision of the T-test by matching the proper method to the data. And yes, there can be separate indications of significance (or not) between the variances and means without one voiding the other. I find it useful to do the F test first for 2 reasons: first to select the proper T-test, and second to test the equality of the variances which might reveal meaningful differences.  

    0
    #101225

    S.K.Bagri
    Member

    Dear Mr. Brad,
    t-test is applied for hypothesis test of mean when population standard deviation is unknown.
    Alternatively, F-test signifys homogenity of two variance among two population.
    Hope it is clear.
    S.K.Bagri.

    0
    #101284

    jediblackbelt
    Participant

    I typically use the Ftest to determine if I am eligible for ANOVA.  If the Ftest shows variances are equal then I use ANOVA for my results.  If I fail to accept the null on Ftest then I use T-test without equal variances. 

    0
    #101289

    Ken Feldman
    Participant

    Oh, young Jedi, do you really use the F test before deciding on ANOVA?  Or do you use Test for Equal Variances dba Bartlett and Levene’s test, which are sort of special cases of the F test?  ANOVA is for more than 2 means and B&L is for more than two variances.  The simple F test is usually for 2 variances as is the 2-sample t.  So which is it?

    0
    #177720

    cait
    Participant

    What I don’t really understand is how to tell if the variances are equal or unequal. I don’t remember what my teacher said!! Does it have to do with the alpha level? Can someone give me an example of when it would be equal or unequal?
     
    HELP!

    0
Viewing 7 posts - 1 through 7 (of 7 total)

The forum ‘General’ is closed to new topics and replies.