SPSS Statistics

 View Only

Homogeneity of variances - Trimmed mean

  • 1.  Homogeneity of variances - Trimmed mean

    Posted Mon November 15, 2021 01:57 PM

    Dear Rick and Jon, 


    I have a question about the test of homogeneity of variances that uses a 5% trimmed mean in the EXAMINE and some GLM commands.


    In these EXAMINE and GLM commands, the results of the original Levene's test of homogeneity of variances are presented, as well as three modifications. One of these modifications is to replace the group means with a trimmed mean (Brown and Forsythe, 1974). 


    Brown, M. B., & Forsythe, A. B. (1974). Robust tests for the equality of variances, Journal of the American Statistical Association, 69(346), 364-367.


    In the aforementioned paper, a 10% trimmed mean was studied.  However, in your Algorithms document, it is stated that a 5% trimmed mean is used by SPSS Statistics.  


    Unfortunately, I could not see any other studies referenced in the Algorithms document suggesting that a 5% trimmed mean should be used instead of the 10% trimmed mean (as studied by Brown and Forsythe, 1974).


    Therefore, I was wondering if I had missed this reference or if there was another reason that the EXAMINE and GLM commands are using a 5% rather than 10% trimmed mean.  Hopefully you can help!


    Many thanks in advance for your time.


    Adam



    ------------------------------
    Adam Lund
    ------------------------------