SPSS Statistics

SPSS Statistics

Your hub for statistical analysis, data management, and data documentation. Connect, learn, and share with your peers! 

 View Only
  • 1.  Wilcoxon-Mann-Whitney rank sum test

    Posted Mon November 07, 2022 02:09 PM
    Again, a question regarding SPSS results of the W-M-W rank sum test:

    The Exact Sig. [2*91-tailed Sig.0] is not adjusted for ties. 

    Can one assume that the Asymp. Sig. (2-tailed) is adjusted for ties?

    Thanks, 
    Cliff

    #SPSSStatistics


  • 2.  RE: Wilcoxon-Mann-Whitney rank sum test

    Posted Mon November 07, 2022 09:15 PM
    From one of our statisticians:

    1. The "2*..." means 2 times the 1-tailed. In small samples, the asymptotic value can be enough different from the exact values that the 1-tailed exact value is much more than half of it, and in some cases maybe larger than the asymptotic value. In general, the exact values are to be preferred, though when not corrected for ties, there's not as much reason to prefer them. The Exact Tests module provides exact results for both 1 and 2-tailed that do correct for ties.
    2. Yes, the asymptotic results do correct for ties.


    ------------------------------
    Rick Marcantonio
    Quality Assurance
    IBM
    ------------------------------



  • 3.  RE: Wilcoxon-Mann-Whitney rank sum test

    Posted Mon November 07, 2022 11:33 PM
    Thanks, Rick. You guys are the greatest. 

    Sent from my iPhone