How does SPSS account for ties in the Mann-Whitney U Test?

Question & Answer

Question

Does the Mann-Whitney U Test in SPSS Statistics account for ties?

Answer

The Mann-Whitney test is available in two different SPSS Statistics procedures: NPTESTS (Analyze>Nonparametric Tests>Independent Samples in the menus) and NPAR TESTS (Analyze>Nonparametric Tests>Legacy Dialogs>2 Independent Samples). In both procedures, the Z test statistic and Asymptotic Significance are corrected for ties. The tie corrections are achieved by assigning the mean rank to all cases with tied values, then incorporating a tie correction into the formula for the Z statistic. See the statistical algorithms for details (Help>Algorithms>Nonparametric Tests Algorithms>Mann-Whitney Test for NPTESTS and Help>Algorithms>NPAR TESTS>Mann-Whitney U Test for NPAR TESTS).

In both procedures the Exact Significance shown as [2*(1-tailed Significance)] is not corrected for ties. If you have the Exact Tests module, you can get Exact Significance (2-tailed) and (1-tailed) values in NPAR TESTS. These exact results are corrected for ties. These results come from algorithms developed by the makers of StatXact software.

Related Information

Need more help? Visit the IBM Support Forum

[{"Product":{"code":"SSLVMB","label":"IBM SPSS Statistics"},"Business Unit":{"code":"BU059","label":"IBM Software w\/o TPS"},"Component":"Not Applicable","Platform":[{"code":"PF025","label":"Platform Independent"}],"Version":"Not Applicable","Edition":"","Line of Business":{"code":"LOB10","label":"Data and AI"}}]

Historical Number

22228

Was this topic helpful?

Document Information

Modified date:
16 April 2020

UID

swg21477001

Tips