You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Mann-Whitney Test Results produced for all pairs -- not just two
Purpose
To more quickly produce a table showing all Mann-Whitney pairs that are statistically significant.
Use-case
I have six groups and scales for 10 dependent variables. I want to use Mann-Whitney tests to identify which group pairs are significantly different on all 10 scales.
Is your feature request related to a problem?
Highly inconvenient to have to test every group pair separately and to transfer the results seperately
Is your feature request related to a JASP module?
Equivalence T-Tests
Describe the solution you would like
Currently, I have to "turn off" all but two groups to check just two at a time.
Describe alternatives that you have considered
Currently, one can test every group pair separately and to transfer the results separately to a spreadsheet
Additional context
Attached are two tables I created, laboriously testing every pair. It would obviously be ideal if JASP produced something like it for me automatically!
One table shows Mann Whitney results for each group and a selected reference group. The other produces a report identifying which group pairs are significantly different, flagging them * p < 0.05, ** p < 0.01, *** p < 0.001.
Ideally, whenever the Mann-Whitney option is selected, the output would include as many extra tables as necessary to show the results for each pair.
Or, as in the Wilcoxon Signed Rank Paired test, users could select the variable pairs to be tested.
Or, the table could be generated like that in my second example, where every group is numbered and the group pairs that are significant are shown in a separate column and flagged for the level of statistical significance.
The text was updated successfully, but these errors were encountered:
Here is another example of how results could be reported for multiple groups.
tomtomme
changed the title
Mann-Whitney test tables identifying results for multiple pairs
[Feature Request]: Mann-Whitney test tables identifying results for multiple pairs
Jan 22, 2025
Description
Mann-Whitney Test Results produced for all pairs -- not just two
Purpose
To more quickly produce a table showing all Mann-Whitney pairs that are statistically significant.
Use-case
I have six groups and scales for 10 dependent variables. I want to use Mann-Whitney tests to identify which group pairs are significantly different on all 10 scales.
Is your feature request related to a problem?
Highly inconvenient to have to test every group pair separately and to transfer the results seperately
Is your feature request related to a JASP module?
Equivalence T-Tests
Describe the solution you would like
Currently, I have to "turn off" all but two groups to check just two at a time.
Describe alternatives that you have considered
Currently, one can test every group pair separately and to transfer the results separately to a spreadsheet
Additional context
Attached are two tables I created, laboriously testing every pair. It would obviously be ideal if JASP produced something like it for me automatically!
One table shows Mann Whitney results for each group and a selected reference group. The other produces a report identifying which group pairs are significantly different, flagging them * p < 0.05, ** p < 0.01, *** p < 0.001.
Example 1: Mann Whitney Example 1.pdf
Example 2: Mann Whitney Example 2.pdf
Ideally, whenever the Mann-Whitney option is selected, the output would include as many extra tables as necessary to show the results for each pair.
Or, as in the Wilcoxon Signed Rank Paired test, users could select the variable pairs to be tested.
Or, the table could be generated like that in my second example, where every group is numbered and the group pairs that are significant are shown in a separate column and flagged for the level of statistical significance.
The text was updated successfully, but these errors were encountered: