Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Reporting the results of a wilcoxon matched pair test

  1. Apr 24, 2012 #1
    I've done 3 wilcoxon matched pair tests and was wondering if someone could give me some advise on the most logical / correct way to report the results.

    test 1:
    pre post
    1.98 1.56
    1.09 2.49
    2.18 2.94
    1.13 1.29
    1.23 2.38

    result: Z = -1.438, two tail p = 0.138

    So I was thinking of reporting that...
    "A wilcoxon matched pairs test was performed to determine if the increase was significant. The mean (median?) gain (M=0.61, SD=0.74, N=5) was greater than zero, Z=-1.48, two tail p=0.138, but not enough to provide evidence that there was statistically significant improvement"

    test 2:
    Pre Post
    39.0 41.4
    30.7 31.1
    58.7 54.6
    48.2 50.9
    53.4 58.2

    result: Z = -0.94, two tail p = 0.345

    "A wilcoxon matched pairs test was performed to determine if the increase was significant. The mean (median?) gain (M=1.24, SD=3.368, N=5) was slightly greater than zero, Z=-0.94, two tail p=0.345, providing no evidence that there was statistically significant improvement"

    test 3:
    pre post
    29.8 52.6
    23.7 34.0
    18.1 57.7
    12.5 48.2
    16.1 35.3

    "A wilcoxon matched pairs test was performed to determine if the increase was significant. The mean (median?) gain (M=25.52, SD=12.051, N=5) was greater than zero, Z=-2.023, two tail p=0.043, providing evidence that there was statistically significant improvement"

    Questions:
    1. should I specify in my description the range given that test 1 looks worse than test 2, but a 0.61 mean change is actually alot better for this test than a 1.24 mean change?
    2. given that wilcoxon compares the medians should I not talk about the mean change at all?
    3. although I can't say that wilcoxon shows that the treatment between pre and post was effective given that there is no control, am I ok saying that wilcoxon showed statistically significant improvement in test 3?
    4. should I be saying something about Confidence Intervals in the description despite wilcoxon not calculating these?
    5. what is Z? am I referring to it correctly or should I be calling it something else in the reporting of the result?
     
  2. jcsd
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook

Can you offer guidance or do you also need help?
Draft saved Draft deleted



Similar Discussions: Reporting the results of a wilcoxon matched pair test
Loading...