New Reply

reporting the results of a wilcoxon matched pair test

 
Share Thread Thread Tools
Apr24-12, 06:23 PM   #1
 

reporting the results of a wilcoxon matched pair test


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?
PhysOrg.com
PhysOrg
science news on PhysOrg.com

>> Bird's playlist could signal mental strengths and weaknesses
>> Minus environment, patterns still emerge: Computational study tracks E. coli cells' regulatory mechanisms
>> Bacterium uses natural 'thermometer' to trigger diarrheal disease, scientists find
New Reply
Thread Tools


Similar Threads for: reporting the results of a wilcoxon matched pair test
Thread Forum Replies
Need help with Wilcoxon rank-sum test Set Theory, Logic, Probability, Statistics 0
Help! Logical calculation of pair comparisons from analysis of existing results General Math 3
Wilcoxon rank-sum test Set Theory, Logic, Probability, Statistics 0
Estimating p-value for Wilcoxon Signed-Rank Test Set Theory, Logic, Probability, Statistics 1
physics I test results Academic Guidance 8