Skip to main content
Cross Validated

Return to Question

appended answer 192272 as supplemental
Source Link
whuber ♦
whuber ♦
  • 333.6k
  • 63
  • 792
  • 1.3k

I have a question about Wilcoxon test in SPSS. For a work in my university i have to test 2 results of metrics applied in 2 algorithm (A2-NSGA3 and NSGA2), the results are as follows:

https://gist.github.com/fernandosg/d8dd0ffd34338962902b A2 and NSGA2 (in the url are the two files).

I follow this tutorial in previous test this: https://statistics.laerd.com/spss-tutorials/wilcoxon-signed-rank-test-using-spss-statistics.php, and in all test,the metric test matches with the criterion that, the min of the sum of range positive and negative, are compared to the critic value, where in this case, with my N are 30, is 151.

The result was this: In the sum of range positive was : 150 In the sum of range negative was : 315. The min was 150 Is suppose that, if the min was 150 and my critical value are 151, the hypothesis null was rejected, but in SPSS, the hypothesis was accepted. In my results showing the P-value, and was 0.09 ¿Why this happened? I read before, that if N was more large that the number of rows on the tables of the critical value, is recommended to calculate the p-value and determine with p-value and .05 (my significance level), but in previous test, with the same N, the criteria of accepted or not,was with the critical value. ¿Why is this different?.

Thanks for read and respond this newbie question, i added the url of the tutorial inside the gist, that i follow, with the previous test Tutorial (within the tutorial is the critical values table.)

I was reading the documentation online and comparing my previous tests, and i see that, the criterion for check if the hypothesis is accepted or not, was the p-value. I read that if N is greater that 20, is suppose to i have to computing the value of the mean of W, the standard deviation and the standarized Z-test. After that, in the table of Normal Distribution you get the p value. If in the report of SPSS, i get Z = -.556, i see that i get in P 0.57854809034163, ¿but how i get this value in the table of normal distribution? The table is this : https://www.math.ucdavis.edu/~soshniko/135a/materials/standardnormaltable.pdf

I have a question about Wilcoxon test in SPSS. For a work in my university i have to test 2 results of metrics applied in 2 algorithm (A2-NSGA3 and NSGA2), the results are as follows:

https://gist.github.com/fernandosg/d8dd0ffd34338962902b A2 and NSGA2 (in the url are the two files).

I follow this tutorial in previous test this: https://statistics.laerd.com/spss-tutorials/wilcoxon-signed-rank-test-using-spss-statistics.php, and in all test,the metric test matches with the criterion that, the min of the sum of range positive and negative, are compared to the critic value, where in this case, with my N are 30, is 151.

The result was this: In the sum of range positive was : 150 In the sum of range negative was : 315. The min was 150 Is suppose that, if the min was 150 and my critical value are 151, the hypothesis null was rejected, but in SPSS, the hypothesis was accepted. In my results showing the P-value, and was 0.09 ¿Why this happened? I read before, that if N was more large that the number of rows on the tables of the critical value, is recommended to calculate the p-value and determine with p-value and .05 (my significance level), but in previous test, with the same N, the criteria of accepted or not,was with the critical value. ¿Why is this different?.

Thanks for read and respond this newbie question, i added the url of the tutorial inside the gist, that i follow, with the previous test Tutorial (within the tutorial is the critical values table.)

I have a question about Wilcoxon test in SPSS. For a work in my university i have to test 2 results of metrics applied in 2 algorithm (A2-NSGA3 and NSGA2), the results are as follows:

https://gist.github.com/fernandosg/d8dd0ffd34338962902b A2 and NSGA2 (in the url are the two files).

I follow this tutorial in previous test this: https://statistics.laerd.com/spss-tutorials/wilcoxon-signed-rank-test-using-spss-statistics.php, and in all test,the metric test matches with the criterion that, the min of the sum of range positive and negative, are compared to the critic value, where in this case, with my N are 30, is 151.

The result was this: In the sum of range positive was : 150 In the sum of range negative was : 315. The min was 150 Is suppose that, if the min was 150 and my critical value are 151, the hypothesis null was rejected, but in SPSS, the hypothesis was accepted. In my results showing the P-value, and was 0.09 ¿Why this happened? I read before, that if N was more large that the number of rows on the tables of the critical value, is recommended to calculate the p-value and determine with p-value and .05 (my significance level), but in previous test, with the same N, the criteria of accepted or not,was with the critical value. ¿Why is this different?.

Thanks for read and respond this newbie question, i added the url of the tutorial inside the gist, that i follow, with the previous test Tutorial (within the tutorial is the critical values table.)

I was reading the documentation online and comparing my previous tests, and i see that, the criterion for check if the hypothesis is accepted or not, was the p-value. I read that if N is greater that 20, is suppose to i have to computing the value of the mean of W, the standard deviation and the standarized Z-test. After that, in the table of Normal Distribution you get the p value. If in the report of SPSS, i get Z = -.556, i see that i get in P 0.57854809034163, ¿but how i get this value in the table of normal distribution? The table is this : https://www.math.ucdavis.edu/~soshniko/135a/materials/standardnormaltable.pdf

Bumped by Community user
Bumped by Community user
Source Link
fsg
fsg
  • 1
  • 1

Problem with Wilcoxon test in SPSS

I have a question about Wilcoxon test in SPSS. For a work in my university i have to test 2 results of metrics applied in 2 algorithm (A2-NSGA3 and NSGA2), the results are as follows:

https://gist.github.com/fernandosg/d8dd0ffd34338962902b A2 and NSGA2 (in the url are the two files).

I follow this tutorial in previous test this: https://statistics.laerd.com/spss-tutorials/wilcoxon-signed-rank-test-using-spss-statistics.php, and in all test,the metric test matches with the criterion that, the min of the sum of range positive and negative, are compared to the critic value, where in this case, with my N are 30, is 151.

The result was this: In the sum of range positive was : 150 In the sum of range negative was : 315. The min was 150 Is suppose that, if the min was 150 and my critical value are 151, the hypothesis null was rejected, but in SPSS, the hypothesis was accepted. In my results showing the P-value, and was 0.09 ¿Why this happened? I read before, that if N was more large that the number of rows on the tables of the critical value, is recommended to calculate the p-value and determine with p-value and .05 (my significance level), but in previous test, with the same N, the criteria of accepted or not,was with the critical value. ¿Why is this different?.

Thanks for read and respond this newbie question, i added the url of the tutorial inside the gist, that i follow, with the previous test Tutorial (within the tutorial is the critical values table.)