0
$\begingroup$

I need some help in understanding the wilcoxon signed rank test. I am using this test on comparing two analysers. We ran the same samples on two analysers for the PT (prothrombin time), the results were non parametric. I ran the wilcoxon signed rank test on the QI macro application and the column that works out the difference between the data sets (each point) shows the difference as being 1 or less than 1 on all data points. However the test still rejects the null hypothesis that the data between the two data sets is not different. When comparing another test, the differences are less than 0.1.

In another data set (again non-parametric), the test was able to reject the null hypothesis even though the differences between the data sets are >1. Would someone be able to explain why this would be? Am I using the incorrect test?

Improve this question
$\endgroup$
2
  • $\begingroup$ "Results" and "data" cannot be parametric or non-parametric. Those are characteristics of models, not data. $\endgroup$
    – Peter Flom
    Commented Apr 19 at 11:17
  • $\begingroup$ What about these results confuses you? (Also, in your second paragraph, do you mean "unable" rather than "able"? That would make a bit more sense in context). $\endgroup$
    – Peter Flom
    Commented Apr 19 at 11:25

0

Reset to default

Browse other questions tagged or ask your own question.