Hi!
New to here with limited stats background.
I'm trying to test statistically significant gender differences for a ordinal outcome variables. Would comparing males and females in the same sample be considered two independent samples or two dependent samples? Since males = total sample - females, does that mean they're dependent samples? Does this mean I would use the Wilcoxon signed rank test on stata? What does the provided "Z" value mean on stata? How do I interpret it and report it?
Help much appreciated!! Thank you!!
Dependency means that if you knew one value of one group, it would tell you something about another value of the other group. This is typically the case when two values are obtained from the same subject: if the first value (say, measured before an intervention) is very high, the other values of this same subject (measured after the intervention) are also expected to be high.
In your case, when males and females are sampled independently, the samples are independent.
It would be dependent if you sampled, for instance, brothers and sisters, and if the measurement was influenced by some genetic or family background. In that case, if a male had a for instance a particularily high value, you would expect that his sister also hase a relatively high value.
Gender would be your independent variable here, so you should consider it as independent samples.
Wilcoxon Rank Sum Test (WRST) needs samples be drawn from independent populations, while Wilcoxon Signed Rank Sum Test (WSRST) requires that the populations be paired.