I have 3 hypothesis to test:

H0: Female subordinates rate themselves the same as their supervisors.
HA: Female subordinates themselves lower than (to the left of) their supervisors.,

H0: Female subordinates rate themselves the same as male subordinates rate themselves.
HA: Female subordinates rate themselves lower (to the left of) than male subordinates rate themselves.

H0: Female supervisors rate subordinates the same as male supervisors rate subordinates.
HA: Female supervisors rate subordinates higher (to the right of) than male supervisors rate subordinates.

I have 5 coulmns of data.

Column 1: ratings of all 35 supervisors of their male subordinates
Column 2:ratings of male subordinate of themselves
Column 3:ratings of all 35 supersivors of their female subordinates
Column 4:ratings of female subordinates of themselves
Column 5:index: indicating male or female supervisor. 1=male, 2=female.

I am having trouble in identifying which nonparametric test to use. Its either Wilcoxon rank sum test, sign test, or the Wilcoxon signed rank sum test.

Can anyone help me identify which statistic test to use for each of the 3 hypothesis? And which coulmns of data to focus on for each hypothesis.