I’m analyzing data in SPSS using the Mann-Whitney U test to compare two groups, and I stumbled upon something that got me wondering. For one of my dependent variables (DV1), Group 1 has a lower mean rank, and the Z value is negative, which makes sense. But for DV2, Group 1 has a higher mean rank, yet the Z value is still negative. Both results are statistically significant.
I thought a positive Z should indicate that Group 1 has higher ranks than Group 2. Does SPSS reverse group codes internally or something? When reporting these results, should I keep the negative Z value in the table, even though it feels counterintuitive to the mean values?
After digging deeper, I realized that the sign of the Z value doesn’t necessarily indicate the direction of the effect. It’s more about the probability of observing the difference between the two groups. So, a negative Z value doesn’t always mean Group 1 has lower ranks than Group 2.
But I’m still curious – has anyone else encountered this issue? How did you handle it in your reports? Any clarification or tips would be appreciated!