# Using Mann Whitney U to verify relation between analitical values and mortality

#### tenrousei

##### New Member
Hello!

I'm new to the forum, and currently trying to do a paper on a retrospective study I did. I don't have much knowledge in statistics, and I was hoping someone could be so kind as to tell me if the way I did this is correct or if my conclusions cannot be taken or the methodology is incorrect.

I was trying to verify the relation between analytic values (p.ex troponin value - continuous variable) and mortality (categorical, "yes" or "no" variable). I ran the normality test and the values were not normally distributed (p<0,05 in the kolmogorov-smirnov test, right?) so I applied the Mann-Whitney U independent-samples test, chose the "Troponin value" as my test field and mortality as my group. It gave me a p<0,05, rejecting the null hypothesis that the troponin value was equal in both groups. Being that the mean troponin was superior in the group that died, i can assume that, in this sample, there was a correlation between higher troponin values and mortality? or is this something I cannot say based on this data?
If I just analyze normal vs. elevated troponin with mortality through a crosstab with chi-square, i do not get a significant p, and with t-test i get an almost significant one (P=0,054).

So I'm having a little trouble interpreting how much i can conclude from the mann whitney U test.

If someone can shed a little light, i would be very grateful!
Thanks.
Ana