# High Mann-Whitney U number good or bad?

#### visualist

##### New Member
Hello all!

I'm quite the newbie with statistics so I'm probably asking stupid questions, but I've been googling now for quite a while without success.

I've analysed newspaper stories and one of the variables is how many photographs they contained. I have two distinct paper groups from two different locations and now I'm trying to see whether I kind find significant differences between the groups and locations. None of my variables show a normal distribution but are majorly skewed.

I performed a Mann-Whitney U test and it shows significant difference in one group (p < .001, and a weak relationship in the other p = .056).

But as I started writing this I noticed the Mann-Whitney U value is really high. Is this a sign that I've done something wrong? Or is it supposed to be that high? I've been looking at other people's articles and the U is often much much lower.

The range of values is min = 0 max = 13
(many stories without any photos, e.g. 0)

This is what I get from SPSS:

-- location A n = 447 (group 1 n = 182, mean = 1.21; group 2 n = 265, mean = 1.63)
Mann-Whitney U = 21658.000
Wilc. W = 38311.000
Z = -1.914
p = .056

-- location B n = 514 (group 1 n = 345, mean = .58; group 2 = 169, mean = 1.99)
Mann-Whitney U = 13737.000
Wilc. W = 73422.000
Z = -10.565
p = .000

If someone has time to give me even a yes/no answer it would be awesome. Then I would know which way to go from here.

Last edited: