The example to the right is data on reaction times after
drinking either water or alcohol. The reaction times by
group were not normally distributed so an independent t- test
could not be used. The Mann-Whitney test is more
appropriate. It tests the hypothesis that the two
distributions are the same. All the data is
ordered and ranked from the fastest to the
slowest irrelevant of group. The sum of the ranks
for each group is used to calculate a test statistic.
If there is no difference between the groups the
sum of the ranks will be similar. SPSS does all the
ranking for you so you don’t need to worry about
that.