Mann-Whitney U Test:
Used when there is –
Hypothesis predicts a difference
An unrelated design (independent groups)
Interval or ordinal data
Step one –
Justify your choice of test
Step two –
Work out your hypothesis
Step three –
Rank your data (if some data is the same number, find average rank of no. of
ranks these numbers would have taken up)
Step four –
Calculate N 1 (no. of individual data in condition 1) and N 2 (for condition 2)
Step five–
Calculate T 1 (sum of the ranks in smaller condition) and T 2 (for larger condition)
Step five–
N 1 ( N 1 +1) N 2 (N 2 +1)
Calculate U 1using U 1 = N 1 N 2+ −T 1 and U 2 using U 2= N 1 N 2+ −T 2
2 2
, select smaller value to be observed value of U then find critical value and
determine if significant using table
Used when there is –
Hypothesis predicts a difference
An unrelated design (independent groups)
Interval or ordinal data
Step one –
Justify your choice of test
Step two –
Work out your hypothesis
Step three –
Rank your data (if some data is the same number, find average rank of no. of
ranks these numbers would have taken up)
Step four –
Calculate N 1 (no. of individual data in condition 1) and N 2 (for condition 2)
Step five–
Calculate T 1 (sum of the ranks in smaller condition) and T 2 (for larger condition)
Step five–
N 1 ( N 1 +1) N 2 (N 2 +1)
Calculate U 1using U 1 = N 1 N 2+ −T 1 and U 2 using U 2= N 1 N 2+ −T 2
2 2
, select smaller value to be observed value of U then find critical value and
determine if significant using table
