Research methods - Mann-Whitney U Test

Question 1

is the mann-whitney u test parametric or non-parametric

Answer 1

a non-parametric test of difference

Question 2

what is the test statistic for the mann-whitney u test

Answer 2

U

Question 3

when is a mann-whitney u test used

Answer 3

used for ordinal level data (can also be used for interval but not nominal) and independent measures design

Question 4

what is ordinal data and why is it good or bad

Answer 4

data placed in rank order
meaningful comparison can be made
can state one value is higher than the other but cannot assume more than that

Question 5

what is interval data and why is it good

Answer 5

measurements are on a standardised, universal scale of fixed units and distances between the points are equal
highest level of measurement

Question 6

is there absolute zero in interval data

Answer 6

no, it is just another point along the scale

Question 7

describe the steps for the mann-whitney u test

Answer 7

1) rank the data across both samples from smallest to largest (if some are the same rank in order to begin then work out mean of the ranks given)
2) Add up the ranks separately for each condition - totals are r1, r2 etc
3) use the formula provided to work out u1 and u2 and find the calculated value of U (smaller one)
4) compare with the critical value of U to establish significance (must be less than or equal)
5) write a significance statement

Question 8

what is needed to find the critical value of U

Answer 8

n1 and n2, number of ps in each condition then find where they intersect

Question 9

what must be included in the significance statement

Answer 9

calculated value of u, critical value of u, significance level (0.05), n values, are results significant, accept or reject null hypothesis and what does this suggest