In: Statistics and Probability
5) A research group has collected information about response times in humans. They want to compare the response times of people early in the workday with response times of participants late in the workday. The response times appear below. Since response time over the population is not normally distributed they use a Mann-Whitney U test. Compute U for these data and evaluate whether there is a significant difference between groups.
Early day 300 275 328 400 280 335 310
Late in the day 325 310 295 375 380 410 390
Early day | Ranking | Late in the day | Ranking |
275 | 1 | 295 | 3 |
280 | 2 | 310 | 5.5 |
300 | 4 | 325 | 7 |
310 | 5.5 | 375 | 10 |
328 | 8 | 380 | 11 |
335 | 9 | 390 | 12 |
400 | 13 | 410 | 14 |
n1 = 7 | = 42.5 | n2 = 7 | = 62.5 |
If the response time of people early day in the work and late in the day of the work both are same. Then this Hypothesis is assumed as H1.
If the response time of people early day in the work and late in the day of the work both are not same. Then this Hypothesis is assumed as H2.
+ = 42.5 + 62.5 = 105
, (where n= n1+n2)
Hence, the above given data follows the Mann-Whitney U test,
From Mann-Whitney U test
So U = Min (U1,U2) = Min (34.5,14.5)
Therefore U = 14.5
Then there is a significant difference between groups. which is equal to
14.5 - 14.5 = 0.05