Exercise1:
The following data show the required operation time (minutes) of 2 surgeons for the same surgical procedure. Analyze whether there is a statistical difference in operation time between 2 surgeons.
Doctor A（11 Cases）：39, 33, 48, 34, 43, 38, 32, 42, 36, 44, 31
Doctor B（10 Cases）：51, 40, 54, 45, 35, 49, 47, 53, 41, 69
Exercise2:
Data show the effect of 2 therapeutic drugs on survival time (weeks) in a disease animal model. Analyze whether there is a statistical difference in survival time between 2 drugs.
Drug A：34, 5, 11, 5, 20, 24, 8, 14, 14, 13, 14, 41, 13, 15, 19, 12
Drug B：8, 14, 8, 8, 8, 11, 25, 23, 9, 12, 13
Comparative Verification Test
Mann-Whitney U test:
Sample 1: StaFinale, <0.01; SPSS, 0.006; EZR, 0.00619.
Sample 2: StaFinale, NS (>0.2); SPSS, 0.212; EZR, 0.0215.
Sample 3: StaFinale, 0.056; SPSS, 0.057; EZR, 0.0571.
Sample 4: StaFinale, untestable data (P value derived from Z score is 0.0358); SPSS, 0.036; EZR, 0.0358.
Samples 1 and 2 have a small number of cases with the same ranking. Statfinale uses U-test Table 2 which does not provide precise P values.
Sample 3 has a small number of cases without ties. Statfinale uses U-test Table 1 which provides precise P values.
Reports of SPSS & EZR
The SPSS report format is the same for 1, 2, 3, and 4.
Samples 1, 2, and 3 were calculated using a formula without using U-test tables. Sample 4 had a small number of cases (5 and 3 cases) and a tie. StatFinale did not allow testing, but SPSS and EZR showed it to be significant (miscalculation? Same as the P value derived from the Z value!).
※Click pics to enlarge.
