3.3. Statistical comparisons
Due to the absence of normally distributed data, it would be erroneous to employ ANOVA methods to make statistical comparisons. Several transformations were tried in the original data, such as log and Box–Cox transformations (Berthouex and Brown, 2002) to ‘‘normalize’’ the data; however, except in a couple of cases, it was not feasible to achieve normality for all the data. Therefore, the non-parametric Mood’s median test was employed to perform statistical comparisons. The median values and the results of the non-parametric statistical comparisons are included in Table 3. According to the test, the following are deduced:
1. Psychiatric hospitals had the lowest HMWUGR and university hospitals had the highest HMWUGR among all hospitals.
2. The birth and general hospitals had statistically similar
HMWUGR in both the public and private sector.
3.3. Statistical comparisons
Due to the absence of normally distributed data, it would be erroneous to employ ANOVA methods to make statistical comparisons. Several transformations were tried in the original data, such as log and Box–Cox transformations (Berthouex and Brown, 2002) to ‘‘normalize’’ the data; however, except in a couple of cases, it was not feasible to achieve normality for all the data. Therefore, the non-parametric Mood’s median test was employed to perform statistical comparisons. The median values and the results of the non-parametric statistical comparisons are included in Table 3. According to the test, the following are deduced:
1. Psychiatric hospitals had the lowest HMWUGR and university hospitals had the highest HMWUGR among all hospitals.
2. The birth and general hospitals had statistically similar
HMWUGR in both the public and private sector.
การแปล กรุณารอสักครู่..
