Comparing O’Brien test using mean, median, symmetric and asymmetric trimmed mean
The assumption of homogeneity of variances can be tested with O’Brien procedure since it is compatible with standard analysis of variance (ANOVA). The test is very sensitive to non-normality which lead to find the alternative technique by replacing the original test with the robust measure of locati...
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Published: |
Malaysian Mathematical Sciences Society
2022
|
| Summary: | The assumption of homogeneity of variances can be tested with O’Brien procedure since it is compatible with standard analysis of variance (ANOVA). The test is very sensitive to non-normality which lead to find the alternative technique by replacing the original test with the robust measure of location. This study will investigate the behaviour of O’Brien test on robustness by using the usual variances on mean, median, symmetric trimmed and asymmetric trimmed mean under normal distribution and skewed normal distributions. The results show that O’Brien test is robust when data is distributed at normal and skewed normal under certain simulation study condition. |
|---|