Table 3.
Data set | % Gaps in data set | Simulate gaps | P-value | Conf. interval | |
---|---|---|---|---|---|
Lower | Upper | ||||
Dunn | 14.759 | yes | <0.01** | <0.01 | 0.036 |
Dunn | 14.759 | no | <0.01** | <0.01 | 0.036 |
Edwards | 14.268 | yes | <0.01** | <0.01 | 0.036 |
Edwards | 14.268 | no | <0.01** | <0.01 | 0.036 |
Liu | 9.149 | yes | <0.01** | <0.01 | 0.036 |
Liu | 9.149 | no | <0.01** | <0.01 | 0.036 |
Wang | 13.092 | yes | 0.026* | 0.014 | 0.044 |
Wang | 13.092 | no | 0.022* | 0.011 | 0.039 |
Notes: SOWHAT by default propagates the exact number and position of gaps present in the original data set into all simulated data sets. Suppressing this feature, thereby excluding gaps from subsequent analyses, did not change the outcome of any SOWH tests examined here. * indicates -values less than 0.05; ** indicates less than 0.01.