Model | Proportion of datasets for which null hypothesis is rejected | ||||||
---|---|---|---|---|---|---|---|
t test (individual observations) | Wilcoxon (individual observations) | t test (means) | Wilcoxon (means) | LMM | GEE | Datta and Satten method | |
1 | 0.512 | 0.496 | 0.047 | 0.043 | 0.049 | 0.079 | 0.053 |
2 | 0.966 | 0.961 | 0.688 | 0.641 | 0.692 | 0.772 | 0.699 |
3 | 0.500 | 0.496 | 0.031 | 0.042 | 0.037 | 0.067 | 0.053 |
4 | 0.781 | 0.960 | 0.359 | 0.529 | 0.367 | 0.448 | 0.706 |
5 | 0.431 | 0.412 | 0.041 | 0.030 | 0.050 | 0.144 | 0.065 |
6 | 0.986 | 0.981 | 0.723 | 0.604 | 0.770 | 0.927 | 0.806 |
7 | 0.405 | 0.412 | 0.021 | 0.029 | 0.034 | 0.124 | 0.065 |
8 | 0.892 | 0.960 | 0.569 | 0.568 | 0.625 | 0.786 | 0.735 |
Table summarizes the results (proportion of datasets for which the null hypothesis is rejected) when clusters contain observations from a single group only; data were obtained from 10,000 simulated datasets for each of the eight models described in Table 1 and for each of the analysis methods in Table 2. A significance level of 5% (i.e. p < 0.05) was used to determine whether to reject the null hypothesis in all cases.