Table 1 Performance assessments of direct estimation and multiple imputation under misspecification of the missing data mechanism based on 100 simulated data sets with 16 genes and 6 replicates per dataset
From: Multiple imputation and direct estimation for qPCR data with non-detects
| Â | Direct estimation | Multiple imputation | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
Bias | MSE | Bias | MSE | |||||||||
| Â | Logit | |||||||||||
\(\theta\) | − 0.024 | 0.003 | 0.025 | 0.072 | 0.114 | 0.156 | − 0.024 | 0.003 | 0.023 | 0.072 | 0.114 | 0.155 |
\(\sigma ^2_\theta\) | − 0.001 | 0.000 | 0.000 | 0.000 | 0.001 | 0.001 | 0.020 | − 0.001 | 0.014 | 0.000 | 0.001 | 0.004 |
| Â | Probit | |||||||||||
\(\theta\) | − 0.022 | 0.003 | 0.024 | 0.071 | 0.114 | 0.153 | − 0.024 | −  0.003 | 0.023 | 0.072 | 0.113 | 0.155 |
\(\sigma ^2_\theta\) | − 0.002 | −  0.001 | 0.000 | 0.000 | 0.001 | 0.001 | − 0.020 | −  0.001 |  0.014 | 0.000 | 0.001 | 0.004 |
| Â | Cloglog | |||||||||||
\(\theta\) | − 0.025 | 0.003 | 0.026 | 0.007 | 0.114 | 0.156 | − 0.025 | 0.002 | 0.022 | 0.072 | 0.114 | 0.156 |
\(\sigma ^2_\theta\) | − 0.001 | 0.000 | 0.000 | 0.000 | 0.001 | 0.001 | − 0.022 | 0.000 | 0.014 | 0.000 | 0.001 | 0.004 |