Table 3.
The effect of measurement error in an exposure on the probability of rejecting the null hypothesis (H0) for the test for statistical interaction.
| Coefficient for true effect of interaction (β3) | Ratio for true effect of interaction (β3 + β2)/β2 | Probability of rejecting H0 for test of interaction | ||||
| σu2 = 0 | σu2 = 1 | σu2 = 2 | σu2 = 4 | σu2 = 9 | ||
| 0.0 | 1.0 | 5% | 5% | 5% | 5% | 5% |
| 0.5 | 1.5 | 87% | 54% | 38% | 26% | 16% |
| 1.0 | 2.0 | 100% | 97% | 87% | 67% | 41% |
| 2.0 | 3.0 | 100% | 100% | 100% | 98% | 82% |
True values of coefficients β1 = 1 (binary genotype), β2 = 1 (continuous exposure), β3 (interaction term) given in first column. No confounding present, β4 = 0. Simulations based on 10,000 simulations of 1000 observations.