Table 5.
Error Rates When the Variances Model for Ωijk Is Inappropriate
|
Error
Rate when Total No. of
Offspring Isa |
|||||
| Offspring per Family | 120 | 240 | 480 | 960 | 1,920 |
| When Parental Genotypes Are Available | |||||
| 1 | 5.9 | 5.6 | 5.2 | 5.1 | 5.2 |
| 2 | 6.3 | 5.2 | 5.5 | 5.3 | 5.8 |
| 4 | 6.3 | 5.2 | 5.5 | 5.3 | 5.6 |
| 8 | 6.9 | 6.6 | 6.2 | 5.9 | 5.5 |
| When Parental Genotypes Are Not Available | |||||
| 2 | 7.0 | 5.7 | 5.8 | 5.2 | 5.7 |
| 4 | 6.5 | 6.6 | 6.1 | 6.3 | 6.2 |
| 8 | 8.6 | 7.0 | 6.2 | 6.1 | 5.5 |
a
Proportion of simulations exceeding the nominal .05 significance level when there is a linked dominant major gene (θ=0, h2=.3) with equally frequent alleles. The model for variances in Ωj did not include a dominance variance component, so that in larger families the error rate is high. This major gene also introduced skewness in the phenotype distribution, violating multivariate normality.