In the fully reduced rank (FARR) model all specific variances are constrained to be zero, as this FARR structure deals solely with the factor component of the model.
In the present study, we therefore started from a model with all variance components (and all fixed effects) and then built simpler nested models, by removing each time a different variance component.
The complete model, including all variance parameters, was re-estimated on each set of 10 folds and predictions were computed for the phenotypes in the remaining 11th fold.
We present parameter estimates, 95% confidence intervals, and goodness of fit statistics for 3 models: the full model (ACE), a model in which all variance was attributable to genetic and specific environmental factors (AE), and a model in which all variance was produced by common and specific environmental factors (CE).
To generate data under the assumption that the Michaelson model explains all the variance in the mutation rate, we sampled from a Poisson distribution with expected values Z(x).
Under the heterogeneous variance MTC models, all between-trial variances are allowed to take on different values.
In this model, all factors had estimated variances and covariances, but the factor means were fixed to 0. Second, we tested for weak factorial (metric) invariance, holding all factor loadings equal across time.
The paper provides several simulated examples, a small empirical example to demonstrate proof of concept, and a larger empirical example based on many experimental conditions and large samples that demonstrates that the individual models capture virtually all the variance in aggregate first choices traditionally modeled in discrete choice experiments.
Test of F in the ANOVA analysis is used for evaluating model variance in residual variance.
The model explained 34.9% of all variance.
