Quantifying idiosyncratic and shared contributions to stimulus evaluations

A fundamental psychological problem is identifying the idiosyncratic and shared contributions to stimulus evaluation. However, there is no established method for estimating these contributions and the existing methods have led to divergent estimates. Moreover, in many studies participants rate the stimuli only once, although at least two measurements are required to estimate idiosyncratic contributions. Here, participants rated faces or novel objects on four dimensions (beautiful, approachable, likeable, dangerous) for a total of ten blocks to better estimate the preferences of individual raters. First, we show that both intra-rater and inter-rater agreement – measures related to idiosyncratic and shared contributions, respectively – increase with repeated measures. Second, to find best practices, we compared estimates from correlation indices and variance component approaches on stimulus-generality, evaluation-generality, data preprocessing steps, and sensitivity to measurement error (a largely ignored issue). The correlation indices changed monotonically and nonlinearly with more repeated measures. Variance component analyses showed large variability in estimates from only two repeated measures, but stabilized with more measures. While there was general agreement among approaches, the correlation approach was problematic for certain stimulus types and evaluation dimensions. Our results suggest that variance component estimates are more reliable as long as one collects more than two repeated measures, which is not the current norm in psychological research, and can be implemented using mixed models with crossed random effects. Recommendations for analysis and interpretations are provided.

Estimation of breeding values for mean and dispersion, their variance and correlation using double hierarchical generalized linear models

SummaryThe possibility of breeding for uniform individuals by selecting animals expressing a small response to environment has been studied extensively in animal breeding. Bayesian methods for fitting models with genetic components in the residual variance have been developed for this purpose, but have limitations due to the computational demands. We use the hierarchical (h)-likelihood from the theory of double hierarchical generalized linear models (DHGLM) to derive an estimation algorithm that is computationally feasible for large datasets. Random effects for both the mean and residual variance parts of the model are estimated together with their variance/covariance components. An important feature of the algorithm is that it can fit a correlation between the random effects for mean and variance. An h-likelihood estimator is implemented in the R software and an iterative reweighted least square (IRWLS) approximation of the h-likelihood is implemented using ASReml. The difference in variance component estimates between the two implementations is investigated, as well as the potential bias of the methods, using simulations. IRWLS gives the same results as h-likelihood in simple cases with no severe indication of bias. For more complex cases, only IRWLS could be used, and bias did appear. The IRWLS is applied on the pig litter size data previously analysed by Sorensen & Waagepetersen (2003) using Bayesian methodology. The estimates we obtained by using IRWLS are similar to theirs, with the estimated correlation between the random genetic effects being −0·52 for IRWLS and −0·62 in Sorensen & Waagepetersen (2003).

How to deal with genotype uncertainty in variance component quantitative trait loci analyses

SummaryDealing with genotype uncertainty is an ongoing issue in genetic analyses of complex traits. Here we consider genotype uncertainty in quantitative trait loci (QTL) analyses for large crosses in variance component models, where the genetic information is included in identity-by-descent (IBD) matrices. An IBD matrix is one realization from a distribution of potential IBD matrices given available marker information. In QTL analyses, its expectation is normally used resulting in potentially reduced accuracy and loss of power. Previously, IBD distributions have been included in models for small human full-sib families. We develop an Expectation–Maximization (EM) algorithm for estimating a full model based on Monte Carlo imputation for applications in large animal pedigrees. Our simulations show that the bias of variance component estimates using traditional expected IBD matrix can be adjusted by accounting for the distribution and that the calculations are computationally feasible for large pedigrees.