INTRODUCTION

That IQ is a highly heritable trait has been widely reported. Rather less well known are recent reports in major scientific journals such as those announcing that the heritability of controllable life events is 53 percent among women and 14 percent among men (Saudino et al. 1997), while the heritabilities of inhibition of aggression, openness to experience, and right-wing authoritarianism are respectively 12, 40, and 50 percent (Pedersen et al. 1989; Bergeman et al. 1993; McCourt et al. 1999). It seems that milk and soda intake are in part heritable, but not the intake of fruit juice or diet soda (de Castro 1993).

These reported heritabilities are parameter estimates obtained in structural modeling of measures taken on pairs of siblings – prototypically, identical (monozygotic) twins and fraternal (dizygotic) twins, some reared together and others reared apart. The models are of the linear random effects type, in which an observed trait – a phenotype – is expressed in terms of latent factors – genetic and environmental – whose prespecified cross-twin correlations differ by zygosity and rearing status. Estimation is by maximum likelihood applied to the phenotypic variances and covariances. Heritability, the key parameter of interest, refers to the proportion of the variance of the phenotype that is attributable to the variance of the genetic factors.

Regarding these studies, various issues arise. Those that I will touch on here include identification, nonnegativity constraints, alternative estimators, pretest estimation, conditioning of the design matrix, multivariate analyses, and the objectives of structural modeling.