Analysis of family resemblance. II. A linear model for familial correlation.
Year: 1974
Type: article
Abstract: Wright [1] developed path analysis to describe familial correlation under a linear model. This remains its principal use, despite extension to nonfamilial data and interactive systems [2, 3]. His seminal paper on the intelligence quotient has been surprisingly neglected by human geneticists, perhaps because values had to be assumed for some unknown parameters [4]. Recently this indeterminacy was resolved by combining path analysis with the concept of an index and a theory of hypothesis testing [5]. In this form it appears to be a powerful tool to analyze family resemblance and group differences in man. Since our work depends on path analysis, which, like so much in population genetics, is the creation of Sewall Wright, we take pleasure in dedicating this paper to him. THEORY In this section we describe the necessary theories of estimation and tests of hypotheses (e.g., see [6]) which can be used for any model; these will be used for the models we develop in later sections. The traditional approach to familial correlation is to estimate heritability under assumptions that are not tested. A more fruitful alternative emphasizes tests of hypotheses. For a phenotype influenced by familial environment and/or subject to assortative mating, in an organism where opportunities for critical experimenta-tion are severely limited, it would be reckless to suppose that any particular model is true, whereas it may be feasible to show that a particular model is false. Our interest in familial correlation is to test hypotheses for which we require a distribution theory. Let rN be an estimated intra- or interclass correlation based on N pairs of observations, and let p be its expected value. Fisher [7] showed that for an interclass correlation the bias-corrected z transform,
Source: Europe PMC (PubMed Central)
Authors: D C Rao, N E Morton, S Yee
Language: English
Subfield: Genetics
Domain: Life Sciences
Open Access status: green
OpenAlex ID: W228858247