Published online by Cambridge University Press: 01 August 2014
The use of twins as a research tool is still proving its worth both by the number of scientists engaged in twin research and by the significance of their results. Old, familiar research designs are finding new applications, and new research designs are appearing. Of greatest current interest are the epidemiological studies made possible by the assembly and the aging of large numbers of twins in twin registries. As an outgrowth partly of the twin registries, partly of conceptual and mathematical progress, new methods have emerged for diagnosis of twin types and for analysis of twin data. One line of development started with the questionnaire method of zygosity diagnosis and has given rise most recently to zygosity diagnosis by principal component analysis. Another line started with probability calculations and has led to the use of generalized distance and noncentral chi-square. The appropriateness of these methods in different contexts needs to be critically considered. Also of importance are the psychologists' new methods of extracting genetic “factors”. The greatest weakness of twin studies, long recognized, is their dependence on the assumption that DZ pairs provide an adequate control on the environmental differences within MZ pairs. This may be valid with respect to environmental influences that are highly self-selected. It is debatable for self-selected influences that differ among families, and clearly untenable for most influences imposed by the social environment peculiar to twins.