Moreover, as norms and attitudes vary, so do the research efforts and data collection mechanisms. The enormous variation of definitions hamper definitive cross-cultural analysis of data on family violence. The discrepancies in the numbers of articles on spouse abuse and child abuse in the international literature could be a reflection of the trend in the United States, where child abuse was identified as a significant social and family problem 10 years before spouse abuse was. Methods of data collection and analysis in the United Kingdom and Canada are similar to those used in the United States. Limitations, also similar to those found in American research, include the failure to use comparison groups in surveys; the use of small, nonrepresentative samples; and the use of samples drawn from a single source or based only on officially recognized cases. Social-structural models of family violence are used more widely in other countries than in the United States, possibly because this other research began after the intraindividual models had fallen into disrepute in the United States. The dominant paradign used by researchers in Europe and countries in the Third World does not attempt to locate the problem in bad people, but focuses instead on social relations or structures. Seventy-three references are listed.