Methodological issues in cross-cultural large-survey research on violence

Researchers concur on the importance of a cross-cultural perspective for understanding and preventing violence, but not on the particular analytic strategy that should inform such a perspective. This paper examines the actual and potential contributions of one of the dominant analytic strategies in cross-cultural research on violence: research based on large-survey data. I review the major sources of large-survey data, and evaluate these data and the research based on them. I then suggest directions for future work that could compensate for limitations of existing data and research. In particular, I argue for two innovations: (1) development of new sources of cross-cultural survey data on violence based on sub-national sampling units, and (2) multilevel strategies that would balance within-society and between-society analyses and examine intervening mechanisms linking macro-social factors with violent outcomes.