Some Theoretical and Methodological Considerations for Measuring Intrasocietal Conflict

Publisher Website
Google Scholar
Abstract
Among previous attempts to measure lntrasocietal conflict, the specification of the measurement model has remained implicit. This paper attempts to explicate the principal design considerations which enter into the construction of such a model (i.e., causal versus effect indicators, uni- versus multidimensions, single versus multiple indicators, correlated versus noncorrelated error terms, and additive versus multiplicative relationships). The statistical and theoretical implications of these alternative features are discussed in terms of the current literature dealing with the national measurement of conflict.