ABSTRACT
In social science, we typically work with measures that are laden with errors. eories are generally couched in terms of constructs or phenomena that are unobserved or unobservable but nevertheless have empirical implications. It is these unobservable indicators that we actually analyze. Hence, an important preliminary task for an analyst is to specify the assumed relationship of observable indicators to the underlying phenomenon that is to be measured. Better still, one can test hypotheses about these relations. e focus of this volume is on cross-cultural methods; this brings with it further complexities for deriving adequate social measurement in that we cannot expect to necessarily see the same empirical patterns of observations across cultures (for instance, on a set of related questionnaire items measuring a single attitude) even where the underlying phenomenon is in fact the same. Absent some resolution of this problem, comparisons of the true dierences in, say, attitudes or beliefs across
cultures are problematic because of the conation of such dierences as may exist with dierences in the ways in which the observable indictors “behave” across the very same cultural boundaries.
