ABSTRACT
The most common technique of measurement involves the combined use of several individual indicators to build a summary score, or index. In practice, the techniques described for aggregating scores across a number of indicators range from a simple additive and unweighted procedure to the kind of estimation techniques considered by Robert M. Hauser and Arthur S. Goldberger. From a practical and especially predictive standpoint, the amount of bias of a particular estimation technique may be of less concern than its variability. The greater the proportion of "known" quantities to "unknowns", the greater is one's ability to reject alternative auxiliary theories linking the measured variables with unmeasured ones. The most striking and significant development in social science methodology concerns the study of measurement models, or, the causal analysis of unobserved variables. In most measurement situations one is faced with the task of selecting measurable indicators of a given variable from a larger pool of all possible indicators.
