Research in the Foundations and Practice of Social Measurement

University Of California-Irvine, Irvine CA

Investigators

Abstract

Social measurement entails discovering the nature of knowledge shared by a group of individuals who possess common backgrounds, cultural knowledge, or social norms, i.e., knowledge defined by social conventions. Unlike most physical and psychophysical measurements, social measurements need not have objective external criteria. Therefore cultural knowledge and social norms must be inferred from the pattern of responses among respondents to test items posed by researchers. Previous research has developed a collection of formal models and methods for social measurement known as cultural consensus theory. The models bear a close relationship to models in psychometric test theory; however, unlike most applications of test theory, the researcher does not assume prior knowledge of the "answer key" for the test items. The theory consists of a family of knowledge aggregation models for questionnaire data that permits the simultaneous estimation of the competence or knowledge of each respondent and the consensus correct answer to each question. Models exist for particular testing formats such as true-false, multiple-choice, matching, and rank order for the case of a single homogeneous culture. This research project will take cultural consensus theory to the next level by developing new models to deal with more complex response formats, and to extend the models to the multiculture case where the respondents can be partitioned into subgroups. The new models will then be applied to a set of empirical studies concerning reliable differences in the perceptual structure of color, in particular across different language groups and between different individuals within a group. State-of-the-art Markov Chain Monte Carlo estimation methods will be developed for these new models. The advances in theory and method that will emerge from this research will improve the quality of social measurement, both for research questions related to the perceptual structure of color and for other measurement questions that depend upon aggregating the judgements of respondents.