Analyzing Cultural Consensus with Proportional Reduction in Error (PRE): Beyond the Eigenvalue Ratio. Issue 2 (May 2016)

• Record Type: Journal Article
• Authors: Lacy, Michael G.; Snodgrass, Jeffrey G.
• Is Part Of: Field methods. Volume 28:Issue 2(2016:May)
• Page Start: 153
• Page End: 169
• Publication Date: 2016-05
• Subjects: cultural sharing -- cognitive models -- consensus analysis -- statistical prediction and error reduction; Ethnology -- Methodology -- Periodicals; Anthropology -- Methodology -- Periodicals; Anthropologie -- Méthodologie -- Périodiques; 306.072
• DOI: 10.1177/1525822X15596919
• Languages: English
• ISSNs: 1525-822X

Abstract

This article provides an alternate method to assess the fit of the cultural consensus model (CCM) of Romney and colleagues to the responses of a group of informants about a domain of knowledge, and thus also to evaluate the extent of shared knowledge within a group. Criteria for judging the existence of singular culture have been articulated previously by examining the relative size of the first versus second eigenvalue from the factor analysis used to estimate the informant competencies. By contrast, our new procedure examines whether a particular cultural consensus answer key and set of competency scores can predict group members' responses, allowing one to assess the degree of cultural sharing in terms of the model's predictive power as compared to guessing. In comparison to previously established CCM approaches for assessing cultural sharing, our approach is intuitively interpretable while remaining based on well-established statistical procedures.