Premium
Confirmatory factor analysis of the Sense of Community Index and development of a Brief SCI
Author(s) -
Long D. Adam,
Perkins Douglas D.
Publication year - 2003
Publication title -
journal of community psychology
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.585
H-Index - 86
eISSN - 1520-6629
pISSN - 0090-4392
DOI - 10.1002/jcop.10046
Subject(s) - confirmatory factor analysis , psychology , construct validity , construct (python library) , likert scale , index (typography) , social psychology , reliability (semiconductor) , factor analysis , psychometrics , developmental psychology , structural equation modeling , statistics , computer science , mathematics , world wide web , programming language , power (physics) , physics , quantum mechanics
The Sense of Community Index (SCI) is the most widely used measure of the construct, despite the lack of confirmation of its intended dimensions or subscales. Yet psychometric appraisals of the SCI have never used the proper tool for evaluating an established dimensional construct: confirmatory factor analysis (CFA). The Block Booster Project data set (of residents of 47 street blocks in Brooklyn and Queens, New York) used in developing the original SCI in 1985 was used here in its reassessment, along with a 1‐year follow‐up survey expanded to 61 blocks. The Sense of Community Index reanalysis using CFA yielded poor model fit for McMillan and Chavis' (1986) original theoretical formulation as well as for a single‐factor index, prompting development of a Brief SCI. The eight‐item, three‐factor BSCI showed good model fit with CFA, reliable construct validity in multilevel correlational analyses, and it differentiated street block neighborhoods reliably based on intragroup agreement, while retaining the profile of a cognitive‐perceptual construct, which does not crossover with other popular community psychology constructs such as place attachment and community satisfaction. The authors, however, recommend future uses of the BSCI employ a 5‐point Likert‐type response format to increase the measure's variability, sensitivity, and internal reliability. © 2003 Wiley Periodicals, Inc. J Comm Psychol 31: 279–296, 2003.