Revisiting the Sense of Community Index: A Confirmatory Factor Analysis and Invariance Test

Abstract

The Sense of Community Index (SCI) has been the most widely used measure of Psychological Sense of Community (PSC), despite abundant evidence it does not adequately measure its intended dimensions. Alternative model specifications are rarely tested beyond their initial appearance in the literature, and the use of the SCI as a 1‐dimensional scale has been promoted without testing for invariance across groups. We use a sample of students from a small Northeastern state college to test the measurement properties of the SCI and to test for invariance across semesters of college and across waves of panel data. Neither the original 1‐ or 4‐factor solutions to the SCI nor an alternative 4‐factor solution provide an adequate fit to the data. Evidence of measurement invariance was mixed, with invariance found across waves of data, but not across semesters of school. Future research using PSC should employ measures other than the SCI.

How many factors does the sense of community index assess?

“…To summarize the discussion of our findings, this study provides consistent evidence with Flaherty and colleagues () that the SCI assessment tool should no longer be considered to assess the four dimensions of SOC as posited by McMillan and Chavis (). In an early study of the structure of the index, Chipuer and Pretty () discussed their findings of various models, with three different samples, and suggested “that the use of the SCI as a unidimensional measure may be the most appropriate until the items making up the SCI are reformulated to reflect the four underlying dimensions as conceptualized” (p. 654).…”

“…This finding of a unidimensional scale is consistent with recommendations to use the SCI for a unitary, rather than multiple, factor model (Chipuer & Pretty, ; Proescholdbell et al., ). However, this is inconsistent with Flaherty and colleagues’ () research results, which did not support the one‐ or four‐factor model, perhaps because the measure they used had the negatively worded items.…”

“…They concluded that using both positively and negatively worded items causes method bias in responses. However, this rationale does not seem to be supported by findings from previous studies, in which both positively and negatively worded items did not load (see Flaherty et al., ; Long & Perkins, ). Further investigation into SCI factor structure is needed.…”

Communities of Place? New Evidence for the Role of Distance and Population Size in Community Attachment

“…4 Research on boomtowns and disasters has studied extensively the effect of local disruption on community satisfaction and attachment, but are limited in their scope to generalize across communities not experiencing large disruption (Brown, Dorius, and Krannich 2005;Brown, Geertsen, and Krannich 1989;Cope et al 2013Cope et al , 2015. 5 Although there is evidence that population size is associated with social ties, it does not appear to directly link to community attachment (Flaherty and Brown 2010) 6 Besides some community sociological work, community psychology scholars have also contributed a substantial literature on community sentiment (see Crowe 2010;Flaherty, Zwick, and Bouchey 2014;McMillan and Chavis 1986). In addition, recent sociological studies have begun to incorporate theory and measures from community psychology literature into their work (see Cope et al 2013). One explanation for the apparent paucity of evidence may be that place-based indicators are often underexamined.…”

“… Besides some community sociological work, community psychology scholars have also contributed a substantial literature on community sentiment (see Crowe ; Flaherty, Zwick, and Bouchey ; McMillan and Chavis ). In addition, recent sociological studies have begun to incorporate theory and measures from community psychology literature into their work (see Cope et al ).…”