On the fit of models implied by unit-weighted scales. Issue 11 (1st November 2020)

• Record Type: Journal Article
• Authors: Beauducel, André; Hilger, Norbert
• Is Part Of: Communications in statistics. Volume 49:Issue 11(2020)
• Page Start: 3054
• Page End: 3064
• Publication Date: 2020-11-01
• Subjects: Factor analysis -- Factor score estimates -- Unit-weighted sum scales; Mathematical statistics -- Periodicals; Mathematical statistics -- Data processing -- Periodicals; Digital computer simulation -- Periodicals; 519.5
• DOI: 10.1080/03610918.2018.1532517
• Languages: English
• ISSNs: 0361-0918

Abstract

Abstract: Unit-weighted scales are often computed in order to estimate the scores of individuals on common factors. Previous work has shown that unit-weighted scales imply a model that can be evaluated. The algebraic framework for this approach was presented here. The relationship between the fit of the model implied by unit-weighted scales, the scale length, and the item inter-correlations was evaluated algebraically and by means of a simulation study. The standardized root mean square residual (SRMR) excluding the elements in the main-diagonal of the item inter-correlation matrix can be used for the fit evaluation of the unit-weighted scale model.