Bias of Coefficient afor Fixed Congeneric Measures with Correlated Errors

By Tenko Raykov. 2001

316 citations recorded by the source corpus.

View Paper: https://www.semanticscholar.org/paper/caa8237166262a4ea5aa315c32cbd8e6a50d28a4

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Statistics and Probability

Citations in the corpus

Best Alternatives to Cronbach's Alpha Reliability in Realistic Conditions: Congeneric and Asymmetrical Measurements (2016, background): “In fact the exact opposite is the case, as was shown by Sijtsma (2009), and its application in such conditions may lead to reliability being heavily overestimated (Raykov, 2001).”
Commentary on Coefficient Alpha: A Cautionary Tale (2009, background): “Much less attention has been paid to the assumption of uncorrelated errors and the effect of violation of this assumption on coefficient alpha (Green & Hershberger, 2000; Zumbo & Rupp, 2004), although a number of researchers have discussed this assumption (e.g., Becker, 2000; Fleishman & Benson, 198”
Correction for Unreliability of Multifactor Measures: Comparison of Alpha and Parallel Forms Approaches (2002, unknown): “Similar problems with the use of alpha have been noted more recently (Komaroff, 1997; Murphy & DeShon, 2000; Raykov, 1998, 2001).”
Interpretaciones del Coeficiente Alpha de Cronbach (2005, background): “…Zumbo & Lalonde, 1993; Osburn, 2000), a la presencia de correlaciones entre los errores de los ítems (Zimmerman et al., 1993; Komaroff, 1997; Raykov, 2001), a la cantidad de componentes (Niemi, Carmines & McIver, 1986; Cortina, 1993; Zimmerman et al., 1993), a la presencia de error…”
The Effects of Parceling Unidimensional Scales on Structural Parameter Estimates in Structural Equation Modeling (2006, background): “The one issue that may influence model estimation without enormous influence on the structural coefficients is when error terms are slightly correlated, as correlated error terms affect a measure’s reliability (e.g., Cronbach’s α; see Green & Hershberger, 2000; Raykov, 1998, 2001).”
Attempted Validation of the Scores of the VARK: Learning Styles Inventory With Multitrait–Multimethod Confirmatory Factor Analysis Models (2010, methodology): “We obtained the reliability of the scores of VARK subscales with the following equation (Raykov, 2001b): (2) where rs is the reliability of the scores of a subscale, lj are the factor loadings of the items of the subscale, var(h) is the variance of the subscale’s factor, and var(ej) are the error…”
A Note on Structural Equation Modeling Estimates of Reliability (2010, unknown): “A number of researchers have suggested the use of SEM for estimating reliability (eg, Bentler, 2009 ; Bollen, 1989 ; Brunner & Sub, 2005 ; Green & Hershberger, 2000 ; Green & Yang, 2009 ; Joreskog, 1971 ; McDonald, 1999 ; Raykov, 1997a , 1997b , 1998 , 2000 , 2001 ; Raykov & Shrout, 200”
A Test for Change of Composite Reliability in Scale Development (2003, methodology): “Indeed, research over more than 30 years has provided ample evidence about deficiencies of (e.g., Lord & Novick, 1968; McDonald, 1999; Novick & Lewis, 1967; Raykov, 1997a, 1997b, 1998a, 2001b; Zimmerman, 1972).”
Estimating ω h for Structures Containing Two Group Factors: Perils and Prospects (2007, background): “…2007 APPLIED PSYCHOLOGICAL MEASUREMENT the bias depending on whether the indicators are multidimensional, are not tau equivalent, and/or have correlated measurement errors (Zinbarg et al., 2005; for related discussions, see Komaroff, 1997; Raykov, 1998a, 2001; Zimmerman, Zumbo, & Lalonde, 1993).”
Use Omega Rather than Cronbach’s Alpha for Estimating Reliability. But… (2020, background): “…it is true that α tends to underestimate reliability when the essential tau-equivalence assumption is violated, research shows that in some circumstances, such as when the random measurement errors are correlated across items, α can overestimate reliability (see McNeish, 2018; Raykov, 1997a, 2001).”