Multiple-Component Measurement Instruments in Heterogeneous Populations: Is There a Single Coefficient Alpha?

This note confronts the common use of a single coefficient alpha as an index informing about reliability of a multicomponent measurement instrument in a heterogeneous population. Two or more alpha coefficients could instead be meaningfully associated with a given instrument in finite mixture setting...

Descripción completa

Detalles Bibliográficos
Publicado en:Educational & Psychological Measurement Vol. 79; no. 2; pp. 399 - 413
Autores principales: Raykov, Tenko, Marcoulides, George A., Harrison, Michael, Menold, Natalja
Formato: Artículo
Publicado: Sage Publications Inc. Apr2019
Materias:
Acceso en línea:Ver este registro en EBSCOhost
fields @attributes:
  recordID: 1
pdfLink:
plink: https://search.ebscohost.com/login.aspx?direct=true&db=ssf&AN=135387976&site=ehost-live
header:
  @attributes:
    shortDbName: ssf
    uiTerm: 135387976
    longDbName: Social Sciences Full Text (H.W. Wilson)
    uiTag: AN
  controlInfo:
    bkinfo:
    jinfo:
      jid:
        00131644
        EPM
      jtl: Educational & Psychological Measurement
      issn: 00131644
      maglogo: Y
    pubinfo:
      dt: Apr2019
      vid: 79
      iid: 2
      pid: 344
      pub: Sage Publications Inc.
    artinfo:
      ui:
        135387976
        10.1177/0013164417733305
      ppf: 399
      ppct: 14
      formats:
      tig:
        atl: Multiple-Component Measurement Instruments in Heterogeneous Populations: Is There a Single Coefficient Alpha?
      aug:
        au:
          Raykov, Tenko
          Marcoulides, George A.
          Harrison, Michael
          Menold, Natalja
        affil:
          Michigan State University, East Lansing, MI, USA
          University of California, Santa Barbara, CA, USA
          University of North Carolina at Chapel Hill, Chapel Hill, NC, USA
          GESIS - Leibniz Institute for the Social Sciences, Mannheim, Germany
      su:
        Psychometrics
        Reliability (Personality trait)
        Empirical research
        Demographic characteristics
        Statistical correlation
        Finite element method
        High performance computing
        Latent structure analysis
        Mathematics
        Research methodology
        Research methodology evaluation
        Evaluation
      sug:
        subj:
          Psychometrics
          Reliability (Personality trait)
          Empirical research
          Demographic characteristics
          Statistical correlation
          Finite element method
          High performance computing
          Latent structure analysis
          Mathematics
          Research methodology
          Research methodology evaluation
          Evaluation
      keyword:
        coefficient alpha
        latent class
        measuring instrument
        mixture
        reliability
        unobserved heterogeneity
        coefficient alpha
        latent class
        measuring instrument
        mixture
        reliability
        unobserved heterogeneity
      ab: This note confronts the common use of a single coefficient alpha as an index informing about reliability of a multicomponent measurement instrument in a heterogeneous population. Two or more alpha coefficients could instead be meaningfully associated with a given instrument in finite mixture settings, and this may be increasingly more likely the case in empirical educational and psychological research. It is argued that in such situations explicit examination of class-invariance in the alpha coefficient must precede any statements about its possible value in the studied population. The approach permits also the evaluation of between-class alpha differences as well as point and interval estimation of the within-class alpha coefficients. The method can similarly be used in situations with (a) known class membership when distinct (sub)populations are investigated while their number is known beforehand and membership in them is observed for studied persons, as well as (b) in settings where only the number of latent classes is known for a population under investigation. The outlined procedure is illustrated with numerical data.
      pubtype: Academic Journal
      doctype: Article
      src: R
    language: English
    refInfo:
    copyright:
      @attributes:
        flag: N
    holdings:
      @attributes:
        islocal: N