Article Version of Record

Estimation quality and required sample sizes in three-level contextual analysis models

Author(s) / Creator(s)

Kerkhoff, Denny
Nussbeck, Fridtjof W.

Abstract / Description

In multilevel analysis, Level-1 predictors that also explain variance at a higher level are called contextual predictors. In the multilevel manifest covariate model, the Level-2 component is modeled as the average of the Level-1 predictor scores within a cluster. In the multilevel latent covariate model, the predictor is decomposed into two latent variables at Level-1 and Level-2. Performance conditions of these modeling approaches for three-level models are largely unexplored. We investigate the two approaches’ performance with respect to bias, coverage, and power in a three-level random intercept model. Results reveal differences in estimation quality and required sample sizes. We provide sampling recommendations for both approaches.

Keyword(s)

hierarchical linear model three-level model sample sizes parameter estimation bias power coverage contextual variable

Persistent Identifier

Date of first publication

2023-06-30

Journal title

Methodology

Volume

19

Issue

2

Page numbers

133–151

Publisher

PsychOpen GOLD

Publication status

publishedVersion

Review status

peerReviewed

Is version of

Citation

Kerkhoff, D. & Nussbeck, F. W. (2023). Estimation quality and required sample sizes in three-level contextual analysis models. Methodology, 19(2), 133-151. https://doi.org/10.5964/meth.9775
  • Author(s) / Creator(s)
    Kerkhoff, Denny
  • Author(s) / Creator(s)
    Nussbeck, Fridtjof W.
  • PsychArchives acquisition timestamp
    2023-11-23T11:52:10Z
  • Made available on
    2023-11-23T11:52:10Z
  • Date of first publication
    2023-06-30
  • Abstract / Description
    In multilevel analysis, Level-1 predictors that also explain variance at a higher level are called contextual predictors. In the multilevel manifest covariate model, the Level-2 component is modeled as the average of the Level-1 predictor scores within a cluster. In the multilevel latent covariate model, the predictor is decomposed into two latent variables at Level-1 and Level-2. Performance conditions of these modeling approaches for three-level models are largely unexplored. We investigate the two approaches’ performance with respect to bias, coverage, and power in a three-level random intercept model. Results reveal differences in estimation quality and required sample sizes. We provide sampling recommendations for both approaches.
    en_US
  • Publication status
    publishedVersion
  • Review status
    peerReviewed
  • Citation
    Kerkhoff, D. & Nussbeck, F. W. (2023). Estimation quality and required sample sizes in three-level contextual analysis models. Methodology, 19(2), 133-151. https://doi.org/10.5964/meth.9775
    en_US
  • ISSN
    1614-2241
  • Persistent Identifier
    https://hdl.handle.net/20.500.12034/9144
  • Persistent Identifier
    https://doi.org/10.23668/psycharchives.13664
  • Language of content
    eng
  • Publisher
    PsychOpen GOLD
  • Is version of
    https://doi.org/10.5964/meth.9775
  • Is related to
    https://doi.org/10.23668/psycharchives.12944
  • Keyword(s)
    hierarchical linear model
    en_US
  • Keyword(s)
    three-level model
    en_US
  • Keyword(s)
    sample sizes
    en_US
  • Keyword(s)
    parameter estimation bias
    en_US
  • Keyword(s)
    power
    en_US
  • Keyword(s)
    coverage
    en_US
  • Keyword(s)
    contextual variable
    en_US
  • Dewey Decimal Classification number(s)
    150
  • Title
    Estimation quality and required sample sizes in three-level contextual analysis models
    en_US
  • DRO type
    article
  • Issue
    2
  • Journal title
    Methodology
  • Page numbers
    133–151
  • Volume
    19
  • Visible tag(s)
    Version of Record
    en_US