On the exploratory road to unraveling factor loading non-invariance

A new multigroup rotation approach

Kim De Roover*, Jeroen K. Vermunt

*Corresponding author for this work

Research output: Contribution to journalArticleScientificpeer-review

1 Downloads (Pure)

Abstract

Multigroup exploratory factor analysis (EFA) has gained popularity to address measurement invariance for two reasons. Firstly, repeatedly respecifying confirmatory factor analysis (CFA) models strongly capitalizes on chance and using EFA as a precursor works better. Secondly, the fixed zero loadings of CFA are often too restrictive. In multigroup EFA, factor loading invariance is rejected if the fit decreases significantly when fixing the loadings to be equal across groups. To locate the precise factor loading non-invariances by means of hypothesis testing, the factors' rotational freedom needs to be resolved per group. In the literature, a solution exists for identifying optimal rotations for one group or invariant loadings across groups. Building on this, we present multigroup factor rotation (MGFR) for identifying loading non-invariances. Specifically, MGFR rotates group-specific loadings both to simple structure and between-group agreement, while disentangling loading differences from differences in the structural model (i.e., factor (co)variances).

Original languageEnglish
Number of pages19
JournalStructural Equation Modeling
DOIs
Publication statusPublished - 2019

Keywords

  • measurement invariance
  • factor loading invariance
  • multigroup exploratory factor analysis
  • rotation identification
  • COVARIANCE STRUCTURE-ANALYSIS
  • CONFIRMATORY FACTOR-ANALYSIS
  • GENERAL-METHOD
  • SAMPLE-SIZE
  • CRITERIA
  • COMMUNALITY
  • INVENTORY
  • TARGET
  • MODEL
  • POWER

Cite this

@article{4f02d5682e854a8d84d1d45ed83ad62d,
title = "On the exploratory road to unraveling factor loading non-invariance: A new multigroup rotation approach",
abstract = "Multigroup exploratory factor analysis (EFA) has gained popularity to address measurement invariance for two reasons. Firstly, repeatedly respecifying confirmatory factor analysis (CFA) models strongly capitalizes on chance and using EFA as a precursor works better. Secondly, the fixed zero loadings of CFA are often too restrictive. In multigroup EFA, factor loading invariance is rejected if the fit decreases significantly when fixing the loadings to be equal across groups. To locate the precise factor loading non-invariances by means of hypothesis testing, the factors' rotational freedom needs to be resolved per group. In the literature, a solution exists for identifying optimal rotations for one group or invariant loadings across groups. Building on this, we present multigroup factor rotation (MGFR) for identifying loading non-invariances. Specifically, MGFR rotates group-specific loadings both to simple structure and between-group agreement, while disentangling loading differences from differences in the structural model (i.e., factor (co)variances).",
keywords = "measurement invariance, factor loading invariance, multigroup exploratory factor analysis, rotation identification, COVARIANCE STRUCTURE-ANALYSIS, CONFIRMATORY FACTOR-ANALYSIS, GENERAL-METHOD, SAMPLE-SIZE, CRITERIA, COMMUNALITY, INVENTORY, TARGET, MODEL, POWER",
author = "{De Roover}, Kim and Vermunt, {Jeroen K.}",
year = "2019",
doi = "10.1080/10705511.2019.1590778",
language = "English",
journal = "Structural Equation Modeling",
issn = "1070-5511",
publisher = "ROUTLEDGE JOURNALS, TAYLOR & FRANCIS LTD",

}

TY - JOUR

T1 - On the exploratory road to unraveling factor loading non-invariance

T2 - A new multigroup rotation approach

AU - De Roover, Kim

AU - Vermunt, Jeroen K.

PY - 2019

Y1 - 2019

N2 - Multigroup exploratory factor analysis (EFA) has gained popularity to address measurement invariance for two reasons. Firstly, repeatedly respecifying confirmatory factor analysis (CFA) models strongly capitalizes on chance and using EFA as a precursor works better. Secondly, the fixed zero loadings of CFA are often too restrictive. In multigroup EFA, factor loading invariance is rejected if the fit decreases significantly when fixing the loadings to be equal across groups. To locate the precise factor loading non-invariances by means of hypothesis testing, the factors' rotational freedom needs to be resolved per group. In the literature, a solution exists for identifying optimal rotations for one group or invariant loadings across groups. Building on this, we present multigroup factor rotation (MGFR) for identifying loading non-invariances. Specifically, MGFR rotates group-specific loadings both to simple structure and between-group agreement, while disentangling loading differences from differences in the structural model (i.e., factor (co)variances).

AB - Multigroup exploratory factor analysis (EFA) has gained popularity to address measurement invariance for two reasons. Firstly, repeatedly respecifying confirmatory factor analysis (CFA) models strongly capitalizes on chance and using EFA as a precursor works better. Secondly, the fixed zero loadings of CFA are often too restrictive. In multigroup EFA, factor loading invariance is rejected if the fit decreases significantly when fixing the loadings to be equal across groups. To locate the precise factor loading non-invariances by means of hypothesis testing, the factors' rotational freedom needs to be resolved per group. In the literature, a solution exists for identifying optimal rotations for one group or invariant loadings across groups. Building on this, we present multigroup factor rotation (MGFR) for identifying loading non-invariances. Specifically, MGFR rotates group-specific loadings both to simple structure and between-group agreement, while disentangling loading differences from differences in the structural model (i.e., factor (co)variances).

KW - measurement invariance

KW - factor loading invariance

KW - multigroup exploratory factor analysis

KW - rotation identification

KW - COVARIANCE STRUCTURE-ANALYSIS

KW - CONFIRMATORY FACTOR-ANALYSIS

KW - GENERAL-METHOD

KW - SAMPLE-SIZE

KW - CRITERIA

KW - COMMUNALITY

KW - INVENTORY

KW - TARGET

KW - MODEL

KW - POWER

UR - https://app-eu.readspeaker.com/cgi-bin/rsent?customerid=10118&lang=en_us&readclass=rs_readArea&url=https%3A%2F%2Fwww.tandfonline.com%2Fdoi%2Ffull%2F10.1080%2F10705511.2019.1590778

U2 - 10.1080/10705511.2019.1590778

DO - 10.1080/10705511.2019.1590778

M3 - Article

JO - Structural Equation Modeling

JF - Structural Equation Modeling

SN - 1070-5511

ER -