Consistent partial least squares path modeling

Theo K. Dijkstra, Jörg Henseler

Research output: Contribution to journalArticleAcademicpeer-review

292 Citations (Scopus)
371 Downloads (Pure)

Abstract

This paper resumes the discussion in information systems research on the use of partial least squares (PLS) path modeling and shows that the inconsistency of PLS path coefficient estimates in the case of reflective measurement can have adverse consequences for hypothesis testing. To remedy this, the study introduces a vital extension of PLS: consistent PLS (PLSc). PLSc provides a correction for estimates when PLS is applied to reflective constructs: The path coefficients, inter-construct correlations, and indicator loadings become consistent. The outcome of a Monte Carlo simulation reveals that the bias of PLSc parameter estimates is comparable to that of covariance-based structural equation modeling. Moreover, the outcome shows that PLSc has advantages when using non-normally distributed data. We discuss the implications for IS research and provide guidelines for choosing among structural equation modeling techniques
Original languageEnglish
Pages (from-to)297-316
JournalMIS quarterly = Management information systems quarterly
Volume39
Issue number2
Publication statusPublished - 2015

Fingerprint

Information systems
Testing
Partial least squares
Modeling
Monte Carlo simulation
Coefficients
Structural equation modeling
Remedies
Information systems research
Résumé
Hypothesis testing
Inconsistency

Keywords

  • METIS-308271
  • IR-93763

Cite this

@article{eec12e1b61644a25af70a4ad01386334,
title = "Consistent partial least squares path modeling",
abstract = "This paper resumes the discussion in information systems research on the use of partial least squares (PLS) path modeling and shows that the inconsistency of PLS path coefficient estimates in the case of reflective measurement can have adverse consequences for hypothesis testing. To remedy this, the study introduces a vital extension of PLS: consistent PLS (PLSc). PLSc provides a correction for estimates when PLS is applied to reflective constructs: The path coefficients, inter-construct correlations, and indicator loadings become consistent. The outcome of a Monte Carlo simulation reveals that the bias of PLSc parameter estimates is comparable to that of covariance-based structural equation modeling. Moreover, the outcome shows that PLSc has advantages when using non-normally distributed data. We discuss the implications for IS research and provide guidelines for choosing among structural equation modeling techniques",
keywords = "METIS-308271, IR-93763",
author = "Dijkstra, {Theo K.} and J{\"o}rg Henseler",
note = "Open access",
year = "2015",
language = "English",
volume = "39",
pages = "297--316",
journal = "MIS quarterly = Management information systems quarterly",
issn = "0276-7783",
publisher = "Management Information Systems Research Center",
number = "2",

}

Consistent partial least squares path modeling. / Dijkstra, Theo K.; Henseler, Jörg.

In: MIS quarterly = Management information systems quarterly, Vol. 39, No. 2, 2015, p. 297-316.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Consistent partial least squares path modeling

AU - Dijkstra, Theo K.

AU - Henseler, Jörg

N1 - Open access

PY - 2015

Y1 - 2015

N2 - This paper resumes the discussion in information systems research on the use of partial least squares (PLS) path modeling and shows that the inconsistency of PLS path coefficient estimates in the case of reflective measurement can have adverse consequences for hypothesis testing. To remedy this, the study introduces a vital extension of PLS: consistent PLS (PLSc). PLSc provides a correction for estimates when PLS is applied to reflective constructs: The path coefficients, inter-construct correlations, and indicator loadings become consistent. The outcome of a Monte Carlo simulation reveals that the bias of PLSc parameter estimates is comparable to that of covariance-based structural equation modeling. Moreover, the outcome shows that PLSc has advantages when using non-normally distributed data. We discuss the implications for IS research and provide guidelines for choosing among structural equation modeling techniques

AB - This paper resumes the discussion in information systems research on the use of partial least squares (PLS) path modeling and shows that the inconsistency of PLS path coefficient estimates in the case of reflective measurement can have adverse consequences for hypothesis testing. To remedy this, the study introduces a vital extension of PLS: consistent PLS (PLSc). PLSc provides a correction for estimates when PLS is applied to reflective constructs: The path coefficients, inter-construct correlations, and indicator loadings become consistent. The outcome of a Monte Carlo simulation reveals that the bias of PLSc parameter estimates is comparable to that of covariance-based structural equation modeling. Moreover, the outcome shows that PLSc has advantages when using non-normally distributed data. We discuss the implications for IS research and provide guidelines for choosing among structural equation modeling techniques

KW - METIS-308271

KW - IR-93763

M3 - Article

VL - 39

SP - 297

EP - 316

JO - MIS quarterly = Management information systems quarterly

JF - MIS quarterly = Management information systems quarterly

SN - 0276-7783

IS - 2

ER -