Partial least squares path modeling: Time for some serious second thoughts

• Author: Jeffrey R. Edwards,Cameron N. McIntosh,Mikko Rönkkö,John Antonakis
• Date: 01 November 2016
• Published date: 01 November 2016
• DOI: http://doi.org/10.1016/j.jom.2016.05.002

Partial least squares path modeling: Time for some serious second

thoughts

Mikko R€

onkk€

o

a

,

*

, Cameron N. McIntosh

b

, John Antonakis

c

, Jeffrey R. Edwards

d

a

Aalto University School of Science, PO Box 15500, FI-00076, Aalto, Finland

b

Public Safety Canada, 340 Laurier Avenue West, Ottawa, Ontario, K1A 0P8, Canada

c

Faculty of Business and Economics, University of Lausanne, Internef #618, CH-1015, Lausanne-Dorigny, Switzerland

d

Kenan-Flagler Business School, University of North Carolina at Chapel Hill, Campus Box 3490, McColl Building, Chapel Hill, NC, 27599-3490, USA

article info

Article history:

Received 26 March 2016

Received in revised form

17 May 2016

Accepted 25 May 2016

Available online 22 June 2016

Keywords:

Partial least squares

Structural equation modeling

Statistical and methodological myths and

urban legends

abstract

Partialle ast squares(P LS) path modeling is increasingly being promoted as a technique of choice for various

analysis scenarios, despite the serious shortcomings of the method. The current lack of methodological

justification for PLS prompted the editors of this journal to declare that research using this technique is

likely to be desk-rejected (Guide and Ketokivi, 2015). Toprovide clarification on the inappropriateness of

PLS for applied research, we providea non-technical review and empirical demonstration of its inherent,

intractable problems. We show that although the PLS technique is promoted as a structural equation

modeling (SEM) technique, it is simplyregressionwith scalescores and thus has verylimited capabilities to

handle the wide array of problems for which applied researchers use SEM. To that end, we explain whythe

use of PLS weights and many rules of thumb that are commonly employed with PLS are unjustifiable,

followed by addressing why the touted advantages of t he method are simply untenable.

©2016 Elsevier B.V. All rights reserved.

1. Introduction

Partial least squares (PLS) has become one of the techniques of

choice for theory testing in some academic disciplines, particularly

marketing and information systems, and its uptake seems to be on

the rise in operations management (OM) as well (Peng and Lai,

2012; R€

onkk€

o, 2014b). The PLS technique is typically presented as

an alternative to structural equation modeling (SEM) estimators

NON SEQUITUR ©2016 Wiley Ink, Inc.. Dist. By UNIVERSAL UCLICK. Reprinted

with permission. All rights reserved.

*Corresponding author.

E-mail address: mikko.ronkko@aalto.fi(M. R€

onkk€

o).

Contents lists available at ScienceDirect

Journal of Operations Management

journal homepage: www.elsevier.com/locate/jom

http://dx.doi.org/10.1016/j.jom.2016.05.002

0272-6963/©2016 Elsevier B.V. All rights reserved.

Journal of Operations Management 47-48 (2016) 9e27

(e.g., maximum likelihood), over which it is presumed to offer

several advantages (e.g., an enhanced ability to deal with small

sample sizes and non-normal data).

Recent scrutiny suggests, however, that many of the purported

advantages of PLS are not supported by statistical theory or

empirical evidence, and that PLS actually has a number of disad-

vantages that are not widely understood (Goodhue et al., 2015;

McIntosh et al., 2014; R€

onkk€

o, 2014b; R€

onkk€

o and Evermann,

2013; R€

onkk€

o et al., 2015). As recently concluded by Henseler

(2014):“[L]ike a hammer is a suboptimal tool to fix screws, PLS is

a suboptimal tool to estimate common factor models”, which are

the kind of models OM researchers use (Guide and Ketokivi, 2015,p.

vii). Unfortunately, whereas a person attempting to hammer a

screw will quickly realize that the tool is ill-suited for that purpose,

the shortcomings of PLS are much more insidious because they are

not immediately apparent in the results of the statistical analysis.

Although PLS promises simple solutions to complex problems and

often produces plausible statistics that are seemingly supportive of

research hypotheses, both the technical and applied literature on

the technique seem to confound two distinct notions: (1) some-

thing can be done; and (2) doing so is methodologically valid

(Westland, 2015,Chapter 3). As stated in a recent editorial by Guide

and Ketokivi (2015):“Claiming that PLS fixes problems or over-

comes shortcomings associated with other estimators is an indirect

admission that one does not understand PLS”(p. vii). However, the

editorial provides little material aimed at improving the under-

standing of PLS and its associated limitations.

Although there is no shortage of guidelines on how the PLS

technique should be used, many of these are based on conventions,

unproven assertions, and hearsay, rather than rigorous methodo-

logical support. Although OM researchers have followed these

guidelines (Peng and Lai, 2012), such works do not help readers

gain a solid and balanced understanding of the technique and its

shortcomings. This state of affairs makes it difficult to justify the

use of PLS, beyond arguing that someone has said that using the

method would be a good idea in a particular research setting (Guide

and Ketokivi, 2015, p. vii) Therefore, in order to mitigate common

misunderstandings, we clarify issues concerning the usefulness of

PLS in a non-technical manner for applied researchers. In light of

these issues, it becomes apparent that the findings of studies

employing PLS are ambiguous at best and at worst simply wrong,

leading to the conclusion that PLS should be discontinued until the

methodological problems explained in this article have been fully

addressed.

2. What is PLS and what does it do?

A PLS analysis consists of two stages. First, observed variables

are combined as weighted sums (composites); second, the com-

posites are used in separate regression analyses, applying null hy-

pothesis significance testing by comparing the ratio of a regression

coefficient and its bootstrapped standard error against Student’st

distribution. In a typical application, the observed variables are

intended to measure theoretical constructs. In this type of analysis,

the purpose of combining multiple indicators into composites is to

produce aggregate measures that can be expected to be more

reliable than any of their components, and can therefore be used as

reasonable proxies for the constructs. Thus the only difference

between PLS and more traditional regression analyses using sum-

med scales, factor scores, or principal components, is how the in-

dicators are weighted to create the composites. Moreover, instead

of applying traditional factor analysis techniques, the quality of the

measurement model is evaluated by inspecting the correlations

between indicators and composites that they form, summarized as

the composite reliability (CR) and average variance extracted (AVE)

indices.

Although PLS is often marketed as a SEM method, a better way

to understand what the technique actually does is to simply

consider it as one of many indicator weighting systems. The

broader methodological literature provides several different ways

to construct composite variables. The simplest possible strategy is

taking the unweighted sum of the scale items, with a refined

version of this approach being the application of unit weights to

standardized indicators (Cohen, 1990). The two most common

empirical weighting systems are principal components, which

retain maximal information from the original data, and factor

scores that assume an underlying factor model (Widaman, 2007),

with different calculation techniques producing scores with

different qualities (Grice, 2001b). Commonly used prediction

weights include regression, correlation, and equal weights (Dana

and Dawes, 2004). Although not linear composites, different

models based on item response theory produce scale scores that

take into account both respondent ability and item difficulty (Reise

and Revicki, 2014). Outside the context of research, many useful

indices are composites, such as stock market indices that can

weight individual stocks based on their price or market capitali-

zation. Given the large number of available approaches for con-

structing composites variables, two key questions are: (1) Does PLS

offer advantages over more well-established procedures?; and (2)

What is the purpose of the PLS weights used to form the com-

posites? We address these questions next.

2.1. On the “optimality”of PLS weights

Most introductory texts on PLS gloss over the purposes of the

weights, arguing that PLS is SEM and therefore it must provide an

advantage over regression with composites (e.g., Gefen et al., 2011);

however,such worksoften do notexplicitly point out that PLS itself

is also simply regression with composites. Other authors suggest

the weights are optimal (e.g., Henseler and Sarstedt, 2013, p. 566),

but do not explain why and for which specific purpose. As noted by

Kr€

amer (2006):“In the literature on PLS, there is often a huge gap

between the abstract model [...] and what is actually computed by

the PLS path algorithms. Normally, the PLS algorithms are pre-

sented directly in connection with the PLS framework, insinuating

that the algorithms produce optimal solutions of an obvious esti-

mation problem attached to PLS. This estimation problem is how-

ever never defined”(p. 22).

The purpose of PLS weights remains ambiguous (R€

onkk€

o et al.,

2015, p. 77), as various rather different explanations abound (see

Table 1 for some examples). However,none of these works (or their

cited literature) provide mathematical proofs or simulation evi-

dence to support their arguments. Perhaps the most common

argument is that the indicator weights maximize the R

2

values of

the regressions between the composites in the model (e.g. Hair

et al., 2014, p. 16). However, this claim is problematic, for two

main reasons: (a) why maximizing the R

2

values is a good opti-

mization criterion is unclear; and (b) PLS has not been shown to be

an optimal algorithm for maximizing R

2

. In contrast, R€

onkk€

o

(2016a, sec. 2.3) demonstrates a scenario where optimizing indi-

cator weights directly with respect to R

2

produces a 118% larger R

2

value than PLS, thus demonstrating that if the purpose of the

analysis is to maximize R

2

, the PLS algorithm is not an effective

algorithm for this task.

Another common claim is that PLS weights reduce the impact of

measurement error (e.g., Chin et al., 2003, p. 194; Gefen et al., 2011,

M. R€

onkk€

o et al. / Journal of Operations Management 47-48 (2016) 9e2710