When to Use Machine Learning: A Course Project

Kayhan, Varol (vkayhan@usf.edu)

Abstract

The number of institutions offering machine learning courses is on the rise. Supplementary materials that help teach these courses fail to address one of the most important steps of the machine learning process, namely identifying a problem, and determining whether it is appropriate for machine learning. We address this problem by first reviewing frameworks in extant work, then proposing a decision flow to help students determine whether an input-output relationship is appropriate for machine learning. Following the discussion of the steps in the decision flow, we present a course assignment that reinforces the concepts in the decision flow. We conclude by discussing the lessons learned after using this assignment in a graduate class.

Partial least squares is an estimator for structural equation models: A comment on Evermann and Rönkkö (2021)

Schuberth, Florian (f.schuberth@utwente.nl)

Abstract

In 2012 and 2013, several critical publications questioned many alleged PLS properties. As a consequence, PLS benefited from a boost of developments. It is, therefore, a good time to review these developments. Evermann and R¨onkk¨o devote their paper to this task and formulate guidelines in the form of 14 recommendations. Yet, while they identified the major developments, they overlook a fundamental change, maybe because it is so subtle: The view on PLS. As mentioned by Evermann and R¨onkk¨o (2021, p. 1), “[PLS] is a statistical method used to estimate linear structural equation models” and consequently should not be regarded as a standalone SEM technique following its own assessment criteria. Against this background, we explain which models can be estimated by PLS and PLSc. Moreover, we present the Henseler-Ogasawara specification to estimate composite models by common SEM estimators. Additionally, we review Evermann and R¨onkk¨o’s 14 recommendations one by one and suggest updates and improvements where necessary. Further, we address Evermann and R¨onkk¨o’s comments about the latest advancement in composite models and show that PLS is a viable estimator for confirmatory composite analysis. Finally, we conclude that there is little value in distinguishing between covariance-based and variance-based SEM – there is only SEM.

Contributing to the success of PLS in SEM: An action research perspective

Kock, Ned (nedkock@tamiu.edu)

Abstract

We share with Evermann & Rönkkö (2021) the belief that classic composite-based partial least squares path modeling (PLS-PM) presents shortcomings when used to conduct structural equation modeling (SEM) analyses. The shortcomings can be traced back to one fundamental problem, which is that latent variables (LVs) are approximated in PLS-PM as exact linear combinations of their corresponding indicators. In SEM, each LV is in fact a factor; i.e., a linear combination of the indicators and a measurement residual. Our approach to addressing the shortcomings of PLS-PM is rather unique among researchers concerned with quantitative methods. We have employed an action research approach, helping investigators employ SEM in their empirical studies. This has led to our development of a widely used software tool for SEM analyses. We illustrate our action research orientation by discussing three recent methodological developments with which we have been closely involved.

Don’t Throw the Baby Out With the Bathwater: Comments on “Recent Developments in PLS”

Russo, Daniel (daniel.russo@cs.aau.dk)

Abstract

Evermann and Rönkkö aim to present an overview of recent advances in PLS, and while some advances are described with several useful recommendations, we argue that their article does not fully deliver on its promise. In this response, we argue that their position presents an unbalanced view, ignores several methodological advances by IS scholars. We note that several recommendations are so stringent that implementing that there are philosophical and practical differences that are insufficiently taken into account. Further, several studies that highlight the shortcomings of PLS seem to be based on specially designed cases that are not necessarily representative of typical use of PLS. In our response, we call for a more balanced debate that takes into consideration different perspectives and that studies of the performance of PLS are conducted fairly. While we do not disagree with E&R’s recommendations, the implementation of those is challenged by a lack of tool support, and we observe that besides scholars using PLS, editors and reviewers also have a responsibility to be cognizant of methodological advances. We commend E&R for their efforts in studying the limitations of PLS which have spurred several methodological advances, but also caution that we should not ‘throw the baby out with the bathwater,” by discarding PLS for its known limitations.

Needed Developments in the Understanding of Quasi-Factor Methods

Rigdon, Edward (erigdon@gsu.edu)

Abstract

This response to Evermann and Rönkkö (this issue) acknowledges areas of agreement regarding application of partial least squares (PLS) path modeling, but cites substantial disagreements. Researchers are encouraged to also consider generalized structured component analysis and regression component analysis as procedures built around weighted composites rather than common factors. All of these methods are best regarded as quasi-factor methods, which avoid the uncertainty of factor indeterminacy though they still rely on covariance structures similar to those in factor analysis. They remain useful tools for the researcher.

Extraordinary claims require extraordinary evidence: A comment on “recent developments in PLS”

Sarstedt, Marko (sarstedt@lmu.de)

Abstract

Evermann and Rönkkö (2021) review recent developments in partial least squares (PLS) with the aim of providing guidance to researchers. Indeed, the explosion of methodological advances in PLS in the last decade necessitates such overview articles. In so far as the goal is to provide an objective assessment of the technique, such articles are most welcome. Unfortunately, the authors’ extraordinary and questionable claims paint a misleading picture of PLS. Our goal in this short commentary is to address selected claims made by ER using simulations and latest research. Our objective is to bring a positive perspective to this debate and highlight the recent developments in PLS that make it an increasingly valuable technique in IS and management research in general.

Comments on Evermann and Rönkkö (2021): Recent Developments in PLS

Thompson, Ron (thompsrl@wfu.edu)

Abstract

Evermann and Rönkkö (2022) have provided an excellent overview of recent findings relating to the use of Partial Least Squares (PLS). Their overall message is that if researchers decide to use PLS, they need to ensure that they follow best practices to reduce the possibility of obtaining misleading or erroneous results. We generally agree with their assessment, but go further to recommend against the use of PLS. We demonstrate exactly how PLS introduces biases, arguing that the algorithm violates accepted norms for statistical inference. Our final recommendation is for the Editors-in-Chief of top IS research journals to convene a task force to assess the advisability of continuing to accept articles where PLS is used for publication in IS journals.

AIS4C - AIS Candid Conversation on Community Conduct: Panel Report from ICIS 2020

Zinner Henriksen, Helle (hzh.digi@cbs.dk)

Abstract

This report reflects the discussion that took place at a virtual panel at the ICIS 2020 conference. It focuses on a candid conversation on code of conduct (AIS4C) among AIS community members. As our AIS community has evolved, we have grown in size, diversity, and in the scope of member needs, it is important for all stakeholders to understand what is expected as members of this academic community. The panel included those currently serving in AIS committees related to member and research conduct. The objective of the panel was to start a dialogue about what we – as members of the AIS – each hope to gain from our academic interactions, how AIS can help members achieve these goals and help each other achieve desired outcomes. Maintaining good standing in the AIS community protects individuals’ professional reputations and the reputation of the IS discipline as a whole. Understanding what AIS offers its members to accomplish these objectives, allows individuals to leverage fully AIS member services to become more successful researchers and teachers. By situating the panel within the current COVID-disrupted world, the descriptions of desirable behavior among members and the outlining of member services, this panel report is intended to benefit current and future members of AIS.

Recent Developments in PLS

Evermann, Joerg (jevermann@mun.ca)

Abstract

Partial Least Squares path modeling (PLS) is a method for estimating linear structural equation models. Widely used in the information systems (IS) discipline, there has been considerable argument over its relative merits compared to simple summed scores or to covariance-based estimation of structural equation models. This paper comments on recent developments in PLS to ensure that IS researchers have up-to-date methodological knowledge and best practices if they decide to use PLS. The paper briefly reviews the mechanisms of PLS, its well-known properties of PLS, and its usage history in IS research. We briefly revisit a high-impact critique and debate a few years ago to identify the critical arguments around current practices and use of PLS. That critique proved the driver for many advances in the field, which are discussed extensively and used to make 14 recommendations for how and when to use PLS or alternatives.

The ACM/AIS IS2020 Competency Model for Undergraduate Programs in Information Systems - A Joint ACM/AIS Task Force Report

Salmela, Hannu (hannu.salmela@utu.fi)

Abstract

The Association of Computing Machinery (ACM) and the Association of Information Systems (AIS) along with the ISCAP EDSIG, recently released a joint taskforce report IS2020: A Competency Model for Undergraduate Programs in Information Systems. In this paper, the co-chairs of IS2020, the latest Information Systems curriculum guidelines, provide their insight on the problems presented with existing guidelines, illustrate the issue, and share their opinions that led to the release of these latest guidelines.