PLS-SEM for the Uninformed: Using SmartPLS in Your PhD Dissertation

Introduction

Like most students, even after completing most of my required PhD coursework and regardless of what level or type of degree being pursued, I was not able to answer the simple question “what is your dissertation going to be about?”  I played around with a number of topics, all centered on technology, probably technology use, but possibly technology management – I just wasn’t able to commit to a topic.  After all, the dissertation process is a big commitment, so why trudge through something for a year or more on a topic that you really don’t care about?  One thing I was certain about was that I planned on conducting a quantitative study.  It was time to man up, break out the calculator and slide rule, and go for it.  Enough of this ‘Arts’ stuff.  Let’s collect data, analyze it, and draw some real conclusions from objective evidence.

When I was very close to starting my comprehensive exam, I finally decided on a topic that analyzes factors that predict adoption of consumerized devices, more commonly referred to as consumerized information technology (CoIT).  CoIT was a relatively new term that described how workers were starting to use their personal devices they purchased for their own use for work-related tasks, many eventually dumping their company provided IT and using their own devices, own networks, etc. for both personal things and work.  I planned to use the Technology Acceptance Model (TAM) as the theoretical lens for the research.  Not the dozen or so extensions of TAM offered up, including TAM 2 (Venkatesh & Davis, 2000) and UTAUT (Venkatesh, Morris, Davis, & Davis, 2003) when researchers decided that TAM was too slim a model to capture seemingly every nuance of things that could influence acceptance.  I planned to use the two independent variables of perceived ease of use and perceived usefulness and study their effect on the perceived use of CoIT.

In preparation of both my comprehensive exam and the dissertation process, I researched the appropriate quantification methods I would need to analyze my data, including descriptive statistics, multiple linear regression, etc.  In the classroom, for everything ‘quantification’ we used IBM SPSS Statistics software, and thankfully I still owned a license for the software.  All of my data analysis efforts could easily be accomplished using SPSS.  My dissertation plan, to include the use of SPSS as the tool of choice to analyze data, was approved by my dissertation committee.  My dissertation was on track!  Or, so I thought.  After data collection, I refreshed my dissertation committee on the method I planned to use to analyze data using SPSS.  My mentor asked why I wasn’t using PLS-SEM, which was the ‘traditional’ method used to analyze data collected in studies using TAM as the theoretical model.  Of course, I was confused, maybe a little angry (after all, the committee had already approved my use of SPSS!), and completely in the dark.  My first question and only real question was “what is PLS-SEM?”  Of course, I said OK, I should have known better, etc. and went on an Internet adventure to figure out what PLS-SEM was and how I was going to use it, regardless of whether I agreed to using this methodology or not.

References:

Radcliffe, J. (2017). Factors that predict adoption of coit: A quantitative study of workers in the us.  doi:10.13140/RG.2.2.33562.75203

Venkatesh, V., & Davis, F. D. (2000). A theoretical extension of the technology acceptance model: Four longitudinal field studies. Management Science, 46(2), 186-204. doi:10.1287/mnsc.46.2.186.11926

Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technology: Toward a unified view. MIS Quarterly, 27(3), 425-478. Retrieved from http://www.misq.org

Next Post – What PLS-SEM is all about!

Add a Comment

Your email address will not be published. Required fields are marked *