Prof. Dr. André Calero Valdez (short bio)

Institute of Multimedia and Interactive Systems, University of Lübeck, Germany

Objectives:

• The goal of the tutorial is to provide learners with an overview of the partial least squares structural equation modeling technique (PLS-SEM) and how it can be applied in HCI research. For this purpose, participants will learn the SEMinR package that uses free software for data analysis.
• After the tutorial, participants will be able to:
  • Know the basics of SEM modelling using the PLS method.
  • Understand the requirements and benefits of SEM modeling over other statistical approaches.
  • Design HCI studies that are suitable for SEM modelling.
  • Run SEM model analysis using free software.

Content and Benefits:

In this course, you will learn how to apply structural equation modeling (SEM) in your research and how to improve your research design so that your data is suitable for SEM. You will also learn how to build PLS-SEM models and run data analysis using a training data set. You will also learn how to report SEM in your next paper. The tutorial is divided in two parts. First, you will learn the basics and estimate your first model. Second, you will learn advanced concepts, such as bootstrapping your model.
All data analysis will be conducted using the R programming language and the RStudio IDE. All tools used in this tutorial are free software. We will go all steps together and no strong background in R is required. But why should you think about SEM?

Human-Computer interaction (HCI) is a research field that studies the relationship between humans and computers. While HCI focuses on design attributes like performance, efficacy, and accuracy, it also looks at users’ perceptions of technology.
Here, HCI research is heavily informed by behavioral research and applied psychology. It focuses on the traits, beliefs, and perceptions of the examined individuals. Yet, these are unobservable variables or latent variables.

Empirical research in HCI requires rigorous, and versatile methods that can incorporate not only observable variables, but also latent variables, and design variables.

PLS-SEM or partial least squares structural equation modelling is uniquely placed to meet this requirement in that it can simultaneously estimate how these latent variables are constructed from their indicators, and how the variables relate to each other, therefore producing consistent and unbiased results. The dual explanatory and predictive nature of PLS-SEM makes it relevant not only to scientific researchers, but also to practitioners interested in predictive modelling.

PLS-SEM is a powerful method to build and test models for data collected in HCI studies and is used by renowned experts in the field. This method allows the researcher to simultaneously test the validity of their measurement instrument (e.g., the survey) and the relationship between model variables, e.g., perceptions, attitudes, and behavior. A further advantage is that PLS-SEM is comparatively robust to small samples and non-parametric data, both issues that often occur in HCI research.

Target Audience:

This tutorial is aimed at researchers at the Master’s or PhD level that are interested in improving their statistical knowledge. No R programming experience is necessary, but it may be helpful. Participants should know some basic statistics (e.g., p-values, confidence intervals, t-tests).

Additional platform or tool(s) to be used by the tutor:

All exercises will be run using the r-programming language and the RStudio IDE.

List of materials or devices required by the participants:

Participants should bring their laptop and have both R and RStudio installed on their machine.
Here, is a video tutorial to do so: https://www.youtube.com/watch?v=BpMZ2aImc3o