Print Page   |   Contact Us   |   Report Abuse   |   Sign In   |   Join or Create a Guest Account
2018 Preconference Workshop 3
Share |

WORKSHOP 3
A problem-solving approach to Exploratory and Confirmatory Factor Analysis and Structural Equation Modeling

Leader: Aek Phakiti


Overview

This two-day workshop aims to provide an introduction to exploratory and confirmatory factor analysis (EFA and CFA, respectively) and structural equation modeling (SEM) for language assessment research. EFA is a multivariate statistical technique for exploring the factors that account for the variation in participants’ responses to research instruments, such as Likert-type scale questionnaires. Common factors are often discussed as latent variables. Unlike EFA, CFA is typically used to confirm a common factor that underlies an ability or psychological facet, such as motivation or anxiety. CFA is also used as a measurement model in SEM, which establishes relationships (linear or causal; direct or indirect) among latent variables. This workshop stresses the importance of conceptual understandings and interpretations of analysis of outputs, rather than statistical formulae. In this workshop, Statistical Package for Social Sciences (SPSS) and EQS 6.3 will be used. Participants will have an opportunity to practise each analysis after a step-by-step instruction.

Target participants

The target participants are graduate students in language testing and assessment or early career researchers who are interested to know more about EFA, CFA and SEM. Although this workshop is introductory in nature, participants should have a good knowledge of basic statistics, such as descriptive statistics, normal distribution, probability and statistical significance, correlation, regression and reliability analysis. The reason for this is that EFA, CFA and SEM rely on an application of these statistics and terms linked to such statistics (e.g., univariate normality and variance and co-variance) may be cross-referenced.

Day 1

Morning
: Day 1 of this workshop provides an overview of important aspects, considerations and practical guidelines for conducting EFA. This session also includes Principal Component Analysis (PCA). Participants will be given a step-by-step instruction of how to perform EFA using SPSS and what to look for and how to interpret SPSS outputs. Participants will have an opportunity to perform EFA on their own.

Afternoon:
Examples of language testing and assessment research that employs EFA or PCA will be discussed and critiqued. Suggestions of how to report EFA will be provided. In the remaining hours of Day 1 (approximately 1.5 to 2 hours), CFA and SEM will be introduced. CFA will be compared to EFA. CFA takes a confirmatory approach to common factors that underlie an ability or psychological facet, such as motivation or anxiety. Typically, researchers develop a hypothesised CFA model, which is then tested using an empirical data set. SEM will be discussed in relation to correlation, multiple regression and path analysis. Steps for CFA and how to read CFA models will be presented.

Day 2

Morning: This workshop will first revisit what was covered and arising issues from Day 1. Then, it will focus on the methodological basics of CFA and SEM techniques. Practical guidelines for how to perform CFA and SEM will be discussed. EQS program will be introduced to participants as it will be used to perform CFA and SEM. Participants will have an opportunity to familiarise with EQS program and have some initial diagram drawing practice prior to CFA. Participants will be provided with a step-by-step instruction of how to perform CFA and how to interpret EQS outputs, as well as revise and retest an initial CFA (if required). Participants will be advised on how to edit and prepare a CFA model for research reports.

Afternoon:
In the remaining hours of Day 2, SEM will be discussed further in terms of how to apply it for testing complex models in language testing and assessment. CFAs are typically used as measurement models in SEM. The main focus of this workshop is only on testing and interpreting a relationship among latent and observed variables, and the role of non-random errors on parameter estimates. More sophisticated SEM analyses (e.g., simultaneous group comparisons) will be introduced but cannot be performed in this workshop. Examples of language testing and assessment research that employs CFA and SEM will be discussed and critiqued. Finally, how to report CFA and SEM will be recommended.

Reading

Two texts will be provided about one to two months prior to the workshop.

Prior to Day 1:

Phakiti, A. (2018). Exploratory factor analysis. In A. Phakiti, P. De Costa, L. Plonsky, & S. Starfield (Eds.), The Palgrave handbook of applied linguistics research methodologies. London: Palgrave.

Prior to Day 2:

Phakiti, A. (2018). Confirmatory factor analysis and structural equation modeling. In A. Phakiti, P. De Costa, L. Plonsky, & S. Starfield (Eds.), The Palgrave handbook of applied linguistics research methodologies. London: Palgrave.



Aek Phakiti is an Associate Professor in TESOL and the Postgraduate Coursework Coordinator at the Sydney School of Education and Social Work at the University of Sydney. His research focuses on, language assessment, second language acquisition, and research methods in applied linguistics. He was the Vice President of ALTAANZ (Association of Language Testing and Assessment for Australia and New Zealand, 2015-2017) and Associate Editor of Language Assessment Quarterly (2012-2017). Aek is author and editor of numerous books in applied linguistics, including Language Testing and Assessment: Theory to Practice (Bloomsbury, 2018), The Palgrave Handbook of Applied Linguistics Research Methodology (Palgrave, forthcoming, 2018), Quantitative Methods for Second Language Research: A Problem-solving Approach (Routledge, 2018), Research Methods in Applied Linguistics: A Practical Resource (Bloomsbury, 2015); Experimental Research Methods in Language Learning (Bloomsbury, 2014), Teaching Academic Writing (University of Michigan Press, 2009), and Strategic Competence and EFL Reading Test Performance (Peter Lang, 2007). His research has been published in numerous peer-reviewed journals, including Language Testing, Language Learning, Language Assessment Quarterly, and TESOL Quarterly.

 

Association Management Software Powered by YourMembership  ::  Legal