Principal speaker
Dr Daniela Vasco
This is one part of a two-part series, aiming to guide you on how to validate surveys. This first part focuses on basic quantitative techniques for validating survey constructs. We start by revisiting the definition of a survey construct (aka scale, factor). We take a more visual approach (than most), by weaving in basic data visualisation methods, including histograms, to add insights on these techniques. For instance, histograms should be the first step in working with any variable, but strangely, are not always presented alongside measures like Cronbach's alpha. Similarly, heatmaps may reveal patterns of correlations within constructs. Next, we examine Cronbach's alpha and its associated item statistics for validating blocks of survey questions that measure one construct, or several. We introduce the concepts of more sophisticated choices for quantitative validation, such as the well-established method of Factor Analysis, which is part of Structural Equation Models (SEM). Throughout, we comment on how methods may be perceived by reviewers, through critical reviews (by Sijtsma and others).
Preparation: We introduce the concept of constructs. It helps to be familiar with at least one construct (aka scale or factor) in your field.
Reading: Please read sections 1, 2, and 6 of Sijtsma, K. (2009). On the use, the misuse, and the very limited usefulness of Cronbach's alpha. Psychometrika, 74(1), 107-120.
Intended audience: Most useful if you are new to survey constructs, and want to start with simple techniques for validating them. Also useful if you wish to understand published evidence, supporting how constructs have evolved in your field over time. Our approach suits conceptual learners or visual thinkers, with equations provided for reference without dwelling on them. These approaches can be implemented using menus in many stats packages; we provide examples coded in R.
Prerequisite: This builds on general training on validation, such as provided by RED workshop Qualitative Validation of Surveys. For those interested in R, you may wish to first attend Induction in R Thinking and Excellent Graphics in R.
Connection to other workshops: Useful preparation for training on SEM, a more sophisticated method for validating survey constructs. Relevant to training on Surveys and/or analysis of Likert-scaled data.
Format: Online workshop, with small group sessions, including optional use of R
Event categories
RSVP
RSVP on or before Tuesday 9 May 2023 09.04 am, by email red@griffith.edu.au , or via https://events.griffith.edu.au/d/h0qnsz/