Sensometrics tutorials

The Sensometric Society in collaboration with the 15th Pangborn Sensory Science Symposium will organize sensometrics tutorials on the morning of 20 August 2023 before the conference.

An introduction to R for sensory and consumer scientists

Mastering the analysis of data from Implicit Association Tasks with R: A Step-by-Step Guide

Gold medal visualisations: Sensory drivers & preference mapping

Why and how to cluster consumers based on their product-related responses

Space on each tutorial will be limited and early booking is advised.



An introduction to R for sensory and consumer scientists

Instructors: Leah Hamilton, UC Davis, USA
Elizabeth Clark, McCormick & Company
Sébastien Lê, Agrocampus Ouest

After the hard work of planning a sensory study and collecting data, researchers still face the sometimes-daunting hurdle of actually analyzing the results. Many of the most powerful, flexible, and free tools for analyzing large multivariate datasets, including those specifically developed for sensory scientists, are in programming languages like R and Python. These can intimidate researchers without programming experience, even those very confident with the underlying statistics. In this tutorial, we aim to demystify statistical programming by teaching participants how to use R for common multivariate analyses. The R skills covered, including fundamental programming concepts and data importing, cleaning, and summarizing, are necessary for most multivariate statistical analyses.

Participants will get hands-on experience coding in R, gaining familiarity with flexible techniques and packages. This tutorial is not intended as a statistics course and will not dive deeply into the technical details of specific analyses, but by the end participants will conduct a Correspondence Analysis and graph the results. This method allows participants to analyze and visually represent ordinal or categorical data about stimuli, such as data from Check-All-That-Apply tests, open comment questions, existing texts, or many other kinds of sensory surveys.

In this tutorial, we will introduce the audience to the R statistical programming environment and the RStudio Interactive Development Environment (IDE) with the aim of developing sufficient basic skills to conduct multivariate analyses (like Correspondence Analysis) on sensory and consumer datasets. We will provide a learning dataset for the analysis—a set of free response comments and overall liking scores from a central location test on berries. We will teach participants how to import, manipulate, and plot data using user-friendly, “tidy” R programming. All resources used in the tutorial are open-source and will remain available to attendees, including an R script covering the full workflow.

At the end of the tutorial, attendees will be able to prepare raw sensory data for common multivariate analyses or visual representations in R.

Duration
3 hours
Audience
Sensory and consumer scientists who are interested in learning basic multivariate analysis using freely available/open-source software (R/RStudio).
Background
Basic understanding of statistics is helpful but not required. We will send out an email to registered participants detailing some basic setup requirements (R/RStudio software installation).
Laptop
This is a coding workshop, and so we ask all participants to bring a laptop with access to RStudio. We will ask for minimal pre-work (installation of R/RStudio).


Mastering the analysis of data from Implicit Association Tasks with R: A Step-by-Step Guide

Instructors: Donato Cereghetti, Firmenich SA, Switzerland
Christelle Porcherot, Firmenich SA, Switzerland
Isabelle Cayeux, Firmenich SA, Switzerland

This tutorial offers a step-by-step guide to help sensory and consumer scientists master data analysis from Implicit Association Tasks (IATs; Greenwald et al., 1998) using R, a free software environment for statistical computing and graphics. The IAT is a computer-based procedure that measures the strength of cognitive associations between concepts. In this procedure, participants must rapidly and accurately sort stimuli by pressing two response keys on a keyboard. The faster and more accurate the sorting, the stronger the association is assumed to be between the concepts. The IAT has gained popularity in sensory and consumer research for measuring implicit attitudes toward products and assessing the relative strength of associations between products, sensory attributes, and feelings.

Participants will get hands-on experience coding in R, gaining familiarity with flexible techniques and packages. This tutorial is not intended as a statistics course and will not dive deeply into the technical details of specific analyses, but by the end participants will conduct a Correspondence Analysis and graph the results. This method allows participants to analyze and visually represent ordinal or categorical data about stimuli, such as data from Check-All-That-Apply tests, open comment questions, existing texts, or many other kinds of sensory surveys.

In this tutorial, we will introduce the audience to the R statistical programming environment and the RStudio Interactive Development Environment (IDE) with the aim of developing sufficient basic skills to conduct multivariate analyses (like Correspondence Analysis) on sensory and consumer datasets. We will provide a learning dataset for the analysis—a set of free response comments and overall liking scores from a central location test on berries. We will teach participants how to import, manipulate, and plot data using user-friendly, “tidy” R programming. All resources used in the tutorial are open-source and will remain available to attendees, including an R script covering the full workflow.

At the end of the tutorial, attendees will be able to prepare raw sensory data for common multivariate analyses or visual representations in RThe tutorial will begin with an introduction to the IAT, highlighting its significant applications in sensory and consumer research. Participants will then complete an IAT to gain hands-on experience with the procedure. The remainder of the tutorial will guide participants through analyzing the data collected. They will learn how to pre-process data, analyze accuracy and reaction times, and use the d-score, a measure introduced by Greenwald and co-workers (2003), to assess the IAT effect. Participants will also learn visualization techniques for IAT data. By the end of the tutorial, participants will gain familiarity with the IAT procedure, have a comprehensive understanding of the R programming environment and its major packages, and acquire the necessary skills to analyze and visualize IAT data.

Duration
3 hours
Audience
Sensory and consumer scientists interested in using IAT in their research activities.
Background
While intermediate knowledge of statistics is helpful, it is not mandatory.
Laptop
Participants are encouraged to bring their personal computers with the latest version of R installed.


Gold Medal Visualisations: Sensory Drivers & Preference Mapping

Instructors: Gemma Hodgson, Qi Statistics, UK
Joshua Brain, Qi Statistics, UK
Anne Hasted, Qi Statistics, UK

Two common methods in the toolkit of modern sensory and consumer scientists are analysis of the sensory drivers of consumer liking and External Preference Mapping. Both methods connect sensory descriptive information with consumer liking/acceptance data and often give the scientists a lot of insight.

So it’s no surprise to see both analyses being prevalent in modern software packages, yet in our experience, there is often a lack of clear, easy to understand graphical information that aids the user in drawing robust conclusions. We often hear from clients that there is a difficulty in quickly attaining good looking (i.e. aesthetically pleasant) graphics for presentations and reports, that are also easy to set up and scientifically sound. A picture, after all, paints a thousand words!

In the tutorial, we’ll first explain the statistics behind the techniques (interactively!) and discuss some of the common pitfalls, before moving on to visualise the data using two of our in house R Shiny apps designed specifically for this purpose. We want to ensure the science is driving the software, not the other way round!

These apps are written in the R language but require no coding ability for the users. As an attendee, we’ll give you a free six-month trial license of both apps so that you can demo visualisations of your own data after Pangborn.

If you’re confused by either of these two common methods and instead want to work from a reliable roadmap, or you want to be able to win gold medals for your statistical communication skills, then this hands on tutorial is for you.

No prior knowledge of either technique is required, although it may be of benefit.

Duration
3 hours
Audience
Sensory and consumer scientists.
Background
While intermediate knowledge of statistics is helpful, it is not mandatory.
Laptop
Participants need to bring their personal computers


Why and how to cluster consumers based on their product-related responses

Instructors: Fabien Llobell, Lumivero, XLSTAT, France
Evelyne Vigneau, Oniris StatSC, France
John Castura, Compusense Inc., Canada

Consumers are diverse, not only in their demographic and socio-economic characteristics, but also in their perceptions, emotions, values, and attitudes.

It is rare that a product will appeal to “the whole population”, but if a product appeals only to the idiosyncratic tastes of few individuals then it will not be commercially successful. A better product strategy is to target meaningfully large subgroups of consumers who like and want to buy the product. One way to find these subgroups is to cluster consumers based on their product opinions or other relevant data.

This tutorial looks at different ways that consumers can be clustered. We focus on cluster analysis based on four types of product-related data: hedonic data, just-about-right data, projective mapping data, sensory and emotion check-all-that apply (CATA) data.

We will give an overview of how hierarchical agglomerative cluster analysis works, how it differs from k-means analysis, and the complementary of both approaches. Then we will discuss some newer approaches to cluster analysis, including clustering variables around latent variables (CLV), CLUSTATIS, CLUSCATA, b-cluster to analysis, and other approaches that cluster consumers based on more than one type of sensory data. Demos will be provided in XLSTAT and in R. For each method, guidance will be provided for determining the number of clusters.

At the end of this workshop, attendees will have a better understanding of why to cluster consumers based on their product-related responses and how to do it in XLSTAT and R.

Duration
3 hours
Audience
Sensory scientists and statisticians interested in cluster analysis of consumer test data
Background
Basic understanding of statistics is advantageous but not required
Laptop
It is not a coding workshop, so a laptop is not required. Examples will be shown in XLSTAT and R. Instructions for tutorial participants who want to follow the demos on their own laptops can be found here: https://sensometric.org/cluster_tutorial_2023(link active May 15 to September 15, 2023)

View Programme
Supporting Publication

Organisers
 
Sponsors and Exhibitors