Summary

Question 1

What is Correspondence Analysis?

Answer 1

• Visualization of textual data
• Used when some descriptive words are more associated with some products than others
• slide 41

Question 2

What’s the main difference between QDA and QDA variants

Answer 2

For QDA, a single list of attributes is used whereas for QDA variants, panelists rate products according to their own set of attributes

Question 3

When is PCA used and when is MFA used?

Answer 3

PCA is used for QDA
MFA is a weighted PCA used for QDA variants (Free Choice Profile + Flash Profile)

Question 4

What is the difference between Sorting and Mapping?

Answer 4

Sorting: products are grouped into homogenous groups/clusters (panelists can include descriptive words they use to group products)
Napping: products are positioned onto a projective map

Question 5

For Sorting, why shouldn’t a similarity matrix + Multi Dimensional Scaling be used? What should be used instead?

Answer 5

In MDS, information is lost about the individual perception of similarity.
Use raw data and Multiple Correspondence Analysis (MCA)

Question 6

What is used for analysis of Sorting data? What about Napping?

Answer 6

Sorting: Multiple Correspondence Analysis (MCA)
Napping: Multiple Factor Analysis (MFA)

Question 7

When is MFA and HMFA used for Napping?

Answer 7

MFA is used for Napping and HMFA is used for Sorted Napping

Question 8

How do you evaluate relationship between sensory attributes and liking?

Answer 8

Complement the results of the PCA with simple linear regressions

Question 9

Why may linear models not be sufficient for liking? What do you do then?

Answer 9

Linear models may not be enough as some attributes show saturation point (point after which increasing/decreasing further the intensity of the attribute affects liking negatively)
Use quadratic model to consider quadratic effects!

Question 10

What is Preference Mapping?

Answer 10

Tool used to combine sensory data (QDA) and hedonic data

Question 11

As a food technologist, which Preference Mapping method should you use?

Answer 11

External (PrefMap): focus is on the sensory profiles of the products (the liking scores are regressed onto a PCA to determine location accepted by a maximum of consumers)

Question 12

Why would you run an ANOVA after plotting the Internal Preference Mapping?

Answer 12

To determine which products are significantly different in terms of consumer liking

Question 12

For JAR, how many penalties are calculated?
What is used to calculate them?

Answer 12

2 penalties (low and high)
One-way ANOVA is used for calculation

Question 12

What are the 2 points of view in IPM?

Answer 12

1. Average ideal product (calculate the average ideal score for each attribute for all products and consumers)
2. Ideal product for the maximum number of consumers (find a point in the product space where the ideal areas of the maximum number of consumers overlap)

Question 13

What are univariate and multivariate analyses?

Answer 13

Determine for which attributes the products are perceived as significant
Univariate (ANOVA)
Multivariate (decat)

Question 13

Why is visualisation of the data not sufficient in Textual Analysis?

Answer 13

We are not looking at the combination of Stimulus x Subject but at the Product: we use a Contingency Table

Question 13

What is the Chi square used for in Textual Data Analysis?

Answer 13

Used to determine whether there is an association between a product and the words used to describe it

Question 14

What is the v-test used for in Textual Data Analysis?

Answer 14

Transforms p-values into more informative statistics
Positive value suggests a word is particularly associated with a product

Question 15

What is Sorted Napping?

Answer 15

slide 57