1

Taking into account the consumers experience in a free sorting task : question and data treatments

P.Faye, P.Courcoux, E.M Qannari - ONIRISA.Giboreau - Institut Paul Bocuse

D.Dubois, LAM - INCAS3

PETRA-INCA3 Workshop – Toulouse Le Mirail, 15/03/12

22

� In sensory sciences, several studies on the impact of expertise onto perception

� Trained panel versus novices (Faye et al., 2004 ; Lelièvre et al., 2008 ; Chollet et al., 2011)

� Professionals versus novices (Giboreau et al., 2001 ; Soufflet et al., 2004 ; Parr et al., 2007 ; Ballester et al., 2008)

� Free sorting task : relevant procedure to study the perception of consumers (Lawless et al., 1995 ; Faye et al., 2006 ; Cartier et al., 2006)

� Statistical treaments� MDS (Borg I. and Groenen P., 1997; Lawless, 1989; Faye et al., 2004)

� Recent works integrating individual responses as CC-Sort (Qannari et al., 2010) or Distatis (Abdi et al, 2007)

Context

3

Research questions

� Hypothesis

� Experience and knowledge of wine have an impact onto the consumer perception of wine glasses

� Properties that structure the perception of the glasses differbetween the most and the least connoisseurs in wine

� Questions

� How to measure and synthetize the knowledge of consumers ?

� How to study the link between the knowledge and the consumerspractices ?

� How to compare product perception between different groups of consumers ?

4

Experimental procedure

� 30 photos of wine glasses (ARC)� Differences in size, volume, design and usage (Champagne, Red or White wine)

� Presented on a grey tablecloth according to a Williams balanced block design

� 209 wine consumers� Well balanced in consumptionfrequency, age, gender, professional activity

� Questionnaire in two parts� Knowledge in wine and wine tasting� Practices and consumptions habits, expertise self evaluation

� Procedure� Perceptual free sorting task� Verbalization task

Q6. Which of the following grapes varieties are used for the AOC wine from Burgundy? (+)1-Pinot noir 5- Chardonnay2- Gamay 6 -Cabernet Sauvigon3 - Grenache 7- No answer 4- Hermitage

(*) Johnson and Bastian(2007), (+) Frost and Noble (2002)

B2. Compared to other people, I know less about wine.(*)

Strongly agree

AgreeNeither agree

nor disagreeDisagree

Strongly disagree

5

Consumers groups based on the wine knowledge level

26 questions

72 dummy variables

Knowledge Rasch model

mh

mh

e

ePhm δβ

δβ

−

−

+=

1With Pmh probability of good answer

βh performance of subject hδm difficulty of the question m.

Knowledge score by

consumer

� Segmentation of consumers, “a priori”, based on their wine knowledge level

� Rasch model (Boomsma et al.,2000)

� Item response theory (IRT), application in psychometrics

� Right/wrong questions (0/1)

� Probability of correct answer modelled according to two factors : subjects ability (“score”) and item difficulty

� Integration of the difficulty of items in deriving the subjects scores : more efficient than summing the right answers

6

Consumers groups characterization

20 declarative questions

26 questions

72 dummy variables

PracticesConsumptions habits, Expertise self evaluation

Knowledge

Correspondance

Multiple Analysis

3 groups of consumers

����Quantitative scores

����Qualitative group variable

1

Characterization of the

groups of consumers

Link between knowledge

and practices

2

3

Illustrative variables

Rasch model

7

Ranked centered Rasch model’s wine knowledge score by consumer

-4

-3

-2

-1

0

1

2

3

4

subjects

Cen

tere

d sc

ore

Consumers groups based on Raschscore

41 non connoisseurs (20%)17% of right answers

41connoisseurs (20%)75 % of right answers

127 intermediates (60%)42% of right answers

1

8

Correspondance Multiple Analysis and Rasch score

2

-1.0 -0.5 0 0.5 1.0

-1.50

-0.75

0

0.75

Axis 2 - 9.13 %

very know ledgeable about w ine

weekly consumption

Consumption exceptionnaly

Consumption occasionaly

consumption more than once a week

consumption 5/6 times a week

daily consumption

no cellar

cellar

10-20 bottles

1-5 bottles

> 20 bottles

5-10 bottles

no wine purchase

very occasional purchase

occasional purchase

frequent purchase

very frequent purchase

no training in oenology

training in oenology

CONNOISSEURS

INTERMEDIATES

NON CONNOISSEURSnot know ledgegeable at all

not know ledegeable about w ine

no opinion on their know ledge about w ine

know ledgeable about w ine

RASCH SCORE

Axis 1 - 11.57 %

-1.0 -0.5 0 0.5 1.0

-1.50

-0.75

0

0.75

Axis 2 - 9.13 %

RASCH SCORE

CONNOISSEURS

INTERMEDIATES

NON CONNOISSEURS

Axis 1 - 11.57 %

9

•Wine professionals (cellarmen and wine grower) and consumers trained in wine tasting,•30-50 years old•Quite frequent wine consumption•Wine cellar•Purchase in specialized wine shops Purchase criteria : price, vintage, grape•Quite confident in their knowledge on wine.

� 3 groups of consumer with different knowledge level

� Differently characterized in terms of practices and habits

� Comparison of non connoisseurs and connoisseurs’ perception

� Correlation between Rasch’s score and the second axis of MCA (R=-0.65)

� Link between practices, consumption habits and knowledge on wine

� Validation of the groups based on knowledge

Consumers groups characterization

- 4

- 3

- 2

-1

0

1

2

3

4

subjects

Cen

tere

d sc

ore

•Students, workers or employees•Occasional consumption of wine •No trained to wine tasting,•Purchases in supermarket •Purchase criteria : Price•Not confident in their wine expertise.

no particular characteristics

Intermediates

ConnoisseursNon connoisseursRanked centered Rasch score

3

10

Non connoisseurs Intermediates Connoisseurs

5

10

15

20

25

Num

ber

of g

roup

s

5.97.17.9

Non connoisseurs Intermediates Connoisseurs

5

10

15

20

25

Num

ber

of g

roup

s

Non connoisseurs Intermediates Connoisseurs

5

10

15

20

25

Num

ber

of g

roup

s

5.97.17.9

Non connoisseurs Intermediates

0

0.2

0.4

0.6

0.8

1

Adj

uste

d R

and

Inde

xConnoisseurs

0.20

0.28

0.36

Non connoisseurs Intermediates

0

0.2

0.4

0.6

0.8

1

Adj

uste

d R

and

Inde

xConnoisseurs

0.20

0.28

0.36

Consumer sorts’variability

� Connoisseurs : less categories, more consensual

� Non connoisseurs : more categories, less consensual

Number of categories Within groups agreement

11

Consumer groups perceptions

( )2/1

2)(2)()()( )(/)()(

−= ∑∑∑∑

h ij

hij

h ij

hij

hij

h XdXdfStress δFree sorting task

N individualdissimilaritymatrices

N consumers

N partitions of P glasses

INDSCAL non metric (Borg I. and Groenen P., 1997)

Group configuration (w)Glass configuration (X)

By coupe of product (i,j)

0 : same group

1: different group

ΣΣΣΣ

[1 4] [2] [3 5]

CNC

2)()( )()( jaiah

ah

hij XXwXd −= ∑with

� Correlation between the axes and the terms generated by each group of consumers

ΣΣΣΣ

Aggregated Matrix

δij

Non connoisseurs

Aggregated Matrix

δij

Connoisseurs

12

stress = 0.013 (3 dimension)

Indscal glass’configuration

30

29

28

27

26

25

24

2322

21

20

19

18

1716

15

14

13

12

1110

9

8

7

6

5

4

3

2

1

-0,8

-0,6

-0,4

-0,2

0,0

0,2

0,4

0,6

-1,0 -0,8 -0,6 -0,4 -0,2 0,0 0,2 0,4 0,6

Axis 1

Axi

s 2

20

28

9 10

11 14

19

25

7

1

4

8

12

1315

1617

1821

22

26

27

3 2

56

23

30

29

24

Champagne glasses

Wine glasses

ClassicalFor great wines

Wine aeration

Legend

ConnoisseursNon connoisseurs

Common terms

ModernReturn of aroma

Round

Angular

Contemporary

12

3

4

5

6

7

8

910

11

1213

14

15

16

17

1819

20

21

22

23

24

25

26

27

28

2930

-0,6

-0,5

-0,4

-0,3

-0,2

-0,1

0,0

0,1

0,2

0,3

0,4

-1,0 -0,8 -0,6 -0,4 -0,2 0,0 0,2 0,4 0,6

Axis 1

Axi

s 3

7

12

5

6

8

9

11

12 13

14

15

16

17

18

20

30

21

24

25

26

27

28

29

4

19

3

10

22 23

Small

Beautiful

Normal

Banal

Old-fashioned

Standard

Not appreciatedFor bar

Elegant

For tasting

For entertaining friends

13

Weight of the axes

� Non connoisseurs weight lessthe axis 3

� 2 underlying dimensions

� Connoisseurs weight the axes 2 and 3 at the samelevel

� 3 underlying dimensions0,2

0,25

0,30

0,35

0,40

0,45

0,50

0,55

axis 1 axis 2 axis 3w

eigh

t

non connoisseurs

connoisseurs

14

Synthesis

�Glasses description�Non connoisseurs : descriptive properties (shape and design) and qualifying

�Connoisseurs : usage properties � inference of usage based on physical characteristics of the glasses

�Glasses configuration�Axis 1 : usage properties for both groups (Champagne / Wine)�Axis 2 : usage properties for connoisseurs, physical properties for non connoisseurs

�Axis 3 � Poor description by non connoisseurs (few terms)� Difference of weights between the both groups of consumers :

small weight for non connoisseurs� More specific to connoisseurs as principally based an « usage »

properties

1515

Conclusions

� Validation of the hypothesis

� Impact of previous experience onto consumers perception

� Different type of properties depending of the consumers

� Statistical treatments

� Complementarity of the statistical treatments

� Rasch model

� Good index to synthetize the score knowledge taking into account the difficulty of the questions

� Multiple Correspondance Analysis

� Characterization of the groups

� Link with the external variables (score)

� Individual Scaling (Indscal)

� Integration of the subjects’variability in a product oriented analysis

� Comparison of groups/subjects’perception based on the same product configuration

� Integration of subjects’previous experience in the consumer studies

16

Abdi H., Valentin D., Chollet S., Chrea C. (2007). Analyzing assessors and products in sorting tasks : DISTATIS, theory and applications. Food Quality and Preference, 18, 627-640.

Ballester J., Patris B., Symoneaux R., Valentin D. (2008). Conceptual vs perceptual wine spaces : Does expertise matter? Food Quality and Preference, 19, 267-276

Borg I., Groenen P. (1997). Modern Multidimensional scaling. Theory and applications. Springer series in statistics, Springer Verlag, New York.

Boomsma A., Van Duijn M.A.J., Snijders T.A.B. (2000). Essays on Item Response Theory. Springer-Verlag, New York.

Cartier R., Rytz A., Lecomte A., Poblete F., Krystlik J., Belin E., Martin N. (2006). Sorting procedure as an alternative to quantitative descriptive analysis to obtain a product sensory map. Food Quality and Preference, 17, 562-571.

Chollet S., Lelièvre M., Abdi H., Valentin D. (2011). Sort and beer: Everything you wanted to know about the sorting task but did not dare to ask. Food Quality and Preference, 22, 507-520.

Faye P., Brémaud D., Durand Daubin M., Courcoux P., Giboreau A., Nicod H. (2004). Perceptive free sorting and verbalization tasks with naive subjects: an alternative to descriptive mapping. Food Quality and Preference, 15, 781-791.

Faye P., Brémaud D., Teillet E., Courcoux P., Giboreau A., Nicod H. (2006). An alternative to external preference mapping. Food Quality and Preference, 17, 604–614

Frost M.B, Noble A.C. (2002). Preliminary study of the effect of knowledge and sensory expertise on liking for red wines. American Journal of Enology and Viticulture, 53:4, 275-284.

Giboreau A., Navarro S., Faye P., Dumortier J. (2001). Sensory evaluation of automotive fabrics :the contribution of categorization tasks and non verbal information to set-up a descriptive method of tactile properties. Food Quality and Preference, 12, 311-322.

References (1/2)

17

Johnson T.E., Bastian S.E.P. (2007). A preliminary study of the relationship between Australian wine

consumers’wine expertise and their wine purchasing and consumption behaviour. Australian Journal

of Grape and Wine Research, 13, 186-197

Lawless H. (1989). Exploration of fragrance categories and ambiguous odors using multidimensional

scaling and cluster analysis. Chemical Senses, 14 (3), 349-360

Lawless H., Sheng N., Knoops Stan S.C.P. (1995). Multidimensional scaling of sorting data applied to

cheese perception. Food Quality and Preference, 6, 91-98.

Lelièvre M., Chollet S., Abdi H., Valentin D. (2008). What is the validity of the sorting task for describing

beers ? A study using trained and untrained assessors. Food Quality and Preference, 19, 697-703.

Parr W., Green J.A., White K.G., Sherlock R.R. (2007). The distinctive flavour of New Zeland Sauvignon

blanc: sensory characterisation by wine professionals. Food Quality and Preference, 18, 849-861.

Qannari E.M., Cariou V., Teillet E., Schlich P.: SORT-CC (2010). A procedure for the statistical treatment of

free sorting. Food Quality and Preference, 21, 302–308

Soufflet I., Calonnier M., Dacremont C. (2004). A comparison between industrial experts’ and novices’

haptic perceptual organization: a tool to identify descriptors of the handle fabrics. Food Quality and

Preference, 15, 689-699.

References (2/2)