INTERNATIONAL ISO

STANDARD 11035
First edition
1994-12-15

Sensory analysis - Identification and

) EW

-
selection of descriptors for establishing a

96
83
.ai VI

-
sensory Profile by a multidimensional
f3
4 d
eh E

99 d
.it R

-1 3 2
approach
ds P

35 / 93
ar RD

10 st
-1 s/si
nd DA

iso d
7 5 a n d:
/ r
7 6 da
76 /st ar
sta AN

bd l o g nd

Analyse sensorielle - Recherche et selection de descripteurs pour

8 a ta ta
ST

20 c a l s

Maboration dun Profil sensoriel, par approche multidimensionnelle

7 - a i / ul
eh

dc h. F
(
iT

-8 i te
9b s.
44 ar d
nd
sta /
s:/
tp
ht

Reference number
PS0 11035: 1994(E)
ISO 11035:1994(E)

Foreword

ISO (the International Organization for Standardization) is a worldwide

federation of national Standards bodies (ISO member bodies). The work
of preparing International Standards is normally carried out through ISO
technical committees. Esch member body interested in a subject for
which a technical committee has been established has the right to be
represented on that committee. International organizations, governmental
and non-governmental, in liaison with ISO, also take part in the work. ISO
collaborates closely with the International Electrotechnical Commission
(IEC) on all matters of electrotechnical standardization.
) EW

-
96
83
Draft International Standards adopted by the technical committees are
.ai VI

-
f3
4 d
circulated to the member bodies for voting. Publication as an International
eh E

99 d
.it R

-1 3 2
Standard requires approval by at least 75 % of the member bodies casting
ds P

35 / 93

a vote.
ar RD

10 st
-1 s/si
nd DA

iso d
7 5 a n d:
/ r

International Standard ISO 11035 was prepared by Technical Committee

7 6 da
76 /st ar
sta AN

bd l o g nd

lSO/TC 34, Agricultural food products, Subcommittee SC 12, Sensory

8 a ta ta
ST

analysis.
20 c a l s
7 - a i / ul
eh

dc h. F
(
iT

Annexes A, B and C of this International Standard are for information only.

-8 i te
9b s.
44 ar d
nd
sta /
s:/
tp
ht

0 ISO 1994
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced
or utilized in any form or by any means, electronie or mechanical, including photocopying and
microfilm, without Permission in writing from the publisher.
International Organization for Standardization
Case Postale 56 l CH-1 211 Geneve 20 l Switzerland
Printed in Switzerland
 0 ISO ISO 11035:1994(E)

Introduction
An Overall sensory Profile describes the sensoty attributes of a product.

A “product” is characterized by several Parameters. Some of these have

a Single dimension (e.g. the diameter of a ball, the weight of a sachet, etc.)
and others have several dimensions (e.g. the shape of a product, the tex-
ture of meat, etc.); a sensory Profile requires monodimensional quantities
to lead to a measurement of intensity.

Consequently, the evaluation of a complex sensory quantity needs meth-

) EW

6-
odology which is founded on identification of appropriate descriptors. This

39
-8
work is given to a trained Panel who describe their perceptions both
.ai VI

4 df3
qualitatively (nature of the Stimulus) and quantitatively (intensity of each
eh E

99 d
.it R

-1 3 2
stimulus).
ds P

35 / 93
ar RD

10 st
-1 s/si

If the aim is to appreciate all the attributes, an “Overall sensory Profile” is

nd DA

iso d
7 5 a n d:
/ r

built; if it concerns the evaluation of only flavour, odour, texture or ap-

7 6 da
76 /st ar
sta AN

bd l o g nd

pearance, a “partial Sensor-y Profile” is then elaborated.

8 a ta ta
ST

20 c a l s
7 - a i / ul

In both cases, the choice of descriptors is the preliminary Phase which

eh

dc h. F
(

determines the quality of the sensory Profile.

iT

-8 i te

There are several ways to establish a list of descriptors, for example:

9b s.
44 ar d
nd

a) leave each assessor to choose and use, for the final Profile, his/her
sta

own descriptors (free-choice Profile);

/
s:/
tp
ht

b) use descriptors common to all the members of the Panel, either

1) by suggesting existing descriptors, on condition that the relevante
of the descriptors for the product has been checked and that the
assessors have been trained on these descriptors (generally with
the help of reference products), or

2) by the creation of descriptors by all the membe rs of the Panel after

individual or COIlective work.

This tan be done by a consensus method (See, for example, ISO 6564)
or by the method described in this International Standard, which is char-
acterized by the elaboration of a list of descriptors convenient for the
product studied and which guarantees, as far as possible, the exhaustivity
of this list, and allows verification of the relevante and independence of
each descriptor, and if they are monodimensional.
NOTES

1 Drawing up a Sensor-y Profile of a product is a complex procedure and the user

of this international Standard needs to know that although this method gives sat-
isfactory results, it requires a large investment in preparation time, calculation and
number of training Sessions.
ISO 11035:1994(E) 0 ISO

2 This International Standard requires a basic knowledge of multidimensional

analysis [in particular, a minimal knowledge of Principal Components Analysis
(PCA) and Hierarchie Ascending Classification (HAC)].

) EW

6-
39
-8
.ai VI

4 df3
eh E

99 d
.it R

-1 3 2
ds P

35 / 93
ar RD

10 st
-1 s/si
nd DA

iso d
7 5 a n d:
/ r
7 6 da
76 /st ar
sta AN

bd l o g nd
8 a ta ta
ST

20 c a l s
7 - a i / ul
eh

dc h. F
(
iT

-8 i te
9b s.
44 ar d
nd
sta /
s:/
tp
ht
INTERNATIONAL STANDARD 0 ISO ISO 11035:1994(E)

Sensory analysis - Identification and selection of

descriptors for establishing a sensory Profile by a
multidimensional approach

) EW

6-
39
1 Scope tell the nature of the differentes in terms of sen-

-8
.ai VI

4 df3
sory perception.
eh E

99 d
.it R

-1 3 2
This International Standard describes a method for
ds P

35 / 93
ar RD

identifying and selecting descriptors which tan then

10 st
-1 s/si

be used for drawing up the sensory Profile of a prod- 2 Normative references

nd DA

iso d
7 5 a n d:
/ r
7 6 da
76 /st ar

uct.
sta AN

bd l o g nd

The following Standards contain provisions which,

8 a ta ta
ST

20 c a l s

lt describes the different stages in the process for through reference in this text, constitute provisions
7 - a i / ul
eh

dc h. F
(

setting up tests through which a complete description of this International Standard. At the time of publi-
iT

of the sensory attributes of a product tan be obtained: cation, the editions indicated were valid. All Standards
-8 i te

are subject to revision, and Parties to agreements

9b s.
44 ar d

- from a qualitative Point of view, by defining by based on this International Standard are encouraged
nd

means of descriptors all the perceptions for dis- to investigate the possibility of applying the most re-
sta

tinguishing one product from others of the same cent editions of the Standards indicated below.
/
s:/
tp

type; Members of IEC and ISO maintain registers of cur-

ht

rently valid International Standards.

- from a quantitative Point of view, by evaluating the
intensity of each descriptor (stronger or weaker ISO 4121:1987, Sensory analysis - Methodology -
impression analysed by an assessor on one el- Evaluation of food products by methods using
ement of the Overall perception). scales.

The so-called “Sensor-y Profile” method tan be used: ISO 5492:1992, Sensor-y analysis - Vocabulary.

- to define a production Standard; identification of ISO 6564:1985, Sensory analysis - Methodology -

the nature of the differentes makes it easier to Flavour Profile methods.
grasp the issue;
ISO 6658: 1985, Sensor-y analysis - Methodology -
- to improve or develop products; General guidance.

- to study the influence of the ageing of products ISO 8586-1 :1993, Sensory analysis - General guid-
and also of the conditions of storage and preser- ante for the selection, training and monitoring of
vation; it is thus possible to determine those assessors - Part 1: Selected assessors.
characteristics which vary and to what extent;
ISO 8586-2: 1994, Sensor-y analysis - Genera/ guid-
- to compare a product with those of the Same type ante for the selection, training and monitoring of
already on the market; it is therefore possible to assessors - Part 2: Experts.
ISO 11035:1994(E) 0 ISO

ISO 8589: 1988, Sensor-y analysis - General guidance Examples: flavour Profile, texture Profile, appearance
for the design of test rooms. Profile and odour Profile.

3 Definitions 4 Principle
For the purposes of this International Standard, the
Identification and selection of a set of relevant de-
definitions given in ISO 5492, and the following defi-
scriptors giving maximum information on the Sensor-y
nitions apply.
attributes of the product under analysis, in Order to
establish a Sensor-y Profile.
3.1 descriptor: A term referring the assessor to an
element of the perception of the product. The prop- The various stages in the methodology are given be-
erties of the descriptor (relevante of the product, low (see figure 1):
monodimensional) shall be such that it tan be used
to produce an evaluation on a scale of intensity (sweet - training of the Panel,
flavour of sucrose, for example).
- preparation of a list of descriptive terms,
3.2 Overall sensory Profile: The use of descriptive
terms in evaluating the Sensor-y attributes of a Sample - reduction of the list of terms,
and the intensity of each attribute.
- choice of reference products,
3.3 partial sensory Profile: The use of descriptive
) EW

6-
terms in evaluating the sensory attributes of a Sample - training,

39
-8
and the intensity of each attribute by one or by several
.ai VI

4 df3
eh E

sensory inputs. - use of the Profile. 99 d

.it R

-1 3 2
ds P

35 / 93
ar RD

10 st
-1 s/si
nd DA

iso d
7 5 a n d:
/ r
7 6 da
76 /st ar
sta AN

bd l o g nd

Meeting of a Panel - Choice of samples

8 a ta ta
ST

20 c a l s
7 - a i / ul
eh

dc h. F
(
iT

Identification of descriptive terms

-8 i te

I
9b s.
44 ar d
nd
sta

List of descriptive
/
s:/

terms
tp
ht

i
Choice of referencek)
by descriptor

Repeatability tests

Figure 1 - Stages in the identification and selection of descriptors for establishing a sensory Profile
0 ISO ISO 11035:1994(E)

5 General test requirements pacity of each assessor to repeat the evaluation shall
be verif ied.
The general directives concerning the methodology
of sensory analysis described in ISO 6658 shall be For conventional methods of selection, training and
followed. testing, consult ISO 8586-1 for assessors and
ISO 8586-2 for experts. However, these methods
shall be adapted to the Profile concerned (Overall pro-
5.1 Test room file, flavour Profile, texture Profile, etc.).

See ISO 8589 for the characteristics of the room in lt is important that the assessors regularly attend
which the tests are to be performed. sensory analysis training Sessions. Regular attend-
ante at the classes also gives an indication of the
5.2 Apparatus motivation of the assessors.

The apparatus should be selected by the Panel leader

6.3 Role of Panel leader
according to the nature of the product or products to
be analysed, the number of samples, etc., and shall The role of the Panel leader is of utmost importante
have no influence on the results. in the selection of the Panel, its training and main-
taining the motivation of the assessors.
If Standard apparatus meets the requirements of the
test, it shall be used. The Panel leader should also be able to conduct and
) EW

6-
coordinate a work group, taking into account the

39
-8
.ai VI

6 Methodoloy opinions of each participant and harmonizing them.

4 df3
eh E

99 d
.it R

-1 3 2
ds P

35 / 93
ar RD

6.1 Number of assessors 6.4 Identification of the largest possible

10 st
-1 s/si
nd DA

number of descriptive terms

iso d
7 5 a n d:
/ r
7 6 da
76 /st ar

A minimum number of six assessors is required in

sta AN

bd l o g nd

Order to allow for individual differentes.

8 a ta ta

The aim of this Stage in the identification of descrip-

ST

20 c a l s

tors is not to neglect any aspect of the product and

7 - a i / ul
eh

dc h. F

In Order to have a group of six to ten assessors per-

(

to avoid bias due to the influence of an individual on

iT

manently available, it is recommended that two or

the group. lt also serves as an apprenticeship for
-8 i te

three times this number be trained and coached. . identifying and evaluating the nature of the differ-
9b s.
44 r d

ences.
da
an

6.2 Training of the Panel

/st
s:/

6.4.1 Selection of products

tp

The quality of the sensory Profile depends on the

ht

quality of the Panel. lt is important that the assessors A series of similar products should be selected (three
be correctly trained and coached before selection ac- or four per Session) which, when tasted, will enable
cording to their ability to recognize and evaluate the the assessors to distinguish all the possible qualitative
Stimuli pertaining to the types of products for which differentes which tan be detected in the product for
a sensory Profile is to be established. which the Profile is to be drawn up.
When selecting the assessors, their creativity and To obtain these differentes between products, fac-
their ability to express themselves are the first criteria tors such as ageing, proportions of ingredients during
to be taken into account. In fact, it is important for the manufacture, or the duration of cer-tain stages in
assessors to have an extensive vocabulary and to be manufacture tan be varied. Another possibility would
at ease in the use of this vocabulary in Order to prod- be to examine similar rival products.
uce a simple and easily understood description of the
product to facilitate communication between the vari-
ous Parties concerned. 6.4.2 Generation of descriptors in tasting booths

Training is carried out on the types of products for In Order for the assessors to achieve the necessary
which the Profile is to be established or on their concentration for individual identification of the de-
component Parts (identification of the components scriptive terms, without trying to make do with the
and evaluation of the perceived intensities). All these terms given by others, the assessors should each be
preliminary Sessions contribute to the training of the in tasting booths (see ISO 8589) under the usual
assessors. When this training period is over, the ca- conditions for sensory evaluation of products.
ISO 11035:1994(E) 0 ISO

In the initial Sessions (at least four), the assessors are - quantitative terms, such as too much, too little,
presented with the product for which the Profile is to strong, wea k, etc.;
be made as well as the series of selected samples.
The assessors are asked to generate the maximum - terms describing the product in its own terms,
number of terms (descriptors) to describe all the sen- such as “bready taste” for bread (except for cer-
sations produced by these products, whether visual, tain cases of prepared or converted products
tactile, olfactory or gustatory (in the case of an Overall where the odour or the aroma of a constituent re-
Profile) and to note down on the form provided all the mains, for example the aroma of Vanillin or vanilla
terms which occur to them. (See annex A for a in vanilla ice cream);
specimen form.)
- irrelevant terms such as “acid” when describing
an odour.
6.4.3 Group discussion
The Panel leader explains to the assessors why these
The assessors then discuss in a group and compare terms are considered to be unsuitable for the in-
their perceptions under the guidance of the Panel tended purpose, which is to identify and describe the
leader who should encourage each of them to analyse nature of the perception and the combinations of dif-
the different components of the perception of the ferences between products.
products.
6.6 First reduction of the number of
These components shall be expressed by an appro- descriptors
priate descriptor (e.g. bitterness, acidity, smoky fla-
) EW

-
96
83
vour, etc.). The descriptors resulting from the preliminary sorting
.ai VI

-
f3
4 d
arc generally too great in number; in further tasting
eh E

The identification of terms ceases once the assessors 99 d

.it R

-1 3 2
Sessions, terms are eliminated which do not appear
ds P

35 / 93

have exhausted their vocabulary on the product. The

ar RD

10 st

suitable for describing or differentiating products from

-1 s/si

group usually manages, in several Sessions, to gen-

nd DA

a Sensor-y Point of view.

iso d
7 5 a n d:
/ r

erate without difficulty more than a hundred different

7 6 da
76 /st ar
sta AN

bd l o g nd

descriptive terms. In order to make this reduction, it is necessary to

8 a ta ta
ST

20 c a l s

make Sure that the assessors have well understood

7 - a i / ul

All the descriptive terms are then collected together

eh

dc h. F
(

the meaning of each descriptor.

iT

at this Stage, including Synonyms.

The assessors are presented with different variants
-8 i te

See an example of an application to sandwich bread

9b s.

of the product and are asked, for each of the de-

44 ar d

in annex B.
nd

scriptors used, to judge the perceived intensity by al-

sta

locating it a mark on a scale from 0 to 5, specifying

/
s:/

that zero (0) is equivalent to an absence of perception

tp
ht

6.5 Preliminary sorting of descriptors for the property considered.

This sorting is performed during the initial Sessions NOTE 3 It is possi ble to check if ther .e are differentes or
by the Panel leader during group discussions and in not in the p erceptio n of the pro ducts by using triangular
the presence of the samples. tests.

The following are gradually eliminated from the Figure2 presents the type of scale which is usuable
discussion: for each descriptor in the first reduction.

- hedonistic terms, such as pleasant, fine, Consult ISO 4121 for methods using scales and cat-
appetizing, good, etc.; egones.

Not perceived Weak Rather weak Average Rather strong Strong

I I I I I I 1
Descriptor 0 1 2 3 4 5

Figure 2 - Scale of intensity for the first reduction of descriptors

4
0 ISO ISO 11035:1994(E)

In Order to reduce at this Stage the number of de- a low perceived intensity but which are mentioned
scriptors, they are initially classified by the geometric often (see ref.[ll in annex C).
mean M, which is the Square root of the product of
the frequency, F, and the relative intensity, 1, of each The classification of descriptors according to the size
descriptor: of this mean makes it possible to eliminate a number
of descriptors whose geometrical means are relatively
M = ,/FmI low.

EXAMPLE
where

is the number of times the descriptor is Tables 1 and 2 give the sequence of calculations in
F
mentioned over the total number of times the case of 5 products tested with 9 descriptors by
it is possible to mention that descriptor, 18 assessors.
expressed as a percentage; and The total number of times each descriptor is men-
is the sum of the intensities given by the tioned is, in this case, 90 (5 products by 18
I
whole Panel for a descriptor over the assessors).
maximum possible intensity for this de- The total possible intensity per descriptor is 450
scriptor, expressed as a percentage. (maximum of 5 on the scale of intensity for 5 products
with 18 assessors = 5 x 5 x 18).
This method of calculation makes it possible to take
into account, in the same way, descriptors which are In this example, the calculations (table3) show that
) EW

6-
39
rarely mentioned but which are very important in the descriptors classed in the lowest two positions

-8
.ai VI

4 df3
terms of the perceived intensity and descriptors with (D3 and 08) could be eliminated.
eh E

99 d
.it R

-1 3 2
ds P

35 / 93
ar RD

10 st

1 - Calculation of the frequency, F, of mentioning each descriptor

-1 s/si

Table
nd DA

iso d
7 5 a n d:
/ r
7 6 da
76 /st ar

Descriptors
sta AN

bd l o g nd

Product
8 a ta ta

Dl D2 D3 D4 D5 D6 D7 08 D9
ST

20 c a l s
7 - a i / ul
eh

Pl 12 8 0 9 8 17 17 1 12
dc h. F
(
iT

P2 17 17 0 15 16 9 4 1 16
P3 2 12 0 4 8 0 1 1 3
-8 i te
9b s.

P4 7 1 3 5 8 14 14 1 4
44 ar d

P5 1 9 0 6 14 2 3 2 3
nd
sta

Number of mentions 39 47 3 39 54 42 39 6 38
/
s:/
tp

F = Numberltotal number (%) 43,3 52,2 3,3 43,3 60,O 46,7 43,3 6,7 42,2
ht

Table 2 - Calculation of the relative intensity, 1, of each descriptor

Descriptors
Product
Dl D2 D3 D4 D5 D6 D7 08 D9
Pl 69 43 0 16 27 64 31 5 19
P2 43 33 0 30 52 44 9 3 33
P3 3 25 0 13 42 2 2 1 11
P4 36 8 10 6 8 37 50 5 5
P5 4 19 0 30 78 5 11 11 7
Perceived intensity per descriptor 155 128 10 95 207 152 103 25 75
I = intensityltotal intensity (%) 34,4 28,4 2,2 21,l 46,0 33,8 22,9 5,6 16,7
ISO 11035:1994(E) 0 ISO

Table 3 - Classification of descriptors by importante (by geometric mean)

Descriptors
Parameter
Dl D2 03 D4 D5 D6 D7 D8 D9
I 0,344 0,284 0,022 0,211 0,460 0,338 0,229 0,056 0,167
F 0,433 0,522 0,033 0,433 0,600 0,467 0,433 0,067 0,422
iV 0,386 0,385 0,027 0,302 0,525 0,397 0,315 0,061 0,265
As a percentage 38,6 38,5 2,7 30,2 52,5 39,7 31,5 6,l 26,5
Classification of descriptors 3 4 9 6 1 2 5 8 7

See also an example of the application to sandwich representation space and by masking the differentes.
bread in annex B (tables B.2 to B.4). These should be eliminated from the search for de-
scriptors.

6.7 Second reduction in the number of Care should be taken that there is a good distribution
descriptors by multidimensional analysis and of the products in the four quadrants of the projection
variance analysis planes. ) EW

6-
The second reduction makes it possible to group to-

39
-8
gether Synonyms (correlated positively) or antonyms
.ai VI

6.7.2 Rules for reducing the number of

4 df3
eh E

(correlated negatively) and eliminate descriptors dekriptors 99 d

.it R

-1 3 2
which make very little contribution to showing differ-
ds P

35 / 93
ar RD

10 st

ences between the products tested in a sensory pro- Three rules have to be applied in Order to reduce the
-1 s/si
nd DA

file .
iso d
7 5 a n d:

number of terms.
/ r
7 6 da
76 /st ar
sta AN

bd l o g nd
8 a ta ta
ST

a) Delete terms which do not best characterize the

20 c a l s

6.7.1 General principles

7 - a i / ul

product space and/or the differentes between

eh

dc h. F
(
iT

Multidimensional analysis allows evaluation of the products. If a descriptor maintains a constant

relative importante and the contribution of descriptors value for all of the products evaluated, it may be
-8 i te
9b s.

in differentiating products since it allows all the prod- kept provided that it is always desirable to find
44 ar d

again that characteristic at a given intensity (e.g.

nd

ucts to be visualized at the Same time as well as the

sta

correlations between the descriptors. By identifying in quality control).

/
s:/

the closeness of the descriptors and the weight at-

tp

In a multidimensional analysis, a weak contribu-

ht

tached to each descriptor for distinguishing between

products, it is possible to eliminate some of them or tion of the descriptors in relation to the relevant
to group them together. axes is interpreted as a term which is only slightly
appropriate or a descriptor which maintains a
The interpretation of a Principal Components Analysis constant value.
(PCA) thus permits the following to be noted:
lt is best to examine the contribution of each de-
a) the relative importante of each axis (linear com- scriptor on all the axes so as not to eliminate a
bination of the descriptors) in relation to the total descriptor which is independent of the others and
cloud (% inertia): which would constitute by itself an axis which is
relevant but of weak inertia.
b) the contribution of each descriptor in relation to
the main axes (relative contribution); b) Delete synonymous terms, since they will appear
close together in the PCA and in the associated
c) the “quality” of the representation of each el- Ascending Hierarchical Classification (AHC) (see
ement on the Chosen plane (cosine Square), which refs. [2]-[6]).
indicates whether or not the element is close to
the plane of projection of the product/descriptor c) Replace by a Single descriptor, to be defined with
space. the assessors, two terms which are sensory op-
posites (correlation close to -1 and descriptors
Certain products tan make too great a contribution to opposite in relation to the centre of gravity of the
the inertia by overpowering the others in the Chosen descriptor space).

6
0 ISO ISO 11035:1994(E)

Before eliminating a strongly correlated term or a term how the assessors use the scale. The variance analysis en-
anti-correlated to another, it is necessary to ensure sures the similarity of the mean of the values of each
that it really does belong to the Same sensory product for one descriptor. The equality of the means given
by assessor and by product tan be tested by applying a
continuum. For example, “sweet” and “acid” in the
two-factor variance analysis (“product” factor and
case of ripening of a fruit do not belong to the same “assessor” factor).
sensory continuum, even if a diminution of acidity
may be related to an increase in sweetness. If only differentes between products are of interest,
descriptor Dl tan be eliminated (provided that sam-
In practice, the corresponding scores of synonymous
ples have been taken of all the products within the
descriptors are added up if a second processing of the
range in Order to make Sure that this descriptor will
data table is desired. Consequently, a Single term will
not play any part in the differentes).
be kept which has to be redefined with all the
assessors. The advantage of the second processing In the case of quality control, for example, it would
is that it tan check that no information has been lost. be important to keep Dl in Order to be Sure of finding
A “product” space or a “descriptor” space largely again this characteristic with the same intensity in all
identical to that resulting from the first processing of the products.
should be obtained.
Descriptors D2 and D3 are always in the Same ratio
The final number of descriptors shall be at most 15 in of intensity; they are therefore closely correlated.
Order to obtain an operational Profile, i.e. one enabling
an assessor to evaluate several samples of products 6.8 Choice of reference products or
) EW

6-
in less than one hour. If the number of descriptors is
substances

39
-8
too high, what one hopes to gain in finesse is lost in
.ai VI

4 df3
the accuracy of the measurement.
eh E

99 d
If the number of descriptors is reduced, a definition
.it R

-1 3 2
ds P

35 / 93

of each descriptor shall be given, understood by all

EXAMPLE
ar RD

10 st

the assessors, which shall be kept for easy reference.

-1 s/si
nd DA

iso d
7 5 a n d:

A suitable stable reference product shall also be as-

/ r

A simplified example given in table4 l) helps to illus-

7 6 da
76 /st ar
sta AN

bd l o g nd

trate two important Points in the reduction of terms signed to each of these descriptors.
8 a ta ta
ST

20 c a l s

from the data table by grouping the intensities given

7 - a i / ul

A pure Chemical is not necessarily a relevant refer-

eh

dc h. F

by the Panel for each product and each descriptor.

(

ence substance for defining a descriptor. For exam-

iT

In this example it tan be seen that descriptor Dl does ple, the assessor must be able to extract from a
-8 i te

complex Sensation the Stimulus which echoes the

9b s.

not discriminate between products since the mean

44 ar d

of the perceived intensities is similar (assuming that descriptor (e.g. bitterness of a coffee, astringency of
nd
sta

all the scales are identical). a fruit, etc.).

/
s:/
tp

lt is important to use reference products which are

ht

NOTE 4 The power of discrimination of each descriptor

should be verified for the different products by a ranking stable or reproducible in time. The choice of these
test and/or a variance analysis. The ranking test permits reference products tan be difficult, as it is a question
verification of the discrimination of the products no matter of reconciling appropriateness and ease of use.

Table 4 - Example of a table of results for the whole Panel

Descl ptors
Dl D2 D3 Di sm. 4,
11 60 20
13 45 15
12 75 25
12 30 10
11 6 2
13 15 5

1) This example is not connected to tables 1 to 3.