SIST ISO 6658:2018
Sensory analysis -- Methodology -- General guidance
Sensory analysis -- Methodology -- General guidance
ISO 6658:2017 gives general guidance on the use of sensory analysis. It describes tests for the examination of foods and other products by sensory analysis, and includes some general information on the techniques to be used if statistical analysis of the results is required.
Generally these tests are intended only for objective sensory analysis. However, if a test can be used for determining preference in hedonic test, this is indicated.
A hedonic test aims to determine the acceptability of the products and/or to determine preferences among two or more products by the specified consumer population. The methods are effective for determining whether a perceptible preference exists (difference in degree of liking), or whether no perceptible preference exists (paired similarity test). General guidance for hedonic tests is given in ISO 11136.
Analyse sensorielle -- Méthodologie -- Lignes directrices générales
ISO 6658:2017 fournit des lignes directrices générales sur la méthodologie de l'analyse sensorielle. Il décrit des essais destinés à l'examen des produits alimentaires et d'autres produits par analyse sensorielle et donne des informations générales relatives aux techniques à utiliser si une analyse statistique des résultats est requise.
En général, ces essais sont uniquement destinés à une analyse sensorielle objective. Toutefois, s'il est possible d'utiliser un essai pour déterminer la préférence au cours d'un essai hédonique, cela est précisé.
Un essai hédonique vise à déterminer l'acceptabilité des produits et/ou à déterminer les préférences d'une population de consommateurs spécifique entre deux produits ou plus. Les méthodes sont efficaces pour déterminer s'il existe une préférence perceptible (différence dans le degré d'appréciation), ou s'il n'existe aucune préférence perceptible (essai de similitude par paires). Les lignes directrices générales pour les essais hédoniques sont données dans l'ISO 11136.
Senzorična analiza - Metodologija - Splošne smernice (ISO 6658:2017)
Ta dokument podaja splošne smernice za uporabo senzorične analize. Opisuje preskuse za pregledovanje živil in drugih proizvodov z uporabo senzorične analize ter vključuje nekaj splošnih informacij o tehnikah, ki se uporabljajo, če je treba izvesti statistično analizo rezultatov.
Ti preskusi so v splošnem namenjeni samo za objektivno senzorično analizo. Vendar če je preskus mogoče uporabiti za določevanje prednostnih izbir v hedoničnem preskusu, je to navedeno.
Namen hedoničnega preskusa je določiti sprejemljivost proizvodov in/ali prednostne izbire
med dvema ali več proizvodi s pomočjo določene skupine potrošnikov. Metode so učinkovite pri
določevanju, ali zaznavna prednostna izbira obstaja (razlika v stopnji všečnosti) ali
zaznavna prednostna izbira ne obstaja (preskus parne podobnosti). Splošno navodilo za hedonične preskuse je podano v
standardu ISO 11136.
General Information
Relations
Standards Content (Sample)
SLOVENSKI STANDARD
01-oktober-2018
1DGRPHãþD
SIST ISO 6658:2011
6HQ]RULþQDDQDOL]D0HWRGRORJLMD6SORãQHVPHUQLFH ,62
Sensory analysis -- Methodology -- General guidance
Analyse sensorielle -- Méthodologie -- Lignes directrices générales
Ta slovenski standard je istoveten z: ISO 6658:2017
ICS:
67.240 6HQ]RULþQDDQDOL]D Sensory analysis
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.
INTERNATIONAL ISO
STANDARD 6658
Third edition
2017-07
Sensory analysis — Methodology —
General guidance
Analyse sensorielle — Méthodologie — Lignes directrices générales
Reference number
©
ISO 2017
© ISO 2017, Published in Switzerland
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form
or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior
written permission. Permission can be requested from either ISO at the address below or ISO’s member body in the country of
the requester.
ISO copyright office
Ch. de Blandonnet 8 • CP 401
CH-1214 Vernier, Geneva, Switzerland
Tel. +41 22 749 01 11
Fax +41 22 749 09 47
copyright@iso.org
www.iso.org
ii © ISO 2017 – All rights reserved
Contents Page
Foreword .v
Introduction .vi
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 General requirements . 1
4.1 Basic information . 1
4.2 Statement of objectives . 2
4.3 Choice of test . 2
4.4 Choosing and training assessors . 3
4.5 Material to be tested . . 3
4.6 Test room . 4
4.7 Planning and conduct of the test . 4
5 Test methods . 5
5.1 General . 5
5.2 Discrimination tests . 6
5.2.1 General. 6
5.2.2 Paired comparison test. 6
5.2.3 Triangle test . 7
5.2.4 Duo-trio test . 8
5.2.5 Two-out-of-five test . 8
5.2.6 “A — not A” test . 9
5.2.7 Tetrad test . 9
5.3 Use of scales and categories .10
5.3.1 General considerations .10
5.3.2 Interpretation of results .10
5.3.3 Classification .11
5.3.4 Grading .11
5.3.5 Ranking .11
5.3.6 Rating and scoring .12
5.4 Descriptive tests.12
5.4.1 General.12
5.4.2 Qualitative sensory profile .13
5.4.3 Quantitative descriptive profile .13
5.4.4 Consensus profile .15
5.4.5 Free-choice profile.15
5.4.6 Flash profile .16
5.4.7 Deviation from reference profile.16
5.4.8 Temporal dominance of sensations (TDS).17
6 Analysis of results.17
6.1 General .17
6.2 Discrimination tests .17
6.2.1 General.17
6.2.2 Paired comparison test (see ISO 5495) .18
6.2.3 Triangle test (see ISO 4120) .19
6.2.4 Duo-trio test (see ISO 10399) .19
6.2.5 Two-out-of-five test .19
6.2.6 “A — not A” test (see ISO 8588) .19
6.2.7 Tetrad test .19
6.2.8 Treatment of “no difference” responses in discrimination tests .19
6.2.9 Systematic effects .20
6.2.10 Sequential approach (see ISO 16820) .20
6.3 Tests using scales and categories .20
6.3.1 General.20
6.3.2 Classification .20
6.3.3 Grading (see ISO 4121) .20
6.3.4 Ranking (see ISO 8587) .20
6.3.5 Rating .21
6.3.6 Scoring .21
6.4 Analytical or descriptive tests .21
7 Test report .21
Annex A (informative) Statistical terms .23
Bibliography .25
iv © ISO 2017 – All rights reserved
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. Each 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.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the
different types of ISO documents should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www .iso .org/ directives).
Attention is drawn to the possibility that some of the elements of this document may be the subject of
patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of
any patent rights identified during the development of the document will be in the Introduction and/or
on the ISO list of patent declarations received (see www .iso .org/ patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation on the voluntary nature of standards, the meaning of ISO specific terms and
expressions related to conformity assessment, as well as information about ISO’s adherence to the
World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT) see the following
URL: w w w . i s o .org/ iso/ foreword .html.
This document was prepared by Technical Committee ISO/TC 34, Food Products, Subcommittee SC 12,
Sensory analysis.
This third edition cancels and replaces the second edition (ISO 6658:2005), which has been technically
revised. The following changes have been made:
— the definition of “sensory analysis” was updated;
— a tetrad test was added to the discrimination tests as 5.2.7;
— the descriptive test in 5.4 was expanded by new methods;
— Clause 6 was updated;
— Table A.1 was deleted from Annex A;
— the Bibliography was updated and expanded.
Introduction
This document constitutes a general introduction to the methodology of sensory analysis and should be
read before undertaking the more detailed test procedures described in other International Standards.
It covers the general area of methodology and is intended to fulfil the following functions:
— to provide a brief background of the essential features of methods of sensory analysis for the user
of specific tests;
— to provide details of general requirements, procedures and interpretation of results common to all
or most tests;
— to provide sufficient guidance on requirements, procedures and interpretation of results for
the different specific tests to allow choice of the most appropriate procedure(s) for solution of a
particular problem.
It comprises three main aspects, covered in Clauses 4, 5 and 6.
It is essential that Clause 4 “General requirements” be read first. Clause 5 “Methods of test” describes,
in a general manner, all the main tests. Clause 6 is concerned with some general principles of data
collection and analysis of sensory data and also briefly covers general principles of statistical treatment
of the results.
vi © ISO 2017 – All rights reserved
INTERNATIONAL STANDARD ISO 6658:2017(E)
Sensory analysis — Methodology — General guidance
WARNING — This document does not purport to address all of the safety problems, if any,
associated with its use. It is the responsibility of the user to establish appropriate safety and
health practices and to ensure compliance with any national regulatory conditions.
1 Scope
This document gives general guidance on the use of sensory analysis. It describes tests for the
examination of foods and other products by sensory analysis, and includes some general information
on the techniques to be used if statistical analysis of the results is required.
Generally these tests are intended only for objective sensory analysis. However, if a test can be used for
determining preference in hedonic test, this is indicated.
A hedonic test aims to determine the acceptability of the products and/or to determine preferences
among two or more products by the specified consumer population. The methods are effective for
determining whether a perceptible preference exists (difference in degree of liking), or whether no
perceptible preference exists (paired similarity test). General guidance for hedonic tests is given in
ISO 11136.
2 Normative references
The following documents, in whole or in part, are normatively referenced in this document and are
indispensable for its application. For dated references, only the edition cited applies. For undated
references, the latest edition of the referenced document (including any amendments) applies.
ISO 5492, Sensory analysis — Vocabulary
3 Terms and definitions
For the purposes of this document, the terms and definitions given in ISO 5492 and the following apply.
ISO and IEC maintain terminological databases for use in standardization at the following addresses:
— IEC Electropedia: available at http:// www .electropedia .org/
— ISO Online browsing platform: available at http:// www .iso .org/ obp
3.1
sensory analysis
science involved with the assessment of the organoleptic attributes of a product by the senses
4 General requirements
4.1 Basic information
This clause covers the general requirements common to all situations encountered in sensory analysis.
The information basic to these requirements is as follows.
a) The human response to one stimulus cannot be isolated from previous experiences or from other
sensory stimuli received from the environment. Nevertheless, influences arising from these two
sources can be controlled and the effect standardized.
b) Variability in sensory response is inherent in any group of people used for testing and is
unavoidable; this can arise from inconsistencies within an individual, and through physiological
and psychological differences between individuals. However, with training, such a group can show
highly consistent individual responses. Recognition of these factors is important in the analysis of
results.
c) Systematic biases in sensory experiments involving human response can result in misleading
data and incorrect interpretation that can be difficult to identify. The factors that can result in
bias should be identified and controlled as far as possible by appropriate experimental design and
conduct of the tests.
d) The validity of the conclusions drawn from the results is dependent upon the test used and the way
it is conducted, including the questions that have been asked.
4.2 Statement of objectives
There are three main types of objective, as follows:
a) those in which the aim of the test is to categorize, rank or describe the product(s);
b) those in which the aim is to distinguish between two or more products; here it is important to
distinguish between the need to know
i) if there is a difference at all,
ii) how great is the magnitude of the difference,
iii) the direction (or quality) of that difference,
iv) the influence of that difference, e.g. with regard to preference, or
v) if all or only part of a population is detecting a difference;
c) those in which reassurance is sought that products are sufficiently similar to be used
interchangeably.
In sensory analysis, a given problem frequently requires appreciable discussion or thought before an
appropriate test is selected. This is because the initial concept of the problem may require clarification.
4.3 Choice of test
The choice of appropriate test depends largely on the nature of the test objective, but also needs to take
account of factors associated with the product, the assessors, the test environment, and the desired
level of analytical precision and statistical confidence in the conclusions. The action that would occur
based on the outcome of the test should be determined in advance.
For each test, an attempt is made in Clause 5 to give guidance as to its relevance. Preliminary tests may
be necessary to confirm the applicability of a given test.
Because of sensory fatigue and the effects of adaptation, only a limited number of samples can be
assessed during a session, depending on the nature of the test and the type of product. Some of these
effects can be moderated by appropriate rinse procedures and recovery between samples.
While the use of control samples is essential in most cases, their use naturally limits the number of
samples that can be assessed during any given session.
The statistical plan should always be determined before starting the tests. This is especially
recommended if the number of samples to be evaluated requires more than one session. Details of
statistical plans should be selected from specialized texts. Whatever test method is used, the sequential
testing approach described in ISO 16820 should be considered whenever it is desirable to keep the
number of samples or the number of assessors to a minimum.
2 © ISO 2017 – All rights reserved
4.4 Choosing and training assessors
A sensory analysis panel constitutes a true “measuring instrument”, and consequently the results of
the analyses conducted depend on its members. The recruitment of persons willing to participate in a
panel, therefore, needs to be carried out with care and should be considered as a real investment, both
in time and financially. Management support in the organization is necessary if it is to be effective.
Sensory assessment may be made by three types of assessor: “sensory assessors”, “selected assessors”
or “expert sensory assessors”. Assessors can be “naive assessors” who do not have to meet any precise
criterion of selection or training, or people who have already taken part in some sensory tests (“initiated
assessors”). “Selected assessors” are assessors who have been selected and trained for the particular
sensory test. “Expert sensory assessors” are selected assessors who have been selected and trained
for a variety of sensory analysis methods and who are able to make consistent and repeatable sensory
assessments of products in one or several categories.
The selection and training methods to be employed depend on the tasks and methods that it is intended
to give to the “selected assessors”. Procedures for training assessors for descriptive tests are different
from those for training assessors in discrimination tests.
Detailed procedures and methods for selection and training of assessors are given in ISO 8586. It
should be noted that these methods sometimes only constitute a way of choosing the better candidates
amongst those who are available, rather than to satisfy predetermined criteria. Also, the selection
of assessors for their ability to discriminate and describe foods and other products is quite different
from that used for preference tests. The former tasks require selection and training, whereas the latter
requires only that the panel be representative of a specified sector of the population, for example, a
group of consumers.
If a selection procedure is to be carried out, some important criteria for choosing assessors are as
follows:
a) general ability to perform the specific sensory task;
b) availability with respect to normal employment;
c) motivation (willingness and interest);
d) good health (including the absence of specific allergies or treatment with medications) and good
dental and general hygienic condition.
The performance of “selected assessors” and “expert sensory assessors” should be monitored regularly
to ensure that the criteria by which they were initially selected continue to be met (see ISO 11132).
4.5 Material to be tested
The nature of the product to be tested determines the experimental protocol of the test, and may
also have an influence on the type of test that is required to satisfy the test objectives. For example,
a protocol in which foods are to be consumed hot will need to take into account the cooling rate of
the product and the likely effect on sensory attributes, and the changes in sensory attributes that may
occur in keeping the product hot prior to testing.
Methods of preparation and presentation of samples should be appropriate for the product and to the
problem concerned.
EXAMPLE 1 A product that is normally consumed hot is prepared in the usual manner and tested hot; however,
elevated temperatures can be used in some circumstances to increase the ease with which some flavours can be
evaluated.
EXAMPLE 2 A product that is normally consumed in discrete pieces is not homogenized in order to retain
textural characteristics. Care is needed, however, to ensure maximum uniformity between sub-samples for each
assessor; this includes similar portion size and uniformity of composition.
General principles for product sampling (in accordance with International Standards relating to
the product under test) should be applied for test samples. In all cases, documentation of sample
identification codes or lot numbers is necessary. Valid conclusions can be drawn for a product as a
whole only if the samples tested are representative.
Carriers may sometimes be used for tests relating to the evaluation of products for which direct tasting
is not feasible (see ISO 5497), for example food ingredients.
Lighting conditions should be specified when appearance is being assessed. When the test concerns
only differences in flavour, the effect of colour differences may be partially masked by the use of lighting
conditions that minimize the colour difference.
Containers should be chosen so as not to affect the test or the product. These may include washable
ceramic or glass containers, or disposable plastic or paper containers, but shall not transfer chemical
materials that could result in taint. In particular, washable containers should be washed only in odour-
and taint-free detergents and rinsed in water, and polymeric and paper containers, including insulated
containers used for hot or cold samples, should be odour- and taint-free.
Palate cleansers may be used by the assessors between samples and between sessions, but care should
be taken to ensure that they do not influence the flavour of products to be assessed. Still and carbonated
water and bland foods (for example, unsalted crackers) may be used between samples and between
sessions. Checks on the water supply are desirable to ensure that it is bland and safe to consume. For
particular purposes, deionized water, glass-distilled water, low mineral content spring water, carbon-
filtered water or boiled tap water may be used, but it shall be noted that they are likely to have different
flavours.
4.6 Test room
Sensory analysis shall be conducted in a dedicated test room (see ISO 8589 for guidance). The aim shall
be to create for each assessor a separate environment with minimum distraction, so that each assessor
can quickly adjust to the nature of the new task(s). Extraneous activities, including preparation of the
samples, should not be allowed during the tests, as these can lead to biased results. The room should be
at a comfortable temperature and should be ventilated with odour-free air; limited airflow is desirable
to avoid excessive temperature fluctuations. Persistent odours, such as tobacco or cosmetics, shall not
be allowed to contaminate the environment of the test room.
Sound should be restricted. A low background noise is usually more tolerable than a fluctuating level
of noise. Conversation is more distracting than background noise. Interruptions cause the greatest
distraction.
It is usually helpful to have control over both the colour and the intensity of the lighting, although
coloured lights rarely succeed in completely masking differences in appearance.
Surfaces shall be non-absorbent and designed to facilitate a high standard of hygiene. The dimensions
of the tasting booths are important; very low ceilings and very narrow booths can be oppressive or can
give rise to a feeling of claustrophobia. Comfortable seating is necessary.
If provision is made for computerized data acquisition, then this should be implemented safely,
hygienically, and in a way that does not compromise sensory judgement.
4.7 Planning and conduct of the test
The planning and operation of the test are determined by the objectives of the programme, the test
chosen, and practical constraints associated with the use of human subjects. In particular, it is important
to recognize the biases that might be inherent in the chosen test, and to operate the test in such a way
as to minimize the effects of any bias. The potential biases can originate from both psychological and
physiological sources.
4 © ISO 2017 – All rights reserved
The most serious psychological bias results from assessors interacting to influence each other’s
judgements, and should be minimized by the use of individual booths or adequate separation of the
assessors. Moreover, strict management of the activities of the assessors is necessary.
The manner and order of presentation of the samples are important aspects of the test, and can introduce
psychological biases. For example, the samples should be coded by random three-digit numbers, and
the codes should be varied for each test. The order of assessment can also be a source of bias and, in
general, the order should be specified. With a small number of samples and assessors, the order can be
balanced so that every possible order occurs an equal number of times. In larger experiments, the order
can be balanced or randomized.
Physiological biases are frequently associated with the nature of the test samples. In particular,
adaptation to a specific flavour stimulus can occur on repeated exposure to that stimulus, and fatigue
can occur when chewing solid foods. Both factors can impose an upper limit on the number of samples
to be assessed in a session. Expectoration of samples may be recommended with trained panels, but
loss of information on specific sensory attributes may result.
Hunger and satiety can influence an assessor’s performance, and, if panels are held too frequently,
performance may deteriorate. If it is possible, assessors should be asked to refrain from smoking and
from consuming snacks such as coffee for 1 h before a test. Assessors shall not carry any foreign odours
into the session, for example tobacco or cosmetic odours, as these could influence the responses of
other assessors.
The time of day at which the test is conducted is important. The schedule should take into account
local customary mealtimes since performance is generally considered optimum at mid-morning and
mid-afternoon. Assessors suffering from emotional upsets, colds and other illnesses should be excluded
from tests until they recover.
The collation of the results comprises three aspects:
— checking that all data have been recorded accurately, either on computer or manually;
— verification that any additional relevant information which may aid or cast doubt on the interpretation
of the results has been noted;
— checking that the assessors are motivated to continue participating if further testing is planned.
5 Test methods
5.1 General
The most commonly used tests are divided into three groups:
a) discrimination tests used to determine the probability of difference or similarity between products;
b) tests using scales and categories to estimate the order or size of differences or the categories or
classes to which samples should be allocated;
c) descriptive tests used to characterize, both qualitatively and quantitatively, the specific sensory
attributes present in a sample (see 5.4).
[27]
Most of these tests with some modifications can be used for consumer testing .
For the number of assessors, refer to the corresponding standards, taking into consideration α or β
risk depending on the purpose of the test. Alternatively, sequential analysis (see ISO 16820) may allow
a decision to be made after fewer trials of the test than would be required by conventional approaches
that use a predetermined number of assessments.
5.2 Discrimination tests
5.2.1 General
Before performing a discrimination test, the objective of the test shall be specified:
— to show that a significant difference between two products exists;
— to show that a significant similarity between two products exists.
In the first case, a difference test is used. In the second case, a similarity test is used.
The following tests are most commonly used to determine the probability of difference or similarity
between samples:
a) paired comparison test (see 5.2.2);
b) triangle test (see 5.2.3);
c) duo-trio test (see 5.2.4);
d) two-out-of-five test (see 5.2.5);
e) “A — not A” test (see 5.2.6);
f) tetrad test (see 5.2.7).
For all these tests, there are different manners of analysing the results.
5.2.2 Paired comparison test
5.2.2.1 Definition
This is a test in which samples are presented in pairs for comparison and detection of a perceptible
sensory difference or a similarity on the basis of some defined criteria.
See ISO 5495 for details.
5.2.2.2 Application
The paired comparison test is recommended
a) to determine if a perceptible difference exists (objective of difference) in a particular attribute (e.g.
sweetness), or to determine if similarity exists in that attribute,
b) to select, train and monitor the performance of the assessors,
c) to compare two products in terms of preference in the context of consumer tests.
Advantages of the test over other discrimination tests are simplicity and less sensory fatigue.
The disadvantage of the method of paired comparisons is that, as the number of samples to be compared
increases, the number of inter-comparisons required rapidly becomes impracticable.
5.2.2.3 Procedure
The assessors receive a set of two samples (i.e. a pair). They designate the sample that they consider to
be the most intense regarding the attribute under consideration, even if this choice is based only on a
guess. One of the samples (of the pair) may be a control. The number of times each sample is selected is
counted and the significance is determined by reference to a statistical table.
6 © ISO 2017 – All rights reserved
It is necessary to determine, prior to carrying out the test
a) what is the objective of the test (is it a difference or similarity test?),
b) what is the most appropriate sensitivity, and
c) whether the statistical test that follows is to be
i) one-sided (i.e. the test supervisor expects a particular direction of difference and the
alternative hypothesis corresponds to the existence of a difference in that direction), or
ii) two-sided (i.e. the test supervisor has no expected direction of difference and the alternative
hypothesis corresponds to a difference in either direction).
Questions of difference and preference should not be combined: the recruitment criteria for panels are
different for these questions.
5.2.2.4 Analysis of results
See 6.2.2.
5.2.3 Triangle test
5.2.3.1 Definition
This is a discrimination test involving three coded samples, two of which are identical, presented
simultaneously. The assessors are asked to select the odd sample.
See ISO 4120 for details.
5.2.3.2 Application
The triangle test is recommended
a) when the nature of the difference is unknown, and
b) for the selection and training of assessors.
The test should not be used for the determination of preference. Some disadvantages of the test are that
— it is uneconomical for the assessment of a large number of samples,
— it may be more affected by sensory adaptation than the paired comparison test using samples with
strong sensory attributes,
— if the nature of the difference is known, it is statistically less efficient than some other tests, and
— the method is applicable only if the products are as homogeneous as possible.
5.2.3.3 Procedure
The assessors are each presented with one set of three coded samples, two of which are identical, and
are asked to select the odd sample.
Samples should be presented an equal number of times in each of the two sets of three distinct
permutations of order, which are
BAA ABA AAB
ABB BAB BBA
5.2.3.4 Analysis of results
See 6.2.3.
5.2.4 Duo-trio test
5.2.4.1 Definition
This is a discrimination test in which the reference sample is presented first. It is followed by two
samples, one of which is identical to the reference sample and which the assessors are asked to identify.
See ISO 10399 for details.
5.2.4.2 Application
This duo-trio test is used to determine if there is a sensory difference or similarity between a given
sample and a reference. It is especially suitable when the reference sample is well known to the
assessors, for example a sample of regular production.
If there are strong after-sensations, this test is less suitable than the paired comparison test (5.2.2) or
the “A — not A” test (5.2.6).
5.2.4.3 Procedure
The assessors are first presented with the identified reference sample. This is followed by two coded
samples, one of which is identical to the reference sample. The assessors are asked to identify this sample.
5.2.4.4 Analysis of results
See 6.2.4.
5.2.5 Two-out-of-five test
5.2.5.1 Definition
This is a discrimination test involving five coded samples, two of which are of one type and three of
another. The assessors are asked to group the two sets of samples.
5.2.5.2 Application
The two-out-of-five test is recommended to establish a difference more economically than other tests
(the method is statistically more efficient).
The disadvantages of this test are similar to those of the triangle test (5.2.3). It is more strongly affected
by sensory fatigue and memory effects but has greater statistical power. Its principal use is in visual,
auditory or tactile applications.
5.2.5.3 Procedure
The assessors are each presented with one set of five coded samples and are told that two are of one
type and three of another. The assessors are asked to group the two sets of samples.
When the number of assessors is less than 20, the order of presentation should be selected at random
from the following 20 distinct permutations:
AAABB BBBAA AABAB BBABA ABAAB BABBA BAAAB ABBBA AABBA BBAAB ABABA BABAB BAABA
ABBAB ABBAA BAABB BABAA ABABB BBAAA AABBB
8 © ISO 2017 – All rights reserved
5.2.5.4 Analysis of results
See 6.2.5.
5.2.6 “A — not A” test
5.2.6.1 Definition
This is a test in which a series of samples, which may be “A” or “not A”, is presented to the assessors
after they have learnt to recognize sample “A”. The assessors are asked to indicate which sample is “A”.
See ISO 8588 for details.
5.2.6.2 Application
This test is a discrimination test, which may be used for the assessment of samples having variations of
appearance or leaving a persistent after-sensation.
It is especially useful when strictly similar repeated samples cannot be obtained.
5.2.6.3 Procedure
The assessors are presented with samples one at a time. The assessors are first presented with the
reference sample “A” several times, until they can recognize it. They are then given several samples, each
of which may be “A” or “not A”, at random, and they have to determine which they are. An appreciable
time interval (for example 2 min to 5 min) should be allowed between receipt of samples, and only a few
samples should be examined during one session.
5.2.6.4 Analysis of results
See 6.2.6.
5.2.7 Tetrad test
5.2.7.1 Definition
This is a discrimination test involving four coded samples, two of which are of one type and other two
of another. The assessors are asked to group the two sets of samples.
See ASTM E3009-15.
5.2.7.2 Application
This test may be used to determine whether a perceptible sensory difference exists between samples of
two products or to estimate the magnitude of the perceptible difference, for example, when a change is
made in ingredients, processing, packaging, handling, or storage.
5.2.7.3 Procedure
The assessors are each presented with one set of four coded samples, two from one group and two from
another group, and are asked to sort the samples into two groups of two samples based on similarity
using either a specified or unspecified approach.
Samples should be presented an equal number of times in each of the two sets of three distinct
permutations of order, which are
AABB ABAB ABAB
BBAA BABA BAAB
5.2.7.4 Analysis of results
See 6.2.7.
5.3 Use of scales and categories
5.3.1 General considerations
See ISO 4121 for details.
Measurement methods in sensory analysis might seek to decide the categories, classes or grades to
which samples should be allocated. They may also express numerical estimates of the magnitude of
perceived attributes of samples or of differences between samples.
There is no direct relationship between the response scale used to elicit numbers and the measurement
scale that corresponds to the values recorded. Thus, the same method of obtaining numbers (response
scale) can lead to values whose measurement scale is only ordinal (unequal intervals) or is on an interval
scale (equal intervals). With an ordinal measurement scale, the size of the difference between two
values cannot be assumed to reflect the difference between the perceived intensities. Nor can the ratio
of t
...
INTERNATIONAL ISO
STANDARD 6658
Third edition
2017-07
Sensory analysis — Methodology —
General guidance
Analyse sensorielle — Méthodologie — Lignes directrices générales
Reference number
©
ISO 2017
© ISO 2017, Published in Switzerland
All rights reserved. Unless otherwise specified, no part of this publication may be reproduced or utilized otherwise in any form
or by any means, electronic or mechanical, including photocopying, or posting on the internet or an intranet, without prior
written permission. Permission can be requested from either ISO at the address below or ISO’s member body in the country of
the requester.
ISO copyright office
Ch. de Blandonnet 8 • CP 401
CH-1214 Vernier, Geneva, Switzerland
Tel. +41 22 749 01 11
Fax +41 22 749 09 47
copyright@iso.org
www.iso.org
ii © ISO 2017 – All rights reserved
Contents Page
Foreword .v
Introduction .vi
1 Scope . 1
2 Normative references . 1
3 Terms and definitions . 1
4 General requirements . 1
4.1 Basic information . 1
4.2 Statement of objectives . 2
4.3 Choice of test . 2
4.4 Choosing and training assessors . 3
4.5 Material to be tested . . 3
4.6 Test room . 4
4.7 Planning and conduct of the test . 4
5 Test methods . 5
5.1 General . 5
5.2 Discrimination tests . 6
5.2.1 General. 6
5.2.2 Paired comparison test. 6
5.2.3 Triangle test . 7
5.2.4 Duo-trio test . 8
5.2.5 Two-out-of-five test . 8
5.2.6 “A — not A” test . 9
5.2.7 Tetrad test . 9
5.3 Use of scales and categories .10
5.3.1 General considerations .10
5.3.2 Interpretation of results .10
5.3.3 Classification .11
5.3.4 Grading .11
5.3.5 Ranking .11
5.3.6 Rating and scoring .12
5.4 Descriptive tests.12
5.4.1 General.12
5.4.2 Qualitative sensory profile .13
5.4.3 Quantitative descriptive profile .13
5.4.4 Consensus profile .15
5.4.5 Free-choice profile.15
5.4.6 Flash profile .16
5.4.7 Deviation from reference profile.16
5.4.8 Temporal dominance of sensations (TDS).17
6 Analysis of results.17
6.1 General .17
6.2 Discrimination tests .17
6.2.1 General.17
6.2.2 Paired comparison test (see ISO 5495) .18
6.2.3 Triangle test (see ISO 4120) .19
6.2.4 Duo-trio test (see ISO 10399) .19
6.2.5 Two-out-of-five test .19
6.2.6 “A — not A” test (see ISO 8588) .19
6.2.7 Tetrad test .19
6.2.8 Treatment of “no difference” responses in discrimination tests .19
6.2.9 Systematic effects .20
6.2.10 Sequential approach (see ISO 16820) .20
6.3 Tests using scales and categories .20
6.3.1 General.20
6.3.2 Classification .20
6.3.3 Grading (see ISO 4121) .20
6.3.4 Ranking (see ISO 8587) .20
6.3.5 Rating .21
6.3.6 Scoring .21
6.4 Analytical or descriptive tests .21
7 Test report .21
Annex A (informative) Statistical terms .23
Bibliography .25
iv © ISO 2017 – All rights reserved
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. Each 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.
The procedures used to develop this document and those intended for its further maintenance are
described in the ISO/IEC Directives, Part 1. In particular the different approval criteria needed for the
different types of ISO documents should be noted. This document was drafted in accordance with the
editorial rules of the ISO/IEC Directives, Part 2 (see www .iso .org/ directives).
Attention is drawn to the possibility that some of the elements of this document may be the subject of
patent rights. ISO shall not be held responsible for identifying any or all such patent rights. Details of
any patent rights identified during the development of the document will be in the Introduction and/or
on the ISO list of patent declarations received (see www .iso .org/ patents).
Any trade name used in this document is information given for the convenience of users and does not
constitute an endorsement.
For an explanation on the voluntary nature of standards, the meaning of ISO specific terms and
expressions related to conformity assessment, as well as information about ISO’s adherence to the
World Trade Organization (WTO) principles in the Technical Barriers to Trade (TBT) see the following
URL: w w w . i s o .org/ iso/ foreword .html.
This document was prepared by Technical Committee ISO/TC 34, Food Products, Subcommittee SC 12,
Sensory analysis.
This third edition cancels and replaces the second edition (ISO 6658:2005), which has been technically
revised. The following changes have been made:
— the definition of “sensory analysis” was updated;
— a tetrad test was added to the discrimination tests as 5.2.7;
— the descriptive test in 5.4 was expanded by new methods;
— Clause 6 was updated;
— Table A.1 was deleted from Annex A;
— the Bibliography was updated and expanded.
Introduction
This document constitutes a general introduction to the methodology of sensory analysis and should be
read before undertaking the more detailed test procedures described in other International Standards.
It covers the general area of methodology and is intended to fulfil the following functions:
— to provide a brief background of the essential features of methods of sensory analysis for the user
of specific tests;
— to provide details of general requirements, procedures and interpretation of results common to all
or most tests;
— to provide sufficient guidance on requirements, procedures and interpretation of results for
the different specific tests to allow choice of the most appropriate procedure(s) for solution of a
particular problem.
It comprises three main aspects, covered in Clauses 4, 5 and 6.
It is essential that Clause 4 “General requirements” be read first. Clause 5 “Methods of test” describes,
in a general manner, all the main tests. Clause 6 is concerned with some general principles of data
collection and analysis of sensory data and also briefly covers general principles of statistical treatment
of the results.
vi © ISO 2017 – All rights reserved
INTERNATIONAL STANDARD ISO 6658:2017(E)
Sensory analysis — Methodology — General guidance
WARNING — This document does not purport to address all of the safety problems, if any,
associated with its use. It is the responsibility of the user to establish appropriate safety and
health practices and to ensure compliance with any national regulatory conditions.
1 Scope
This document gives general guidance on the use of sensory analysis. It describes tests for the
examination of foods and other products by sensory analysis, and includes some general information
on the techniques to be used if statistical analysis of the results is required.
Generally these tests are intended only for objective sensory analysis. However, if a test can be used for
determining preference in hedonic test, this is indicated.
A hedonic test aims to determine the acceptability of the products and/or to determine preferences
among two or more products by the specified consumer population. The methods are effective for
determining whether a perceptible preference exists (difference in degree of liking), or whether no
perceptible preference exists (paired similarity test). General guidance for hedonic tests is given in
ISO 11136.
2 Normative references
The following documents, in whole or in part, are normatively referenced in this document and are
indispensable for its application. For dated references, only the edition cited applies. For undated
references, the latest edition of the referenced document (including any amendments) applies.
ISO 5492, Sensory analysis — Vocabulary
3 Terms and definitions
For the purposes of this document, the terms and definitions given in ISO 5492 and the following apply.
ISO and IEC maintain terminological databases for use in standardization at the following addresses:
— IEC Electropedia: available at http:// www .electropedia .org/
— ISO Online browsing platform: available at http:// www .iso .org/ obp
3.1
sensory analysis
science involved with the assessment of the organoleptic attributes of a product by the senses
4 General requirements
4.1 Basic information
This clause covers the general requirements common to all situations encountered in sensory analysis.
The information basic to these requirements is as follows.
a) The human response to one stimulus cannot be isolated from previous experiences or from other
sensory stimuli received from the environment. Nevertheless, influences arising from these two
sources can be controlled and the effect standardized.
b) Variability in sensory response is inherent in any group of people used for testing and is
unavoidable; this can arise from inconsistencies within an individual, and through physiological
and psychological differences between individuals. However, with training, such a group can show
highly consistent individual responses. Recognition of these factors is important in the analysis of
results.
c) Systematic biases in sensory experiments involving human response can result in misleading
data and incorrect interpretation that can be difficult to identify. The factors that can result in
bias should be identified and controlled as far as possible by appropriate experimental design and
conduct of the tests.
d) The validity of the conclusions drawn from the results is dependent upon the test used and the way
it is conducted, including the questions that have been asked.
4.2 Statement of objectives
There are three main types of objective, as follows:
a) those in which the aim of the test is to categorize, rank or describe the product(s);
b) those in which the aim is to distinguish between two or more products; here it is important to
distinguish between the need to know
i) if there is a difference at all,
ii) how great is the magnitude of the difference,
iii) the direction (or quality) of that difference,
iv) the influence of that difference, e.g. with regard to preference, or
v) if all or only part of a population is detecting a difference;
c) those in which reassurance is sought that products are sufficiently similar to be used
interchangeably.
In sensory analysis, a given problem frequently requires appreciable discussion or thought before an
appropriate test is selected. This is because the initial concept of the problem may require clarification.
4.3 Choice of test
The choice of appropriate test depends largely on the nature of the test objective, but also needs to take
account of factors associated with the product, the assessors, the test environment, and the desired
level of analytical precision and statistical confidence in the conclusions. The action that would occur
based on the outcome of the test should be determined in advance.
For each test, an attempt is made in Clause 5 to give guidance as to its relevance. Preliminary tests may
be necessary to confirm the applicability of a given test.
Because of sensory fatigue and the effects of adaptation, only a limited number of samples can be
assessed during a session, depending on the nature of the test and the type of product. Some of these
effects can be moderated by appropriate rinse procedures and recovery between samples.
While the use of control samples is essential in most cases, their use naturally limits the number of
samples that can be assessed during any given session.
The statistical plan should always be determined before starting the tests. This is especially
recommended if the number of samples to be evaluated requires more than one session. Details of
statistical plans should be selected from specialized texts. Whatever test method is used, the sequential
testing approach described in ISO 16820 should be considered whenever it is desirable to keep the
number of samples or the number of assessors to a minimum.
2 © ISO 2017 – All rights reserved
4.4 Choosing and training assessors
A sensory analysis panel constitutes a true “measuring instrument”, and consequently the results of
the analyses conducted depend on its members. The recruitment of persons willing to participate in a
panel, therefore, needs to be carried out with care and should be considered as a real investment, both
in time and financially. Management support in the organization is necessary if it is to be effective.
Sensory assessment may be made by three types of assessor: “sensory assessors”, “selected assessors”
or “expert sensory assessors”. Assessors can be “naive assessors” who do not have to meet any precise
criterion of selection or training, or people who have already taken part in some sensory tests (“initiated
assessors”). “Selected assessors” are assessors who have been selected and trained for the particular
sensory test. “Expert sensory assessors” are selected assessors who have been selected and trained
for a variety of sensory analysis methods and who are able to make consistent and repeatable sensory
assessments of products in one or several categories.
The selection and training methods to be employed depend on the tasks and methods that it is intended
to give to the “selected assessors”. Procedures for training assessors for descriptive tests are different
from those for training assessors in discrimination tests.
Detailed procedures and methods for selection and training of assessors are given in ISO 8586. It
should be noted that these methods sometimes only constitute a way of choosing the better candidates
amongst those who are available, rather than to satisfy predetermined criteria. Also, the selection
of assessors for their ability to discriminate and describe foods and other products is quite different
from that used for preference tests. The former tasks require selection and training, whereas the latter
requires only that the panel be representative of a specified sector of the population, for example, a
group of consumers.
If a selection procedure is to be carried out, some important criteria for choosing assessors are as
follows:
a) general ability to perform the specific sensory task;
b) availability with respect to normal employment;
c) motivation (willingness and interest);
d) good health (including the absence of specific allergies or treatment with medications) and good
dental and general hygienic condition.
The performance of “selected assessors” and “expert sensory assessors” should be monitored regularly
to ensure that the criteria by which they were initially selected continue to be met (see ISO 11132).
4.5 Material to be tested
The nature of the product to be tested determines the experimental protocol of the test, and may
also have an influence on the type of test that is required to satisfy the test objectives. For example,
a protocol in which foods are to be consumed hot will need to take into account the cooling rate of
the product and the likely effect on sensory attributes, and the changes in sensory attributes that may
occur in keeping the product hot prior to testing.
Methods of preparation and presentation of samples should be appropriate for the product and to the
problem concerned.
EXAMPLE 1 A product that is normally consumed hot is prepared in the usual manner and tested hot; however,
elevated temperatures can be used in some circumstances to increase the ease with which some flavours can be
evaluated.
EXAMPLE 2 A product that is normally consumed in discrete pieces is not homogenized in order to retain
textural characteristics. Care is needed, however, to ensure maximum uniformity between sub-samples for each
assessor; this includes similar portion size and uniformity of composition.
General principles for product sampling (in accordance with International Standards relating to
the product under test) should be applied for test samples. In all cases, documentation of sample
identification codes or lot numbers is necessary. Valid conclusions can be drawn for a product as a
whole only if the samples tested are representative.
Carriers may sometimes be used for tests relating to the evaluation of products for which direct tasting
is not feasible (see ISO 5497), for example food ingredients.
Lighting conditions should be specified when appearance is being assessed. When the test concerns
only differences in flavour, the effect of colour differences may be partially masked by the use of lighting
conditions that minimize the colour difference.
Containers should be chosen so as not to affect the test or the product. These may include washable
ceramic or glass containers, or disposable plastic or paper containers, but shall not transfer chemical
materials that could result in taint. In particular, washable containers should be washed only in odour-
and taint-free detergents and rinsed in water, and polymeric and paper containers, including insulated
containers used for hot or cold samples, should be odour- and taint-free.
Palate cleansers may be used by the assessors between samples and between sessions, but care should
be taken to ensure that they do not influence the flavour of products to be assessed. Still and carbonated
water and bland foods (for example, unsalted crackers) may be used between samples and between
sessions. Checks on the water supply are desirable to ensure that it is bland and safe to consume. For
particular purposes, deionized water, glass-distilled water, low mineral content spring water, carbon-
filtered water or boiled tap water may be used, but it shall be noted that they are likely to have different
flavours.
4.6 Test room
Sensory analysis shall be conducted in a dedicated test room (see ISO 8589 for guidance). The aim shall
be to create for each assessor a separate environment with minimum distraction, so that each assessor
can quickly adjust to the nature of the new task(s). Extraneous activities, including preparation of the
samples, should not be allowed during the tests, as these can lead to biased results. The room should be
at a comfortable temperature and should be ventilated with odour-free air; limited airflow is desirable
to avoid excessive temperature fluctuations. Persistent odours, such as tobacco or cosmetics, shall not
be allowed to contaminate the environment of the test room.
Sound should be restricted. A low background noise is usually more tolerable than a fluctuating level
of noise. Conversation is more distracting than background noise. Interruptions cause the greatest
distraction.
It is usually helpful to have control over both the colour and the intensity of the lighting, although
coloured lights rarely succeed in completely masking differences in appearance.
Surfaces shall be non-absorbent and designed to facilitate a high standard of hygiene. The dimensions
of the tasting booths are important; very low ceilings and very narrow booths can be oppressive or can
give rise to a feeling of claustrophobia. Comfortable seating is necessary.
If provision is made for computerized data acquisition, then this should be implemented safely,
hygienically, and in a way that does not compromise sensory judgement.
4.7 Planning and conduct of the test
The planning and operation of the test are determined by the objectives of the programme, the test
chosen, and practical constraints associated with the use of human subjects. In particular, it is important
to recognize the biases that might be inherent in the chosen test, and to operate the test in such a way
as to minimize the effects of any bias. The potential biases can originate from both psychological and
physiological sources.
4 © ISO 2017 – All rights reserved
The most serious psychological bias results from assessors interacting to influence each other’s
judgements, and should be minimized by the use of individual booths or adequate separation of the
assessors. Moreover, strict management of the activities of the assessors is necessary.
The manner and order of presentation of the samples are important aspects of the test, and can introduce
psychological biases. For example, the samples should be coded by random three-digit numbers, and
the codes should be varied for each test. The order of assessment can also be a source of bias and, in
general, the order should be specified. With a small number of samples and assessors, the order can be
balanced so that every possible order occurs an equal number of times. In larger experiments, the order
can be balanced or randomized.
Physiological biases are frequently associated with the nature of the test samples. In particular,
adaptation to a specific flavour stimulus can occur on repeated exposure to that stimulus, and fatigue
can occur when chewing solid foods. Both factors can impose an upper limit on the number of samples
to be assessed in a session. Expectoration of samples may be recommended with trained panels, but
loss of information on specific sensory attributes may result.
Hunger and satiety can influence an assessor’s performance, and, if panels are held too frequently,
performance may deteriorate. If it is possible, assessors should be asked to refrain from smoking and
from consuming snacks such as coffee for 1 h before a test. Assessors shall not carry any foreign odours
into the session, for example tobacco or cosmetic odours, as these could influence the responses of
other assessors.
The time of day at which the test is conducted is important. The schedule should take into account
local customary mealtimes since performance is generally considered optimum at mid-morning and
mid-afternoon. Assessors suffering from emotional upsets, colds and other illnesses should be excluded
from tests until they recover.
The collation of the results comprises three aspects:
— checking that all data have been recorded accurately, either on computer or manually;
— verification that any additional relevant information which may aid or cast doubt on the interpretation
of the results has been noted;
— checking that the assessors are motivated to continue participating if further testing is planned.
5 Test methods
5.1 General
The most commonly used tests are divided into three groups:
a) discrimination tests used to determine the probability of difference or similarity between products;
b) tests using scales and categories to estimate the order or size of differences or the categories or
classes to which samples should be allocated;
c) descriptive tests used to characterize, both qualitatively and quantitatively, the specific sensory
attributes present in a sample (see 5.4).
[27]
Most of these tests with some modifications can be used for consumer testing .
For the number of assessors, refer to the corresponding standards, taking into consideration α or β
risk depending on the purpose of the test. Alternatively, sequential analysis (see ISO 16820) may allow
a decision to be made after fewer trials of the test than would be required by conventional approaches
that use a predetermined number of assessments.
5.2 Discrimination tests
5.2.1 General
Before performing a discrimination test, the objective of the test shall be specified:
— to show that a significant difference between two products exists;
— to show that a significant similarity between two products exists.
In the first case, a difference test is used. In the second case, a similarity test is used.
The following tests are most commonly used to determine the probability of difference or similarity
between samples:
a) paired comparison test (see 5.2.2);
b) triangle test (see 5.2.3);
c) duo-trio test (see 5.2.4);
d) two-out-of-five test (see 5.2.5);
e) “A — not A” test (see 5.2.6);
f) tetrad test (see 5.2.7).
For all these tests, there are different manners of analysing the results.
5.2.2 Paired comparison test
5.2.2.1 Definition
This is a test in which samples are presented in pairs for comparison and detection of a perceptible
sensory difference or a similarity on the basis of some defined criteria.
See ISO 5495 for details.
5.2.2.2 Application
The paired comparison test is recommended
a) to determine if a perceptible difference exists (objective of difference) in a particular attribute (e.g.
sweetness), or to determine if similarity exists in that attribute,
b) to select, train and monitor the performance of the assessors,
c) to compare two products in terms of preference in the context of consumer tests.
Advantages of the test over other discrimination tests are simplicity and less sensory fatigue.
The disadvantage of the method of paired comparisons is that, as the number of samples to be compared
increases, the number of inter-comparisons required rapidly becomes impracticable.
5.2.2.3 Procedure
The assessors receive a set of two samples (i.e. a pair). They designate the sample that they consider to
be the most intense regarding the attribute under consideration, even if this choice is based only on a
guess. One of the samples (of the pair) may be a control. The number of times each sample is selected is
counted and the significance is determined by reference to a statistical table.
6 © ISO 2017 – All rights reserved
It is necessary to determine, prior to carrying out the test
a) what is the objective of the test (is it a difference or similarity test?),
b) what is the most appropriate sensitivity, and
c) whether the statistical test that follows is to be
i) one-sided (i.e. the test supervisor expects a particular direction of difference and the
alternative hypothesis corresponds to the existence of a difference in that direction), or
ii) two-sided (i.e. the test supervisor has no expected direction of difference and the alternative
hypothesis corresponds to a difference in either direction).
Questions of difference and preference should not be combined: the recruitment criteria for panels are
different for these questions.
5.2.2.4 Analysis of results
See 6.2.2.
5.2.3 Triangle test
5.2.3.1 Definition
This is a discrimination test involving three coded samples, two of which are identical, presented
simultaneously. The assessors are asked to select the odd sample.
See ISO 4120 for details.
5.2.3.2 Application
The triangle test is recommended
a) when the nature of the difference is unknown, and
b) for the selection and training of assessors.
The test should not be used for the determination of preference. Some disadvantages of the test are that
— it is uneconomical for the assessment of a large number of samples,
— it may be more affected by sensory adaptation than the paired comparison test using samples with
strong sensory attributes,
— if the nature of the difference is known, it is statistically less efficient than some other tests, and
— the method is applicable only if the products are as homogeneous as possible.
5.2.3.3 Procedure
The assessors are each presented with one set of three coded samples, two of which are identical, and
are asked to select the odd sample.
Samples should be presented an equal number of times in each of the two sets of three distinct
permutations of order, which are
BAA ABA AAB
ABB BAB BBA
5.2.3.4 Analysis of results
See 6.2.3.
5.2.4 Duo-trio test
5.2.4.1 Definition
This is a discrimination test in which the reference sample is presented first. It is followed by two
samples, one of which is identical to the reference sample and which the assessors are asked to identify.
See ISO 10399 for details.
5.2.4.2 Application
This duo-trio test is used to determine if there is a sensory difference or similarity between a given
sample and a reference. It is especially suitable when the reference sample is well known to the
assessors, for example a sample of regular production.
If there are strong after-sensations, this test is less suitable than the paired comparison test (5.2.2) or
the “A — not A” test (5.2.6).
5.2.4.3 Procedure
The assessors are first presented with the identified reference sample. This is followed by two coded
samples, one of which is identical to the reference sample. The assessors are asked to identify this sample.
5.2.4.4 Analysis of results
See 6.2.4.
5.2.5 Two-out-of-five test
5.2.5.1 Definition
This is a discrimination test involving five coded samples, two of which are of one type and three of
another. The assessors are asked to group the two sets of samples.
5.2.5.2 Application
The two-out-of-five test is recommended to establish a difference more economically than other tests
(the method is statistically more efficient).
The disadvantages of this test are similar to those of the triangle test (5.2.3). It is more strongly affected
by sensory fatigue and memory effects but has greater statistical power. Its principal use is in visual,
auditory or tactile applications.
5.2.5.3 Procedure
The assessors are each presented with one set of five coded samples and are told that two are of one
type and three of another. The assessors are asked to group the two sets of samples.
When the number of assessors is less than 20, the order of presentation should be selected at random
from the following 20 distinct permutations:
AAABB BBBAA AABAB BBABA ABAAB BABBA BAAAB ABBBA AABBA BBAAB ABABA BABAB BAABA
ABBAB ABBAA BAABB BABAA ABABB BBAAA AABBB
8 © ISO 2017 – All rights reserved
5.2.5.4 Analysis of results
See 6.2.5.
5.2.6 “A — not A” test
5.2.6.1 Definition
This is a test in which a series of samples, which may be “A” or “not A”, is presented to the assessors
after they have learnt to recognize sample “A”. The assessors are asked to indicate which sample is “A”.
See ISO 8588 for details.
5.2.6.2 Application
This test is a discrimination test, which may be used for the assessment of samples having variations of
appearance or leaving a persistent after-sensation.
It is especially useful when strictly similar repeated samples cannot be obtained.
5.2.6.3 Procedure
The assessors are presented with samples one at a time. The assessors are first presented with the
reference sample “A” several times, until they can recognize it. They are then given several samples, each
of which may be “A” or “not A”, at random, and they have to determine which they are. An appreciable
time interval (for example 2 min to 5 min) should be allowed between receipt of samples, and only a few
samples should be examined during one session.
5.2.6.4 Analysis of results
See 6.2.6.
5.2.7 Tetrad test
5.2.7.1 Definition
This is a discrimination test involving four coded samples, two of which are of one type and other two
of another. The assessors are asked to group the two sets of samples.
See ASTM E3009-15.
5.2.7.2 Application
This test may be used to determine whether a perceptible sensory difference exists between samples of
two products or to estimate the magnitude of the perceptible difference, for example, when a change is
made in ingredients, processing, packaging, handling, or storage.
5.2.7.3 Procedure
The assessors are each presented with one set of four coded samples, two from one group and two from
another group, and are asked to sort the samples into two groups of two samples based on similarity
using either a specified or unspecified approach.
Samples should be presented an equal number of times in each of the two sets of three distinct
permutations of order, which are
AABB ABAB ABAB
BBAA BABA BAAB
5.2.7.4 Analysis of results
See 6.2.7.
5.3 Use of scales and categories
5.3.1 General considerations
See ISO 4121 for details.
Measurement methods in sensory analysis might seek to decide the categories, classes or grades to
which samples should be allocated. They may also express numerical estimates of the magnitude of
perceived attributes of samples or of differences between samples.
There is no direct relationship between the response scale used to elicit numbers and the measurement
scale that corresponds to the values recorded. Thus, the same method of obtaining numbers (response
scale) can lead to values whose measurement scale is only ordinal (unequal intervals) or is on an interval
scale (equal intervals). With an ordinal measurement scale, the size of the difference between two
values cannot be assumed to reflect the difference between the perceived intensities. Nor can the ratio
of two values be assumed to reflect the ratio of the perceived intensities. With an interval measurement
scale, larger numerical values correspond to larger perceived intensities (or degrees of pleasure) and
the size of the difference between two values reflects the size of the difference in perceived intensity
of the property being measured. However, a numerical value of zero might not indicate a total absence
of the property and the ratio of two values cannot be assumed to reflect the ratio of the perceived
intensities.
The choice of response scale depends on the objectives of the study and the products being studied. In
any specific case, there may be a choice among several equally good scales. Whatever response scale is
adopted, it should be easy to use, discriminating, unbiased and easily understood by the assessors (see
ISO 4121).
5.3.2 Interpretation of results
5.3.2.1 Quality of the measurements obtained
Irrespective of the response scale, the quality
...
NORME ISO
INTERNATIONALE 6658
Troisième édition
2017-07
Analyse sensorielle — Méthodologie
— Lignes directrices générales
Sensory analysis — Methodology — General guidance
Numéro de référence
©
ISO 2017
DOCUMENT PROTÉGÉ PAR COPYRIGHT
© ISO 2017, Publié en Suisse
Droits de reproduction réservés. Sauf indication contraire, aucune partie de cette publication ne peut être reproduite ni utilisée
sous quelque forme que ce soit et par aucun procédé, électronique ou mécanique, y compris la photocopie, l’affichage sur
l’internet ou sur un Intranet, sans autorisation écrite préalable. Les demandes d’autorisation peuvent être adressées à l’ISO à
l’adresse ci-après ou au comité membre de l’ISO dans le pays du demandeur.
ISO copyright office
Ch. de Blandonnet 8 • CP 401
CH-1214 Vernier, Geneva, Switzerland
Tel. +41 22 749 01 11
Fax +41 22 749 09 47
copyright@iso.org
www.iso.org
ii © ISO 2017 – Tous droits réservés
Sommaire Page
Avant-propos .v
Introduction .vi
1 Domaine d’application . 1
2 Références normatives . 1
3 Termes et définitions . 1
4 Exigences générales . 2
4.1 Informations de base . 2
4.2 Objectifs . 2
4.3 Choix de l’essai . 2
4.4 Choix et entraînement des sujets . 3
4.5 Produit à soumettre à l’essai . 4
4.6 Local d’essai . 4
4.7 Planification et conduite de l’essai . 5
5 Méthodes d’essai . 6
5.1 Généralités . 6
5.2 Essais de discrimination . 6
5.2.1 Généralités . 6
5.2.2 Essai de comparaison par paires . 7
5.2.3 Essai triangulaire . 7
5.2.4 Essai duo-trio . 8
5.2.5 Essai 2 sur 5 . 9
5.2.6 Essai «A» ou «non A». 9
5.2.7 Essai tétrade .10
5.3 Essais utilisant des échelles et des catégories .10
5.3.1 Aspects généraux à prendre en compte .10
5.3.2 Interprétation des résultats .11
5.3.3 Catégorisation .12
5.3.4 Classification .12
5.3.5 Classement par rang .12
5.3.6 Cotation ou notation .13
5.4 Essais descriptifs .13
5.4.1 Généralités .13
5.4.2 Profil sensoriel qualitatif .14
5.4.3 Profil sensoriel quantitatif.14
5.4.4 Profil par consensus.16
5.4.5 Profil libre choix .16
5.4.6 Profil flash .17
5.4.7 Profil par rapport à une référence .17
5.4.8 Dominance temporelle des sensations (DTS) .18
6 Analyse des résultats .18
6.1 Généralités .18
6.2 Essais de discrimination .19
6.2.1 Généralités .19
6.2.2 Essai de comparaison par paires (voir l’ISO 5495) .19
6.2.3 Essai triangulaire (voir l’ISO 4120) .20
6.2.4 Essai duo-trio (voir l’ISO 10399) .20
6.2.5 Essai 2 sur 5 .20
6.2.6 Essai «A» ou «non A» (voir l’ISO 8588) .20
6.2.7 Essai tétrade .21
6.2.8 Traitement des réponses «aucune différence» dans les essais
de discrimination .21
6.2.9 Effets systématiques .21
6.2.10 Approche séquentielle (voir l’ISO 16820) .21
6.3 Essais utilisant des échelles et des catégories .21
6.3.1 Généralités .21
6.3.2 Catégorisation .21
6.3.3 Classification (voir l’ISO 4121) .22
6.3.4 Classement par rangs (voir l’ISO 8587) .22
6.3.5 Essai de cotation .22
6.3.6 Essai de notation .22
6.4 Essais analytiques ou descriptifs .22
7 Rapport d’essai .22
Annexe A (informative) Termes statistiques .24
Bibliographie .26
iv © ISO 2017 – Tous droits réservés
Avant-propos
L’ISO (Organisation internationale de normalisation) est une fédération mondiale d’organismes
nationaux de normalisation (comités membres de l’ISO). L’élaboration des Normes internationales est
en général confiée aux comités techniques de l’ISO. Chaque comité membre intéressé par une étude
a le droit de faire partie du comité technique créé à cet effet. Les organisations internationales,
gouvernementales et non gouvernementales, en liaison avec l’ISO participent également aux travaux.
L’ISO collabore étroitement avec la Commission électrotechnique internationale (IEC) en ce qui
concerne la normalisation électrotechnique.
Les procédures utilisées pour élaborer le présent document et celles destinées à sa mise à jour sont
décrites dans les Directives ISO/IEC, Partie 1. Il convient, en particulier de prendre note des différents
critères d’approbation requis pour les différents types de documents ISO. Le présent document a été
rédigé conformément aux règles de rédaction données dans les Directives ISO/IEC, Partie 2 (voir www
.iso .org/ directives).
L’attention est appelée sur le fait que certains des éléments du présent document peuvent faire l’objet de
droits de propriété intellectuelle ou de droits analogues. L’ISO ne saurait être tenue pour responsable
de ne pas avoir identifié de tels droits de propriété et averti de leur existence. Les détails concernant
les références aux droits de propriété intellectuelle ou autres droits analogues identifiés lors de
l’élaboration du document sont indiqués dans l’Introduction et/ou dans la liste des déclarations de
brevets reçues par l’ISO (voir www .iso .org/ brevets).
Les appellations commerciales éventuellement mentionnées dans le présent document sont données
pour information, par souci de commodité, à l’intention des utilisateurs et ne sauraient constituer un
engagement.
Pour une explication de la nature volontaire des normes, la signification des termes et expressions
spécifiques de l’ISO liés à l’évaluation de la conformité, ou pour toute information au sujet de l’adhésion
de l’ISO aux principes de l’Organisation mondiale du commerce (OMC) concernant les obstacles
techniques au commerce (OTC), voir le lien suivant: w w w . i s o .org/ iso/ fr/ avant -propos .html.
Le présent document a été élaboré par le Comité Technique ISO/TC 34, Produits alimentaires, sous-
comité SC 12, Analyse sensorielle.
Cette troisième édition annule et remplace la deuxième édition (ISO 6658:2005), qui a fait l’objet d’une
révision technique. Les modifications suivantes ont été apportées:
— la définition de «l’analyse sensorielle» a été révisée;
— un essai tétrade a été ajouté aux essais de discrimination, au 5.2.7;
— de nouvelles méthodes d’essai descriptif ont été ajoutées en 5.4;
— l’Article 6 a été révisé;
— le Tableau A.1 a été supprimé de l’Annexe A;
— la Bibliographie a été mise à jour et étendue.
Introduction
Le présent document constitue une introduction générale à la méthodologie de l’analyse sensorielle et il
convient de la lire avant d’appliquer les méthodes d’essai plus détaillées décrites dans d’autres Normes
internationales. Il traite du domaine général de la méthodologie et il est destiné à remplir les fonctions
suivantes:
— fournir à l’utilisateur d’essais spécifiques un bref exposé général sur les caractéristiques essentielles
des méthodes d’analyse sensorielle;
— fournir des détails sur les exigences générales, les modes opératoires et l’interprétation des résultats
communs à tous les essais ou à la plupart d’entre eux;
— constituer un guide suffisant relatif aux exigences, aux modes opératoires et à l’interprétation des
résultats des différents essais spécifiques, afin de permettre de choisir le ou les modes opératoires
les plus appropriés à la résolution d’un problème particulier.
Il comprend trois parties principales, faisant l’objet des Articles 4, 5 et 6.
Il est indispensable de lire en premier l’Article 4 («Exigences générales»). L’Article 5 («Méthodes d’essai»)
décrit les principaux essais d’une manière générale. L’Article 6 traite de certains principes généraux
relatifs à la collecte et à l’analyse de données sensorielles et aborde aussi brièvement les principes
généraux du traitement statistique des résultats.
vi © ISO 2017 – Tous droits réservés
NORME INTERNATIONALE ISO 6658:2017(F)
Analyse sensorielle — Méthodologie — Lignes directrices
générales
AVERTISSEMENT — Le présent document n’a pas pour but de traiter les problèmes de sécurité
qui peuvent être liés à son utilisation. Il est de la responsabilité de l’utilisateur d’établir des
pratiques appropriées de sécurité, de protection de la santé et de garantir la conformité vis-à-
vis des réglementations nationales en vigueur.
1 Domaine d’application
Le présent document fournit des lignes directrices générales sur la méthodologie de l’analyse
sensorielle. Il décrit des essais destinés à l’examen des produits alimentaires et d’autres produits
par analyse sensorielle et donne des informations générales relatives aux techniques à utiliser si une
analyse statistique des résultats est requise.
En général, ces essais sont uniquement destinés à une analyse sensorielle objective. Toutefois, s’il est
possible d’utiliser un essai pour déterminer la préférence au cours d’un essai hédonique, cela est précisé.
Un essai hédonique vise à déterminer l’acceptabilité des produits et/ou à déterminer les préférences
d’une population de consommateurs spécifique entre deux produits ou plus. Les méthodes sont efficaces
pour déterminer s’il existe une préférence perceptible (différence dans le degré d’appréciation), ou s’il
n’existe aucune préférence perceptible (essai de similitude par paires). Les lignes directrices générales
pour les essais hédoniques sont données dans l’ISO 11136.
2 Références normatives
Les documents ci-après, dans leur intégralité ou non, sont des références normatives indispensables à
l’application du présent document. Pour les références datées, seule l’édition citée s’applique. Pour les
références non datées, la dernière édition du document de référence s’applique (y compris les éventuels
amendements).
ISO 5492, Analyse sensorielle — Vocabulaire
3 Termes et définitions
Pour les besoins du présent document, les termes et définitions donnés dans l’ISO 5492 ainsi que les
suivants s’appliquent.
L’ISO et l’IEC mettent à jour les bases de données terminologiques utilisées en normalisation aux
adresses suivantes:
— IEC Electropedia: accessible sur http:// www .electropedia .org/
— ISO Online browsing platform: accessible sur http:// www .iso .org/ obp
3.1
analyse sensorielle
science relative à l’examen des propriétés organoleptiques d’un produit par les organes des sens
4 Exigences générales
4.1 Informations de base
Le présent article traite des exigences générales communes à toutes les situations rencontrées lors
d’analyses sensorielles et dont les principes de base sont les suivants.
a) La réponse humaine à un stimulus ne peut être isolée d’expériences précédentes ou des autres
stimuli sensoriels provenant de l’environnement. Néanmoins, l’influence de ces deux facteurs peut
être contrôlée et ses effets pris en compte.
b) La variabilité de la réponse sensorielle est inhérente à tout groupe de personnes utilisé pour les
essais et elle est inévitable; elle peut provenir d’incohérences propres à un individu et de différences
physiologiques et psychologiques entre les individus. Toutefois, avec de l’entraînement, un tel
groupe de sujets peut donner des réponses individuelles très homogènes. L’identification de ces
facteurs est importante lors de l’analyse des résultats.
c) Des biais systématiques lors d’expériences sensorielles impliquant une réponse humaine peuvent
aboutir à des données fausses et à une interprétation incorrecte qu’il peut être difficile d’identifier.
Il convient d’identifier les facteurs susceptibles d’aboutir à des biais et de les contrôler, dans toute
la mesure du possible, en concevant et en conduisant les essais de manière appropriée.
d) La validité des conclusions tirées des résultats dépend de l’essai utilisé et de la façon dont il a été
conduit, y compris le choix des questions posées.
4.2 Objectifs
Trois principaux types d’objectifs sont exposés:
a) ceux pour lesquels le but de l’essai est de classer par catégorie ou par rang, ou de décrire le(s)
produit(s);
b) ceux dont le but est de différencier deux produits ou plus; il est important ici de faire la distinction
entre la nécessité de savoir:
i) s’il existe bien une différence;
ii) quelle est l’ampleur de la différence;
iii) quel est le sens (ou la nature) de la différence;
iv) quelle est l’influence de cette différence (par exemple sur la préférence);
v) si la totalité ou seulement une partie de la population détecte une différence;
c) ceux dont le but est de vérifier que les produits sont suffisamment similaires pour être utilisés de
façon interchangeable.
En analyse sensorielle, le problème donné nécessite fréquemment une discussion ou une réflexion
sérieuse avant de choisir un essai approprié. Cela est dû au fait que le concept initial du problème peut
nécessiter d’être clarifié.
4.3 Choix de l’essai
Le choix d’un essai approprié dépend en grande partie de l’objectif de l’essai, mais nécessite également
la prise en compte de facteurs liés au produit, aux sujets, à l’environnement d’essai ainsi qu’au niveau
recherché de précision analytique et de confiance statistique des conclusions. Il convient que l’action
pouvant découler des résultats des essais soit déterminée à l’avance.
Pour chaque essai, l’Article 5 s’efforce de donner des lignes directrices pour son domaine d’application.
Des essais préliminaires peuvent s’avérer nécessaires pour confirmer la faisabilité d’un essai donné.
2 © ISO 2017 – Tous droits réservés
Du fait de la fatigue sensorielle et de phénomènes d’adaptation, il n’est possible d’évaluer qu’un nombre
limité d’échantillons au cours d’une session, nombre variable selon la nature de l’essai et le type de
produit. Certains de ces effets peuvent être atténués par des modes opératoires appropriés de rinçage
et de repos entre échantillons.
Bien que des échantillons témoins soient nécessaires dans la plupart des cas, leur utilisation limite
naturellement le nombre d’échantillons qu’il est possible d’évaluer au cours d’une session donnée.
Il convient de toujours déterminer le plan d’expériences avant de commencer les essais. Cette
recommandation prévaut notamment si le nombre d’échantillons à évaluer nécessite plus d’une session.
Il convient de sélectionner les détails des plans d’expérience dans des ouvrages spécialisés. Quelle que
soit la méthode d’essai utilisée, chaque fois qu’il est souhaitable de minimiser le nombre d’échantillons
ou de sujets, il convient d’envisager l’approche séquentielle décrite dans l’ISO 16820.
4.4 Choix et entraînement des sujets
Un jury d’analyse sensorielle constitue un véritable «instrument de mesure» et, par conséquent, les
résultats des analyses effectuées dépendent de ses membres. Le recrutement de personnes désireuses
de faire partie d’un jury nécessite donc d’être effectué avec soin et il convient de le considérer comme
un réel investissement, à la fois en termes de temps et de coût. L’appui de la hiérarchie dans l’entreprise
est nécessaire pour qu’il soit accompli.
L’évaluation sensorielle peut être effectuée par trois types de personnes, à savoir des «sujets», des
«sujets qualifiés» ou des «sujets experts». Les sujets peuvent être des «sujets naïfs», n’ayant pas à
remplir de critère de sélection ou d’entraînement précis, ou encore des personnes ayant déjà pris
part à quelques essais sensoriels («sujets initiés»). Les «sujets qualifiés» sont des sujets qui ont été
sélectionnés et entraînés pour un essai sensoriel particulier. Les «sujets experts» sont des sujets
qualifiés qui ont été sélectionnés et entraînés pour différentes méthodes d’analyse sensorielle et qui
sont capables de réaliser des évaluations sensorielles homogènes et répétables de produits d’une ou
plusieurs catégories.
Les méthodes de sélection et d’entraînement à employer dépendent des tâches et des méthodes
destinées aux «sujets qualifiés». Les modes opératoires utilisés pour entraîner les sujets aux essais
descriptifs sont différents de ceux utilisés pour entraîner les sujets aux essais de discrimination.
Les modes opératoires et les méthodes de sélection et d’entraînement des sujets sont détaillés dans
l’ISO 8586. Il convient de noter que, parfois, ces méthodes ne font que constituer un moyen de choisir
les meilleurs candidats parmi ceux disponibles, plutôt que de satisfaire à des critères préétablis. Il
convient également de noter que la sélection de sujets pour leur aptitude à discriminer et à décrire les
produits alimentaires et d’autres produits est tout à fait différente de celle utilisée pour des essais de
préférence. Dans le premier cas, il est nécessaire d’effectuer une sélection et un entraînement alors que,
dans le second cas, il est uniquement demandé au jury d’être représentatif d’une certaine partie de la
population, par exemple un groupe de consommateurs ciblés.
Si un mode opératoire de sélection doit être mis en œuvre, plusieurs critères s’avèrent importants pour
le choix des sujets:
a) capacité générale à réaliser certaines tâches sensorielles spécifiques;
b) disponibilité par rapport au travail habituel;
c) motivation (volonté et intérêt);
d) bonne santé (y compris l’absence d’allergies particulières ou de traitement médicamenteux
particuliers) et bonnes conditions dentaires et d’hygiène générale.
Pour s’assurer que les «sujets qualifiés» et «sujets sensoriels experts» continuent de remplir les critères
sur lesquels ils ont été sélectionnés, il convient de contrôler régulièrement leurs performances (voir
ISO 11132).
4.5 Produit à soumettre à l’essai
La nature du produit à soumettre à essai détermine le protocole expérimental de l’essai et peut
également avoir une incidence sur le type d’essai requis pour répondre aux objectifs. Par exemple, un
protocole dans lequel les produits alimentaires doivent être consommés chauds nécessite de prendre
en compte la vitesse de refroidissement du produit et son effet probable sur les propriétés sensorielles,
ainsi que les changements des propriétés sensorielles qui peuvent se produire lorsque le produit est
maintenu chaud avant l’essai.
Il convient que les méthodes de préparation et de présentation des échantillons soient adaptées au
produit et au problème posé.
EXEMPLE 1 Un produit qui est normalement consommé chaud est préparé selon la technique habituelle et
dégusté chaud; toutefois, dans certaines circonstances, il est permis d’utiliser des températures élevées afin
d’augmenter la facilité avec laquelle certains goûts peuvent être évalués.
EXEMPLE 2 Il convient de ne pas homogénéiser un produit normalement consommé en morceaux distincts
afin de conserver les caractéristiques de sa texture. Il est cependant nécessaire de s’assurer qu’il y a une
uniformité maximale entre les sous-échantillons destinés à chacun des sujets, ce qui inclut des tailles de portions
similaires et une composition uniforme.
Il convient d’appliquer aux échantillons pour essai les principes généraux d’échantillonnage de produit
(conformément aux Normes internationales portant sur le produit soumis à l’essai). Dans tous les cas,
il est nécessaire d’indiquer les codes d’identification des échantillons ou les numéros de lots. Il n’est
possible de tirer des conclusions valables pour un produit dans son ensemble que si les échantillons
soumis à essai sont représentatifs.
Des supports peuvent parfois être utilisés dans les essais portant sur des produits pour lesquels
l’analyse directe n’est pas possible (voir l’ISO 5497), par exemple des ingrédients alimentaires.
Dans le cas de l’évaluation de l’aspect du produit, il convient de spécifier les conditions d’éclairage.
Lorsque l’essai porte uniquement sur les différences de goût, l’effet des différences de couleur peut être
partiellement masqué en utilisant des conditions d’éclairage qui réduisent au minimum la différence de
couleur.
Il convient de choisir les récipients de manière qu’ils n’affectent pas l’essai ou les produits. Il peut s’agir
de récipients en verre ou en céramique lavables ou de récipients en papier ou en plastique jetables,
mais il ne doit pas se produire de transfert de produits chimiques qui pourraient provoquer des goûts
ou des odeurs parasites. Il convient notamment de laver les récipients lavables uniquement à l’aide de
détergents inodores, ne laissant pas de goût anormal et de les rincer à l’eau; il convient que les récipients
à base de polymères et en papier, y compris les récipients isothermes utilisés pour les échantillons
chauds ou froids, soient également inodores et ne laissent pas de goût anormal.
Les sujets peuvent utiliser des rince-bouche entre les échantillons et entre les sessions, mais il convient
de prendre soin de s’assurer qu’ils n’ont pas d’incidence sur le goût des produits à évaluer. Il est possible
d’utiliser de l’eau plate et de l’eau gazeuse ainsi que des produits neutres (par exemple des biscuits
sans sel) entre les échantillons et entre les sessions. Il est souhaitable de vérifier l’alimentation en eau,
afin de s’assurer qu’elle est insipide et propre à la consommation. Dans certains cas particuliers, il est
possible d’utiliser de l’eau désionisée, de l’eau distillée dans un appareil en verre, de l’eau de source
faiblement minéralisée, de l’eau filtrée sur charbon ou de l’eau du robinet bouillie, mais il faut noter
qu’elles sont susceptibles d’avoir des goûts différents.
4.6 Local d’essai
Les analyses sensorielles doivent être menées dans un local d’essai prévu à cet effet (voir les
recommandations de l’ISO 8589). Le but recherché doit être de créer, pour chaque sujet, un
environnement séparé offrant le moins possible de distractions, de façon que chaque sujet puisse
s’adapter rapidement à la nature de sa (ses) nouvelle(s) tâche(s). Pendant les essais, il convient de
ne pas permettre d’activités annexes, y compris la préparation des échantillons, car elles peuvent
biaiser les résultats. Il convient que la pièce soit à une température confortable et bien ventilée avec
de l’air inodore; un débit d’air limité est souhaitable afin d’éviter des variations de températures trop
4 © ISO 2017 – Tous droits réservés
importantes. Les odeurs persistantes, comme les odeurs de tabac ou de cosmétiques, ne doivent pas
contaminer l’environnement du local d’essai.
Il convient de limiter le bruit. Un bruit de fond assez bas est généralement plus tolérable qu’un bruit
dont le niveau change. Une conversation est plus gênante qu’un bruit de fond. Ce sont les interruptions
qui causent le plus de perturbations.
Il est généralement utile d’avoir un contrôle simultané de la couleur et de l’intensité de l’éclairage, bien
que les lumières colorées ne réussissent que rarement à masquer complètement les différences d’aspect.
Les surfaces doivent être non absorbantes et conçues de manière à garantir un niveau d’hygiène élevé.
Les dimensions des cabines de dégustation sont importantes; des plafonds très bas et des cabines très
étroites peuvent créer une impression d’oppression ou un phénomène de claustrophobie. Des sièges
confortables sont nécessaires.
Si des dispositions sont prises pour une saisie informatisée des données, il convient de le faire en toute
sécurité, dans des conditions d’hygiène et de manière à ne pas compromettre le jugement sensoriel.
4.7 Planification et conduite de l’essai
Les objectifs du programme, en fonction de l’essai choisi et des contraintes pratiques associées à
l’intervention de sujets humains, déterminent la planification et la conduite de l’essai. Il est notamment
important de reconnaître les biais qui peuvent être inhérents à l’essai choisi et de mener l’essai de
manière à réduire au minimum les effets de n’importe quel biais. Les biais potentiels peuvent avoir des
causes à la fois psychologiques et physiologiques.
Les biais psychologiques les plus sérieux proviennent de sujets s’influençant mutuellement et il convient
de les réduire au minimum en utilisant des cabines de dégustation individuelles ou des séparations
adéquates entre les sujets. De plus, il est nécessaire d’organiser de manière stricte les activités des sujets.
Le mode et l’ordre de présentation des échantillons sont des aspects importants de l’essai et peuvent
introduire des biais psychologiques. Il convient, par exemple, de coder les échantillons à l’aide de
nombres à trois chiffres pris au hasard et de changer ces codes à chaque essai. L’ordre d’évaluation
peut également créer des biais et il convient généralement de préciser cet ordre. Avec un petit nombre
d’échantillons et de sujets, l’ordre peut être équilibré de manière que tous les ordres possibles se
présentent un même nombre de fois. Lors d’essais plus importants, l’ordre peut être équilibré ou laissé
au hasard.
Les biais physiologiques sont fréquemment associés à la nature des échantillons pour essai. En
particulier, l’exposition répétée à un stimulus de goût spécifique peut créer une adaptation à ce stimulus
et mâcher des aliments solides peut provoquer de la fatigue. Ces deux facteurs peuvent imposer une
limite supérieure quant au nombre d’échantillons à évaluer par session. Avec des jurys entraînés, il peut
être préconisé de recracher les échantillons, mais il peut en résulter une perte d’informations portant
sur des propriétés sensorielles spécifiques.
L’état de faim ou de satiété peut influencer les performances du sujet et, si les jurys sont sollicités trop
fréquemment, leurs performances peuvent se détériorer. Dans la mesure du possible, il convient de
demander aux sujets de s’abstenir de fumer et de consommer des encas, comme le café, au moins 1 h
avant l’essai. Les sujets ne doivent pas introduire d’odeurs étrangères lors de la session, des odeurs de
tabac ou de cosmétiques par exemple, car cela pourrait influencer les réponses des autres sujets.
L’heure du jour à laquelle se déroulent les essais est importante. Il convient que les horaires de l’essai
tiennent compte des heures de repas locales habituelles, étant donné que l’acuité maximale est
obtenue au milieu de la matinée et de l’après-midi. Il convient que les sujets enrhumés, les sujets se
trouvant sous le coup d’une émotion ou souffrant d’autres maladies soient exclus des essais jusqu’à leur
rétablissement.
La collecte des résultats comporte trois aspects:
— s’assurer que toutes les données ont été enregistrées avec exactitude soit sur ordinateur, soit
manuellement;
— vérifier que toutes les informations pertinentes complémentaires pouvant aider ou, au contraire,
créer un doute sur l’interprétation des résultats ont été notées;
— s’assurer que les sujets sont toujours motivés pour continuer à participer aux essais, si nécessaire.
5 Méthodes d’essai
5.1 Généralités
Les essais les plus courants sont divisés en trois groupes:
a) les essais de discrimination utilisés pour déterminer la probabilité de différence ou de similitude
entre les produits;
b) les essais utilisant des échelles et des catégories, pour estimer l’ordre ou l’ampleur des différences
ou les catégories ou les classes dans lesquelles il convient de répartir les échantillons;
c) les essais descriptifs, utilisés pour caractériser, de manière qualitative et quantitative, les attributs
sensoriels particuliers présents dans un échantillon (voir 5.4).
La plupart de ces essais, moyennant quelques modifications, peuvent être utilisés pour les essais de
[27]
consommateur .
Le nombre de sujets peut être déterminé en se référant aux normes correspondantes, en considérant les
risques α ou β qui sont fonction de l’objectif de l’essai. L’analyse séquentielle (voir l’ISO 16820) constitue
une alternative qui peut permettre de prendre une décision avec un plus petit nombre d’essais qu’avec
une approche conventionnelle qui utilise un nombre prédéterminé d’évaluations.
5.2 Essais de discrimination
5.2.1 Généralités
Avant d’effectuer un essai de discrimination, l’objectif de l’essai doit être précisé:
— montrer l’existence d’une différence significative entre deux produits;
— montrer l’existence d’une similitude significative entre deux produits.
Dans le premier cas, un essai de différence est utilisé. Dans le second, un essai de similitude est utilisé.
Les essais suivants sont couramment utilisés pour déterminer la probabilité de différence ou de
similitude entre les échantillons:
a) essai de comparaison par paires (voir 5.2.2);
b) essai triangulaire (voir 5.2.3);
c) essai duo-trio (voir 5.2.4);
d) essai 2 sur 5 (voir 5.2.5);
e) essai «A» ou «non A» (voir 5.2.6);
f) essai tétrade (voir 5.2.7).
Pour tous ces essais, il y a différentes manières d’analyser les résultats.
6 © ISO 2017 – Tous droits réservés
5.2.2 Essai de comparaison par paires
5.2.2.1 Définition
Au cours de cet essai, les échantillons sont présentés par paires en vue de les comparer et de détecter
des différences sensorielles perceptibles ou une similitude, en se basant sur des critères définis.
Pour les détails, se reporter à l’ISO 5495.
5.2.2.2 Application
L’essai de comparaison par paires est recommandé pour:
a) déterminer s’il existe ou non une différence perceptible (objectif de différence), pour un attribut
particulier (par exemple sucré), ou déterminer s’il existe une similitude pour cet attribut;
b) sélectionner, entraîner et surveiller les performances des sujets;
c) comparer deux produits en terme de préférence, dans le cadre des essais de consommateurs.
Les avantages de cet essai par rapport à d’autres essais de discrimination sont sa simplicité et la
moindre fatigue sensorielle procurée.
L’inconvénient de la méthode des comparaisons par paires est que dès que le nombre d’échantillons
à comparer s’accroît, le nombre d’inter-comparaisons nécessaires devient rapidement impossible à
réaliser.
5.2.2.3 Mode opératoire
Les sujets reçoivent un jeu de deux échantillons (c’est-à-dire une paire). Ils désignent l’échantillon qu’ils
considèrent le plus intense sur l’attribut considéré, même si ce choix n’est qu’une supposition. L’un des
échantillons (de la paire) peut être un témoin. Le nombre de fois où chaque échantillon est sélectionné
est comptabilisé, et la signification est déterminée en se référant à un tableau statistique.
Avant de mener l’essai, il est nécessaire de déterminer
a) son objectif (s’agit-il d’un essai de similitude ou d’un essai de différence?),
b) quelle est la sensibilité la plus adaptée, et
c) si le test statistique qui suit est
i) unilatéral (le superviseur de l’essai s’attend à un sens particulier de la différence et l’hypothèse
alternative correspond à l’existence d’une différence dans ce sens), ou
ii) bilatéral (le superviseur de l’essai ne s’attend pas à un sens particulier de la différence et
l’hypothèse alternative correspond à une différence dans l’un ou l’autre sens).
Il convient de ne pas associer les questions portant sur une différence et une préférence: les critères de
recrutement des jurys sont différents pour ces deux questions.
5.2.2.4 Analyse des résultats
Voir 6.2.2.
5.2.3 Essai triangulaire
5.2.3.1 Définition
Cet essai de discrimination consiste à présenter simultanément trois échantillons codés, dont deux sont
identiques. Les sujets sont priés d’identifier l’échantillon différent.
Pour les détails, se reporter à l’ISO 4120.
5.2.3.2 Application
L’essai triangulaire est recommandé:
a) lorsque la nature de la différence est inconnue; et
b) pour la sélection et l’entraînement des sujets.
Il convient de ne pas utiliser cet essai pour la détermination de préférences. L’essai présente les quelques
inconvénients suivants:
— il est d’un coût assez élevé lorsqu’il faut évaluer de nombreux échantillons;
— l’incidence de l’adaptation sensorielle peut être plus importante sur cet essai que sur l’essai de
comparaison par paires, dans le cas d’échantillons ayant des attributs sensoriels prononcés;
— si la nature de la différence est connue, il est statistiquement moins efficace que certains autres essais;
— la méthode n’est applicable que si les produits présentés sont d’aspect indiscernable.
5.2.3.3 Mode opératoire
Un jeu de trois échantillons codés, parmi lesquels deux sont identiques, est présenté aux sujets qui sont
priés de sélectionner l’échantillon différent.
Il convient de présenter les échantillons un nombre égal de fois dans chacune des positions
correspondant aux deux jeux de trois présentations qui sont:
BAA ABA AAB
ABB BAB BBA
5.2.3.4 Analyse des résultats
Voir 6.2.3.
5.2.4 Essai duo-trio
5.2.4.1 Définition
Il s’agit d’un essai de discrimination au cours duquel l’échantillon de référence est présenté en premier.
Il est suivi de deux échantillons dont l’un est identique à l’échantillon de référence et que les sujets sont
priés d’identifier.
Pour les détails, se reporter à l’ISO 10399.
5.2.4.2 Application
L’essai duo-trio est utilisé pour déterminer s’il existe une différence ou une similitude sensorielle
entre un échantillon donné et un échantillon de référence. Cet essai est spécialement adapté lorsque
l’échantillon de référence est bien connu des sujets, par exemple un échantillon d’une production
régulière.
S’il y a des sensations résiduelles prononcées, cet essai est moins adapté que l’essai de comparaison par
paires (5.2.2) ou que l’essai «A» ou «non A» (5.2.6).
8 © ISO 2017 – Tous droits réservés
5.2.4.3 Mode opératoire
L’échantillon de référence identifié est présenté en premier aux sujets. Deux échantillons codés, dont
l’un est identique à l’échantillon de référence, sont ensuite présentés. Les sujets sont priés d’identifier
l’échantillon identique à l’échantillon de référence.
5.2.4.4 Analyse des résultats
Voir 6.2.4.
5.2.5 Essai 2 sur 5
5.2.5.1 Définition
Cet essai de discrimination comprend cinq échantillons codés, dont deux sont d’un type et trois d’un
autre. Les sujets sont priés de regrouper les échantillons en deux jeux.
5.2.5.2 Application
L’essai 2 sur 5 est recommandé pour établir une différence de façon plus économique qu’en effectuant
d’autres essais (la méthode est statistiquement plus efficace).
Cet essai présente des inconvénients similaires à ceux de l’essai triangulaire (5.2.3). Il est plus fortement
affecté par la fatigue sensorielle et par les effets de mémorisation mais il a une plus grande puissance
statistique. Il est principalement utilisé dans des applications tactiles, visuelles ou auditives.
5.2.5.3 Mode opératoire
Un jeu de cinq échantillons codés est présenté aux sujets, en leur indiquant que deux sont d’un type et
trois d’un autre. Les sujets sont priés de regrouper les échantillons en deux jeux.
Lorsque le nombre de sujets est inférieur à 20, il
...
Questions, Comments and Discussion
Ask us and Technical Secretary will try to provide an answer. You can facilitate discussion about the standard in here.
Loading comments...