CEN/TR 15645-3:2008

SLOVENSKI STANDARD
01-april-2008
Papir, karton in lepenka v stiku z živili - Umerjanje za preskus neustreznega vonja
ali okusa - 3. del: Suha živila
Paper and board intended to come into contact with foodstuffs - Calibration of the off-
flavour test - Part 3: Dry food
Papier und Pappe vorgesehen für den Kontakt mit Lebensmitteln - Kalibrierung für die
Geschmacksprüfung - Teil 3: Trockene Lebensmittel
Papiers et cartons destinés à entrer en contact avec les denrées alimentaires -
Etalonnage des essais de flaveur atypique - Partie 3 : Aliments secs
Ta slovenski standard je istoveten z: CEN/TR 15645-3:2008
ICS:
67.250 Materiali in predmeti v stiku z Materials and articles in
živili contact with foodstuffs
85.060 Papir, karton in lepenka Paper and board
2003-01.Slovenski inštitut za standardizacijo. Razmnoževanje celote ali delov tega standarda ni dovoljeno.

TECHNICAL REPORT
CEN/TR 15645-3
RAPPORT TECHNIQUE
TECHNISCHER BERICHT
January 2008
ICS 67.250; 85.060
English Version
Paper and board intended to come into contact with foodstuffs -
Calibration of the off-flavour test - Part 3: Dry food
Papiers et cartons destinés à entrer en contact avec les Papier und Pappe vorgesehen für den Kontakt mit
denrées alimentaires - Etalonnage des essais de flaveur Lebensmitteln - Kalibrierung für die Geschmacksprüfung -
atypique - Partie 3 : Aliments secs Teil 3: Trockene Lebensmittel
This Technical Report was approved by CEN on 13 August 2007. It has been drawn up by the Technical Committee CEN/TC 172.
CEN members are the national standards bodies of Austria, Belgium, Bulgaria, Cyprus, Czech Republic, Denmark, Estonia, Finland,
France, Germany, Greece, Hungary, Iceland, Ireland, Italy, Latvia, Lithuania, Luxembourg, Malta, Netherlands, Norway, Poland, Portugal,
Romania, Slovakia, Slovenia, Spain, Sweden, Switzerland and United Kingdom.
EUROPEAN COMMITTEE FOR STANDARDIZATION
COMITÉ EUROPÉEN DE NORMALISATION
EUROPÄISCHES KOMITEE FÜR NORMUNG
Management Centre: rue de Stassart, 36  B-1050 Brussels
© 2008 CEN All rights of exploitation in any form and by any means reserved Ref. No. CEN/TR 15645-3:2008: E
worldwide for CEN national Members.

Contents Page
Foreword.4
Introduction .5
1 Scope .6
2 Normative references .6
3 Terms and definitions .6
4 Principle.8
5 Materials and reagents.8
5.1 General.8
5.2 Test substance.8
5.3 Spiking compound.8
5.4 Solvent .8
5.5 Silicone (optional).9
5.6 Saturated sodium chloride solution .9
6 Equipment .9
6.1 General.9
6.2 Dessicators.9
6.3 Glass-weighting flasks.9
6.4 Glass rods .9
6.5 Plastic spoons.9
7 Preparation of calibration samples.10
7.1 General.10
7.2 Stability of calibration samples.10
7.3 Preparation of saturated sodium chloride solution .10
7.4 Preparation of icing sugar test portions .10
7.5 Preparation of spiking solutions.11
7.6 Preparation of calibration samples.11
8 Sensory tests .13
8.1 General test conditions.13
8.2 Sensory panel .13
8.3 Sample presentation .13
8.4 Sensory evaluation of off-flavour.14
9 Test procedure for sensory panel.14
9.1 General.14
9.2 First step: Assessment before training.15
9.3 Second step: Training of the panel.15
9.4 Third step: Assessment after training.17
10 Data collection and reporting.17
10.1 General.17
10.2 Data format .17
11 Data analysis.18
Annex A (informative) Specifications of recommended dessicators .19
Annex B (informative) Experimental design for sample evaluation before training (Rep 1) .20
Annex C (informative) Experimental design for sample evaluation before training (Rep 2) .21
Annex D (informative) Example of evaluation form for use during assessments (first step and
third step) .22
Annex E (informative) Example of evaluation form for use during training (second step).23
Annex F (informative) Experimental design for sample evaluation after training (Rep 1) .24
Annex G (informative) Experimental design for sample evaluation after training (Rep 2).25
Annex H (informative) Example of data entry sheets for the first step .26
Annex I (informative) Example of data entry sheets for the third step.30
Annex J (informative) Mood’s median test.34
Annex K (informative) Tukey's Quick (Pocket) Test.37
Annex L (informative) Method of chemical analysis .38
Bibliography.41

Foreword
This document (CEN/TR 15645-3:2008) has been prepared by Technical Committee CEN/TC 172 “Pulp,
paper and board”, the secretariat of which is held by DIN.
Attention is drawn to the possibility that some of the elements of this document may be the subject of patent
rights. CEN [and/or CENELEC] shall not be held responsible for identifying any or all such patent rights.
Introduction
Paper and board, intended to be in contact with food, may have characteristic off-flavours that can migrate via
the airspace to the food packed in it. The purpose of testing the off-flavour of food in contact with paper and
board is to establish whether the material to be tested possesses an inherent off-flavour when kept at room
temperature.
In order to gain reliable results from the sensory evaluation, the performance of a sensory panel assessing the
off-flavour within test material needs to be validated. This can be implemented through a training procedure by
using spiked calibration samples prepared according to the given instructions.
This document consists of:
 protocol to prepare the calibration samples (spiked icing sugar) for sensory evaluation of off-flavour;
 description of the training procedure for a sensory panel in the use of the calibration samples;
 instructions for sensory evaluation of calibration samples before and after training.
This guide is meant to be used in connection with the European Standard EN 1230-2. The guidance given in
this document is only a recommendation. Please note that the calibration samples, the preparation of which is
described, can be applied also in other ways than described in this document.
This guide has been devised and collaboratively tested in the context of the EU research project
CALIBSENSORY (Growth programme, Measurement and Testing activity, GRD2-2000-30015) and it is the
sole responsibility of its authors. It does in no way represent the views of the Commission or its services.
Published results of the project are available at http://www1.kcl.fi/euproj/calib.html .
1 Scope
This Technical Report specifies a written protocol to prepare calibration samples for assessment of off-flavour
(given by benzaldehyde) in a test substance representative of dry food products (icing sugar). Essentially, this
is meant to simulate the transfer of off-flavours from paper and board to a dry food product.
This Technical Report also specifies how to train the panel in the use of the calibration samples.
The general outline of the test procedure consists of sensory assessment of off-flavour in samples without
formal training by a selected panel, followed by training of the panel, and finally sensory assessment of the
samples after training by the same sensory panel.
2 Normative references
The following referenced documents are indispensable for the application of this document. For dated
references, only the edition cited applies. For undated references, the latest edition of the referenced
document (including any amendments) applies.
ISO 6658, Sensory analysis — Methodology — General guidance
ISO 8586-1, Sensory analysis — General guidance for the selection, training and monitoring of assessors —
Part 1: Selected assessors
3 Terms and definitions
For the purposes of this document, the following terms and definitions apply.
3.1
assessor
any person taking part in a sensory test
[see ISO 5492:1992]
3.2
calibration procedure
protocol of calibration samples and written instructions to train selected assessors with calibration samples i.e.
calibrate the panel
3.3
calibration sample
test substance that is spiked with spiking compound at a certain concentration
3.4
control sample
a hidden reference sample served blind coded to the assessors among the calibration samples, and prepared
according the procedure of the calibration samples but without the spiking compound
3.5
flavour
complex combination of the olfactory, gustatory and trigeminal sensations perceived during tasting
NOTE The flavour may be influenced by tactile, thermal, painful and/or kinaesthesic effects
[see ISO 5492:1992]
3.6
multicomparison test
test where the assessor is asked to give a rating of the intensity of the difference in taste between test
portions for analysis and a known reference sample
3.7
off-flavour
a typical flavour often associated with deterioration or transformation of the product
[see ISO 5492:1992]
3.8
reference sample
calibration sample without any spiking compound.
NOTE This is presented to the assessors labelled as "Reference" and served to the assessors before tha calibration
samples.
3.9
selected assessor
assessor chosen for his/her ability to perform a sensory test
[see ISO 5492:1992]
3.10
spiking compound
volatile chemical compound having a specific flavour
NOTE In this case benzaldehyde is the selected spiking compound
3.11
spiking method
method for spiking the test substance with spiking compounds
3.12
taint
taste or odour foreign to the product
[see ISO 5492:1992]
3.13
taste
sensations perceived by the taste organ when stimulated by certain soluble substances
[see ISO 5492:1992]
NOTE For simplicity, taste and flavour are used as synonyms in the European Standard EN 1230-2, though this is
not exactly in accordance with ISO 5492. The same regards taint and off-flavour.
3.14
test portion
portion of the test sample, which is directly tested by the assessor
[see ISO 5492:1992]
3.15
test substance
substance to be assessed by the assessor
NOTE It may be the food product intended to be packed, or a suitable simulant that may absorb compounds from the
packaging materials. In this case icing sugar is the selected test substance.
4 Principle
A validation of the training process for a sensory panel has to be conducted with freshly made calibration
samples. Spiking solutions shall be prepared by diluting the spiking compound (benzaldehyde) in triacetin,
until required concentrations are reached. Calibration samples have to be prepared by exposure of icing sugar
to the different spiking solutions by an absorption procedure, using glass desiccators. The panel performance
in evaluating the off-flavour intensity within the calibration samples has to be determined before and after a
formal training step. The effectiveness of training can then be determined.
The test procedure consists of 3 steps:
a) first step: assessment of calibration samples before training;
b) second step: training procedure;
c) third step: assessment of calibration samples after training.
Detailed instructions for general test procedure, sample preparation and sensory evaluation are provided
within this Technical Report, and must be carefully followed to ensure validity of results.
5 Materials and reagents
5.1 General
Icing sugar is the test substance used to absorb different concentrations of the spiking compound diluted with
the solvent. In this Technical Report, benzaldehyde shall be used as the spiking compound and triacetin as
the solvent.
Every substance and reagent shall be certified for use in foodstuffs in accordance with European Standards
and free from off-flavours.
5.2 Test substance
Icing sugar that is free from off-flavour is used as the test substance. It should be stored in an odourless
environment, with storage conditions recommended as follows:
 Temperature:  18 °C to 25 ºC;
 Relative Humidity:  20 % to 60 %.
5.3 Spiking compound
Food grade benzaldehyde with purity of ≥ 99 % is used as the spiking compound.
5.4 Solvent
Triacetin (glyceryl triacetate) shall be used as a solvent for diluting the spiking compound. The CAS number of
triacetin is 102-76-1. Recommended purity of 97 % to 100 %.
5.5 Silicone (optional)
Silicone may be used to seal the cap of the glass-weighting flasks. If silicone is used, it should be chemically
pure and odourless (or very weak odour). Note that the use of silicone is optional.
5.6 Saturated sodium chloride solution
Pure NaCl (sodium chloride) has to be used to create batches of saturated sodium chloride solutions. NaCl
shall be free from off-flavours.
6 Equipment
6.1 General
All equipment used shall be free from odour and only in use for sensory analysis (see ISO 6658). Minimum-
odour cleaning agents for cleaning shall be used. Equipment and glassware shall be cleaned between
sessions. Any glassware used shall be washed with minimum-odour cleaning agents and dried by placement
in an oven at a temperature > 100 °C and left to evaporate for at least 2 h. All other equipment (including
dessicators) shall be washed with minimum-odour cleaning agents and left to dry at room temperature.
6.2 Dessicators
The absorption process has to be carried out in a glass dessicator fitted with a porcelain plate. One dessicator
is used per concentration of the spiking compound, five in total. Ensure the same dessicator is used to
make the same concentration of sample at each session. Specifications for the recommended dessicator
are given in Annex A.
6.3 Glass-weighting flasks
Glass-weighting flasks with ground glass cap that fits into an internally ground glass body.
Dimensions: diameter 50 mm
height 35 mm
If these flasks are not available, glass flasks of similar dimensions may be used, as long as they do not
introduce off-flavours or absorb odour.
Each test portion should be prepared and served in a separate glass-weighting flask. Six glass flasks are
required for each assessor on the panel. E.g. if there are eight assessors on the panel, then 48 flasks have to
be required.
NOTE Five extra glass-weighting flasks should be required for the spiking solutions. If possible, these should be of
identical dimensions, i.e. 50 mm diameter × 35 mm height.
6.4 Glass rods
Glass rods to stir the samples.
6.5 Plastic spoons
Each sample should be assessed using a separate spoon. At each session, every assessor will need one
plastic spoon for each sample.
7 Preparation of calibration samples
7.1 General
In general, six samples will be needed to be prepared. These are summarised, along with their concentrations
of benzaldehyde, see Table 1.
During the assessment sessions (first and third step), all six calibration samples will be needed to be
prepared. During the training session (second step), only four of the samples will be needed to be prepared
(reference, S1, S2 and S4). This will impact on the amount of icing sugar test portions that is needed to be
prepared, as well as which spiking solutions are needed to be made.
Table 1 — Samples and their concentrations of benzaldehyde (ppm = mg benzaldehyde/g sugar)
sample benzaldehyde concentration first step second third step
determined by the chemical step
analyses
ppm
reference 0 x x x
S0 (control sample) 0 x x
S1 3,22 ± 0,19 x x x
S2 6,44 ± 0,26 x x x
S3 14,63 ± 0,60 x x
S4 30,32 ± 1,39 x x x
7.2 Stability of calibration samples
The concentration of benzaldehyde significantly decreases in the calibration samples within 3 d of preparation.
Sample preparation shall begin 24 h prior to assessment (see details above).
If the dimensions of the used dessicator and the porcelain plate are not exactly the same as described in
Annex A, it is strongly recommended that the concentrations of benzaldehyde in calibration samples are
checked by chemical analyses (see Annex L) to be within the limits presented in Table1.
7.3 Preparation of saturated sodium chloride solution
NaCl has to be used to create batches of saturated sodium chloride solutions. One batch is required for each
dessicator (five batches in total). For each batch, heat 1 l of distilled water up to 90 ºC, add 500 g of pure NaCl
and stir using a magnetic stirrer until dissolved. Once it has cooled to room temperature, the solution has to be
placed in the dessicators, providing a saturated solution of NaCl standing over excess salt. A relative humidity
of 75 % will prevail in the dessicators with the aid of this solution.
7.4 Preparation of icing sugar test portions
First individual test portions of icing sugar shall be prepared. For each session, as many icing sugar samples
as you need (one 4,8 g sample of icing sugar for each sample for each assessor) have to prepared.
a) prior to measure test portions of icing sugar, the container with the 470 g of icing sugar for that session
has to be shaken for approximately 30 s for homogenisation;
b) weigh (4,8 ± 0,02) g of icing sugar in each glass-weighting flask;
c) the glass-weighting flasks have to be closed immediately;
NOTE The containers may be sealed by smearing silicone around the cap (optional);
d) surface of the sample is levelled out by gently knocking each container on the work surface (table).
Keep the icing sugar samples ready, but do not place them in the dessicators until after the following step.
7.5 Preparation of spiking solutions
Once the icing sugar is ready, the spiking solutions for each concentration should be prepared.
NOTE 1 The mixtures should be prepared immediately prior to introducing the glass-weighting flasks with the icing
sugar in the dessicators.
The spiking solutions have to be prepared in glass-weighting flasks (the same size and specifications as the
glass-weighting flasks used for the icing sugar).
The spiking compound (benzaldehyde) has to be dissolved and diluted in the solvent (triacetin), (3 ± 0,003) g
total weight. The amount of benzaldehyde and triacetin for each calibration sample, along with the final
concentrations, can be found in Table 2.
Table 2 — Concentration of spiking compound in solvent ((3 ± 0,003) g total weight)
sample benzaldehyde triacetin benzaldehyde in
solvent
g g
%
reference and S0 0,000 3,000 0
S1 0,045 2,955 1,5
S2 0,090 2,910 3
S3 0,210 2,790 7
S4 0,480 2,520 16
NOTE 2 The reference and control sample (S0) should be made simultaneously in the same dessicator, so only one
3 g batch of solution (triacetin only) is required to make both of these samples.
As soon as the spiking solution is ready, the container for each spiking solution by placing the lid on the glass-
weighting flask should be closed.
7.6 Preparation of calibration samples
For each session, fresh calibration samples have to be prepared. Prepare each set of calibration samples
using a separate dessicator for each concentration, following the steps below:
a) prepare the saturated NaCl solution as described in 7.3, pour the solution in the dessicators;
b) prepare the icing sugar samples as described in 7.4;
c) prepare the spiking solutions as described in 7.5;
d) closed container with the spiking compound has to be placed immediately in the centre of the porcelain
plate inside the dessicator. Dessicators should be labelled to identify the concentration inside.
Ensure the same dessicator is used to make the same concentration of sample at each session;
 each concentration of spiking solution has to be placed in a separate dessicator;
 one spiking solution only will be needed for the reference and control samples, which have to be
made in the same dessicator;
e) the containers with icing sugar are opened (lids removed), and placed in the dessicator, arranged around
the walls of the dessicator;
 for the spiked samples (S1, S2, S3, S4), up to fifteen icing sugar samples can be placed in one
dessicator (flasks may not be stacked);
 for the non-spiked samples (the reference and control samples), both sets of test portions (maximum
of 15 each, 30 in total) can be made simultaneously in the same dessicator (flasks are allowed to be
stacked);
f) the glass-weighting flask with the spiking compound is opened and is immediately followed by closure of
the dessicator, which is stored for 24 h at room temperature (20 ± 1) °C.
After the 24 h storage period:
g) dessicator is opened and the containers with the spiking compound and the icing sugar samples are
immediately closed. This has to be done for each dessicator, one at a time;
h) bottles containing the icing sugar are then coded with the appropriate 3-digit blinding codes (as identified
by the dessicator) from the experimental design, using an odourless marker. This has to be done one
dessicator at a time to ensure samples are correctly labelled.

Key
1 Dessicator
2 Container with the spiking solution
3 Porcelain plate
4 Container with the icing sugar
5 Sodium chloride saturated solution
Figure 1 — Arrangement of icing sugar samples in the dessicator
During the entire preparation procedure samples should be handled with care to avoid any movement that
might disturb the sample. Furthermore following completion of the spiking procedure the lids on the spiked
calibration samples should not be removed until ready for dilution with water and the sensory assessment that
follows.
The calibrated samples should be evaluated within 2 h of removing the spiked samples from the dessicator.
Immediately prior to sensory assessment: the spiked icing sugar samples should be diluted by adding
1,2 ml of mineral water to each sample using a pipette. Each lid should be replaced after the water has been
added.
Each sample then must be stirred for 30 s. It may save some preparation time to ask the assessors to
perform this task. In this case, for one sample at a time, instruct the assessors to remove the lid, stir for 30 s
then replace the lid. This has to be done for each sample. Once all samples have been stirred, the sensory
assessment may begin as described in clause 8.
NOTE The instructions in Annex D and Annex E assume that the assessors should stir the samples themselves. If
this is not the case, remove this from their instructions.
8 Sensory tests
8.1 General test conditions
Basic information on general requirements, test methods and analysis of results in sensory evaluation is given
in ISO 6658 as a general guidance.
For the design of, and the conditions in, the room in which tests are conducted ISO 8589 should be consulted.
In general, the evaluation shall be carried out at room temperature in a quiet, well-ventilated room that is free
from odours.
The containers and equipment used shall be odourless and should not have any influence on the test results.
8.2 Sensory panel
Guidelines concerning the selection and training of assessors are given in ISO 8586-1. The assessors shall
be in good health and shall not suffer from the common cold at the time of the test.
Only selected, trained assessors shall be used in the sensory evaluations. The availability of assessors, their
interest and motivation, and capacity to concentrate should be ascertained prior the training procedure. The
recommended panel size is from 10 assessors to 12 assessors. A minimum of 8 assessors should be used,
and due to the restricted amount of icing sugar, a maximum of 15 assessors is imposed. The panel members
may not change during the test procedure. If the panel members are not equivalent at each step, the
missing assessors may not be replaced with other assessors.
8.3 Sample presentation
Except during the training procedure, all samples must be served to the assessors in randomised order
following an experimental design. Suggested designs are provided in the Annexes.
NOTE 1 For each session, ensure that panellist numbers are consistent (e.g. panellist number ‘1’ is the same person
each time).
Panellists should be provided with water (preferably warm) and dry crackers to use as palate cleansers
between samples.
Prior to serving the samples, the calibration samples have to be prepared and diluted with water as described
in 7.5. Samples have to be presented with their lids on.
Each assessor must receive the reference plus five test samples presented in separate glass-weighting flasks.
Samples are labelled with 3-digit random numbers and served in random order according to the appropriate
experimental design. The presentation order will thus be different for each assessor.
Once the assessors have received the samples, they should be instructed to stir the samples as follows
(unless the technician has stirred the samples prior to serving), using a separate spoon for each sample:
For one sample at a time, remove the lid, stir for 30 s, then replace the lid. Do this for each sample. Once all
samples have been stirred, the sensory assessment may begin.
NOTE 2 Assessors should ensure that the presentation order does not change while they stir the samples.
8.4 Sensory evaluation of off-flavour
The perceived off-flavour of benzaldehyde within the samples should be evaluated by using the multiple
comparison method. The reference sample must be available at all times, and the five test samples are
presented following the design order.
Assessors shall be instructed to taste the reference sample, then taste the samples in the order provided
(using a different spoon for each sample). Assessors should cleanse their palate between each sample tasted
by rinsing the mouth with water (preferably warm water) and/or eating a piece of dry cracker.
Each assessor has his/her own sample set and an evaluation form. Evaluation forms for each step of the test
procedure can be found in the Annexes, as described in clause 9.
For each sample, the assessor records the intensity of the off-flavour relative to the reference sample using
the 0 to 4 scale shown below. Assessors are allowed to re-taste the reference sample if needed.
0 = no perceptible off-flavour
1 = off-flavour just perceptible (difficult to define)
2 = weak off-flavour
3 = clear off-flavour
4 = strong off-flavour
In addition to the whole numbers (0, 1, 2, 3, 4), ½ scores may be given but only within the range of 0 to 4 (i.e.
0,5; 1,5; 2,5 and 3,5). The reference sample is defined as 0 and a blind-coded control sample is presented
among the other samples.
9 Test procedure for sensory panel
9.1 General
The test procedure for the sensory panel consists of three steps:
a) first step: assessment before training (two sessions);
b) second step: training of the panel (one session);
c) third step: assessment after training (two sessions).
Due to the constraints of sample preparation, consecutive sessions have to take place at least 2 d apart.
Depending on how the assessments are scheduled, this means that at most, two sessions should take place
in one week.
9.2 First step: Assessment before training
The first step of the test procedure is the sensory assessment of samples before training by a selected
sensory panel. Two replicate sessions are required for this step. The replicate sessions should not be closer
than on sequential days.
For each replication:
 Prepare the test samples (S0 to S4) and the reference sample, as described in clause 7. Note that the
sample preparation process must begin 24 h prior to assessment;
 Present the samples to assessors following the procedure as described in 8.3. Each assessor has to
make individual assessments from their own sample set coded with three-digit blinding codes;
 Samples should be presented to the assessors following a randomised experimental design. Suggested
designs for each replicate, including sample blinding codes, are given in Annex B and Annex C;
 Assessors shall evaluate the samples as described in 8.4. An example of a questionnaire, which includes
instructions for the participants, is shown in Annex D;
 Data may also be collected using a computerised data capture systems. If using a computerised data
capture system, a 5-point category scale from 0 to 4 should be used, which allows ½ scores to be
recorded;
 Data from paper or computerised evaluation forms could be transferred to the Excel Data Entry Sheet
(Annex I) provided, as described in clause 10.
9.3 Second step: Training of the panel
The second step of the test procedure is the training of the sensory panel. The training should take place in a
single session. During the training process all samples should be evaluated by the same selected assessors
as in the first step. If the panel members are not all present in each step, the missing assessors may not be
replaced by other assessors and should be disregarded from the training procedure.
The samples used for training are shown in Table 3. Prepare the samples following the procedure described
in clause 7. Samples should be presented labelled as shown in Table 3.
Table 3 — Samples and labels used for training
sample label shown
reference reference
S1 A
S2 B
S4 C
The training procedure consists of three activities.
a) evaluation of samples by assessors;
b) table discussion and teaching by panel leader;
c) re-tasting of samples and consolidation of learning.
The assessor should receive only one 4,8 g portion of each sample to be used for the training session. Ensure
that enough of each sample is saved to complete the three activities.
Activity 1 Evaluation of samples by assessors
 Present the four samples to the assessors. Present the samples to each assessor in the same order:
reference, A, B, C.
 Each assessor should make individual assessments from their own sample set. The assessors evaluate
the samples as described in 8.4, rinsing the mouth with warm water or eating a piece of dry cracker
between the samples.
 An example of an evaluation form that can be used during the training session, which includes
instructions for the participants, is shown in Annex E.
 For this training step, data should be collected using paper ballots, as the assessors will be working
together in the same room.
Activity 2 Table discussion and teaching by panel leader
 Round table discussions related to the calibration samples have to be performed immediately after the
evaluation. The panel leader has to initiate discussion amongst the assessors regarding the perceived
intensity of each samples, and the scores given by the assessors.
 Assessors should be trained to the reference points by teaching them which score should be given for
each calibration sample. The panel leader has to use the information in Table 4 to guide the discussion.
Table 4 — Teaching the association between intensity perception and score for each sample
labelled off-flavour intensity perception off-flavour intensity off-flavour
sample score description
on scale 0 to 4
reference no perceptible off-flavour 0

A off-flavour just perceptible (difficult to define) 1

B weak off-flavour (can define) 2

C strong off-flavour 4
Activity 3 Re-tasting of samples and consolidation of learning
 The assessors should re-taste the samples and learn to associate the perceived intensity of each sample
with the intended score.
 During this activity the samples should be evaluated in a random but in a fixed order (same order for all
assessors, as in activity 1) and should be discussed as follows:
1) Assessors should taste the sample labelled ‘Reference’ to familiarise themselves with the flavour of
the pure test material. This sample should be given a score of 0.
2) Assessors should taste the calibration sample A. This sample should be given a score of 1.
3) Assessors should taste the calibration sample B. This sample should be given a score of 2.
4) Assessors should taste the calibration sample C. This sample should be given a score of 4.
 Further discussion should take place after the samples have been re-tasted, and the assessors should
use this activity to consolidate their learning and come to consensus regarding the scores.
 At the end of the training session an agreement of the scores relating to the different flavour intensities
should be reached in order to calibrate the panel.
9.4 Third step: Assessment after training
The third step of the test procedure is the sensory assessment of samples after training. The samples should
be evaluated by the same selected assessors as in the first and second step. Two replicate sessions are
required for this step.
For each replication:
 Prepare the test samples (S0 to S4) and the reference sample, as described in clause 7. Note that
sample preparation process shall begin 24 h prior to assessment.
 Present the samples to assessors following the procedure as described in 8.3. Each assessor has to
make individual assessments from their own sample set coded with three-digit blinding codes.
 Samples should be presented to the assessors following a randomised experimental design. Suggested
designs for each replicate, including sample blinding codes, are given in Annex F and Annex G.
 Assessors have to evaluate the samples as described in 8.4. The same questionnaire should be used as
in the first step, an example of which is shown in Annex D.
 As in the first step, data may also be collected using a computerised data capture systems. If using a
computerised data capture system, a 5-point category scale from 0 to 4 should be used, which allows ½
scores to be recorded.
 Data from paper or computerised evaluation forms could be transferred to the Excel Data Entry Sheet
(Annex I) provided, as described in clause 10.
10 Data collection and reporting
10.1 General
Sensory assessment of the samples will take place in the first and third step. Using the questionnaire provided
in Annex D, assessors have to record their scores on the evaluation form in the same order in which the
samples are presented to them (i.e. following the suggested experimental design).
10.2 Data format
There are four separate data entry worksheets (all within the same Excel file).
 Assessment data for the first step (before training) Rep 1 should be entered in the column "Score
given" in the worksheet named "Before - Rep 1".
 Assessment data for the first step (before training) Rep 2 should be entered in the column "Score
given" in the sheet named "Before - Rep 2".
 Assessment data for the third step (after training) Rep 1 should be entered in the column "Score
given" in the sheet named "After - Rep 1".
 Assessment data for the third step (after training) Rep 2 should be entered in the column "Score
given" in the sheet named "After - Rep 2".
11 Data analysis
Analysis is applied both the ‘before’ and ‘after’ training results. The aim of the analysis is to evaluate the
panel’s accuracy (median scores) for each samples as well the panel’s ability to discriminate between the
adjacent samples using Mood’s Median Test.
The Mood's Median Test performs a hypothesis test of the equality of population medians. The median test
provides a non-parametric alternative to the usual one-way analysis of variance. The median test is
particularly robust against outliers and errors in data. This test reports the median values for each sample.
Confidence intervals around the median are used to evaluate significant differences between adjacent pairs of
samples. Results are summarised in tables in order to help readability and interpretation.
The before and after training results are then compared to evaluate the effectiveness of panel training.

Annex A
(informative)
Specifications of recommended dessicators
Five large glass dessicators (each 300 mm of nominal diameter) fitted with porcelain plate (300 mm in
diameter) are required for the dry food test.
The shape of the dessicator should be as shown. Specific dimensions that the dessicator should have are
given below.
Figure A.1 — Recommended dessicator
Table A.1 — Dimensions of the dessicator
measurement parameter recommended dessicator
Total volume (volume of cap + volume of body) 15,5 l
Total height of dessicator + cap 335 mm
Diameter of holes in porcelain plate 5 mm
Number of holes in porcelain plate 96

Annex B
(informative)
Experimental design for sample evaluation before training (Rep 1)
Design – Before training
This design is based on a balanced 5 present 5 Williams Design with 10 combination.
The combinations have then been repeated and randomised for the two different sessions.

sample blinding code product description
number
1 240 S0 (Control)
2 751 S1 (0,045 g)
3 311 S2 (0,09 g)
4 556 S3 (0,21 g)
5 328 S4 (0,48 g)
serving position (order)
1st
Panellist 2nd 3rd 4th 5th
1 240 328 751 556 311
2 751 240 311 328 556
3 311 751 556 240 328
4 556 311 328 751 240
5 328 556 240 311 751
6 311 556 751 328 240
7 556 328 311 240 751
8 328 240 556 751 311
9 240 751 328 311 556
10 751 311 240 556 328
11 311 240 328 751 556
12 556 751 240 311 328
13 328 311 751 556 240
14 240 556 311 328 751
15 751 328 556 240 311
Annex C
(informative)
Experimental design for sample evaluation before training (Rep 2)
Design – Before training
This design is based on a balanced 5 present 5 Williams Design with 10 combination.
The combinations have then been repeated and randomised for the two different sessions.

sample blinding code product description
number
1 472 S0 (Control)
2 817 S1 (0,045 g)
3 973 S2 (0,09 g)
...