HCMC University of Technology and Education

PRINCIPLES OF GOOD PRACTISE

1 Phạm Thị Hoàn, PhD.

hoanpt@hcmute.edu.vn

Ho Chi Minh City, 2019 26/2/2020

2 Outline

1 Introduction
2 The Sensory Testing Environment
3 Test Protocol Considerations
4 Experimental Design
5 Panelist Considerations
6 Tabulation and Analysis
7 Factors affecting sensory measurements

26/2/2020
3 1 Introduction

26/2/2020
4 2 The Sensory Testing Environment
 The sensory facility should be located close to potential judges but not in
the middle of areas with extraneous odors and/or noise.

 the sensory facility should be on the ground floor of a building and

that the area should be near the entrance to the complex.

 Panelists should enter and exit the facility without passing through the
preparation area or the office areas of the facility.

 it is not a good idea to have panelists wandering through the sensory area
where they may pick up information about projects or other panelists.
26/2/2020
5

26/2/2020
6
Sample preparation area
7
Testing booths

• Light
• Humidity
• Temperature
• Noise
• Strange smell
• Ability to circulate ...
 A briefing room or an office
8
9 3 Test Protocol Considerations
3.1 Sample Serving Procedures
3.2 Sample Size
3.3 Sample Serving Temperatures
3.4 Serving Containers
3.5 Carriers
3.6 Palate Cleansing
3.7 Swallowing and Expectoration
3.8 Instructions to Panelists
3.9 Randomization and Blind Labeling
26/2/2020
10 3.1 Sample Serving Procedures
 The sensory specialist should be very careful to standardize all serving
procedures and sample preparation techniques except the variable(s) under
evaluation.

 the sensory specialist should pay careful attention to the following areas
when writing the test protocol and when performing the study: the visual
appearance of the sample, sample size and shape, and sample serving
temperature.

 the sensory specialist should decide which serving containers should be

used, whether the sample should be served with a carrier, how many
samples should be served in a session, whether the panelists should rinse
their mouths between samples, whether samples are to be expectorated or26/2/2020

swallowed.
11 3.2 Sample Size
If the samples are evaluated in a discrimination test and the
appearance of the sample is not the variable under evaluation
then the samples should appear identical.

If it is not possible to standardize the appearance exactly, a

sequential monadic serving order may be used.

Sample size affected the intensity scores assigned to textural

attributes by panelists, even when the panelists were unaware
of the sample size differential. 26/2/2020
 3.2 Sample Size
12
These results make it very important that the sensory specialist
specifies the sample size and shape used in study, since the
possibility exists that a different sample size may have led to
different results.

Deciding on the sample size depends on the purpose of the

study, the normal portion size, the number of attributes have to
evaluate, the easy manipulation of the size.

However, a reasonable balance between cost associated with

the product, storage, and preparation in relation to the sample
26/2/2020

size should be maintained.

13 3.3 Sample Serving Temperatures
 The serving temperature of the product must be specified in the test
protocol.
 For example:
Meats: sand baths heated in an oven to a fixed temperature (usually 50◦C) are
used. Small glass beakers or ceramic crucibles used as holding dishes sit
embedded in the sand baths and these in turn hold the samples to be tested.
Fluid milk can be served at 15◦C instead of the more
usual 4◦C to enhance the perception of volatile flavors.
Ice cream should be tempered at –15◦C to –13◦C for
at least 12 h before serving since scooping is difficult
if the ice cream is colder. At higher temperatures the
ice cream would melt. It is also usually best to scoop
ice cream directly from the freezer immediately before
serving rather than to scoop the portions and store these 26/2/2020

in a freezer.
14 3.3 Sample Serving Temperatures
 When samples are served at ambient temperatures the sensory specialist
should measure and record the ambient temperature during each session.

 The temperatures of samples that are served at non-ambient temperatures

should be checked at the time of serving to ensure that the specified
temperatures were achieved.

 If samples are to be held for an extended period, so need to determine if

the holding period leads to changes in the sensory attributes of the product.

 if products are to be held at elevated temperatures for any period the

sensory specialist should also monitor potential microbial growth that could
26/2/2020
compromise the safety of the panelists.
15 3.4 Serving Containers
 The sensory specialist should choose the container that is
most convenient, yet the choice of container should not
negatively affect the sensory attributes of the product.

 Can use disposable containers.

26/2/2020
16 3.5 Carriers
 “Carriers” usually refer to materials that form a base or vehicle for the food
being tested, but may more broadly be considered as any other food
that accompanies the one being tested so that they are ingested (and
tasted) together.

 If there are serious consequences from missing a difference (Type II error)

then the use of a carrier that could potentially mask differences is not
recommended.
17 3.5 Carriers
 If on the other hand, a false alarm or detection of a false positive
difference (Type I error) poses serious problems, then the obscuring of a
difference by the carrier is less detrimental. The degree of realism added
by the carrier may complicate the situation, but it could prevent the
detection of a difference that might be meaningless to consumers. The
sensory specialist should discuss with the client whether the degree of
realism in the test is a concern.

 This can be very informative about the size of the perceivable difference
as well as the nature of the interactions between the carrier and the food
to be tested.

 In some cases it may be advisable to do the test both with and without the
carrier if time and resources permit.
18 3.6 Palate Cleansing
 The goal of palate cleansers should be to aid in the removal of residual
materials from previous samples.

 Ex. roast beef slices will help undo the effects of high tannin in red wine
samples.

 Lucak and Delwiche (2009) evaluated the effects of a range of palate

cleansers (chocolate, pectin solution, table water crackers, warm water,
water, and whole milk) on foods representing various tastes and mouthfeel
effects such as jelly beans (sweet), coffee (bitter), smoked sausage (fatty),
tea (astringent), spicy tortilla chip (pungent), mint (cooling), and applesauce
(non-lingering). They found that table water crackers were the only palate
cleanser effective across all representative foods. 26/2/2020
19 3.7 Swallowing and Expectoration
 In most analytical sensory tests, swallowing is avoided and samples are
expectorated. This is assumed to provide less carry-over or unwanted
influence of one product to the next. Also swallowing high-fat products can
add unnecessary calories to panelists’ diets.

 in consumer testing where acceptability is being measured, swallowing

and post-ingestion effects can affect consumers’ opinions on the products.

 There is no effect of swallowing versus expectoration on thresholds (Kelly

and Heymann (1989)).

26/2/2020
20 3.7 Swallowing and Expectoration
 A time–intensity evaluation of Yerba mate infusions by Calviño et al.
(2004) found that swallowing versus expectoration did not affect the
perceived bitterness intensity of the infusion but that spitting did increase
the rate of decay of the sensation.

 One advantage of swallowing in analytical sensory testing is the

stimulation of sensory receptors in the throat.

26/2/2020
21 3.8 Instructions to Panelists
 These should be very clear and concise.

 It is frequently desirable to give the instructions on how to perform the

sensory evaluation both verbally, before the panelists enter the booth
area, and in written form on the score sheet.

26/2/2020
22 3.9 Randomization and Blind Labeling
 Samples should be blind labeled with random threedigit codes
to avoid bias

Do not use:
• Single digit numbers
• Consecutive letters
• Same codes at consecutive sessions

26/2/2020
23 3.9 Randomization and Blind Labeling
 sample order should be randomized to avoid artifacts due to
order of presentation.

26/2/2020
24 4 Experimental Design
 At the beginning of any project the sensory specialist and all the
parties that are requesting the study should define the objective of the
study.

 Next, the sensory specialist should meticulously scrutinize the study

step by step.

 it is a good idea to write down a “skeleton” statistical analysis.

26/2/2020
25 See the tables 3.2 and 3.3 (p. 67, 68)

Step-by-step instructions for setting up the

sensory tests

26/2/2020
26 5 Panelist Considerations
 Incentives : a thing that motivates or encourages one to do something
A guideline for motivating participation is the concept of the token
incentive.

For employees, students, or academic staff who participate in a sensory

test, there are other reasons to become involved, such as positive
feelings from helping out in the testing program.

26/2/2020
27 5 Panelist Considerations

 Use of Human Subjects: Sensory specialists should be very aware of the

health and safety of their panelists.
 Panelist Recruitment : The sensory specialist must make sure that the
people who are recruited know what is expected of them during the study.

the sensory specialist may also need to make sure that each panelist
voluntarily signs an informed consent form prior to participation in the study.

26/2/2020
28 5 Panelist Considerations
 Panelist Selection and Screening
it is not necessary to have only the most highly discriminating panelists
at the outset of training.

To screen for panelists the sensory scientist should create a battery of tests
that are appropriate to the products to be evaluated and the general tasks
required of the panelists.

26/2/2020
29 5 Panelist Considerations

 Training of Panelists
The amount of training required is dependent on the task and the level of
sensory acuity desired. For most descriptive tests extensive and in-depth
training is necessary. For many discrimination tests, only minimal training is
involved.

 Panelist Performance Assessment

The performance of trained panelists used over long periods of time may
fluctuate, as the panelists become more or less motivated to participate and to
concentrate on the task at hand during evaluation sessions. Also if
people do not participate for awhile due to transfers, vacations, leaves-of-
26/2/2020
absence, etc., their performance may deteriorate and require re-training.
 6 Tabulation and Analysis
30  The advantages of computerization of the sensory booth include:
(a) the speed of receiving test results
(b) a ready interface between the data entry system and statistical and
graphing programs
(c) a reduction in the errors involved in data entry (“key punching”)

 Disadvantages include
(a) consumers may be unfamiliar with computers and ill-at-ease with
using the system. Their concentration may shift to the response system
rather than the products
(b) errors in use may go undetected if data are analyzed “automatically,”
e.g., without inspection
(c) computer programs may not be flexible enough to handle variation in
experimental designs or requirements for specific scale types.
26/2/2020
31 7. Factors affecting sensory measurements

Unlike instruments, human judgements can easily be affected by

psychological or physiological factors. The sensory professional
must be aware of these factors and ensure that the chosen
procedure and experimental design eliminate or reduce such bias.

1. Psychological factors
2. Physiological factors
3. Cultural factors
26/2/2020
 7.1 Psychological factors
32
1. Expectation error (Lỗi thông tin/mong đợi)
2. Suggestion effect (Hiệu ứng đề xuất/dự đoán-lỗi độc lập)
3. Distraction error (Lỗi phân tâm, xao nhãng)
4. Stimulus and logical error (Lỗi kích thích và logic)
5. Halo effect and proximity error (Hiệu ứng halo)
6. Habituation (Thói quen)
7. Order effect (Hiệu ứng trật tự trình bày mẫu)
8. Contrast and convergence effects (Hiệu ứng tương phản
và hội tụ)
9. Central tendency error (Lỗi xu hướng trung tâm)
26/2/2020

10.Motivation error (Lỗi thiếu nhiệt tình)

33 6.1 Psychological factors
Expectation error: Knowledge of experimental objectives,
or the samples to be evaluated, can influence an assessor’s
judgement. People tend to find what they expect to find.

✔ Provide assessors with the minimum amount of information

required to perform the test.

✔ Code samples. Use codes such as random three-digit

numbers and not letters or colours.

26/2/2020
34

Suggestion effect
Comments or noises made out loud, e.g. urghh! or
Mmmm! can influence sensory judgements.

✔ Isolate assessors during sample evaluation, e.g. use

of sensory booths.

✔ Discourage assessors from discussing samples

before or after evaluation unless instructed to do so.

26/2/2020
35
Distraction error
Assessors can be easily distracted from the task in hand,
either by stimuli in the test environment, e.g. radios and
other conversations, or by personal preoccupations, e.g.
time pressure or domestic issues.

✔ Ensure test area is quiet.

✔ Create an environment that encourages professionalism
amongst the assessors.
✖ Prohibit the use of electronic devices, e.g. mobile phones
during testing
26/2/2020
36 Stimulus and logical error
Stimulus error occurs when assessors use additional
information to make a judgement about the samples under
assessment. When this stimulus is also logically
associated with one or more of the characteristics under
evaluation, it is called logical error.

✔ Ensure sample characteristics are consistent and/or

mask irrelevant differences where possible, e.g. use of
coloured lighting, blindfolds, nose clips and ear defenders
where appropriate.
26/2/2020
37 Halo effect and proximity error
Judgements concerning the rating of one attribute may
influence the ratings of other attributes when assessors are
asked to judge several attributes at once. This is more likely
with untrained assessors.

✔ Where possible, evaluate one, or at least a limited number

of attributes, at a time.
✔ Where possible and appropriate, use trained assessors.
✔ Where appropriate, randomise the order of attribute
evaluation if several attributes have to be rated at once.

26/2/2020
39 Habituation
When assessors score similar products on a regular
basis, e.g. on quality panels, they can develop a habit of
assigning similar scores each time rather than scores
which truly represent the samples.

✔ Vary products or introduce spiked samples from time

to time.

26/2/2020
40
Order effect
The score assigned to a sample can be influenced
by the sensory character of the preceding product.

✔ Randomise or balance the order of presentation of

samples.

✔ For affective tests, use a dummy sample in position

one.

26/2/2020
41
Contrast and convergence effects
If two products in the sample set are strikingly different,
assessors may exaggerate their ratings of this difference
(contrast). If similar products are rated as part of a widely
varying sample set, then the difference between them
may be rated smaller than it actually is (convergence).

✔ Randomise or balance the order of presentation of

samples.

✔ Consider removing outlying samples from the sample

set. 26/2/2020
42
Central tendency error
When using scales, assessors tend to avoid the extremes
and confine their ratings to the middle of the scale. This is
more likely to occur with untrained assessors or when
assessors are not familiar with the product range.

✔ Train assessors in the use of the scale and expose them

to a wide product range where possible.

✔ Use a large enough scale to differentiate between the

products, particularly with untrained assessors.
26/2/2020
43
Motivation error
A motivated panellist will learn better and, ultimately,
perform more reliably. If assessors do not respect the
panel leader or product manufacturer, they may rate
samples based on how they feel. This can be an issue
when using employee panels.

✔ Respect assessors.
✔ Give regular feedback to assessors.
✔ Carry out sessions in a professional manner.

26/2/2020
44 7.2 Physiological factors

1. Adaptation
2. Perceptual interactions between stimuli
3. Physical condition

26/2/2020
 7.2 Physiological factors
45

 Adaptation
Continued exposure to a stimulus results in a decrease in sensitivity to
that stimulus and/or a change in sensitivity to other stimuli.

✔ Limit the number of samples presented.

✔ Ensure appropriate time intervals between samples to allow the sensory
system to recover;
✔ Ensure assessors take adequate breaks between single and sets of
samples; the length of break will vary dependent on sample and test type.
✔ Provide assessors with appropriate palate cleansers, which ensure
removal of any sample lingering in the oral cavity, e.g. milk rather
than water may be needed for some spicy compounds.
26/2/2020
46  Perceptual interactions between stimuli
Certain stimuli can interact to cause the following:
• Enhancement (potentiation): The presence of one substance
increases the perceived intensity of another, e.g. salt increases
perceived intensity of chicken flavour.
• Synergy: The intensity of a mixture is greater than the intensity of
the sum of the individual components, e.g. sweetness and sourness
impact on strawberry flavour.
• Suppression: The presence of one substance decreases the
perceived intensity of another, e.g. sourness reduces peach flavour.

✔ Where appropriate, employ careful experimental design to ensure

that
the effects of combined and individual stimuli are understood.
26/2/2020
47  Physical condition
Health and nutritional disorders, together with the drugs prescribed
to treat them, can affect sensory performance. Age and stress can also
impact on sensory acuity, as can the time of day.

✔ Screen assessors prior to testing or remove assessor data if medical

conditions or associated drugs affect the sensory performance.
✔ Instruct assessors to refrain from eating for at least an hour before
sensory sessions.
✔ Schedule sessions for around a similar time each day – preferably
between 10 and lunch.
✔ Monitor assessor’s performance to highlight changes in sensory ability
that may occur due to variation in physical state, e.g. age, hormonal
state mood, etc.
26/2/2020
48 7.3 Cultural factors
 When working with assessors from different cultures or geographical
location, the sensory professional needs to be aware of the impact that
cultural effects can have on sensory data.

✔ Be sensitive to coding issues.

✔ Clarify translations of sensory scales or questionnaires into other
languages, e.g. the use of back translation.
✔ Be aware of cultural tendencies – these will have an impact on many
aspects of sensory testing such as products, protocols, scale use and
feedback.
✔ Build up information on cultural norms from different cultures or
countries.
26/2/2020
Thank you !