FT 209: SENSORY EVALUATION OF

FOODS
Components
❖ Part I: Sensory Evaluation (Ms. Alice
Gomezulu)

❖ Part II: Product Development (Prof. Msuya)

 Prerequisites
❖ MTH 101

❖ MTH 201
 Learning outcomes
Upon successful completion of the course students should be able to:
❖ Understand the principles of physiology upon which methods of

sensory testing are based

❖ Understand the importance sensory evaluation in the product

development and quality control and quality assurance

❖ Understand the principles of good sensory practices

❖ Understand the test methods involved in discrimination,

descriptive analysis and consumer testing of food products
❖ Sensometrics (Analysis and interpretation of sensory data)
COURSE CONTENT
1. Introduction
2. Sensory attributes
3. Our Senses
❖ How human sense work

❖ The sense of taste

❖ The of smell

❖ The sense of touch, hearing and vision

COURSE CONTENT CONT..
3. Principles of Good Practice
❖ Sensory testing environment

❖ Test protocol considerations

❖ Experimental Design

❖ Panelist considerations

❖ Tabulation and Analysis

4. Factor Affecting Sensory Measurements

COURSE CONTENT CONT..
5. Methods for Sensory Testing
❖ Difference (Discriminatory) test

✔ Paired comparison test

✔ Multiple Comparison test

✔ Triangle test

✔ Duo trio test

COURSE CONTENT CONT..
❖ Acceptance and Preference testing
✔ Paired comparison test
✔ Hedonic test
✔ Ranking test

❖ Descriptive Sensory Analysis

6. Sensometrics (Analysis and interpretation of

sensory data)
 INTRODUCTION
What is sensory evaluation?
❖ A scientific method used to evoke, measure,

analyze and interpret those responses to

products as perceived through the senses of
sight, smell, touch, taste and hearing

❖ It is a science of judging and evaluating the

quality of a food by the use of the senses, i.e.
taste, smell, sight, touch and hearing
INTRODUCTION CONT…..
❖ The principles and practices of sensory evaluation
involve each of the four activities mentioned in this
definition

❖ “To evoke,” sensory evaluation gives guidelines for

the preparation and serving of samples under
controlled conditions so that biasing factors are
minimized
 INTRODUCTION CONT…..
❖ For example, people in a
sensory test are often
placed in individual test
booths so that the
judgments they give are
their own and do not
reflect the opinions of
those around them
INTRODUCTION CONT…..
❖ Samples are labeled
with random numbers
so that people do not
form judgements based
upon labels, but rather
on their sensory
experiences
INTRODUCTION CONT…..
❖ “To measure,” sensory evaluation is a quantitative
science in which numerical data are collected to
establish lawful and specific relationships between
product characteristics and human perception

❖ Sensory evaluation is a science of measurement;

Like other analytical test procedures, sensory
evaluation is concerned with precision, accuracy,
sensitivity, and avoiding false positive results
(Meiselman, 1993)
 INTRODUCTION CONT…..
❖ The 3rd process in sensory evaluation is analysis
❖ Proper analysis of the data is a critical part of sensory
testing

❖ Data generated from human observers are often highly

variable

❖ There are many sources of variation in human

responses that cannot be completely controlled in a
sensory test
 INTRODUCTION CONT…..
❖ Examples include the mood and motivation of the
participants, their innate physiological sensitivity to
sensory stimulation, and their past history and
familiarity with similar products

❖ In order to assess whether the relationships observed

between product characteristics and sensory responses
are likely to be real, and not merely the result of
uncontrolled variation in responses, the methods of
statistics are used to analyze evaluation
 INTRODUCTION CONT…..
❖ The 4th process in sensory evaluation is the
interpretation of results

❖ A sensory evaluation exercise is necessarily an

experiment

❖ In experiments, data and statistical information are

only useful when interpreted in the context of
hypotheses, background knowledge, and implications
for decisions and actions to be taken
 INTRODUCTION CONT…..
❖ Conclusions must be drawn that are reasoned
judgments based upon data, analyses, and results

❖ Conclusions involve consideration of the method, the

limitations of the experiment, and the background
and contextual framework of the study
 INTRODUCTION CONT…
❖ It is a field of science in it own right that is and
requires an extensive skill set drawn from a range of
disciplines e.g.
◻ Biological sciences,
◻ Psychology,
◻ Experimental design and Statistics

❖ Sensory scientists will often be required to work with

other specialists from these areas
INTRODUCTION CONT…
❖ Additional challenges are presented by working
with a human ‘measuring instrument’ that is highly
variable
❖ Sensory evaluation can be divided into two
categories of testing:
✔ Objective and
✔ Subjective
 INTRODUCTION CONT…
❖ In objective testing, the sensory attributes of a
product are evaluated by a selected or trained panel

❖ In subjective testing, the reactions of consumers to

the sensory properties of products are measured

❖ The power of sensory evaluation is realized when

these two elements are combined
INTRODUCTION CONT…
❖ It reveals insights into the way in which sensory
properties drive consumer acceptance and
emotional benefits (Preference mapping)
Why use human to collect sensory data?

❖ Human perception of foods and consumer

products are results of complex sensory and
interpretation process

❖ Perceptions of such stimuli as conducted by human

nervous system are difficult or impossible to
predict from instrumental measure
 Why use human to collect sensory data?

❖ In many cases, instruments lack sensitivity of

human nervous system e.g. smell

❖ Most importantly, instrumental assessments give

values that miss an important perceptual process

❖ The perceptual process is an interpretation of sensory

experience by the human brain prior to responding
 Why use human to collect sensory data?

❖ The brain lies interpose between sensory inputs and

generation of responses that form our data

❖ Sensory experience is interpreted, evaluated while

integrating multiple inputs and finally judgement are
rendered as our data

❖ Thus there is a chain of perception rather than simply

stimulus and response
Why use human to collect sensory data?

❖ Another reason is based on the fact that, only

human sensory data provide the best model for
how consumers are likely to perceive and react
to food product in real life
Factors that affect sensory perception

❖ Individual differences – genetics

❖ Individual differences – age (children/elderly)
❖ Context
❖ Consumers’ attitudes to food
❖ Consumers’ expectations
❖ Patterns of consumption
 Factors
i) Health of Evaluator:
❖ The evaluator should be physically and mentally in

good health

❖ The sensitivity for evaluator in respect of sense of

smell and taste should be normal

❖ He/she should not be suffering from any disease

 Factors
ii) Age: Evaluators should preferably be in the age
group of 18-50 years

❖ Persons of younger age are unable to properly

interpret and communicate the sensory results,
whereas at older age the memory decreases

❖ Sharp memory of evaluator is considered highly

useful in judging products, particularly for quality
control applications
 Factors
iii) Interest and Motivation:
❖ Sensory evaluation work is very time consuming and

sometimes fatigue also.

❖ The evaluator should, therefore, have interest and be

motivated
 Factors
iv) Adaptation:
❖ Continuous exposure of evaluator to a particular

stimulus, particularly at high concentration for long

time, leads to decrease in his sensitivity(also called as
fatigue)
❖ It is therefore desirable either to give sufficient time

between the samples or use taste sanitizers, such as

brine solutions, fruits and mild acids
❖ The taste sanitizers improve the taste sensitivity or
bring it back to normal level
 Factors
v) Sampling:
❖ The sample should be representative of the lot

❖ Care shall be taken that no loss of flavour occurs and

no foreign tastes or odours are imparted during the
sample presentation

❖ The sample should be drawn from a bulk lot in such a

way that body and texture characteristics are not
changed
Factors
vi) Sample Numbers and Quantity:
❖ For economic efficiency of sensory testing, larger the number

of samples per session, the better it will be

❖ However, the size and quantity of sample for each judging

session vary from product to product

❖ Normally 5-8 samples with average intensity of flavour for
each sitting are optimum
❖ Number of samples may be increased in case of odorous

material of low intensity

❖ The amount of each sample should be about 25-50 ml or gm,

which is sufficient for one full sip or bite

Factors
vii) Score Card:
❖ The evaluation card should be simple, brief, easy to

follow and all important sensory attributes included

in it
❖ It should be clearly printed and the matter should

be arranged in logical sequence

❖ Terminology used shall be clear and understandable

Factors
viii) Miscellaneous Factors:
❖ The temperature of serving should be close to that

recommended for each product

❖ The test should be carried out preferably one hour

before or after lunch
Factors
◻ Use of materials, which are likely to vitiate results,
such as smoking, chewing gum and taking
intoxicants by the evaluator should have a time
lapse of at least 30 minutes before the test

◻ Use of strong odoriferous substances such as

cosmetics, flavours, hair oil should be avoided by
the evaluator as well as in the testing area
READING ASSIGNMENT:
ROLES OF SENSORY EVALUATION
❑ In recent years, the role of sensory evaluation has
changed significantly from a service provider used
to supply data to key players in:

Product development
Quality assurance
Marketing
Research and development
 Role of Sensory Evaluation in Product
Development

❖ Definition of product development

❖ Defined as creation of products with new or different
characteristics that offer new or additional benefits to
the consumer

❖ It may involve modification of an existing product or

its presentation, or formulation of an entirely new
product that satisfies a newly defined customer want or
market niche (www.businessdictionary.com)
 Roles of SE in PD cont….
What is a new product?
❖ New-to-the-world of products e.g. Penicillin

❖ New to product lines e.g. A dairy company starting to

produce orange juice

❖ Additions to existing product lines e.g. Additional

colours, sizes, flavours etc.
 Roles of SE in PD cont….
❖ Improvements and revisions to existing products
e.g. New and improved”

❖ Cost reductions e.g. Similar performance - lower

cost
 Roles of SE in PD cont….
❖ A typical product development process?

Product
Idea developme Scale up Launch
nt
 Roles of SE in PD cont….
❖ Product development is expensive
❖ Most newly developed products introduced in the
market fail, even if you have worked professionally
❖ Only about 10% of the products launched in the
grocery stores succeed (AC Nielsen, 2006)
❖ Product lifecycles are becoming shorter
❖ The survivors are those companies that have most
knowledge and experience in product development
 Role of SE in PD cont..
❖ Stages in a product development process

Stage I Stage II Stage III Stage IV

Product
Product Launch &
implementatio Product
strategy and Evaluation
n and commercialization
definition
marketing

Sensory Evaluation
Role of SE in PD cont..
❖ Stage I: Product strategy and definition

Consumer needs
& market Product concept and
Strategic plan
opportunity design specifications
assessment

Consumer focus groups

Descriptive profiling
Affective consumer tests
(Moskowitz, Beckley & Resurreccion, 2006)
 Role of SE in PD cont..
❖ Stage II: Product implementation and marketing

Prototype Product Market strategy &

Benchmarking
development optimization market testing

Descriptive
Descriptive
Descriptive profiling
profiling Affective
profiling Affective
consumer test
consumer test
Role of SE in PD cont..

◻ Stage III & IV: Commercialization & Launch

Scale-up & trial production Product launch & evaluation

Discrimination test
Descriptive profiling?
 Role of Sensory evaluation in Quality
Assurance
❖ One of the most usual meanings is define the quality
as “in conformity with consumer’s requirements and
acceptance

❖ Determined by their sensory attributes, chemical

composition, physical properties, and level of
microbiological and toxicological, contaminants, shelf
life, packaging and labelling
 Role of Sensory evaluation in Quality
Assurance
❖ Quality Assurance—All those planned or systematic
actions necessary to provide adequate confidence that
a product or a service will satisfy given needs

❖ Quality assurance defines the function of quality

control and its programs and procedures

❖ QA aims at achieving lower defect levels

 Role of Sensory evaluation in Quality
Assurance
❖ For quality assurance, sensory evaluation goal is to
determine if manufactured products meet
specifications set by the product development process

❖ For quality assurance purposes, unlike product

development, preference is not relevant

❖ Rather, the program measures conformity to standards

without regard to the personal preferences of the
panelists
 Role of Sensory evaluation in Quality
Assurance
❖ It is also important for quality management sensory
panelists to identify as precisely as possible the nature
of any deviation from the control

❖ To do that, training is specifically focused on

identifying deviations from set standards

❖ Panelists should also be able to identify possible

causes of any deviations found
 Role of Sensory evaluation in Quality
Assurance
❖ This enables aggressive corrective actions to be taken

❖ However, the integration of sensory evaluation into the

areas of QC/QA has achieved only moderate success
throughout the food and beverage industries

❖ Most QC sensory programs have not been as

successful as anticipated
 Role of Sensory evaluation in Quality
Assurance
❖ This is reasonable because many manufacturing
plants operate under the old axiom, if it s not broke,
don’t fix it

❖ The challenge is for QC sensory professionals that

understand the benefits of a well run sensory
program to continue to enlighten industry
management by demonstrating the efficiency and
practicality of a sensory quality model
Role of Sensory evaluation in marketing

❖ Sensory evaluation information can be vital in

making marketing decisions

❖ Data from sensory and consumer testing can inform

the society/market about product profiles,
preferences and acceptability

❖ It can also ensure that sensory properties work in

synergy with brand communication and advertising
Role of SE in Research and Development

❖ Sensory and consumer testing is widely employed

in the research field to investigate new technologies
to aid product development and to understand
consumer behavior
How to Conduct a Sensory Evaluation Study

❖ The sensory scientist in a research program or

individual study should be able to plan and conduct a
sensory study in order to achieve the required
objectives

❖ In a team, sensory specialists should be full partners

and play an active role in:
✔ Developing the research goals,
✔ Choosing the appropriate attributes to be analyzed,
✔ Selecting appropriate methods
How to Conduct a Sensory Evaluation Study?

Setting the experimental designs, which finally will

be used to answer the questions at hand
Furthermore, the scientists should be able to
interpret the obtained findings and suggest the
required actions

❖ The steps are as follows:

How to Conduct a Sensory Evaluation Study?

1. Problem definition
❖ Sensory evaluation begins by defining the nature of
the project and what is expected from the samples
Is it a new product development or product
improvement;
Is it about ingredients, equipment, or process
change; or product matching?
Does the product vary only in one or several
attributes?
Is there any attribute that may introduce sensory
preferences
 How to Conduct a Sensory Evaluation Study?

❖ Importance??
Enable to use appropriate test method and appropriate result
interpretation

2. Definition of test objective

The sensory scientists should be able determine the test objective:
Is it overall difference, attribute differences, relative preference,
acceptability, etc.
❖ Advice:

Avoid answering too many questions in one single test to

minimize possible errors
 How to Conduct a Sensory Evaluation Study?

3. Designing of the test

❖ It involves:

Selection of the test method

Selecting and training judges
Designing the accompanying score sheet
Specifying the criteria for sample preparation and
presentation
Data analysis methods
❖ Consider principles of good sensory practices (PGSP)
in each step in order to get good result
How to Conduct a Sensory Evaluation Study?

4. Conducting the test (Test proper)

❖ This follows after designing of the test

❖ The sensory scientist should be in-charge of test

❖ By ensuring that all the requirements are met as per

test design and protocol (i.e. Principles of good
conduct).
 How to Conduct a Sensory Evaluation Study?

5. Statistical data analysis

❖ Appropriate statistical data analysis should follow

conducting a test

❖ Using appropriate techniques and software

❖ Proper presentation of the results

❖ Preferably in in tabular and/or graphical forms

How to Conduct a Sensory Evaluation Study?

6. Interpreting and reporting results

❖ Last task is to interpret the results and draw

inference of the study based on the stated

objectives
❖ Recommendations are made for action where deem

important or necessary
❖ The conclusions and recommendations should be
stated clearly and concisely
PRINCIPLES OF GOOD SENSORY
PRACTICES
 PRINCIPLES OF GOOD PRACTICES

❑ A sensory scientist must control many variables in order to

get reliable results from a sensory measurement

❑ However, for convenience, the variables are categorized and

discussed into the following groups:
❖ Sensory Testing Environment
❖ Test Protocol Considerations
❖ Experimental Design
❖ Panelist Consideration
❖ Data Entry and Analysis
 1. SENSORY TESTING ENVIRONMENT

a. Location
❖ Facility should be close to the potential judges

❖ Not in the middle of areas with extraneous odours and/or

noise such as machines, markets and shops, production lines
and cafeteria kitchens e.g.
✔ Meat processing plant vs smokehouse
✔ Winery vs bottling noise

❖ Booth area must be easily accessible to the panelists

 ELEMENTS CONT…

❖ The traffic pattern of the panelists should be kept in

mind

✔ Panelists should enter and exit the facility without

passing through the preparation area

❖ Aim~ prevent information to panelist that may bias

their responses
 ELEMENTS CONT…

b. Evaluation Area
i. Room
❖ The sensory evaluation facility would need an evaluation
area for judgements of products and often times rooms are
used

❖ It is important to remember that, evaluation conducted in

a quite, uninterrupted manner the likelihood of success is
increased
 ELEMENTS CONT…
ii. The booth
❖ Importance~ Panelists should not
influence each other

❖ Not available?~ Use tables

❖ Should be arranged in such a way

that participants should not face
each other

❖ Temporary booths constructed

using boxes may be used instead
 ELEMENTS CONT…

iii. General Design

❖ If sensory evaluation is an integral
part of product development and
quality assurance cycle in the plant
or company; then permanent
sensory area should be constructed
Sensory testing environment cont…

General Design Factors

i. Appearance
❖ The colour and lightning in the testing area
should be planned to allow adequate viewing
of samples while minimizing interference

❖ Well- balanced daylight-type fluorescent bulbs

are recommended
 Sensory testing environment cont…

ii. Partitioning
❖ Evaluation area should encompass a
Discussion area,

Booth area,

Preparation area and

Waiting room for the panelists

 Sensory testing environment cont…

The Booth
❖ The booth area is the heart of the operation
❖ It should be:
Clean and professional looking

Neutral or no distracting colour

3-25 number (Space available?)

Equipped with electrical outlet

 Sensory testing environment cont…

Isolated from the preparation area

Comfortable but not too casual

Discouraged of sink installation as they might be a

source of odour contamination

The entire area should be constructed with materials

that are easy to clean and wash
 Sensory testing environment cont…

iii. Climate control (Temp, air circulation and Humidity)

❑ As far as climate control is concerned; the booth and
discussion areas should :
❖ Be climate controlled and odour free

❖ Have excellent ventilation

❖ Have minimized odour from preparation

❖ Have no extraneous odours from cleaning supplies

 Sensory testing environment cont…

❖ Be as noise free and distraction possible~

Silence please signs)

❖ Have comfortable temperature and RH for

panelists

❖ Have adequate light

❖ The lightning should be even and shadow free

 2. Test Protocol Considerations
i. Equipment for Sample Preparation
❖ Equipment used for sample preparation and presentation
should be carefully selected to reduce the introduction of
biases and errors

❖ Equipment that imparts extra odours or flavours and likely to

affect the sensory properties of the product should be
avoided

❖ Wooden materials should also be avoided for cutting or as

mixing utensils
 Test Protocol Considerations
ii. Sample serving procedure
❖ Standardization of all serving procedures
and sample preparation techniques is very
important to reduce the biases during
sample serving

❖ Failure to standardize procedure may lead

to some panelists pick out the samples that
are different based on attribute or
conditional differences leading to bias error
 Test Protocols cont…
iii. Sample size
❖ The sample size used in evaluation should be specified as different
sample sizes may lead to different results

❖ For instance, evaluating texture intensity of two same products with

different sizes may affect the scores with large pieces scoring higher
than smaller ones

❖ Sample size affect intensity of score especially texture [1 cm vs 10

cm thick]

❖ Therefore, sample size and shape used in the study should be

specified
 Test Protocols cont…
❖ Consider the following when deciding sample size to serve:
Purpose of the study

Size of the normal portion size for the product

Size of the normal mouthful size of the product

How many attributes are to be evaluated?

Possibility of manipulating the size of the product without

difficulties
 Test Protocols cont…

It is also advised to specify the minimum

amount to be eaten (sometimes)

Reasonable balance between costs associated

with products manufacturing, storage and
preparation in relation to sample size to be
served should be maintained
 Test Protocols cont…

iv. Sample serving temperature

❖ Serving temperature of the product must be specified
in test protocol

❖ Product should be served in respective containers e.g

warmed containers

❖ In dairy products e.g. milk, the temperature should

be above those of storage (15°C vs 4°C) to enhance
perception of volatile flavor
 Test Protocols cont…
❖ The temperature of samples served at ambient (room/
ordinary temp) temperature should be measured, recorded
and maintained in each session

❖ For sample served at non-ambient;

The serving temperature should be specified
Method of maintaining that temperature e.g. thermos flask,
water bath, refrigerators, freezer etc.

❖ The temperature should be checked at the time of serving to

ensure that specified temperatures were specified
 Test Protocols cont…

v. Serving containers
❖ No rigid rules that guides choice of containers for
sensory evaluation
❖ Since different conditions exist in different facilities

❖ The use of disposable containers are preferred to

permanent ones

❖ It minimizes purchasing cost of containers and washing

time
 Test Protocols cont…

❑ The sensory panelist should consider the following

when choosing and labelling the serving containers
for sensory evaluation:
❖ Choose container that is most convenient

❖ Container should not affect the sensory attributes of

the products

❖ Choose and ensure that, the ink used to label does

not impart an aroma
 Test Protocols cont…
vi. Randomization and blind
labelling

❖ Samples should be blind

labeled with random
three-digit codes to avoid
bias

❖ Sample order should be

randomized to avoid error
due to order of presentation
 Test Protocols cont…
vii. Palate Cleansing
❖ The goal of palate cleansing is to remove residual materials
from previous samples and thus do away with the carry-over
effect

❖ A range of palate cleansers have been identified:

Chocolate (dark)
Pectin solution
Table water crackers (Type of biscuits)
Warm water
Water
 Test Protocols cont…

Whole milk

Fish flesh-Effective binder of certain

odor/aroma

❖ Studies had shown that table water cracker was

the most effective across different foods
 QUESTION

How is the sense of smell cleared

while selecting a perfume?
 Test Protocols cont…

viii. 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

❖ However, in some consumer test swallowing is

acceptable
 Test Protocols cont…

❖ Some studies have found no significant effects

between swallowing and expectoration

❖ One advantage of swallowing in analytical sensory

test is the stimulation of sensory receptors in the
throat

❖ Important in some products and flavor systems

(Throat burn in pepper samples)
 Test Protocols cont…

❖ The instruction to technicians and staff should

also be very clear and preferably should be in
written form

❖ There should be no communication gaps

between the sensory specialists and the people
performing the study
 Test Protocols cont…
ix. Instruction to Panelists
❑ It is desirable to give the instructions on how to perform the
sensory evaluation which should be:

❖ Very clear and concise

❖ Both verbally, before the panelists enter the booth and in

written form on the score sheet

❖ Should be pre-tested by having someone unfamiliar with

sensory testing
 3. Panelist considerations

x. Incentives
❖ Some incentive to motivate people to participate in a
sensory study is usually necessary

❖ A guideline for motivating participation is the

concept of the token

❖ By "token", it means that the incentive is just enough

to get the person to participate in the evaluation
 Panelist considerations cont…

❖ But not so much that it is the sole reason

for participation

❖ The over paid individuals may have little

or no motivation

❖ But they are just "in it for money"

 Panelist considerations cont…

❖ Common token incentives include snacks

❖ This can serve as social or coffee-break time

for employees or staff

❖ Small gifts for repeated testing and free

company products are also common incentives
 Panelist considerations cont…

❖ For very high levels of participation, larger

gifts or social activities such as luncheon or a
holiday part is given

❖ One of the most important incentive for

participant is the management recognition
 Panelist considerations cont…

Use of Human Subjects

❖ Sensory specialists should be very aware of the
health and safety of their panelists

❖ The panelists are human subjects

❖ The specialists should know and follow the

guidelines that constrain the use of human
subject
 Panelist considerations cont…
❖ Stipulated in the declaration of Helsinki (Moris, 1966)
❖ The guidelines principally state the following:
i. It is essential the subjects give voluntary consent

ii. The subject should have the legal capacity to give the
consent

iii. The subject should be able to exercise free power of choice

about participating in the study

iv. The study should yield fruitful results for the good of society
 Panelist considerations cont…

v. The researchers should protect the rights and welfare of all

the subjects

vi. The risks to the subjects associated with the study do not
outweigh either the potential benefits to the participants or the
expected value of the knowledge sought to society

vii. Each person participating in the study had the right of the
adequate and informed consent without undue dues
 Panelist considerations cont…
Panelist recruitment
❖ The sensory specialist must make sure that the people who
are recruited know what is expected of them during the
study

❖ Small factories and also research institutions often select

their food assessors among the employers

❖ And very often personnel engaged in laboratory work will

be candidates
 Panelist considerations cont…

❖ Much information about time and product categories

should be made available to the assessor during
recruitment

❖ Individualinterview can be used to assess

commitment and motivation

❖ They must know what is expected from them during

the study
 Panelist considerations cont…
❖ Qualifications that should be emphasized during the
recruitment of assessors for a trained panel are:

i. Commitment and motivation

❖ Assessors should be committed and motivated to participate
in a trained panel

❖ Regardless of how well potential assessor performs, if he or

she is unable to attend training or evaluation sessions he or
she is of no use in the panel
 Panelist considerations cont…

ii. Sound health

❖ Assessors must be in good general health without
any defects affecting sensory perception

❖ Any physiological or health restrictions must be

documented

❖ E.g. allergies, false teeth, migraines, as these may

affect their participation in certain tests
 Panelist considerations cont…

iii. Sensory acuity

❖ Assessors should have at least normal sensory acuity
with regard to:

Detecting stimuli

Discriminating between stimuli

Recognizing and describing stimuli

 Panelist considerations cont…

❖ Some people may be very discriminating in general,

but have one or two problems
❖ Also, many average panelists will improve markedly
with training
❖ Therefore it is not necessary to have only the most
discriminating panelists at the outset of training
❖ Nevertheless, in some circumstances it may be
desirable to select assessors with high sensitivity to
certain stimuli e.g. taint assessment
 Panelist considerations cont…
iv. Personality
❖ Generally, an assessor should
Be able to work as a team, be good listener and
communicator

Be able to concentrate, train and learn, as well as be capable

of independent judgment

Have cosmopolitan preferences i.e. freedom from prejudices

in respect of a particular food product or free from food
neophibia
 Panelist considerations cont…
Training of panelists
❖ Panelists need to be trained for successful sensory evaluation
study

❖ However, the amount of training required is dependent on

the task and level of sensory acuity desired

❖ Objective test~ extensive education is required

❖ Subjective test~ no or only minimal training is required

 Panelist considerations cont…

❖ For most descriptive tests~ extensive and in-depth

training is necessary

❖ Whereas minimal to moderate training is required for

a discrimination test such as the orientation of the
panelist to task and procedures

❖ No or only minimal training is required for

subjective test methods
 Panelist considerations cont…

❖ During training phase, panelists should realize that

sensory work is difficult, requires attention and
concentration and thus they should work as a team
 4. Data Entry and Analysis
Data Entry and Analysis
❖ The collected sensory data are not useful unless analyzed

❖ Principles to be considered when explore the feasibility of

different data entry system and analysis are:

All tests used in different situations can be processed with

different specific software system

Must be planned for this from the first day of recruitment

 4. Data Entry and Analysis cont…

❖ If the computerized system has to be purchased,

then he/she consider the time to become
comfortable with using the system e.g. Unscrambler
or Panel Check Software

❖ The computerization of the sensory booth (in high

level settings) is common
METHODS FOR SENSORY
TESTING
 INTRODUCTION

❖ Several different sensory evaluation methods have

been developed

❖ The primary concern of any sensory evaluation

specialist is to insure that the test method is
appropriate to answer the questions being asked
about the product in the test
 Overview

❖ The questions are:

Is there a difference? (Discrimination)

What is the nature and magnitude of difference?

(How are they different?)

Is the difference acceptable? (Concern consumer

likes and dislikes)
 Overview

❖ For this reason, tests are usually classified according

to their purpose and most valid

❖ Three types of sensory testing are commonly used

❖ Each with different goal and each using participants

selected using different criteria
 Overview

❖ Therefore, the experimenter should be thoroughly

familiar with the advantages and disadvantages of
each method

❖ The most practical and efficient method should be

selected for each situation

❖ It should be noted that, no one method can be used

universally
 Overview

❖ The experimenter must precisely define the purpose

of the test and the information he wants to acquire
 SENSORY TEST METHODS

❖ Broadly, there are two major types of

sensory tests

Objective tests (Discrimination and

Descriptive)

Subjective tests (Affective or consumer)

Sensory Evaluation Methods

Sensory Evaluation
Methods

Objective Subjective

Discriminatio Affective/consu
Descriptive
n mer
 OBJECTIVE TEST

❖ Objective tests provides objective data on the

sensory properties of products

❖ Also known as product-oriented tests

❖ They are carried out by trained assessors

 Sensory test methods cont…

❖ The objective test is further classified into two

classes:

1. Discrimination tests

❖ Determine whether there are sensory differences

between the samples

❖ Attempt to answer question 1; Is there a difference?

 Sensory test methods cont…

2. Descriptive tests
❖ Identify the nature of a sensory difference
and/or the magnitude of the difference

❖ Attempt to answer question 2; What is the

nature and magnitude of the difference?
 Discrimination or difference tests

Introduction
❖ They are just simple difference testing procedures

❖ Merely attempt to answer weather any perceptible

difference exists between two or more types of
products

❖ These tests usually take the form of a forced choice

procedure
 Discrimination tests cont…

❖ Participants are asked to select one choice from

among a set of products in which only one is
physically different from some standard sample

❖ Often used as a preliminary step to consumer /

affective testing
 Discrimination test cont…

❖ The tests are often used when the samples are

considered to be confusable i.e their differences
are not obvious but need to be investigated

❖ They are commonly used in the following

circumstances:
Screening and training panelists

Determining sensitivity thresholds

 Discrimination test cont…

Quality assurance/quality control e.g. screening

raw materials for consistency

Investigating the effect of ingredient/process

changes e.g. for cost reduction or supply change

Preliminary assessment
 Discrimination test cont…

❖ Typically, discrimination test requires 25-40

participants

❖ Must have been screened for their sensory acuity

to common product differences and are familiar
with the test procedures

❖ A replicate test is often performed while

respondents are in sensory test facility
 Discrimination test cont…

❖ Data analysis is simple and is usually based on the

statistics of frequencies and proportions (counting
the rights and wrong)

❖ From the test results, the differences is inferred

based on the proportions of persons who are able to
choose a test with product correctly from among a
set of similar or control products
 Discrimination test cont…

❖ Statistical tables derived from binomial

distribution give the minimum number of correct
responses needed to conclude significance as a
function of the number of participant

❖ Thus a sensory technician merely needs to count

answers and refers to a table to give a simple
statistical conclusion,
 Discrimination test cont…

❖ Results can be easily and quickly reported

❖ Tables of exact probabilities from the binomial

distribution and of minimum numbers of correct
judgments necessary to reject the null hypothesis
under the binomial distribution are found in any
statistical and sensory evaluation books
 Discrimination test methods

❖ Basically there are three types of discrimination

test that are used most often

Triangle tests

Duo–trio tests

Paired comparison tests

 Triangle test

i. Introduction and objective

❖ The triangle test is a three-alternative test in which

two products are from the same batch while a third
product is different (odd sample)

❖ Judges would be asked to pick the odd sample from

among the three
 Triangle test

❖ The objective is to determine if a difference

exists between two samples

❖ This method is also called the 3 - AFC (3-

Alternative Forced Choice) test
 Triangle test cont…
❖ The ability to discriminate the differences would be inferred
from consistent correct choices above the level expected by
chance

❖ The probability of a panelist selecting a sample by chance is

1/3

❖ Thus the null hypothesis states that, the probability (Pt) of

making a correct selection when there is no perceptible
difference between the samples is one in three (Ho: Pt = 1/3)
 Triangle test cont…

❖ The alternative hypothesis states that, if

there is a perceptible difference between the
samples the population would select the odd
sample correctly more frequently than in one
third (Ho: Pt>1/3)
 Triangle test cont…
❖ The method is very useful in quality control work to
ensure that samples from different production lots are the
same

❖ It is also used to determine if ingredient substitution

or some other change in manufacturing results in a
detectable difference in the product

❖ The triangle test is often used for selecting panelists

 Triangle test cont…

i. Number of panelists
❖ Generally, 20 – 40 panelists are used for triangle
tests, although few up to 12 can be used when the
differences are large and easy to identify

❖ Larger panels are more discriminating and are

commonly used when the differences are very small
or when the aim of the test is to determine similarity
 Triangle test cont…

ii. Procedure
Description of panelists’ task

❖ Panelists are presented with three samples and told

that two samples are the same and one is different

❖ They are asked to assess the samples in the order

provided and determine which sample is ‘the odd
one out of three’
 Triangle test cont…
❖ They may also be asked to describe the difference

❖ Appropriate palate cleansers should be used between each

sample

❖ Samples are labelled with three-digit codes (blind coded)

❖ All the code numbers on the sample presented to each

panelist must be different, even though two of the samples
are identical
Questionnaire for Triangle test
 Triangle test cont…

iii Experimental design (Presentation of samples)

❖ The two different samples (A and B) are presented to the
panelist in set of three

❖ The test is counterbalanced for the identity of the odd

sample (both ABB and BAA used) and its position in tasting
(ABB, BAB, BBA)

❖ Panelist receive either two A’s and one B, or Two B’s and
one A making six possible serving orders
 Triangle test cont…

Six possible serving orders for triangle test

Panelist Order of sample presentation
First Second Third
1 457 (A) 149 (A) 840 (B)
2 149 (A) 840 (B) 457 (A)
3 840 (B) 457 (A) 149 (A)
4 404 (B) 840 (B) 457 (A)
5 840 (B) 457 (A) 404 (B)
6 457 (A) 404 (B) 840 (B)
 Triangle test cont…

❖ Each order should be presented an equal number of

times, n for a balanced serving order

❖ This is possible however; only if there are six, or multiple

of six panelists (6n)

❖ Alternatively, the order can be randomized so that each

panelist has an equal chance of receiving any of the six
possible serving orders
 Triangle test cont…

❖ The samples are presented simultaneously in the

order selected for each panelist so that the panelist
can evaluate the samples from left to right

❖ Restating of the sample is allowed

❖ It is good practice to use each possible presentation

order an equal number of times with 24–30 assessors
 Triangle test cont…

❖ However, the absolute number chosen depends on

the overall aim and the significance level selected

❖ Larger panels are more discriminating and are

commonly used when the differences are very small
or when the aim of the test is to determine similarity
 Triangle test cont…

iv. Statistical data analysis

❖ Analyses for difference tests utilize the binomial
approximation of the Normal distribution for
discrete outcomes

❖ The total number of responses correctly identifying

the ‘odd’ sample is counted and compared to
statistical tables
 Triangle test cont…

❖ The table states the minimum number of correct

identifications required (at different levels of
significance) before a significant difference can be
concluded from the test

❖ The total number of correct responses must exceed

the critical minimum value from the table.
 Discriminator Theory

❖ There are two major assumptions behind the

discriminator theory
Assumption 1: Discriminator and Non- discriminator
assumption

❖ This assumes that, there are two kinds of subjects:

Discriminators, who recognize true differences and select
the correct "odd" sample

Non-discriminators, who see no difference and guess

 Discriminator Theory cont..

Assumption 2: Correct and incorrect Non-

discriminator assumption
❖ This assumes that the Non-discriminators include people
who guess correctly and people who guess incorrectly

❖ The best estimate of the number of non-discriminators

who guess correctly is based on the chance performance
level
 Discriminator Theory cont..

Inference:
❖ The total number of correct responses by the panel
reflects the sum of the discriminators and the fraction
of non-discriminators who guessed correctly

❖ This leads to the following set of equations:

 Discriminator Theory cont..

Let, N = Total number of panelists

D = Number of discriminators
XD = Number of non- discriminators (XD = N – D)
C = Total number of panelists answering correctly
For any given triangle test; the following relationship should
apply:
N = D + XD; two equations and two unknowns
C = D + 1/3XD; N and C are unknown
C = D + 1/3(N – D); simplified to a single equation
 Discriminator Theory cont..

Worked Example:
❖ A triangle test was carried out with 45 judges and 21
chose the odd sample correctly. Test (p < 0.05), if
there exist significant difference between products.

❖ How many people were actual discriminators?

 Discriminator Theory cont..

Solution
N = 45
C = 21
Solving for D
C = D + 1/3(N – D)
21 = D + 1/3(45 – D)
21 = 2/3D + 15
6 = 2/3D
Thus; D = 9
 Discriminator Theory cont..

❖ Only nine people out of 45 (21% of the sample)

were likely to have detected the difference

❖ The other 12 who correctly chose the odd sample

likely guessed
 Class activity

A triangle test was carried out with 40 judges and 20

chose the odd sample correctly. Test (at 0.05%) if
there was a significant difference between the
samples and calculate how many people were actual
discriminators?
DUO- TRIO TEST
 DUO- TRIO TEST

Introduction and objective

❖ An alternative test to triangle test. Panelists also receive
the three samples simultaneously

❖ One sample is marked references (R) and this sample is

the same formulation as one of the two coded samples

❖ The panelists are asked to pick the coded sample that is

most similar to reference (R)
 Duo- trio test cont…

Objective
❖ Main objective~ to determine if a difference exists
between two samples

❖ However, the direction of the difference is not

indicated i.e the sensory specialist will only know the
samples are perceptibly different but not in which
attribute(s) the sample differed
 Duo- trio test cont…

Probability and hypothesis

❖ The probability of a panelist selecting a sample by
chance is 1/2

❖ Therefore, the null hypothesis states that the long-run

probability (Pt) of making a correct selection when
there is no perceptible difference between the samples
is one in two
(Ho: Pt=1/2).
 Duo- trio test cont…

❖ The alternative hypothesis is that if there is a

perceptible difference between the samples the
population would match the reference and the sample
correctly more frequently than one in two times
(Ha: Pt>1/2).
 Duo- trio test cont…

❖ The test has the same applications as the triangle

test but is less efficient because the probability of
selecting the correct sample by chance is 50%

❖ It can be used in place of the paired comparisons

test if no characteristic is specified

❖ In the duo-trio and the triangle tests the panelist

bases his judgment on any difference he can detect
Duo- trio test cont…
 Duo- trio test cont…

Procedures
❖ Assessors are presented with three samples, two blind
coded and one labelled as a ‘reference’

❖ They are asked to assess the reference sample, followed

by the two coded samples in the order provided and
determine which is the most similar (or different) to the
reference
❖ Appropriate palate cleansers should be used after
each sample
 Duo- trio test cont…

Experimental designs
❖ There are two formats to the duo trio test, namely

i. Constant reference duo-trio test and

ii. Balanced reference duo trio test.

❖ From the panelists point of view the two formats are

identical however to the sensory specialist the two
formats differ in the sample(s) used as reference
 Duo- trio test cont…

i. Constant reference duo-trio test

❖ All panelists receive the same formulation as the
reference

❖ And thus the format has two possible serving orders

(RA AB, RA BA) which should be counterbalanced
across all panelists (Table 1)
 Duo- trio test cont…

Table 1: Two possible serving orders for a Constant

reference duo-trio test

Panelist No. Order of sample preparation

First Second Third

1 RA 257 (A) 841 (B)

2 RA 841 (B) 257 (A)

 Duo- trio test cont…

❖ The test seems to be more sensitive especially if the

panelists had prior experience with the product

❖ For example, if product Y is the current available product

which is familiar to the panelists

❖ And Z is a new reformulated/modified product then a

constant reference duo-trio test with product Y as
reference would be the method of choice
 Duo- trio test cont…

ii. Balanced Reference Duo-trio test

❖ Half of the panelists receive one formulation as the
reference and the other half of the panelists receive the
other sample formulation as the reference

❖ This makes four possible serving orders (RA BA, RA AB,

RB AB, RB BA) which should be counterbalanced across the
all panelists (Table 2)
 Duo- trio test cont…

Table 2: Four possible serving orders for a Balanced

reference duo-trio test
Panelist No. Order of sample preparation

First Second Third

1 RA 257 (A) 841 (B)

2 RA 841 (B) 257 (A)

3 RB 841 (B) 257 (A)

4 RB 257 (A) 841 (B)

 Duo- trio test cont…

❖ The method is used when both products are

unfamiliar to the panelists

❖ Or when there is not a sufficient quantity of the more

familiar product to perform a constant reference duo
trio test

❖ a minimum of 32 assessors
 Duo- trio test cont…

❖ although the absolute number depends on the overall

aim and the significance level selected

Data analysis
❖ The total number of correct responses (correctly
identifying the sample that was the same as or different
to the reference) is counted and compared to statistical
tables
 Duo- trio test cont…

❖ The total number of correct responses must exceed the

critical minimum value from the table

Conclusion
❖ From a duo-trio test, the conclusion is that a significant
difference does OR does not exist between the two samples

❖ In either case, the significance level of the test, e.g. p <0.05,

must also be stated
 Duo- trio test cont…

EXAMPLE
❖ A duo-trio test was used to determine if methional could be
detected when added to cheddar cheese in amount of 0.125 ppm
and 0.250 ppm. Each tray had a control sample marked R and
two coded samples, one with methional added and one with no
methional.

❖ The duo-trio test was used in preference to the triangle test

because less tasting is required to form a judgment. This fact is
important when tasting a substance with a lingering aftertaste,
such as methional.
 Duo- trio test cont…

❖ The test was performed on two successive days using eight

judges

❖ Each day the judges were presented with two trays. One
tray contained a sample with 0.125-ppm methional two
control samples and the other contained a sample with
0.250-ppm methional and two control samples

❖ A total of 16 judgments were made at each level. The

results are shown in the following table (Table 3)
Table 3: Duo- trio test on cheddar cheese containing 0.125 and
0.250 ppm methional, test at 1% and 5% level of significance

Level of methional added, ppm

 Duo- trio test cont…

❖ 0.125 ppm = 10 out of 16 correct judgments

❖ 0.250 ppm = 14 out of 16 correct judgments

❖ Consult Statistical table, for 16 judges in a two-sample test

❖ This chart shows that 14 correct judgments are significant at the

1% level, but 10 correct judgments are not significant, even at the
5% level

❖ The conclusion is that methional added to cheddar cheese can be

detected at the 0.250ppm level but not at the 0.125-ppm level
PAIRED COMPARISONS TESTS
 Simple Paired Comparisons Test

Introduction and Objective

❖ The main objective of this test is to determine if a
difference exists between two samples with regard to a
specified attribute, e.g. sweetness, hardness and
intensity

❖ The two samples are presented to the panelist

simultaneously and panelist is asked to identify the
sample that is higher in the specified sensory attribute
 Simple paired comparison test cont….

❖ The panelist must clearly understand what the

sensory specialist means by the specified dimension

❖ Therefore, panelist should be trained to identify the

specified sensory attribute as well as be trained on
how to perform the task as described in questionnaire
Simple paired comparison test cont….
 Simple paired comparison test cont….

❖ Partly due to the fact that the panelist attention is

directed to a specific attribute, this test is very
sensitive to differences

❖ This method has applications similar to the triangle

test

❖ Fewer samples are required and there is less tasting,

but the statistical efficiency is not as great
 Simple paired comparison test cont….

❖ The probability of a panelist selecting a sample by

chance is 50%

❖ The null hypothesis states if the underlying

population cannot discriminate between the samples
then the probability of choosing sample A (that is the
PA) is equal to the probability of choosing sample B
(PB)
[ Ho; PA=PB=1/2]
 Simple paired comparison test cont….

❖ The alternative hypothesis is that if the underlying

population can discriminate the samples based on the
specified sensory attribute then the sample higher in the
specified dimension (say A) will be chosen more often as
higher in intensity of the specified dimension than the
other sample (say B)
[ HA; PA>1/2]

The test is one tailed since the experimenter knows which

sample is supposed to be higher in the specified dimension
 Simple paired comparison test cont….

Procedure
❖ Panelists are presented with two blind coded samples

❖ They are asked to assess the samples and determine

which of the two has the greatest intensity of a
specified attribute

❖ Assessors may be pre-trained on the attribute,

depending on the test objectives
 Simple paired comparison test cont….

❖ Appropriate palate cleansers should be used after each

sample

❖ Ideally, the samples should vary only in intensity of the

attribute in question, although practically this is very
hard to achieve

❖ If there are too many differences between samples, an

overall discrimination test should be used, e.g. triangle test
 Simple paired comparison test cont….

❖ The paired comparison test is rapid and easy to

use

❖ It can also be used for assessing preference

between two samples, in which case it is referred
to as a paired preference test and the question
asked is which sample is preferred
 Simple paired comparison test cont….

Experimental design:
❖ Samples are presented in pairs

❖ There are two possible orders of presentation (AB, BA)

which should be randomized across panelist with equal
number of panelist receiving either sample A or sample B
first (Table 4)

❖ A minimum of 30 assessors should be used, although some

texts vary in terms of their recommendation
Table 4 Two possible serving orders for a
paired comparison test
 Simple paired comparison test cont….

Data analysis
❖ By use of binomial statistical Table

❖ Number of corrected judgment is compared with

values in statistical table

❖ Level of significance should be stated

 Simple paired comparison test cont….

Example
❖ To determine if there was a difference in sweetness between
pineapples canned in liquid sucrose at 45 Brix or in 52%
invert syrup at 45 Brix, a paired comparisons test was used.

❖ Pineapples from each treatment were served in coded dishes

to 20 judges. Half the judges were asked to taste one sample
first, the others to taste it second.
❖ Twelve judges chose the 52% invert syrup samples as
sweeter. Test if there was a significant difference at 5%.
 Simple paired comparison test cont….

❖ Twelve judges chose the 52% invert syrup samples

as sweeter. According to Statistical Chart, in a
two-sample difference test one sample must be
selected 15 times out of 20 judgments to be
significantly different at the 5% level of
significance.

❖ The conclusion is that there is no detectable

difference in sweetness between the two
treatments.
 Simple paired comparison test cont….

Multiple paired comparisons test

❖ When there are more than two samples to be evaluated,
each must be compared with every other sample

❖ This test is called a multiple paired comparisons test

❖ The results of multiple paired comparisons cannot be

analyzed by using the same statistical tables as simple paired
comparisons

❖ Data are analyzed by analysis of variance

DESCRIPTIVE ANALYSIS
 DESCRIPTIVE ANALYSIS
Introduction
❖ DA is a sensory method that provides quantitative

descriptions of products, based on perceptions of a

group of qualified subjects

❖ The most generally useful and highly informative class

of sensory tests

❖ It attempt to provide a quantitative specification of all

the sensory attributes of a food or product
 Descriptive analysis cont….
❖ They allow the sensory scientist to:
Obtain complete sensory description of products

Identify the underlying ingredients and process

variable

Determine which sensory attributes are important

to acceptance
 Descriptive analysis cont….
Number and Characteristics of Panelists
❖ A generic descriptive analysis would usually have

between 8 and 12 panelists.

❖ Training on:

The use of reference standards

Agree on the meaning of the attributes used.
Use a quantitative scale for intensity which allows the
data to be statistically analyzed
❖ Do not ask panelist about their hedonic responses to the

products.
 Descriptive analysis cont….
Uses of descriptive test
i. Useful in any situation where a detailed specification of the
sensory attributes of a single product or comparison of the
sensory differences among several products is desired

ii. It can indicate exactly how in the sensory dimension the

competitor's product is different from yours

iii. Shelf life testing especially if judges are well trained and
are consistent over time.
 Uses of DA cont….
iv. Product development to measure how close a new
introduction is to the target or to assess suitability of
prototype products

v. Useful for troubleshooting customer complaints (But

very expensive)

vi. The data are also used for correlation with consumer
judgment for purposes of building predictive models of
factors driving like and dislikes.
 Uses of DA cont….
vi. Since they are quantitative and analytic in nature, the
sensory specifications are also sometimes examined for
correlation with instrumental measures of food properties.

❖ Descriptive should never be used with consumers

❖ Aim is for panelists to become an instrument, which will
provide quantitative data and actionable results and
recommendations
❖ Because in all descriptive methods, the panelists should be
trained to be consistent and reproducible
 Components of DA

❖ There are several components,

❖ However, only the qualitative and quantitative
components will be discussed

1. The qualitative components

Deal with characteristics of the products
It involves those perceived sensory parameters that
define the products or sensory profile
Are referred to as attributes, descriptors, or terminology
 Components of DA cont…
❖ The terms are diverse and may mean different concept to
different people

❖ Therefore, thorough training of all panelists to fully

understand and use them in a similar way is an important step
in descriptive analysis

❖ Selection of sensory attributes and corresponding definition

should be related to the real chemical and physical properties
of the products e.g viscosity in liquid foods
 Components of DA cont…
2. The quantitative components
Deals with intensity of the attributes
It defines the degree to which each of the qualitative
components is present
It involves the assignment of numerical values to a
measurement scale
Panelists should thoroughly be trained on the use of
the scale across all samples and time in a similar
ways
 Descriptive Test Methods

❖ There are several methods

Flavour profile,
Texture profile,
Quantitative descriptive analysis,
Time - intensity
free choice profiling
 Flavour profile (FP)

❖ The FP studies the overall flavour and individual

detectable flavour components of the food system.

❖ It involves the analysis of perceived aroma and

flavor characteristics of a product, their intensities,
order of appearance, aftertaste, and overall
impression
 Flavour profile (FP) cont……

❖ Panelists are selected based on:

A physiological test for taste discrimination
Taste intensity discrimination
Olfactory discrimination, and description.
❖ They should be subjected to thorough training for their
performance to be reproducible
❖ However actual test is on individual basis
❖ TP uses standardized terminology to describe the textural
parameters of any product.
 Texture profile (TP)
❖ Panelists are selected based on their ability to discriminate known
textural differences (like mechanical, geometrical, moisture etc.) in
specific products that the panel is to be trained for.

❖ For example, if the training is for analysis of liquid/semi-liquid foods

such as porridge.

❖ Then panelists are exposed to a wide range of products from this

category to provide a wide range of reference.

❖ Introduced to the underlying textural principles involved in the

structure of these products.
 Quantitative Descriptive Analysis (QDA)

❖ The most commonly used descriptive profiling

method

❖ Developed to correct some of the observed problems

connected to FP and TP

❖ It provides a complete description of sample

differences
 QDA cont..
❖ It relies heavily on:
Statistical methods to determine terms to be used,
procedures
And panelists to be used for analysis of specific
products

❖ Panelists are selected based on their ability to

discriminate differences in sensory parameters among
samples of the specific products to be trained and tested
for
 QDA cont..
❖ The training requires the use of products and
ingredients reference for terms generation

❖ Development of the consistent terms is the primary

objective during term generation

❖ However, panelist are allowed to develop their own

terms using their own understandable language
 Procedures for doing QDA
❖There are four major steps:
i. Selection of panelists
ii. Training
iii. Determination of panelist reproducibility
during training
iv. Actual test
 1. Selection of Panelists
❖ Panelist selection is a key component in making a good
panel and getting successful results

❖ The panelists are going to act as an instrument for

describing product attributes and perceived differences

❖ Observe principles of good sensory practices (GSP)

❖ Small number of panelists (8-12) is required

 Panel members cont…

❖ Must be able to demonstrate that results are reliable and

valid (screening process must be thorough and
appropriate)
❖ They can perceive differences at better than chance
between products of the same type that they will evaluate
in the actual test
❖ Poor screening will have a major impact on results
❖ There will be doubt about panelists ability to describe
differences or rate intensities with confidence
Panel members cont…
❖ If screening is thorough fewer panelists can be used
without loss of information.

❖ Not every volunteer will be suitable

 2. Training
❖ The main aim of panel training is to make panelists to have
a similar understanding of:
The terminology meaning
Scale to be used
Procedures for testing.

i. Terminology development
❖ Panelist sit and develop their own descriptors/terminology
and reach consensus
❖ Use Their own language they understand
 QDA cont…

ii. Descriptive Scale

❖ In addition to terminology generation, panelists also

develop as scale for measuring intensity.

❖ They can develop either structured line scale or

unstructured line scale.
 QDA cont…
3. Determining panelist reproducibility during
training
❖ A subgroup of samples to be used for the actual test
is served to the panelists in duplicate or triplicate
❖ Data are analyzed to determine performance of the
panel in two aspects: Discrimination ability and
reproducibility
❖ p*MSE plots are used to assess panel performance
(Figure 12)
 QDA cont…
❖ If not ok consider further training in the use of
descriptors is required to some panelist

❖ However, in an event that all panelists fail to

reproduce, then they all need to go back to the
training session
 4. Actual test
❖ Each judge performs the actual product evaluations
individually.

❖ Samples are usually served monodically where all

attributes are evaluated before next sample is served

❖ Or simultaneously where all samples are served

together and attribute evaluated at a time across
QDA cont…
❖ Panelists are asked to evaluate the sample and rate
the intensity of the attributes according to a scale
provided
 QDA cont…

❖ All good sensory practices such as use of booth,

sample coding, randomization etc. should be
applied during actual evaluation.

❖ Data are analyzed statistically by using analysis

of variance and multivariate statistical techniques
e.g. PCA, PLSR and DA
 SUBJECTIVE/CONSUMER TEST

❖ Consumer test is another important type of

sensory methods used to assess subjective
responses to a product.

❖ The investigator is interested in whether the

consumers like the product, prefer it over
another product or find the product acceptable
based on its sensory characteristics.
 Consumer test cont…
❖ Consumer testing is, therefore, a key part of the
product development process and can also be
used to develop sensory-related brand
positioning, communication and advertising

❖ Finding the right participant is a key to a

successful consumer test
 Consumer test cont…
❖ Selection of representative sample of panelists to
participate in the test must be taken into
consideration.

❖ Moreover, the panelist should probably be frequent

user of the product

❖ And they need no training as they approach product

in different frame of mind from the average
consumer.
 Consumer test cont…

❖ In general, employees should not be recruited for

consumer testing as they may be biased as a
consequence of product knowledge.

❖ Since the variability in personal preferences is

usually quite high, large numbers of participants are
usually needed (N> 100, as a rough rule of thumb).
 Consumer test cont…

❖ With such larger samples, it is possible to look for

segments or groups of consumers with different
preference patterns, rather than simply looking at
overall means for different products or other
measures of central tendency.
Approaches to consumer test
❖ There are two main approaches to consumer
sensory testing;

i. Measurement of preference (choice)

i. Measurement of acceptance (rate liking)

 I. Measurement of Preference
❖ In preference measurement the consumer has a
choice.

❖ One product is to be chosen over one or more

other products usually a paired test to see which is
liked better

❖ Preference or choice data are usually analyzed by

means of binomial distribution statistics
 Preference cont….
❖ It is widely believed that preference data are more
sensitive than rated acceptability,
❖ Because two products can get the same acceptability
rating on a category scale, but there might be a slight
preference for one over the other.
❖ Furthermore, preference tests tell little or nothing about
the overall level of acceptance.
❖ This is because one product might be preferred over the
other, but both might be unacceptable
 1. Simple Paired Preference testing
a) Introduction and objective
❖ Preference tests are choices involving comparisons

between two products or among several products

❖ If there are two product, this is known as paired

preference test.

❖ It is the simplest and most popular type of test that

looks at the appeal of product to consumer.
 Paired preference cont….
❖ The probability of selection of a specific product is one
chance in two.

❖ The null hypothesis states that when the population does

not have a preference for one product over the other,
consumer will pick each product an equal number of times
,

❖ Ho: P(A)=P(B)=1/2
 Paired comparison cont….
❖ The alternative hypothesis states that when the population
have a preference for one product over the other,
consumer will pick one product more times than the other

❖ Ha: P(A)≠P(B) ≠ 1/2

❖ The test is a two tailed since prior to the study the

experimenter does not know which sample will be
preferred by the consumer population
 Paired comparison cont….
b) Procedure
❖ The participants receive two coded samples

❖ The two samples are presented simultaneously and

the panelists is asked to identify the sample that is
preferred
❖ To simplify data analysis and interpretation, subject
must make a choice (Forced choice)
❖ Response with no preferences or equally preferred is
not an option
 Paired comparison cont….
c) Experimental design
❖ The paired preference test has two possible serving

order (AB, BA)

❖ The sequence should be randomized across panelist

with an equal number of panelist receiving either
sample A or sample B first.
 Paired comparison cont….
d) Data analysis
❖ Data analysis can be used based on 2 tailed binomial

(1 and 2)

❖ The number of correct responses is compared to the

values in statistical table
 Example
❖ In a paired preference test using 45 panelists, 24
panelists preferred sample A. Find if there was
consumer preference for one sample.

❖ From the table the minimum value for 45 panelists at

5% is 30
 II. Measurement of Acceptance
❖ In measurement of acceptance the consumer
panelists rate their liking for product on a scale
❖ It is used to determine the degree of consumer
acceptance for product.
❖ Provides an indication of the magnitude
acceptability of products.
❖ Acceptance measurement can be done on a single
products
❖ Do not require a comparison to another product
 Acceptance cont…
❖ The main goal of this test is to assess the appeal of
product to consumer on a sensory basis
❖ That is to get the consumers reaction on the basis of
appearance, aroma, taste, flavor, mouth feel and
texture.
❖ A scale that measures the sensory appeal of a
product has a distinct advantage over a simple
choice
❖ It provides some information on whether the
product is liked or not in some absolute sense
 1. Hedonic scaling: Quantification of
Acceptability
a) Introduction and objective
❖ This is the most popular acceptance measurement

❖ The main objective is to determine the level of

liking of one or more products.

❖ E.g. Ascertaining how much consumers like a new

product
 Hedonic cont…
❖ The most common hedonic scale is the 9-point
hedonic scale
❖ Category scale ranging from like extremely , through
neither like nor dislike, to dislike extremely with
varying number of categories are used.
❖ Responses on this scale are usually assigned values
from 1 to 9
❖ 1 for dislike extremely and 9 for like extremely
Hedonic cont…
Hedonic scale
9-Like extremely
8-Like very much
7-Like moderately
6-Like slightly
5-Neither like nor dislike
4-Dislike slightly
3-Dislike moderately
2-Dislike very much
1-Dislike extremely
Hedonic cont…
❖ Panelists indicate their degree of liking for each
sample by choosing the appropriate category
 Hedonic cont…
b) Procedure
❖ Participants receive coded samples of several

products

❖ And asked to evaluate the coded samples for degree

of liking, on hedonic scale

❖ They do this by checking a category on the scale

which ranges from like extremely to dislike
extremely
 Hedonic cont…
c) Experimental design (Presentation of samples)

❖ The samples are presented in identical containers, coded

with 3-digit random numbers

❖ Each sample must have a different number

❖ The sample can be randomized for each panelist or

balance if possible
 Hedonic cont…
❖ Sample may be presented all at once
(simultaneously) or one at once (monodically)

❖ Presenting the samples simultaneously is preferred

as;
It is easier to administer
It allows panelist to re - evaluate the sample
It allows comparison of the sample
 Hedonic cont…
d) Data analysis
❖ Categories are converted to numerical score

❖ Scores for each sample are tabulated and analyzed

by ANOVA

❖ To determine whether significant differences in

mean degree of liking scores exist among the
samples.
Hedonic cont…
❖ Once a significant difference has been found,
multiple comparison test can be carried out to
determine which treatment or population means
differs from each other.