ASTM E3041-17

Standard Guide for Selecting and Using Scales for Sensory Evaluation

SIGNIFICANCE AND USE
4.1 Rating instruments or rating scales are commonly used in many areas such as sensory evaluation, marketing research, experimental psychology, survey research, and economics in which there is interest in quantifying perceptions such as liking, preference, level of purchase interest, intensity of an attribute, degree of difference, or level of agreement with statements. This guide is concerned with the scales that are used to record human responses to physical stimuli rather than measuring physical entities. Many types of rating scales already exist and have been used in the above fields. Specific rating scales each have their own properties, advantages, and disadvantages. Some rating scales are intended for specific applications, while others have broader applications. Some rating scales have been extensively studied and modeled and have well-established properties.  
4.2 Given the overwhelming number of scales available to practitioners when designing research, it is necessary for the researcher to have some knowledge about the scales that are available along with the many considerations that surround their use and applications. This guide will be useful to the sensory researcher who wants to use a scale as a measuring tool for their work. Selecting the right scale is a critical step towards meeting the research objective and making valid conclusions.
SCOPE
1.1 The objective of this guide is to provide information to be reviewed and considered by the sensory and consumer scientist who wants to select and use scales to measure responses from consumers or trained assessors. For ease of reading, the term sensory scientist is used throughout the guide when statements apply to the sensory and consumer scientists.  
1.2 This guide covers a brief definition of scales as well as some fundamental and practical challenges the sensory and consumer scientists should be aware of when using scales. It also provides a list and a description of the most commonly used scales in the field of sensory evaluation and consumer product research along with a classification framework for these scales.  
1.3 The scope of this guide is limited to the sensory and consumer science professional’s selection and use of rating scales when an assessor assigns one symbol/value to one stimulus, to their perception of a stimulus or an internal attitude/opinion. It does not cover:  
1.3.1 Details of analysis of data obtained from rating scales,  
1.3.2 Guidelines for questionnaire design including attribute selection,  
1.3.3 Fundamentals of measurement such as reliability and validity,  
1.3.4 Measurement models used to convert scale responses into measures of unobserved sensory or hedonic quantities,  
1.3.5 Tasks in which the assessor assigns a symbol/value to a group of stimuli,  
1.3.6 Rankings, and  
1.3.7 Multi-item scales.  
1.4 Units—The values stated in SI units are to be regarded as the standard. No other units of measurement are included in this standard.  
1.5 This standard does not purport to address all of the safety concerns, if any, associated with its use. It is the responsibility of the user of this standard to establish appropriate safety, health, and environmental practices and determine the applicability of regulatory limitations prior to use.  
1.6 This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for the Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.

General Information

• Status: Published
• Publication Date: 30-Nov-2017
• ICS: 03.120.30 - Application of statistical methods
• Technical Committee: E18 - Sensory Evaluation
• Drafting Committee: E18.04 - Test Methods

Relations

• Refers: ASTM E253-19 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 15-Oct-2019
• Refers: ASTM E253-18a - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Oct-2018
• Refers: ASTM E253-18 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 15-Jun-2018
• Refers: ASTM E253-17 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-May-2017
• Refers: ASTM E253-16 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Jun-2016
• Refers: ASTM E253-15b - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Dec-2015
• Refers: ASTM E253-15a - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Jun-2015
• Refers: ASTM E253-15 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 15-Jan-2015
• Refers: ASTM E253-13a - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Nov-2013
• Refers: ASTM E253-13 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-May-2013
• Refers: ASTM E253-12a - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Nov-2012
• Refers: ASTM E253-12 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-May-2012
• Refers: ASTM E253-11a - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 01-Dec-2011
• Refers: ASTM E2299-11 - Standard Guide for Sensory Evaluation of Products by Children
• Effective Date: 15-Nov-2011
• Refers: ASTM E253-11 - Standard Terminology Relating to Sensory Evaluation of Materials and Products
• Effective Date: 15-Apr-2011

Overview

The ASTM E3041-17 standard, titled "Standard Guide for Selecting and Using Scales for Sensory Evaluation", offers comprehensive guidance for sensory and consumer scientists on choosing and implementing rating scales for measuring human responses to sensory stimuli. Developed by ASTM, this guide is instrumental for professionals in fields such as sensory evaluation, marketing research, and consumer product testing who aim to quantify perceptions like liking, intensity, purchase interest, or agreement with statements.

Sensory rating scales are essential tools for translating subjective human experiences into measurable data, enabling more effective product development, quality control, and consumer insights. With numerous scale types available, each offering unique advantages and challenges, ASTM E3041-17 assists practitioners in selecting the most appropriate scale to support robust, valid research outcomes.

Key Topics

• Purpose and Scope: The standard focuses on rating scales that capture subjective human responses (e.g., sensory, affective, attitudinal measures), excluding scales for direct physical measurement or multi-item questionnaires.
• Scale Classification: It details definitions and classifications of different scale types such as:
  • Nominal scales (categorical, no order)
  • Ordinal scales (rank order without explicit spacing)
  • Interval scales (ordered with equal spacing, no true zero)
  • Ratio scales (ordered, equal spacing, true zero reference)
• Scale Structure: Discussion of unipolar versus bipolar scales, category versus line (visual analog) scales, and assessor-focused versus product-focused scales.
• Anchoring and Labeling: Emphasizes the importance of clear anchoring points and culturally appropriate labels to minimize response bias and facilitate consistent scale use.
• Considerations in Scale Selection: Highlights factors such as research objectives, assessor training, number of scale points, cultural context, and data entry/analysis methods.
• Response Biases and Consistency: Advises on recognizing potential biases (e.g., tendency to avoid scale extremes, context effects), and stresses the benefit of consistent scale use across studies.

Applications

ASTM E3041-17 is widely applicable in several domains where sensory perception needs to be quantified:

• Consumer Product Testing: To gauge consumer preferences, likability, and purchase intent for products like food, beverages, cosmetics, and personal care items.
• Descriptive Sensory Analysis: For trained panels to evaluate and rate the intensity of specific product attributes (e.g., sweetness, texture) to guide product formulation or quality assurance.
• Quality Control: Enables objective decision-making based on sensory specifications, ensuring consistency in manufacturing and product release.
• Global Consumer Research: Supports cross-cultural studies by recommending best practices for scale adaptation, translation, and use in different regions and populations.

This standard is indispensable for sensory professionals aiming to select scales that yield meaningful, reliable data for statistical analysis, enabling impactful business and research decisions.

Related Standards

Practitioners using ASTM E3041-17 may also benefit from the following related ASTM standards:

• ASTM E253 – Terminology Relating to Sensory Evaluation of Materials and Products
• ASTM E2299 – Guide for Sensory Evaluation of Products by Children and Minors

These related documents provide additional context on terminology, assessor training, and considerations for specific populations, ensuring holistic coverage of sensory evaluation practices.

---

Keywords: ASTM E3041-17, sensory evaluation, rating scales, sensory science, consumer testing, sensory analysis, scale selection, product research, sensory measurement, scale types, quality control, standard guide.

Standards Content (Sample)

This international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for the
Development of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to Trade (TBT) Committee.
Designation: E3041 − 17
Standard Guide for
Selecting and Using Scales for Sensory Evaluation
This standard is issued under the fixed designation E3041; the number immediately following the designation indicates the year of
original adoption or, in the case of revision, the year of last revision. A number in parentheses indicates the year of last reapproval. A
superscript epsilon (´) indicates an editorial change since the last revision or reapproval.
1. Scope priate safety, health, and environmental practices and deter-
mine the applicability of regulatory limitations prior to use.
1.1 The objective of this guide is to provide information to
1.6 This international standard was developed in accor-
be reviewed and considered by the sensory and consumer
dance with internationally recognized principles on standard-
scientist who wants to select and use scales to measure
ization established in the Decision on Principles for the
responses from consumers or trained assessors. For ease of
Development of International Standards, Guides and Recom-
reading, the term sensory scientist is used throughout the guide
mendations issued by the World Trade Organization Technical
when statements apply to the sensory and consumer scientists.
Barriers to Trade (TBT) Committee.
1.2 This guide covers a brief definition of scales as well as
some fundamental and practical challenges the sensory and
2. Referenced Documents
consumer scientists should be aware of when using scales. It
2.1 ASTM Standards:
also provides a list and a description of the most commonly
E253 Terminology Relating to Sensory Evaluation of Mate-
used scales in the field of sensory evaluation and consumer
rials and Products
product research along with a classification framework for
E2299 Guide for Sensory Evaluation of Products by Chil-
these scales.
dren and Minors
1.3 The scope of this guide is limited to the sensory and
consumer science professional’s selection and use of rating 3. Terminology
scales when an assessor assigns one symbol/value to one
3.1 Definitions—See Terminology E253 for definitions re-
stimulus, to their perception of a stimulus or an internal
lating to sensory evaluation of materials and products.
attitude/opinion. It does not cover:
3.2 Definitions of Terms Specific to This Standard:
1.3.1 Details of analysis of data obtained from rating scales,
3.2.1 interval data—data obtained from a scale for which
1.3.2 Guidelinesforquestionnairedesignincludingattribute
numerically identical differences on any part of the scale
selection,
correspond to the same magnitude of difference.
1.3.3 Fundamentals of measurement such as reliability and
3.2.1.1 Discussion—The occurrence of a zero point in
validity,
interval data does not correspond to the complete absence of
1.3.4 Measurement models used to convert scale responses
the characteristic measured. An example of interval data is a
into measures of unobserved sensory or hedonic quantities,
temperatureindegreesFahrenheitwhereeachdegreechangeis
1.3.5 Tasks in which the assessor assigns a symbol/value to
thesamechangeinthermalheatregardlessofpointonthescale
a group of stimuli,
and 0°F does not represent the complete absence of thermal
1.3.6 Rankings, and
energy.
1.3.7 Multi-item scales.
3.2.2 ordered category scale, n—rating instrument in which
1.4 Units—The values stated in SI units are to be regarded
the categories used to encode the responses are ordered by
as the standard. No other units of measurement are included in
magnitude.
this standard.
3.2.3 ordinal data, n—data obtained when items are ordered
1.5 This standard does not purport to address all of the
with respect to magnitude, but the magnitudes of difference
safety concerns, if any, associated with its use. It is the
among successively ordered items are not necessarily equal.
responsibility of the user of this standard to establish appro-
3.2.3.1 Discussion—Examples include ranking, just-about-
right scales, and the Likert scale.
This guide is under the jurisdiction of ASTM Committee E18 on Sensory
Evaluation and is the direct responsibility of Subcommittee E18.04 on Fundamen- For referenced ASTM standards, visit the ASTM website, www.astm.org, or
tals of Sensory. contact ASTM Customer Service at [e-mail]. For Annual Book of ASTM
Current edition approved Dec. 1, 2017. Published January 2018. DOI: 10.1520/ Standards volume information, refer to the standard’s Document Summary page on
E3041-17. the ASTM website.
Copyright © ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United States
E3041 − 17
3.2.4 rating instrument, n—collection of symbols/values, for their work. Selecting the right scale is a critical step
provided by the sensory scientist to the assessor along with towards meeting the research objective and making valid
instructions, from which the assessor chooses to communicate conclusions.
an affective, attitudinal, behavioral, or perceptual response to a
stimulus. 5. Data Properties
3.2.4.1 Discussion—Examples of rating instruments include
5.1 This section concerns the properties of the data obtained
category scales, line scales, and list of CATAterms printed on
from numerically encoding responses obtained from rating
paper or displayed on an electronic device.
scales. Data generated using rating scales are classified by the
type of information supplied, often referred to as “levels of
3.2.5 ratio data, n—data obtained from a scale that has an
measurement.” Note that the levels of measurements outlined
absolute zero point and for which numerically identical differ-
in the following pertain to the data generated rather than to the
ences on any part of the scale correspond to the same
rating instrument itself. The data properties should be consid-
magnitude.
ered when determining which statistical analyses are appropri-
3.2.5.1 Discussion—An example of ratio data is a tempera-
ate.
ture on the Kelvin scale in which each degree change is the
samechangeinthermalheatregardlessofthepointonthescale
5.2 The four levels of measurement are:
and zero represents the complete absence of thermal energy.
5.2.1 Nominal Data—Differentiates samples or assessors
based on arbitrary categories or qualitative classifications. The
3.2.6 scale, n—(1) rating instrument, sometimes referred to
categories or classifications do not have numerical signifi-
as a rating scale, used to encode human responses to stimuli
cance.
numerically, an example of which is an ordered category scale
5.2.1.1 Examples are gender, ethnicity, and religion.
and (2) continuum on which perceptions are quantified with
5.2.2 Ordinal Data—Ordinal data arise from ranking items
specifiedtheoreticalpropertiesthatdependonthetypeofscale,
an example of which is an interval scale. or a set of ordered categories. In either case, the data do not
include information about the relative spacing between these
3.2.6.1 Discussion—In the sense of definition (1), while all
scores.Inotherwords,numericallyidenticaldifferencesonany
scales have several characteristics in common, for example, all
part of the scale are not necessarily identical in magnitude with
have at least two response options and all are used to encode
respect to the variable measured.
responses to stimuli, scales differ in the amount of information
5.2.2.1 Many of the scales presented throughout this guide
they provide per data point. In evaluating the sweetness of
are ordered category scales. Strictly speaking, they do not
products,forexample,CATA(thatis,yes/no)datamayprovide
generate anything more complex than an order of the items or
less information about sweetness than do rank data, in which
sensations being evaluated.
the relative sweetness of a group of products is ordered from
least to most, which in turn provide less information about 5.2.2.2 Example—A five-category “meets expectations”
scale with anchoring points ranging from “much worse than
sweetness than direct intensity ratings obtained from an or-
expected” to “much better than expected” allows each assessor
dered category or line scale.
to categorize items based on how well the items met his or her
expectations. However, the difference between scale categories
4. Significance and Use
is not likely to be interpreted the same among assessors.
4.1 Rating instruments or rating scales are commonly used
5.2.3 Interval Data—Interval data are obtained from a
in many areas such as sensory evaluation, marketing research,
rating instrument that does not have a true zero point even
experimental psychology, survey research, and economics in
though one of the scale point labels may be called “zero” and
which there is interest in quantifying perceptions such as
has numerically identical differences on any part of the scale.
liking, preference, level of purchase interest, intensity of an
In other words, points on the scale are equally spaced such that
attribute, degree of difference, or level of agreement with
the numbers assigned represent actual degrees of difference
statements. This guide is concerned with the scales that are
between samples. Since an interval scale does not have a
used to record human responses to physical stimuli rather than
meaningful zero point, ratio comparisons are not appropriate.
measuring physical entities. Many types of rating scales
However, the numeric differences between values assigned to
already exist and have been used in the above fields. Specific
the categories are meaningful. Differences on an interval scale
rating scales each have their own properties, advantages, and
do have ratio properties.
disadvantages. Some rating scales are intended for specific
5.2.3.1 Example—The Fahrenheit temperature scale is an
applications, while others have broader applications. Some
interval scale. A 5° change represents the same degree of
rating scales have been extensively studied and modeled and
difference at all points on the scale (that is, the difference
have well-established properties.
between5and10°isthesameasthedifferencebetween25and
4.2 Given the overwhelming number of scales available to 30°). However, since the 0 value is arbitrary, it is not
practitioners when designing research, it is necessary for the appropriate to apply ratio comparisons such as “80° is twice as
researcher to have some knowledge about the scales that are hot as 40°.” However, one could say that (212 - 68)/(68 - 32)
available along with the many considerations that surround = (100 - 20)/(20 - 0), which connects a Fahrenheit-based ratio
their use and applications. This guide will be useful to the to a Celsius-based ratio and shows that ratios of differences on
sensoryresearcherwhowantstouseascaleasameasuringtool interval scales are meaningful.
E3041 − 17
5.2.4 Ratio Data—Ratio data are obtained from a rating end. An example of a unipolar scale for sweetness intensity is
instrument that has a true zero point and numerically identical ascaleanchoredat“notatallsweet”atoneendand“extremely
differences on any part of the scale. In other words, points on sweet” at the other end. In addition to rating intensity, unipolar
the scale are equally spaced such that the numbers assigned scales are appropriate for rating amount (for example, amount
represent actual degrees of difference between samples.As the of sauce) and frequency (for example, of consuming a certain
name implies, ratios of these assigned values are meaningful. product) to name a few applications.
5.2.4.1 Examples are length, mass, age, and the Kelvin 6.3.1.3 In a bipolar scale, the endpoints are semantic oppo-
temperature scale. It is appropriate to say that 12 m is twice as sites with an implied or stated midpoint. Examples of a bipolar
long as 6 m. scale are the hedonic scale (see Fig. 1), the just-about-right
5.2.5 The visual look of the rating instrument does not scale (see Fig. 2), the purchase intent scale (see Fig. 3), and
guarantee any property of the data collected with it. Many Likert scales (see Fig. 4). With bipolar scales, careful consid-
rating instruments in sensory science appear to generate data eration shall be given to whether the endpoints are truly
with interval or ratio properties when they do not. The opposites, as the bipolarity implies. For example, a scale
researcher needs an understanding of the data properties to ranging from “much too sour” to “much too sweet” or a scale
choose the appropriate statistical analysis approach (1). with a mid-point of “neither sweet, nor sour” are incorrect
5.2.6 Reliability and Validity—Rating scales do not have because “sour” and “sweet” are both present in many products
inherent reliability and validity. How the panel uses the rating and are not semantic opposites, and therefore, they should be
scale, the experimental procedure, and many other factors separated into two different scales. One of these scales should
impact the reliability and validity of a sensory method. For an range from “not sweet enough” to “much too sweet;” while the
explanation of reliability and validity in the context of sensory other scale should range from “not sour enough” to “much too
studies, refer to the major sensory science publications (2-5). sour.” Bipolar scales can either be balanced (with an equal
number of categories or an equal line length on either side of
6. Classification of Scales
the midpoint) or unbalanced. The BASES scale (Fig. 5)isa
bipolar unbalanced scale. Unbalanced scales are less common
6.1 Classification Based on Objective:
than balanced scales. When translating scale(s) to another
6.1.1 Hedonicscalesareusedwhentheresearchobjectiveis
language,careshouldbetakentoassuresemanticoppositesare
to assess how much assessors like products or samples.
maintained.
6.1.2 Relativescalesareusedwhentheresearchobjectiveis
6.3.2 Scale Continuity:
to assess samples relative to another sample or to an ideal.
6.3.2.1 Scales are category scales or line scales; they can
6.1.3 Attitude scales are used when the research objective is
also be a hybrid of the two.
to assess consumers’ attitudes or opinions.
6.3.2.2 Category scales offer the assessors a limited number
6.1.4 Intensityscalesareusedwhentheresearchobjectiveis
of responses (typically ten or fewer) from which to choose.
to assess the perceived intensity of samples’sensory attributes
More specifically ordered category scales indicate an increas-
or the perceived intensity of the difference between samples.
ing or decreasing degree of a variable labeled with numbers,
6.2 Classification Based on the Objective of the Response—
words, or symbols. The intensity scales in Figs. 6 and 7 are
All scales can be classified according to whether the assessor
examples of ordered category scales.
communicates an internal reaction, attitude, or intention or
whether the assessor communicates the property of an external
product or stimulus. If the response is a function of the person
making the rating, it is an “assessor-focused” scale. Responses
on an assessor-focused scale can change when the researcher
changes the characteristics of the respondent sample. Liking,
attitudes, emotion, and agreement are all assessor-focused
scales.Whentheresponseisafunctionoftheproduct,thescale
is considered a “product-focused scale.” With product-focused
scales, the responses are a function of the product and thus are
not expected to change unless there is variation in the product,
even when the characteristics of the respondent sample are
changed. Product-focused scales are intensity scales, quality
scales, grading scales, and relative to reference rating scales.
6.3 Structural Classification:
6.3.1 Scale Polarity:
6.3.1.1 Scales are either unipolar or bipolar.
6.3.1.2 Aunipolar scale is used to record responses that are
increasing from low (or zero) at one end to high at the other
The boldface numbers in parentheses refer to a list of references at the end of
this standard. FIG. 1 Nine-Point Hedonic Scale
E3041 − 17
7.1.4 The researcher needs to consider the features of
different scales to determine which is the most appropriate for
the needed information. There are risks and benefits that shall
be considered for each situation.
7.1.5 Scales typically used in descriptive analysis and qual-
ity control are intensity and relative scales.
7.1.6 Scales typically used in consumer testing are hedonic
FIG. 2 Just-About-Right (JAR) Scale
scales, behavioral choice, relative, and attitude scales.
7.1.7 Scales are used in many different situations. Some
scalesmaybespecifictoasingleapplication,whileothersmay
be appropriate for multiple applications. For example the
nine-point hedonic scale (see 10.1.1.1) is only used in con-
sumer testing applications. For example Check-All-That-
Apply scales (see 10.3.3) which were originally used in
consumer testing have been more recently used in descriptive
analysis or quality control situations.
7.2 Ease of Use:
7.2.1 When choosing a rating scale, the sensory scientist
needs to consider who the assessors are along with their
FIG. 3 Purchase Intent Scale
cultural background, their cognitive abilities, and if relevant,
their level of sensory training. All these factors impact their
6.3.2.3 Line scales, also known as visual analog scales
understanding of the rating scale and their ability to perform
(VAS) or unstructured line scales, indicate an increasing or
thetask.Itisalsoimportanttoconsiderhoweasythedataentry
decreasing degree of a variable using an anchored continuum
will be and how the resulting data will be analyzed. In general,
forrespondingandarenotrestrictedtoafixedsetofcategories.
the rating scale should be easy to use for the assessors as well
The assessors are free to place a mark anywhere along the line
as for the sensory scientist.
implying, at least theoretically, an infinite number of response
7.2.2 Familiar scales are more comfortable for assessors to
options.The intensity line scale in Fig. 8 is an example of such
use. When developing a full questionnaire/ballot for naïve
a scale.
assessors,itisadvisabletonothavetoomanydifferenttypesof
6.3.2.4 It is also common to find scales that are a hybrid of
scales in a single questionnaire to avoid confusion.
category and line scales. The visual thickness scale in Fig. 9 is
7.2.3 Line scales are better suited when one uses a comput-
anexample.Thesescalesareusedbythepanelinthesameway
erized data collection; this is because data entry from a paper
as a line scale (that is, as a continuum) but the categories
ballot requires measuring responses on the line scale, which
provide anchoring points. They are more often used with a
can be a cumbersome process before actual data entry.
trained panel that was calibrated to the scale and its anchoring
7.2.4 Special considerations are necessary when using rat-
points.
ing scales with children and assessors across cultures and
6.3.2.5 Many attributes (for example, sensory intensity) can
countries, as research has shown that different geographies
be measured using category scales or line scales. The choice is
think about and use scales differently. The International
up to the researcher. See Section 7 for an overview of
Consumer Product Testing across Cultures and Countries,
considerations the researcher should take into account.
MNL55-EB (7) provides guidelines on the use of rating scales
in different countries and Guide E2299 provides guidelines
7. Important Considerations in Scale Selection
when testing with children. All the considerations outlined in
7.1 Objectives of the Research:
7.3 – 7.9 need to be reviewed ahead of time. A preliminary
7.1.1 The main applications presented in this guide are
study to validate the rating scale with the intended population
those that the sensory scientist will typically encounter. They
of assessors is recommended. When interpreting the data from
are: (1) descriptive analysis when the scientist intends to
multiple geographies or cultures or both, it is important to
measure the specific sensory properties of samples; (2) quality
understand if the product acceptance ratings are actually
control when the scientist intends to assess the sensory quality
different or reflect different use of the scale.
ofsamples;and(3)consumertestingwhenthescientistintends
7.3 Number of Points on the Rating Scale:
to assess the consumers’ perception, liking or preference of
samples, or attitudes. 7.3.1 The number of points on the rating scale shall be
7.1.2 The sensory scientist may occasionally need to work sufficienttoallowassessorstoexpresstheirrangeofresponses.
with marketing research scales. These are outside the scope of Toofewscalepointsmayforceassessorstoreportsimilarscale
valuesforstimulithateliciteddifferentresponses.Itisuptothe
this guide as they often are multi-item scales. The Marketing
Scales Handbookseries (6)isacompendiumofsuchpublished sensory scientist to determine what that number should be to
allow separation of responses.
marketing research scales.
7.1.3 The research objective will provide some direction as 7.3.2 The number of discrete scale points shall not be so
to which scale to choose. This guide presents commonly used large that it becomes cumbersome or hinders ease of use on the
scales for each research objective in Sections8–10. part of the assessors or the sensory scientist. When too many
E3041 − 17
FIG. 4 Example of a Likert (Agree-Disagree) Scale
options are used (for example, 3- to 100-point scales). Often
intensityscaleswithlargernumbersofscalepointsdonothave
every scale point labeled. Preston and Colman (8) found that
rating scales with seven, nine, or ten response categories were
better liked by consumers than scales with fewer response
categories.
7.3.4 Note that assessors have different levels of comfort
using the ends of the scale.The number of scale points or scale
length or both should be sufficient to still allow discrimination
for the samples even if assessors tend to avoid the end points
of the scale.
7.3.5 A specific number of scale points cannot be recom-
FIG. 5 The BASES Scale
mended. Research of scale usage has found high correlations
between scales with a different number of points (2); as long as
the scale fits the research objective appropriately, many scales
will perform well if the task is reasonable and clear. When the
researcher is interested in comparing samples, similar trends
are typically observed between samples regardless of the
specific scale used.
7.3.6 Bipolar scales are best used with an odd number of
response categories. These scales should be balanced; have the
same number of response categories on either side of the
mid-point; and allow for a full range of probable responses,
including a neutral response. Unbalanced scales run the risk of
biasing assessors either positively or negatively by suggesting
they focus on one part of the scale.
7.3.7 Unstructured line scales are advantageous because no
FIG. 6 Intensity Ordered Category Scale
consideration needs to be given to the number of points on the
scale. Many unstructured line scales are 15 cm long. However,
consideration needs to be given to the length of the scale.
Analogous to the number of points required for a category
scale, a line scale should be sufficiently long as to allow
discrimination.Consumersmayfinditmoredifficulttouseline
FIG. 7 Intensity Ordered Category Scale
scales consistently across sessions and they may benefit from
the anchors that category scales can provide. Additional
scalepointsareusedrelativetothediscriminationabilityofthe consideration should be given to the device on which the line
assessors, they will likely simplify the scale and use a smaller scale will be displayed and whether it will allow the same
number of the points. degree of discrimination across multiple devices. A line scale
7.3.3 Research to date has not defined a single optimal may not allow the same degree of discrimination when
number of response categories on a rating scale. Most attitude projected on a smaller screen (such as a phone) as when
scales, including Likert scales, have five or seven response projected on a larger screen (such as a computer monitor).
categoriesbecauseitiseasytodifferentiateandunderstandfive Scale length should be consistent for a specific panel through-
to seven different labels. For intensity scales, a wide number of out training and data collection.
E3041 − 17
FIG. 8 Intensity Line Scale
FIG. 9 Visual Thickness Scale
7.4 Scale Anchoring Points: assessor has experienced in this category in the past or it may
7.4.1 Anchoring points are an important part of any scale. consist of a direct comparison to a concrete product provided
They define one or several points on the scale and indicate its in the test (that is, a reference sample). Because all sensory
directionality and they help assessors use the scale in a responses are inherently relative, they are especially prone to
common way.Aminimum of two anchoring points are needed biases (see 7.8).
to anchor the ends of the scale. Anchoring points may also be
7.6 Scale Orientation—The orientation of the rating scale,
used at intermediate points along the scale. Anchoring points
whether it reads from left to right, right to left, up to down, or
indicate whether the scale is unipolar or bipolar. They are
down to up, needs to conform to the cultural norms of the
especiallyimportantwhenthesamequestionandscaleareused
country in which it is used so that it does not introduce any
over multiple time points or multiple sessions—they help the
confusion to the assessors (7). Rating scales can be presented
assessor use the scale more consistently and indicate that the
in either a horizontal or a vertical layout.
response is similar or different to other questions in the same
7.7 Instructions and Training:
study. Maintaining consistent anchoring points from study to
7.7.1 Some scales are more likely to require assessor
study helps the researcher to compare data across studies.
training. The need to train assessors on scale use is determined
7.4.2 Anchoring points may be single words but are often
by the research objective, the test method, and corresponding
word phrases (for example, “none” or “like extremely”); they
bestpractice.Typically,hedonic,behavioralchoiceandattitude
can also be numbers, pictures, physical stimuli, or a combina-
scalesdonotrequiretraining.Relativeandintensityscalesmay
tion of these. While anchoring points are required to define a
or may not require assessor training depending on the pool of
scale, the researcher shall consider whether to label any or all
assessors.
intermediatescalepoints.Whenselectingwordphrases,ensure
7.7.2 Instructions need to be concise but detailed enough
wordingisneutralandbalancedalongtheentirescaletoreduce
that assessors will understand what their task is. In the case of
potential biases (see 7.8.6). Also, the chosen wording should
hedonic scales, general instructions are typically sufficient.
convey the intended distances between anchoring points as
“Liking,” “sweetness,” and “thickness” are relatively unam-
much as possible.
biguous terms.Terms such as “creamy,” “smooth,” and “fresh”
7.4.3 The range of anchoring points needs to match the
may be less clear with respect to what they refer to, either in
range of expected responses. If a rating scale is anchored in
terms of the product itself, or with respect to the assessor’s
such a way that all samples are scored very close to the top (or
experience. In general, the researcher shall include terms that
bottom) of the scale, it may be difficult to discern differences
are as unambiguous as possible, and the sensory scientist
between the samples. For example, if sugar candies are
always needs to interpret data carefully.
evaluated on sweetness with the top of the scale anchored as
7.7.3 In the case of descriptive analysis and quality control
“sweet” instead of “very sweet,” it may become difficult to
in which assessors rate the intensity of specific attributes,
detect differences in sweetness because of the samples all
training is recommended to ensure that assessors understand
receiving high scores.
the attribute being rated and the meaning of the anchoring
7.4.4 Cultural differences need to be taken into account
points on the scale.Ambiguity of terms or attributes may cause
when designing anchoring points on a scale. A direct transla-
assessor confusion when evaluating a perceived intensity.
tion of anchoring points may not be adequate (7). In some
7.7.4 Members of descriptive panels are trained on the
cultures, the use of numbers as anchors may not be appropriate
concept (common definition/criteria) of the attribute being
(7). In some other cultures, there may be reluctance to use
measuredandsometimescalibratedonintensityalongthescale
negative response categories such as “dislike” on a rating scale
depending on the descriptive method used.
(9).
7.5 Relativity of Responses on a Scale—All responses on a 7.8 Response Biases:
scale involve a comparison to an internal frame of reference. 7.8.1 No matter the type of rating scale being used, re-
This frame of reference may consist of all the products that the searchers need to be aware that scale responses are influenced
E3041 − 17
by many other factors.This does not invalidate the use of these recognize specific sensory attributes and quantify the intensity
rating scales, but it does imply that caution needs to be taken of these attributes in samples. Agreement among assessors on
in data interpretation. themeaningofthesensoryattributesandgeneralagreementon
7.8.2 Biases can arise from a biased perception of the the intensity levels of the scale require training of the panel.
stimulus or the way assessors use the scale. This guide is only
8.2 The specific scales that are used in descriptive testing
concerned with biases arising from the way the scale is used.
can be found in Manual on Descriptive Analysis Testing for
Contrast and context effects that impact the perception of the
Sensory Evaluation(13).
stimulus are outside the scope of this guide.
8.3 There are two main types of such scales: intensity
7.8.3 Range and frequency effects constitute a response bias
ordered category scales and intensity line scales.
of ordered category scales that was well documented by
8.3.1 Attribute Intensity Ordered Category Scales—
Parducci (10).Itwasdemonstratedthatassessorshaveabiasto
Attribute intensity ordered category scales present the trained
use the scale categories with equal frequency. Because of this
assessors with a discrete number of options to report their
bias, a product will typically be reported less intense when
perceived intensities for a specific attribute. These scales are
evaluated in the context of other very intense products than it
intensity, unipolar, ordinal, and ordered category scales. Some
wouldhavebeenifithadbeenevaluatedinthecontextofother
very weak products. Therefore, responses to a stimulus are or all of the categories will be labeled with descriptions of the
corresponding intensity levels. In Figs. 6 and 7 two examples
affected by the intensities of the other stimuli in the set when
using such a scale. of such scales are shown.
7.8.4 Similar range effects have been found across different 8.3.1.1 Advantages:
types of scales other than ordered category scales (11);itis, (1) Intensity ordered category scales present a greater
therefore, unclear whether such effects are perception biases or opportunityfortrainingofthepanelthanlinescales(see8.3.2).
scale usage biases. The sensory scientist can provide thorough descriptions of the
7.8.5 Another well-known bias is the one of “end effects.” scale levels or sensory references or both to illustrate these
Assessors can be reluctant to use the ends of the scale to keep levels.
these ends available for a potentially more extreme stimulus (2) These scales are easier for the researcher to use than
presented in the future. This bias essentially reduces the line scales when data collection is done through paper ballots.
numberofscalepointsofacategoryscalebytwo(forexample, Thedatacanbeeasilyenteredintoaspreadsheetbecauseofthe
from nine to seven or from seven to five). This bias can be discrete number of responses.
more or less prominent in different cultures (7).
8.3.1.2 Disadvantages—When using an ordered category
7.8.6 As a general rule of thumb, the sensory scientist needs
scale, one needs to consider carefully the number of categories
to be aware of the fact that the distances on the scale between
provided on the scale. Guidelines outlined in 7.3 should be
responses to stimuli are not necessarily reflective of the
followed. This requires some understanding of the panel
distances between the assessors’ perceptions of these stimuli.
discrimination abilities relative to the sample sets evaluated.
This is because biases affect the assessors’ use of the scale.
When verbal labels are used to identify the intensity levels, the
Therefore, caution needs to be used when analyzing and
sensory scientist needs to ensure that the assessors interpret
interpreting the data generated by the scale (see Section 5).
these labels as reflecting increasing intensity levels.
7.8.7 For a complete review of potential biases for hedonic
8.3.2 Attribute Intensity Line Scales:
scales, see Ref 12.
8.3.2.1 Attribute intensity line scales present the trained
assessors with a potentially infinite number of options to report
7.9 Consistent Use of a Scale—Whatever the scale is, there
their perceived intensities up to the resolution of the measuring
is considerable value in using the same scale in the exact same
system. These scales are intensity, unipolar line scales. While
format across many studies and even at several points in the
the data that these scales generate are typically ordinal data,
same study. This allows the researcher to build a historical
theyareoftenanalyzedasintervaldata.Thenumberofanchors
databaseofresultstowhichanewresultcanbecompared.This
on the scale is often less than for ordered category scales. End
helps the researcher provide context to the new result and
anchors are usually provided; sometimes a middle anchor is
meaningful interpretation. It is recommended that the re-
added. Fig. 8 is an example of such a scale.
searcheruseachosenscaleconsistentlyforagivenapplication.
8.3.2.2 Some lines scales, as illustrated in Fig. 10, may
8. Examples of “Product-Focused” Scales in Descriptive
extendbeyondtheendanchorswiththegoalofminimizingend
Testing
effects (the tendency for assessors to avoid using the extreme
end points of a scale) (14 and 15).
8.1 The scales used in descriptive testing are intensity
scales; they are used to assess the intensity of specific sensory 8.3.2.3 Advantages—Intensity line scales provide the asses-
attributes in a set of samples.Assessors are typically trained to sors more freedom to report their perceived intensities than the
FIG. 10 Line Scale Extending Beyond the End Anchors
E3041 − 17
ordered category scales because they are not restricted to using 9.2.1.2 Disadvantages—This method and scale require ex-
a discrete number of categories.The sensory scientist may also tensive upfront training with the panel to establish the concept
findthemeasiertousebecauseitisnotnecessarytochoosethe
of “normal production” and the corresponding boundaries. It is
number of categories and their labels.
not recommended for complex products for which many
8.3.2.4 Disadvantages—If using a paper ballot, the asses-
variables can affect product quality. It is typically used for raw
sor’s responses need to be converted to numbers by measuring
materials or simple finished products. It is a decision-making
the distance from the left end of the line, which may be done
tool but does not provide any specific attribute difference,
by scanning, digitizing, or, as a last resort, measuring with a
which makes it difficult to identify the possible root cause of
ruler.This can be time consuming.Also, the values themselves
any potential QC issues.
are seldom simple integers, so data entry is more complicated.
9.2.1.3 Consideration—Since this method is a “decision-
making” tool rather than a source of information, it is essential
9. Examples of “Product-Focused” Scales in Quality
tomonitorapanelcarefullytoensureaccurateproductscoring.
Control (QC) Testing
This scale can be adapted to include an additional category for
9.1 To answer questions such as “are the ingredients within
the assessor to use when the assessor is unsure about whether
sensory specification” or “does the product fit the pre-
the sample is in or out of specification. It can also be followed
determined sensory specifications,” the following methods and
by comments to identify potential reasons for “out” ratings.
scales are commonly used and described in 9.2 – 9.6.
Note that a similar binary category scale is used in theA/notA
test in which the two response categories of the scale are “A”
9.2 Binary Category Scales:
and “not A.” This method requires presentation of a reference
9.2.1 The most common binary category scale used for
qualitycontroltestingistheoneusedwiththe“In/Out”method sample (labeled “A”) and the assessor’s task is to indicate for
(Ref 16 and Fig. 11). In this method, production samples are
each test sample whether it matches the reference’s profile
presented to a panel and each assessor indicates whether each
(“A” response) or not (“not A” response). This approach may
sample is “in” (representative of normal production) or “out”
be of interest when the level of training required by the “In
(not representative of normal production). A reference sample
/Out”methodisnotpossible.Thismaybethecaseforcomplex
may or may not be provided along with the test samples; when
finished products for which the concept of “not representative
no reference is provided, the trained assessors compare their
of normal production” may be difficult to capture over the
perception of the test sample to their mental representation of
course of a training.
“normal production;” when a reference is provided, the asses-
9.3 Attribute Intensity Scales—Attribute intensity scales can
sors compare their perception of the test sample to that of the
be used for QC purposes to evaluate the intensity of pre-
reference sample; in either case, this is a relative scale.
established quality attributes and confirm whether a product is
9.2.1.1 Advantages—The task is easy with very short evalu-
within specification limits. The same type of intensity scales
ations. The test results are very easy to interpret with a direct
presented in Section 8 can be used.
decision-making application. There is no need to present and
9.3.1 Advantages—Intensityscalesprovidethedetailedsen-
maintain a reference sample. The number of samples is
minimizedandcanbeaslowasone,whichisadvantageousfor sory profile of a given product. The scale can be used without
a sensory control/standard as long as specification limits for
product categories with lingering flavors or products that can
be evaluated only one at a time. each attribute are very well defined.
FIG. 11 Example of In/Out Binary Scale
E3041 − 17
9.3.2 Disadvantages—Assessors need extensive training scale is that it does not provide direction to assess product
with focus on individual attributes. References shall be used issues. It can also be difficult to calibrate the panel on the size
and panel calibration is essential for each attribute.
of the difference corresponding to each scale point especially
9.3.3 Consideration—Intensity scales are best used for QC the “extreme difference” point.
when the detailed sensory profile of the “control” product is
9.4.4 Consideration—Scale length will vary depending on
known and specific attributes are being assessed. Attributes
the product complexity and range of acceptability. A stable
with high consumer relevance should be selected. Because of
control/standard sample shall be available to use to present to
rigorous training required on each attribute, the number of
the assessors.
attributes of interest is a key consideration.
9.5 Sensory Quality Systems (SQS) Scale (19 and 20)—The
9.4 Difference Scales:
SQS scale is a relative scale measuring the acceptability of a
9.4.1 The degree of difference (DOD) (16 and 17) and
testsamplecomparedtoacontrol(seeFig.15).Itisaten-point
difference from control (DFC) tests typically use relative
scale with four anchors: 1-2 = Reject, 3-5 = Unacceptable, 6-8
category scales to assess the size of the difference between a
= Acceptable, and 9-10 = Match. Additionally, if the test
test sample and a labeled control/standard. They may be run
sampleisrated8orlower,thesampleisnotedforhowitdiffers
with either bipolar (directional difference) or unipolar scales.
from the control (Fig. 16).
The difference may be assessed for the overall product or for a
9.5.1 Advantages—The scale is easy to conceptualize and
specific attribute.
hasbeenadaptedforuseinproductionfacilitiesglobally.Ithas
9.4.1.1 Bipolar—The center of the scale is anchored as “no
application in both plant and laboratory environments and has
difference” compared to the control/standard, and the ends of
been used for shelf-life studies, ingredient replacements, and
the scale are anchored as “extreme difference” for lower and
commercialization scale ups.
higher intensities (Fig. 12). The researcher may also choose to
9.5.2 Disadvantages—The scale requires a control or gold
label intermediate points on the scale with the relative degrees
standardproduct.Withoutpropertrainingandcalibrationusing
of difference (Fig. 13). Bipolar scales are used to capture both
known standards, the scale could be subjective and less
the degree to which the samples differ from the control/
“product focused.”
standard and whether the intensity is stronger or weaker.Atest
run using a bipolar scale may be referred to as “directional.”
9.5.3 Consideration—Propertrainingonthescaleusageand
9.4.1.2 Unipolar—One end of the scale is anchored as “no
lexicon references is necessary before using. Though intended
difference”comparedtothecontrol/standard,andtheotherend
as an overall difference from a control, it could be adapted to
anchored as “extreme difference” from the control standard
evaluate single modalities such as appearance, texture, or
(Fig. 14).The researcher may also choose to label intermediate
flavor.
points on the scale with the relative degrees of difference.
9.6 Sensory Grading:
Unipolar scales are used to capture the intensity of the overall
9.6.1 Sensory quality grading is often used in the commerce
(or attribute) difference only.
of foods such as fish, meat, and produce. Assessors, called
9.4.2 Advantages—An advantage is a rapid assessment if a
“graders,” possess specialized training and significant practical
single overall difference is used. Because management often
knowledge of the product category. For example, graders
chooses some level of difference (for example, slight differ-
typically understand the production and spoilage process that
ence) as a pass or fail threshold, having a range is useful in the
the given food may undergo. Graders and expert graders are
selection process. The scale can also be used to assess specific
extensively trained and often obtain their training on the job
attributes or overall differences from a control/standard, which
from other graders over a long period of time. Selected graders
can be used to identify batch-to-batch variation among prod-
ucts. The training is also less rigorous than when training a may reach an expert level. In some cases, an “official qualifi-
cation” may be obtained. Expertise is attained through experi-
panel for descriptive profiling. An advantage specific to the
bipolar scales is that it provides direction to address product ence for a certain set of products and it cannot be transferred
across product categories. One judgment is enough if the
issues.
9.4.3 Disadvantages—They cannot replace a detailed sen- assessor has achieved expert level (perASTM International or
ISO definition). Graders rate the quality of products based on
sory profile. Also, care shall be taken using these scales in
situations when the product shows inherent variability because very specific guidelines that encompass all sensory aspects of
of production. However, the tests can be adapted to control for the product: visual appearance, odor, flavor,
...