1、Designation: E3041 17Standard Guide forSelecting and Using Scales for Sensory Evaluation1This standard is issued under the fixed designation E3041; the number immediately following the designation indicates the year oforiginal adoption or, in the case of revision, the year of last revision. A number
2、 in parentheses indicates the year of last reapproval. Asuperscript epsilon () indicates an editorial change since the last revision or reapproval.1. Scope1.1 The objective of this guide is to provide information tobe reviewed and considered by the sensory and consumerscientist who wants to select a
3、nd use scales to measureresponses from consumers or trained assessors. For ease ofreading, the term sensory scientist is used throughout the guidewhen statements apply to the sensory and consumer scientists.1.2 This guide covers a brief definition of scales as well assome fundamental and practical c
4、hallenges the sensory andconsumer scientists should be aware of when using scales. Italso provides a list and a description of the most commonlyused scales in the field of sensory evaluation and consumerproduct research along with a classification framework forthese scales.1.3 The scope of this guid
5、e is limited to the sensory andconsumer science professionals selection and use of ratingscales when an assessor assigns one symbol/value to onestimulus, to their perception of a stimulus or an internalattitude/opinion. It does not cover:1.3.1 Details of analysis of data obtained from rating scales,
6、1.3.2 Guidelines for questionnaire design including attributeselection,1.3.3 Fundamentals of measurement such as reliability andvalidity,1.3.4 Measurement models used to convert scale responsesinto measures of unobserved sensory or hedonic quantities,1.3.5 Tasks in which the assessor assigns a symbo
7、l/value toa group of stimuli,1.3.6 Rankings, and1.3.7 Multi-item scales.1.4 UnitsThe values stated in SI units are to be regardedas the standard. No other units of measurement are included inthis standard.1.5 This standard does not purport to address all of thesafety concerns, if any, associated wit
8、h its use. It is theresponsibility of the user of this standard to establish appro-priate safety, health, and environmental practices and deter-mine the applicability of regulatory limitations prior to use.1.6 This international standard was developed in accor-dance with internationally recognized p
9、rinciples on standard-ization established in the Decision on Principles for theDevelopment of International Standards, Guides and Recom-mendations issued by the World Trade Organization TechnicalBarriers to Trade (TBT) Committee.2. Referenced Documents2.1 ASTM Standards:2E253 Terminology Relating to
10、 Sensory Evaluation of Mate-rials and ProductsE2299 Guide for Sensory Evaluation of Products by Chil-dren and Minors3. Terminology3.1 DefinitionsSee Terminology E253 for definitions re-lating to sensory evaluation of materials and products.3.2 Definitions of Terms Specific to This Standard:3.2.1 int
11、erval datadata obtained from a scale for whichnumerically identical differences on any part of the scalecorrespond to the same magnitude of difference.3.2.1.1 DiscussionThe occurrence of a zero point ininterval data does not correspond to the complete absence ofthe characteristic measured. An exampl
12、e of interval data is atemperature in degrees Fahrenheit where each degree change isthe same change in thermal heat regardless of point on the scaleand 0F does not represent the complete absence of thermalenergy.3.2.2 ordered category scale, nrating instrument in whichthe categories used to encode t
13、he responses are ordered bymagnitude.3.2.3 ordinal data, ndata obtained when items are orderedwith respect to magnitude, but the magnitudes of differenceamong successively ordered items are not necessarily equal.3.2.3.1 DiscussionExamples include ranking, just-about-right scales, and the Likert scal
14、e.1This guide is under the jurisdiction of ASTM Committee E18 on SensoryEvaluation and is the direct responsibility of Subcommittee E18.04 on Fundamen-tals of Sensory.Current edition approved Dec. 1, 2017. Published January 2018. DOI: 10.1520/E3041-17.2For referenced ASTM standards, visit the ASTM w
15、ebsite, www.astm.org, orcontact ASTM Customer Service at serviceastm.org. For Annual Book of ASTMStandards volume information, refer to the standards Document Summary page onthe ASTM website.Copyright ASTM International, 100 Barr Harbor Drive, PO Box C700, West Conshohocken, PA 19428-2959. United St
16、atesThis international standard was developed in accordance with internationally recognized principles on standardization established in the Decision on Principles for theDevelopment of International Standards, Guides and Recommendations issued by the World Trade Organization Technical Barriers to T
17、rade (TBT) Committee.13.2.4 rating instrument, ncollection of symbols/values,provided by the sensory scientist to the assessor along withinstructions, from which the assessor chooses to communicatean affective, attitudinal, behavioral, or perceptual response to astimulus.3.2.4.1 DiscussionExamples o
18、f rating instruments includecategory scales, line scales, and list of CATA terms printed onpaper or displayed on an electronic device.3.2.5 ratio data, ndata obtained from a scale that has anabsolute zero point and for which numerically identical differ-ences on any part of the scale correspond to t
19、he samemagnitude.3.2.5.1 DiscussionAn example of ratio data is a tempera-ture on the Kelvin scale in which each degree change is thesame change in thermal heat regardless of the point on the scaleand zero represents the complete absence of thermal energy.3.2.6 scale, n(1) rating instrument, sometime
20、s referred toas a rating scale, used to encode human responses to stimulinumerically, an example of which is an ordered category scaleand (2) continuum on which perceptions are quantified withspecified theoretical properties that depend on the type of scale,an example of which is an interval scale.3
21、.2.6.1 DiscussionIn the sense of definition (1), while allscales have several characteristics in common, for example, allhave at least two response options and all are used to encoderesponses to stimuli, scales differ in the amount of informationthey provide per data point. In evaluating the sweetne
22、ss ofproducts, for example, CATA(that is, yes/no) data may provideless information about sweetness than do rank data, in whichthe relative sweetness of a group of products is ordered fromleast to most, which in turn provide less information aboutsweetness than direct intensity ratings obtained from
23、an or-dered category or line scale.4. Significance and Use4.1 Rating instruments or rating scales are commonly usedin many areas such as sensory evaluation, marketing research,experimental psychology, survey research, and economics inwhich there is interest in quantifying perceptions such asliking,
24、preference, level of purchase interest, intensity of anattribute, degree of difference, or level of agreement withstatements. This guide is concerned with the scales that areused to record human responses to physical stimuli rather thanmeasuring physical entities. Many types of rating scalesalready
25、exist and have been used in the above fields. Specificrating scales each have their own properties, advantages, anddisadvantages. Some rating scales are intended for specificapplications, while others have broader applications. Somerating scales have been extensively studied and modeled andhave well
26、-established properties.4.2 Given the overwhelming number of scales available topractitioners when designing research, it is necessary for theresearcher to have some knowledge about the scales that areavailable along with the many considerations that surroundtheir use and applications. This guide wi
27、ll be useful to thesensory researcher who wants to use a scale as a measuring toolfor their work. Selecting the right scale is a critical steptowards meeting the research objective and making validconclusions.5. Data Properties5.1 This section concerns the properties of the data obtainedfrom numeric
28、ally encoding responses obtained from ratingscales. Data generated using rating scales are classified by thetype of information supplied, often referred to as “levels ofmeasurement.” Note that the levels of measurements outlinedin the following pertain to the data generated rather than to therating
29、instrument itself. The data properties should be consid-ered when determining which statistical analyses are appropri-ate.5.2 The four levels of measurement are:5.2.1 Nominal DataDifferentiates samples or assessorsbased on arbitrary categories or qualitative classifications. Thecategories or classif
30、ications do not have numerical signifi-cance.5.2.1.1 Examples are gender, ethnicity, and religion.5.2.2 Ordinal DataOrdinal data arise from ranking itemsor a set of ordered categories. In either case, the data do notinclude information about the relative spacing between thesescores. In other words,
31、numerically identical differences on anypart of the scale are not necessarily identical in magnitude withrespect to the variable measured.5.2.2.1 Many of the scales presented throughout this guideare ordered category scales. Strictly speaking, they do notgenerate anything more complex than an order
32、of the items orsensations being evaluated.5.2.2.2 ExampleA five-category “meets expectations”scale with anchoring points ranging from “much worse thanexpected” to “much better than expected” allows each assessorto categorize items based on how well the items met his or herexpectations. However, the
33、difference between scale categoriesis not likely to be interpreted the same among assessors.5.2.3 Interval DataInterval data are obtained from arating instrument that does not have a true zero point eventhough one of the scale point labels may be called “zero” andhas numerically identical difference
34、s on any part of the scale.In other words, points on the scale are equally spaced such thatthe numbers assigned represent actual degrees of differencebetween samples. Since an interval scale does not have ameaningful zero point, ratio comparisons are not appropriate.However, the numeric differences
35、between values assigned tothe categories are meaningful. Differences on an interval scaledo have ratio properties.5.2.3.1 ExampleThe Fahrenheit temperature scale is aninterval scale. A 5 change represents the same degree ofdifference at all points on the scale (that is, the differencebetween 5 and 1
36、0 is the same as the difference between 25 and30). However, since the 0 value is arbitrary, it is notappropriate to apply ratio comparisons such as “80 is twice ashot as 40.” However, one could say that (212 - 68)/(68 - 32)= (100 - 20)/(20 - 0), which connects a Fahrenheit-based ratioto a Celsius-ba
37、sed ratio and shows that ratios of differences oninterval scales are meaningful.E3041 1725.2.4 Ratio DataRatio data are obtained from a ratinginstrument that has a true zero point and numerically identicaldifferences on any part of the scale. In other words, points onthe scale are equally spaced suc
38、h that the numbers assignedrepresent actual degrees of difference between samples. As thename implies, ratios of these assigned values are meaningful.5.2.4.1 Examples are length, mass, age, and the Kelvintemperature scale. It is appropriate to say that 12 m is twice aslong as 6 m.5.2.5 The visual lo
39、ok of the rating instrument does notguarantee any property of the data collected with it. Manyrating instruments in sensory science appear to generate datawith interval or ratio properties when they do not. Theresearcher needs an understanding of the data properties tochoose the appropriate statisti
40、cal analysis approach (1).35.2.6 Reliability and ValidityRating scales do not haveinherent reliability and validity. How the panel uses the ratingscale, the experimental procedure, and many other factorsimpact the reliability and validity of a sensory method. For anexplanation of reliability and val
41、idity in the context of sensorystudies, refer to the major sensory science publications (2-5).6. Classification of Scales6.1 Classification Based on Objective:6.1.1 Hedonic scales are used when the research objective isto assess how much assessors like products or samples.6.1.2 Relative scales are u
42、sed when the research objective isto assess samples relative to another sample or to an ideal.6.1.3 Attitude scales are used when the research objective isto assess consumers attitudes or opinions.6.1.4 Intensity scales are used when the research objective isto assess the perceived intensity of samp
43、les sensory attributesor the perceived intensity of the difference between samples.6.2 Classification Based on the Objective of the ResponseAll scales can be classified according to whether the assessorcommunicates an internal reaction, attitude, or intention orwhether the assessor communicates the
44、property of an externalproduct or stimulus. If the response is a function of the personmaking the rating, it is an “assessor-focused” scale. Responseson an assessor-focused scale can change when the researcherchanges the characteristics of the respondent sample. Liking,attitudes, emotion, and agreem
45、ent are all assessor-focusedscales. When the response is a function of the product, the scaleis considered a “product-focused scale.” With product-focusedscales, the responses are a function of the product and thus arenot expected to change unless there is variation in the product,even when the char
46、acteristics of the respondent sample arechanged. Product-focused scales are intensity scales, qualityscales, 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 A unipolar scale is used to re
47、cord responses that areincreasing from low (or zero) at one end to high at the otherend. An example of a unipolar scale for sweetness intensity isa scale anchored at “not at all sweet” at one end and “extremelysweet” at the other end. In addition to rating intensity, unipolarscales are appropriate f
48、or rating amount (for example, amountof sauce) and frequency (for example, of consuming a certainproduct) to name a few applications.6.3.1.3 In a bipolar scale, the endpoints are semantic oppo-sites with an implied or stated midpoint. Examples of a bipolarscale are the hedonic scale (see Fig. 1), th
49、e just-about-rightscale (see Fig. 2), the purchase intent scale (see Fig. 3), andLikert scales (see Fig. 4). With bipolar scales, careful consid-eration shall be given to whether the endpoints are trulyopposites, as the bipolarity implies. For example, a scaleranging from “much too sour” to “much too sweet” or a scalewith a mid-point of “neither sweet, nor sour” are incorrectbecause “sour” and “sweet” are both present in many productsand are not semantic opposites, and therefore, they should beseparated into two different scales. One of these scales shoul
