The Role of Sensory Evaluation in The Food Industry
The Role of Sensory Evaluation in The Food Industry
The Role of Sensory Evaluation in The Food Industry
65
66 J L. Sidel, H. Stone
The basis for these industries supporting sensory Eventually these experts were retiring from companies
evaluation is that it represents a cost-effective resource more quickly than they could be replaced, and this was
with a wide range of applications, and it can provide occurring at a time when there was increased demand
unique product information not readily available from for more experts, as a result of expansion from new
other sources. Continued support from industry will products, new companies, and new plants. Lastly, as
depend on how well sensory evaluation continues to product lines expanded, it was almost impossible for a
satisfy these objectives. single expert to have the knowledge, skills and time
To fully appreciate the role of sensory evaluation in available for evaluating all of the companys products.
the food industry, it is useful to examine its evolution A logical solution for these concerns was to qualify
from early reliance on individual product experts to individuals more rapidly, document and systematize
the development of independent corporate functions the procedure so that it could be repeated at multiple
staffed by one or more qualified sensory professionals. locations, and train more than one or two people at a
Understanding how sensory evaluation has evolved in time to be product experts. The time was appropriate
the food industry is necessary to understanding its role for development of the expert panel (Cairncross &
in todays competitive business environment, and to Sjdstrom, 1950). It should be noted that some com-
anticipating future changes for that role. panies continue to rely on experts in addition to having
panels; however, their use has been modified.
Concurrent with the development of expert panels,
EVOLUTI ON OF SENSORY industry and trade associations, usually in concert with
government agencies, worked to develop product stan-
EVALUATI ON I N THE FOOD dards and specifications (Mosher et al., 1950; Cross
I NDUSTRY et al., 1978). While such standards appeared to serve
useful purposes from a quality control and purchase
Experts and expert panels specification perspective, they were erroneously equated
with sensory evaluation and consumer acceptance. The
All food companies have had and continue to have assumption was that highest quality, i.e. the top grade,
product specialists (or experts) whose responsibilities would be the most preferred product and would com-
are concerned with the quality of raw materials and of mand a premium price. Fortunately, consumers re-
finished products. These specialists have technical sponded independently (McBride & Hall, 1979; Side1
training and extensive practical knowledge about how et al, 1981), unaware of what was best for them, and
raw materials and processing affected finished product there were sufficient failures to force companies to
quality. Company managers assumed (or were told) look for better ways of assessing products before com-
that the experts knowledge and familiarity with ma- mitting to major expenditures without any assurances
terials, processes and products resulted in their having of success.
superior sensory skills, and their decisions about pro- Thus, changing consumer habits, competition, and
ducts would enhance success of the product in the the dramatic increase in the technology of product
marketplace. It is interesting to note that sensory manufacture significantly changed the role of experts
research demonstrated that experts were no more sen- and expert panels. With expanding markets, more
sitive than non-experts (Jones, 1968). Experts also complex and diverse raw material sources, and new
trained others to evaluate products, and that training technology and competition, there was impetus for
was a form of an apprenticeship. Most of the pro- change, stimulated by the increasing difficulty of pro-
cedures and criteria for determining who was qualified duct experts to translate their decisions about products
to evaluate products were not documented and the into market success. Many of the evaluation activities
decision when training was completed was based on the performed by expert panels, which had evolved from
experts opinion. To the extent that the experts pro- the individual product expert, would evolve into special-
duct decisions could be translated to market success, ized panels for discrimination, descriptive, and affec-
such endeavors were actively supported and considered tive evaluations (Amerine et al, 1965).
an important part of the business decision-making
process. Acceptance/preference testing
The success of the product expert raised some practi-
cal problems for management. First, months and some- The military, prompted by the governments effort to
times years were required to obtain the knowledge, provide more acceptable food for armed forces person-
experience and training necessary to qualify an expert. nel, contributed significantly to the development of
Typically, experts were senior staff, who had been with sensory evaluation. Research attention focused on
the company for many years. Often there was no avail- developing reliable and valid preference survey and
able replacement for the expert and management con- acceptance/preference measurement methods (Peryam
sidered that to be an undesirable risk for the company. et a& 1954, 1960).
The Role ofSensoryEvaluation in the Food Industry 67
provided, as well as demonstrate the value and contri- problems. Raw materials, manufacturing practices, and
bution of that information to the success of the com- even geography will influence products; and question-
pany. New funding or increased financial support for naire design, subject qualifications, and test practices
the sensory activity usually will have a negative effect on will influence results. When establishing the para-
the budget of another group within the company; i.e. meters for a test, information from several disciplines is
the money has to come from somewhere. Other groups used, including, for example, physiology, psychology,
aggressively seeking to enlarge their budgets (and im- mathematics and statistics. For any problem there will
portance) may view sensory funds as an incentive to be many options and the sensory professionals task is
seek control of the sensory group. Whatever the situ- to determine the most appropriate action for that
ation, it is necessary for the sensory group to have their problem. Because each company has unique products
own budget, particularly if they are to develop useful and problems, its sensory department often needs to
methods for evaluation of the companys products. develop or modify standard practices and procedures.
Frequently this will require some research, for which a
Organizational location qualified staff and available time and funding will be
necessary.
Failure to develop a coherent business plan and ad- The rapid growth of sensory evaluation in the food
dress the longer term issues has produced some rather industry, the failure (or inability) of some companies to
unusual scenarios, which have severely diminished the include on their staff qualified sensory researchers
impact of sensory information in many companies. By skilled at designing behavioral studies, and the lack of
unusual scenarios we are referring to sensory groups appropriate funding for this type of activity often re-
positioned within, and answering to, a product develop sults in insufficient and inadequate investigation into
ment department, or each product category having its the reliability and validity of the various sensory pro-
own sensory specialists also answering to the same man- cedures that are used. Consequently, companies are
ager as the product specialist. In some organizations unaware of better methods, and do not have staff able
R&D will use sensory evaluation as a means of off- to develop new procedures or to improve those that are
setting negative results from marketing research, or use available.
it as a substitute for or to duplicate marketing research Fortunately this situation is changing, and when eco-
without any appreciation for the consequences. In other nomics allow, more companies are supporting applied
organizations the sensory function is based on a single and basic sensory research. Sensory departments are
method and problems are fitted to the method. held accountable for developing or adopting methods
While our purpose is not to make a list of the abuses for improving the reliability of panelists, question-
to which sensory evaluation is subjected, it is important naires, and evaluation procedures; as well as for im-
to appreciate their impact on developing a meaningful, proving the validity of results and conclusions. As new
useful and unique source of information. For more dis- and more complex products are developed, and new
cussion about the misuses of sensory evaluation, see information about measuring behavior becomes avail-
Pangborn (1980). Although this publication is now able, sensory managers are expected to quickly and
13 years old, it is as current now as it was in 1980. efficiently evaluate that information and incorporate it
into their testing programs.
As companies become more aware of market oppor-
tunities available through sensory research, they are
CURRENT ROLE OF SENSORY more willing to fund those activities.
EVALUATI ON
Service role
Sensory evaluation has a dual role within a company. Its
primary role is that of service to marketing, R&D, and Sensory evaluation is primarily a source for product in-
manufacturing. Of equal importance, yet frequently formation. However, it is also responsible for educating
treated as a lesser role, is its research activity related to and updating managers and others about sensory prac-
perception and the development and refinement of tices and applications, and any new developments. Sen-
methods and procedures appropriate to that com- sory evaluation also could be supervising tests and
panys products. training panels for other company functions.
A variety of sensory methods are currently used in
Research role this service function. The methods are classified as dis-
criminative, descriptive and affective, and each has
The sensory evaluation department is responsible for unique requirements for qualifying subjects and mea-
providing actionable information from reliable and suring responses. There are different classification
valid tests that are related to a business objective. systems; however, most sensory professionals agree
Sensory professionals are seldom faced with routine that discrimination and descriptive tests are sensory
The Role of SensoryEvaluation in the Food Industy 69
analytical-laboratory tests and use subjects qualified ac- Test design and superuision. To assist with the design,
cording to their sensory skills, and that affective tests supervision, and analysis of product evaluation tests for
are used with consumers qualified according to pro- marketing research tests. Assistance from sensory evalu-
duct attitude and usage behaviors. ation is frequently requested for tests involving pro-
Regarding this latter issue, we are reminded of Pang- ducts which have difficult preparation, handling and
borns admonition that Although the distinction servicing instructions, and when large numbers (more
among these four types of measurement [sensitivity, than three) of products are to be evaluated in a central
quantitative, qualitative, and affective] is very clear-cut, location test.
the most frequent errors one sees in the present food
science literature are related to incorrect selection Services to quality assurance and quality control
and/or use of methods. Often, objectives are poorly
defined, methods are confounded or combined in Sexxny specifications. To establish intensity limits for
attempts to short-cut proper procedures, and all too important sensory attributes for products (Stone & Sidel,
often, investigators do not distinguish analytical sensory 1993).
analysis from consumer-type tests (Pangborn, 1980). Training and monikning QA/QC panels and programs.
The following section will describe some important To design, implement, and monitor procedures used
business applications for sensory evaluation. in production for the routine evaluation of ingredients,
in-process and finished product (Nakayama 8c Wessman,
Services to product development 1979).
Distribution product. The objective is to evaluate retail
New product development and product improvement. To product from different geographical locations and pro-
provide sensory information about product differ- duction sites.
ences, attribute intensities, and preferences as a func-
tion of ingredient, packaging and process variables.
Pilot+nt scale-up. To measure product differences as
a function of the changeover from pilot-plant to manu-
AREAS FOR FURTHER
facturing. ADVANCES
Production benchmark.To provide a permanent descrip
tive and affective sensory record for future reference. Research role
Cost reduction. To determine the sensory implications
of lower cost materials and processes. Success for sensory evaluation in business depends on
Product/@-ocessingchange. To assess the sensory effect the usefulness and the cost of the information pro-
of any ingredient, process, or packaging change not vided. This is a powerful incentive to continue develop-
associated with cost reduction or product improve- ing new and useful applications, and to continue to
ment. improve the accuracy and lower the cost for available
Instrument/sensq correlations. To describe the statis- methods. Selected areas for research deserving atten-
tical relationship between instrument and sensory tion are discussed below.
measures (Noble, 1975).
Product stability and she&@. To determine the sensory Affective tests
effect of different storage conditions.
Product optimization. To determine that combination Sensory affective tests provide a direct link between the
of product variables and sensory attributes that will be consumer and other sensory responses, such as the de-
best liked by consumers (Schutz, 1983; Side1 & Stone, velopment process. These tests are an integral part of
1983). any sensory program.
Subject screening criteria and test procedures need
Services to MRD to be improved to assure that results from sensory affec-
tive tests are a valid indicator of consumer preference.
Monitor competition. To evaluate current and new The same may be said for consumer results from mar-
competitive products. This may be done for a limited keting research studies. Improved screening criteria
number of products, or for an entire category (e.g. may allow for fewer participants in a test. Question-
strawberry preserves). naires and scoresheets need to be developed which
Advertising claim suppmt. To provide sensory informa- further minimize, or eliminate, biasing consumer test
tion to support or challenge a sensory-based advertising results. More information is needed about the effects
claim (Side1 & Stone, 1992). of different contexts and usage situations on affective
Identijjing consumer preference groups. To identify scores (Schutz, 1988; McDermott, 1990).
unique groups or clusters of consumers who are Information from sensory affective tests provide
differentiated by sensory preferences. greater opportunities for interaction with marketing
70 J. L. S&l, H. Stone
research, and to integrate the two information sources dures for test validity. For example, a panel that recog-
(Carter & Riskey, 1990). However, there must be a clear nizes that products are different can be taught to
differentiation between sensory objectives and affective associate each product with a specific scaler value (e.g.
tests, and marketing research objectives and tests. Other- told that product A is equal to an intensity score of 2,
wise, there will be confusion and conflict about respon- product B is equal to a 4, and so on). When that scale is
sibilities. Where such conflict has occurred, sensory given an attribute name (e.g. stale), it will appear that
often finds itself absorbed into other business func- the subjects can provide intensity information, when in
tions, effectively reducing their usefulness to product fact they are only providing nominal scale information.
development and quality control; prohibited from In this example the nominal scale has been given nu-
doing affective tests with consumers; or discontinued, merical labels. Although there is no validity to the
because they were viewed redundant with other com- panels scores for stale, their ability to reliably identify
pany testing activities such as marketing research. which product is associated with the 2, 4, and so forth,
Sensory affective tests focus on products; acceptance will be erroneously viewed as a measure of validity. This
is measured as a function of product differences and was a fatal flaw for expert systems, and is regrettably,
similarities; a physical product is present and is evalu- frequently encountered in some descriptive proce-
ated; context for the test is limited to the minimal dures today (Meilgaard et al., 1991). Rather than
amount of information necessary to describe the pro- teach subjects what they need to perceive, it is more
duct (e.g. barbecue-flavored ketchup) and its use (e.g. useful to measure what it is that they perceive.
a snack, a breakfast drink, etc.); and the dependent The use of reference standards for panel training
variable is passive (e.g. preference/acceptance) or is a (Wolfe, 1979; Rainey, 1986) and product testing also
scale that correlates well with the hedonic scale (e.g. requires scrutiny. The literature contains comprehen-
FACT scale; Schutz, 1965). Sensory affective tests are sive lists of references and their respective scores for
frequently used to reduce product alternatives to one scale extremes (Muiioz, 1986); however, there is no evi-
or two that will be studied by marketing research. dence that without them a panel cannot function or
In contrast, the marketing research model focuses that results are unreliable and/or invalid. References
on estimating the purchase intent for various popula- are a source of variability, and if very different from the
tions and population segments; a product may or may test products, are irrelevant or invalid. Training sub-
not be present, and if a product is present the con- jects to provide identical scores for a reference infers
sumer typically evaluates only one or two pairs of prod- that there is one correct response; it does not account
ucts; the consumer may be provided with a broad for individual differences and distributions of re-
context for evaluating the product, and this context sponses, and confuses reliability with validity.
may include components important for measuring pur- There should be little argument or need to further
chase intent; and ballots usually contain many pages of discuss the rationale for classifying descriptive tests as
questions for measuring a range of consumer attitudes. sensory analytical-laboratory methods rather than as
The dependent variable is more cognitive than sensory, consumer methods, as has recently been described by
or is action oriented (e.g. purchase intent) to estimate some researchers (Williams 8c Langron, 1984, Williams
the potential marketability for a product. & Arnold, 1985).
REFERENCES
Time-intensity tests Amerine, M. A., Pangborn, R. M. & Roessler, E. B. (1965).
Principlesof Sensory Evaluation of Food. Academic Press, New
After food is taken into the mouth the perceived inten- York, USA.
sity for one or more of the attributes will change over Anon. (1975). Minutes of Division Business Meeting. Institute
time. Some companies (e.g. those promoting non- of Food Technologists Sensory Evaluation Division, IFT,
nutritive sweeteners) have long recognized the impor- Chicago, IL, USA.
tance of the time-intensity measurement; however, Anon. (1984). Overview. Outstanding symposia in food
most food companies remain unaware of its value, and science & technology. Use of computers in the sensory lab.
are unable to do this research. Although there have Food TechnoL, 38(g), 6688.
been many recent advances in the equipment used to Billmeyer, B. A. & Wyman, G. (1991). Computerized sensory
evaluation system. Food Technd, 45(7), 100-l.
measure time-intensity (Guinard et aL, 19853, and in
Cairncross, W. E. & Sj&rom, L. B. (1950). Flavor profile-
statistical approaches to analyzing those data (MacFie
a new approach to flavor problems. Food TechnoL, 4,
& Liu, 1992), additional research is required for pro-
308-l 1.
cedures to reduce psychological biases during data col- Carter, C. & Riskey, D. (1990). The roles of sensory research
lection (Lawless 8c Clark, 1992), to improve the overall and marketing research in bringing a product to market.
efficiency for obtaining those measurements, and to Food TechnoL, 44(11), 160-2.
provide for more efficient measurement of multiple de- Cross, H. R., Moen, R. & Stanfield, M. S. (1978). Training and
scriptive attributes. testing of judges for sensory analysis of meat quality. Food
Unfortunately, the current literature about time- TechnoL, 32,48-54.
intensity presents an array of results and a dilemma Eggert, J. (1989). Sensory strategy for success in the food
with regard to the treatment of the within and across industry.j. SensoryStudies, 3,161-7.
Gacula, M. D. Jr & Singh, J. (1984). Statistical Methods in
subject variability, which can be so large as to limit
Food and Consumer Research. Academic Press, Orlando, FL,
specific product recommendations. Until time-intensity
USA.
measures demonstrate a direct and practical appli-
Guinard, J.-X., Pangborn, R. M. & Shoemaker, C. F. (1985).
cation, it is appropriate for industry to view them with Computerized procedure for time-intensity sensory mea-
caution. Most importantly, it is necessary to demon- surements.J. Food Sci., 50,543-4,546.
strate the relevance of these measures to the practical Hen&a, R. G. (1972). Simple and effective system for use with
issues of product development and differentiation, response surface methodology. &real Sti. Today, 17( lo),
which will lead to marketplace success. 309-14,334.
The Role of SensoryEvaluation in the Food Zndust?y 73
Hootman, R. C. (1992). Manual on Desniptiue Analysis Testing Acceptance Testing Methodology. Quartermaster Food and Con-
for Sensory Evaluation (ASTM Manual Series: MNL 13). tainer Institute of the Armed Forces, Chicago, IL, USA.
American Society of Testing Materials, Philadelphia, PA, Peryam, D. R., Polemis, B. W., Kamen, J. M., Einhoven, J. &
USA. Pilgrim, F. J. (1960). Food Preferences of Men in t& Armed
Jones, F. N. (1968). The information content of olfactory Forw. Quartermaster Food and Container Institute of the
quality. In Theuries of Odor Mea.surement, ed. Necmi I. Tanyolac. Armed Forces, Chicago, IL, USA.
Robert College Research Center, Bebek, Istanbul, Turkey, Rainey, B. A. (1986). Importance of reference standards in
pp. 133-46. training panelists. J senSo?yStud., 1, 149-54.
Lawless, H. T. & Clark, C. C. (1992). Psychological biases in Schutz, H. G. (1965). A food action rating scale for measur-
time-intensity scaling. Food TechnoL, 46( 11), 81-90. ing food acceptance.J. Food Sci., 30,365-74.
MacFie, J. H. H. & Liu, Y. H. (1992). Developments in the Schutz, H. G. (1983). Multiple regression approach to opti-
analysis of time-intensity curves. Food Technol., 46( ll), 92-7. mization. Food Technol., 37( 1 l), 468, 62.
McBride, R. L. & Hall, C. (1979). Cheese grading versus con- Schutz, H. G. (1988). Multivariate analyses and the measure-
sumer acceptability: An inevitable discrepancy. Aust. J. ment of consumer attitudes and perceptions. Food TechnoL,
Daily Technol., 34, 668. 42(11), 141-4,156.
McDermott, B. J. (1990). Identifying consumers and con- Sidel, J. L. & Stone, H. (1983). Introduction to optimization
sumer test subjects. Food Technol., 44( 11)) 154-8. research-definitions and objectives. Food TechnoL, 37( ll),
Meilgaard, M., Civille, G. V. & Carr, B. T. (1991). Sensory 36-8.
Evaluation Tpchniques. CRC Press, Inc. Boca Raton, FL, USA. Sidel, J. L. & Stone, H. (1992). Sensory evaluation: selecting
Masher, H. A., Dutton, H. J., Evens, C. D. & Cowan, J. C. the right test design. In NAD W&shop N, Product Pe$w-
(1950). Conducting a taste panel for the evaluation of mance Tests. Council of Better Business Bureaus, Inc., New
edible oils. Food Terhnol., 4, 105-9. York, USA, pp. 15-20.
Mmioz, A. M. (1986). Development and application of tex- Sidel, J. L., Stone, H. & Bloomquist, J. (1981). Use and misuse
ture reference sca1es.J. Sensory Stud., 1,55-83. of sensory evaluation in research and quality control. J
Nakayama, M. & Wessman, C. (1979). Application of sensory Dai?y Sk, 64, 2296302.
evaluation to the routine maintenance of product quality. Stone, H. & Sidel, J. L. (1993). Sensory Evaluation Practices
Food Technol., 33 (9), 38-9. (2nd ed.). Academic Press, Inc., San Diego, CA, USA.
Noble, A. ( 1975). Instrumental analysis of the sensory proper- Williams, A. A. & Arnold, G. (1985). A comparison of the
ties of food. Food Technol., 29( 110,56-60. aromas of six coffees characterized by conventional pro-
OMahony, M. (1986). Sensmy Evaluation of Food. Marcel filing, free-choice profiling and similarity scaling methods.
Dekker, Inc, New York, USA. J. Sci. Food Agric., 36,204-14.
Pangborn, R. M. (1980). Sensory science today. Cereal Foods Williams, A. A. & Langron, S. P. (1984). The use of free-
World, 25( lo), 637-40. choice profiling for the evaluation of commercial ports.
Peryam, D. R. (1991). A history of ASTM Committee E-18. J. Sci. Food Agric., 35, 558-68.
ASTM Stand. News., 19(3), 28-35. Wolfe, K A. (1079). Use of reference standards for sensory
Peryam, D. R., Pilgrim, F. J. & Peterson, M. S. (1954). Food evaluation of product quality. Food TechnoZ., 33(9), 43-4.