Caballero B. (ed.) Encyclopaedia of Food Science, Food Technology and Nutrition. Ten-Volume Set

Подождите немного. Документ загружается.

tissue levels of folates. There is evidence of folate

abnormalities associated with affective psychiatric

disorders, and some patients may benefit markedly

from folic acid supplementation. Raised homo-

cysteine levels have recently been reported in the

context of dementia, including Alzheimer’s disease.

A congenital disease associated with changes in the

X chromosome and known as fragile X syndrome has

been effectively treated with large folate supplements.

0022 Tetrahydrobiopterin, another member of the same

pteridine family to which the folates belong, is found in

most mammalian tissues. It plays an important cofac-

tor role in certain enzymes, especially those which

introduce oxygen into aromatic amino acids. Unlike

folic acid, tetrahydrobiopterin can be synthesized de

novo within the body, and is therefore not classified

as an essential vitamin. However, a dietary source of

it may be essential in some abnormal metabolic

states, particularly for some infants. A congenital

lack of dihydropteridine reductase produces

severe neurological defects, attributable to a tissue

deficiency of tetrahydrobiopterin.

Currently Important Public Health Issues

Neural Tube Defects

0023 During the early stages of embryonic development,

failure of the normal closure of the neural tube can

result in permanent abnormalities of the brain, cord,

and/or spine called spina bifida, anencephaly, and

encephalocele, which may be highly debilitating or

fatal. This condition was relatively common in parts

of the UK and Northern Ireland until the 1970s, aver-

aging around 0.4% of human fetuses; it has now fallen

to around 0.1%. Pregnancy termination has usually

been advised following its detection by amniocentesis.

Evidence of possible involvement of vitamins, and par-

ticularly folic acid, was suspected, and then proven by

a large Medical Research Council trial, published in

1991. A 70% reduction in recurrence of neural tube

defects (NTD) in a high-risk group, namely mothers

with a previously affected fetus, was achieved by a

4 mg day

1

folic acid supplement during the critical

early development period. More recently, studies in

Hungary and China obtained substantial reductions

in NTD incidence by smaller folate supplements, of

800 mg and 400 mg respectively. It is now recommended

in the UK that women planning a pregnancy who are in

this high-risk category should receive 4 or 5 mg day

1

and those who are not at high risk should receive

400 mg (0.4 mg) day

1

, for a period which lasts from

before conception until the third month of pregnancy.

In the USA, there has been mandatory fortification of

cereal grain products with folic acid at 1.4 mgg

1

from

the beginning of 1998, one purpose being to reduce the

incidence of NTD. In the UK, mandatory fortification

has not been introduced, but a recent Commission on

Medical Aspects of Food Policy (COMA) report has

recommended folic acid fortification of flour at

2.4 mgg

1

, calculated to be the best compromise be-

tween reduction of NTD births and avoidance of

intakes >1 mg day

1

in people >50 years.

0024NTD, like most pathologies, has multifactorial

risks, and its etiology is not fully understood. An

abnormality in vitamin B

economy has been

reported, and NTD occur frequently in women

who are homozygous for the thermolabile form

of the folate-metabolizing enzyme, methylene tetra-

hydrofolate reductase (MTHF). Recently, a defi-

ciency of folate has also been linked with the

occurrence of hare lip and of facial clefts.

Homocysteine

0025Folate, together with vitamins B

and B

, are all

B vitamins whose supply and status can affect the

circulating concentrations of the sulfur amino acid

homocysteine. Moderately raised levels are an

independent risk factor for several types of vascular

disease, including cardiovascular disease, cerebro-

vascular disease, and deep vein thrombosis. Rela-

tively high intakes and tissue levels of these three

vitamins help to minimize plasma homocysteine

levels. Homocysteine, the precursor of methionine,

requires methyl tetrahydrofolate as the cosubstrate

of vitamin B

-dependent methionine synthase.

In most normal adults, the plasma concentration

of homocysteine is c.5–15 mmol l

1

; moderate hyper-

homocysteinemia is defined as 15–30 mmol l

1

; inter-

mediate as 30–100 mmol l

1

; severe as >100 mmol l

1

Folate, across its physiological range, is a particularly

powerful determinant of plasma homocysteine.

Although there is a widespread expectation that

folate status may influence vascular disease risk, this

has not yet been proven. There is the further per-

ceived complication (see below), that high folate

intakes can increase hemoglobin synthesis, without

preventing neuropathy, in the early stages of vitamin

deficiency. The ‘masking’ of B

-deficiency

anemia might delay diagnosis and treatment, with

irreversible consequences. Therefore, public health

interventions designed to optimize the folate status

of populations need to proceed with caution.

Toxicity

0026Folic acid is generally considered to be nontoxic,

even in large amounts when taken orally, but some

concern has been expressed about the danger of

masking, or even possibly hastening the progress of,

2568 FOLIC ACID/Physiology

the early stages of vitamin B

-deficiency neuropathy,

including that caused by pernicious anemia. Most

workers now consider that masking is the main

hazard. Very high folic acid intakes may have subtle

effects on zinc economy, but this is not now con-

sidered a serious hazard. Reports of mood changes,

irritability, sleeplessness, and gastrointestinal symp-

toms on intakes of 15 mg of pteroylglutamate per

day have been reported, but other observers disagree.

Occasional reports of allergic reactions to folic acid

have occurred, and high doses may be epileptogenic

or reduce the efficacy of anticonvulsant control of

epilepsy in susceptible subjects. In general, water-

soluble vitamin intakes seem to be tolerated by most

people over a very wide range. However, in view of

the current publicity about benefits of folic acid, it

seems prudent to warn that excessive intakes of

any nutrient may produce imbalances which are

potentially risky.

See also: Amino Acids: Metabolism; Anemia (Anaemia):

Megaloblastic Anemias; Carcinogens: Carcinogenic

Substances in Food: Mechanisms; Cobalamins:

Physiology; Coenzymes; Dietary Requirements of

Adults; Drug–Nutrient Interactions; Elderly:

Nutritionally Related Problems; Immunoassays:

Radioimmunoassay and Enzyme Immunoassay; Infant

Foods: Milk Formulas; Infants: Breast- and Bottle-

feeding; Nucleic Acids: Physiology

Further Reading

Bailey LB (ed.) (1995) Folate in Health and Disease. New

York: Marcel Dekker.

Department of Health (1999) Report on Health and Social

Subjects no. 50. Folic Acid and the Prevention of Dis-

ease. Report of the Committee on Medical Aspects of

Food and Nutrition Policy. London: The Stationery

Office.

Food and Nutrition Board. Institute of Medicine. National

Academy of Sciences, Subcommittee on Folate, other B

Vitamins and Choline (1998) Dietary Reference Intakes:

Thiamin, Riboflavin, Niacin, Vitamin B-6, Folate,

Vitamin B-12, Pantothenic Acid, Biotin and Choline.

Washington, DC: National Academy Press.

Gregory JF (1989) Chemical and nutritional aspects of

folate research: analytical procedures, methods of folate

synthesis, stability and bioavailability of dietary folates.

Advances in Food and Nutrition Research 33: 1–101.

Gregory JF (1997) Bioavailability of folate. European Jour-

nal of Clinical Nutrition 51 (supplement 1): S54–S59.

Hawkes JG and Villota R (1989) Folates in foods: reactivity,

stability during processing, and nutritional implications.

CRC Critical Reviews in Food Science and Nutrition 8:

439–538.

Kim Y-I (1999) Folate and carcinogenesis: evidence,

mechanisms and implications. Journal of Nutritional

Biochemistry 10: 66–88.

MRC Vitamin Study Research Group (1991) Prevention of

neural tube defects: results of the Medical Research

Council vitamin study. Lancet 2: 131–137.

Picciano MF, Stokstad ELR and Gregory JF (eds) (1990)

Folic Acid Metabolism in Health and Disease. Contem-

porary Issues in Clinical Nutrition, vol. 13. New York:

Wiley-Liss.

Refsum H, Ueland PM, Nygaard O and Vollset SE (1998)

Homocysteine and cardiovascular disease. Annual

Review of Medicine 49: 31–62.

Zittoun J and Cooper BA (eds) (1989) Folates and Cobala-

mins. Berlin: Springer-Verlag.

FOOD ACCEPTABILITY

Contents

Affective Methods

Market Research Methods

Affective Methods

H L Meiselman and A V Cardello, US Army Soldier

Centre, Natick, MA, USA

Introduction

0001 The study of human behavior with respect to food

includes food selection, food acceptance, and food

consumption. Studies of food acceptability involve

both sensory and affective dimensions of food. Sens-

ory studies relate to the perception of the gustatory,

olfactory, textural, visual, thermal, and other attri-

butes of food. Affective studies concern people’s

evaluative reactions on a like–dislike or hedonic

dimension. Over the past decade, there has been

increased attention to the many variables that influ-

ence food acceptability and a growing awareness

of the complexity of food acceptability. Along

FOOD ACCEPTABILITY/Affective Methods 2569

with these changes have come new sensory and

affective methods for assessing food acceptance, and

greater appreciation for how to apply existing

methods.

Sensory and Affective Judgments

0002 While a model of food selection and consumption

based solely on the sensory properties of foods is

conceivable, the evaluative dimension is a primary

driver of the consumer’s behavioral response to food.

One of the critical questions in food acceptability

research is whether to ask respondents for strictly

sensory judgments, or to ask for evaluative or hedonic

judgments, or both. This distinction between descrip-

tion and evaluation is key in the selection of test

subjects, methods, samples, controls, and other

experimental design considerations. Both the sensory

and affective aspects of food acceptability are well

covered by recent volumes such as Lawless and Hey-

mann, and Meiselman and MacFie. However, the

social, cultural, physiological, and contextual aspects

of the eating situation must also be considered for

their influence in determining choice, acceptance,

and consumption. This article will review and

compare the primary methods for measuring sensory

and affective responses to foods (Figure 1).

Trained, Expert, and Consumer Panels

0003 A trained panel is a group of individuals who have

been selected for training based on criteria related to

their ability to perceive differences in basic sensory

attributes, and then trained in specific sensory

evaluation techniques and the use of a standardized

lexicon. Preselection criteria commonly include sens-

ory threshold testing. Sensory training is usually

conducted in one specific technique, e.g., descriptive

texture analysis, rather than in a variety of tech-

niques. An expert panel is usually a group with exten-

sive formal or informal experience with the product

or the product category, but without formal or

standardized selection and training. Expert panels

are often made up of inhouse experts within an or-

ganization. Trained and expert panels members may

or may not be actual users of the product or product

class. Consumer panels are comprised of people

without special training or experience; a consumer

panel should represent the target population of the

product.

0004 Most professionals agree that affective acceptance

testing should be done only with untrained,

nonexpert consumers. The reason for this is that

trained panelists may well have different preferences

from the average consumer. A number of studies have

compared trained and consumer panels for hedonic

judgments. For example, Shepherd and colleagues

found that ratings of eight tomato soups were differ-

ent for trained and consumer panels. Only the trained

panel showed significant differences in overall prod-

uct ratings, and in general, the trained and consumer

panels preferred different products.

0005On the other hand, the extent to which consumers

can be used to provide sensory data is somewhat

controversial. The reason for this is that the fields of

psychology and psychophysics have had a long and

successful history of using naive subjects (consumers)

to make sensory and perceptual judgments on a wide

range of both simple and complex stimuli. This para-

digmatic approach has been carried over by psycholo-

gists such as Moskowitz to the study of food

acceptability, and methods have been developed to

apply consumer terminology to the study of the sens-

ory dimensions of food (free-choice profiling). How-

ever, traditional food industry approaches focused on

either expert or trained panels to provide such sens-

ory data on foods. At present, it can best be said that a

consumer panel should be used when affective data

are being obtained or when sensory data are to be

generalized to a consumer population. Trained panels

should be used when there is a requirement to maxi-

mize the sensitivity to differences in the product,

when detailed descriptive information is required, or

when comparisons are to be made to instrumental

data. Expert panels, in general, should be restricted

to specific, in-plant, quality control problems or to

commodity areas in which expert judgments have

been shown to be a valid and reliable approach to

obtaining necessary data.

Category Scales

0006Category scales are the most commonly used methods

for assessing food likes and dislikes. They are charac-

terized by a series of labeled points or categories, and

a food item is rated by assigning it to one of several

(usually 3–10) available descriptive categories. One

of the earliest and still most commonly used scales of

food preference and food acceptance is based on the

category method. This is the nine-point hedonic scale

developed by Perham and colleagues in the early

1950s at the US Army Quartermaster Food and Con-

tainer Institute in Chicago which later moved to

Natick, Massachusetts.

0007Empirical evidence indicates that the points on the

nine-point hedonic scale do not represent equal sub-

jective intervals. Jones and colleagues scaled a variety

of verbal phrases for their meaning on an independ-

ent like–dislike scale. These data reveal that the inter-

vals between phrases used as category labels for the

2570 FOOD ACCEPTABILITY/Affective Methods

nine-point hedonic scale are unequal. For example,

the interval between ‘like extremely’ and ‘like very

much’ is larger than the interval between ‘like moder-

ately’ and ‘like slightly.’ Earlier work on this problem

has been reviewed by Schutz and Cardello, and this

nonequivalence of scale intervals demonstrated using

magnitude estimates of liking; this will be discussed in

the next section.

0008 Another problem with category scales is that re-

spondents tend not to use the extreme categories,

because they fear that if they use these categories

to rate a food item and a more extremely liked (or

disliked) item is presented subsequently, they will

have no appropriate category left to assign. As such,

the nine-point hedonic scale can effectively become

a seven-point scale of liking/disliking. Also, many

category scales (such as the nine-point hedonic scale)

have a neutral point. There is some evidence that

eliminating this neutral point may increase the effi-

ciency of the scale and may discourage the compla-

cency in judgment that is provided by this ‘safe’

category.

0009 In spite of these problems, many researchers have

adopted the nine-point hedonic scale for studying

food acceptability because of its ease of use, its ap-

plicability to both preference and acceptance data,

and its good reliability. In addition, category scales

are widely used for rating other dimensions of food

acceptability, such as the sensory intensity of

food attributes, the situational appropriateness

of foods, sensory and/or hedonic expectations toward

foods, and even preference mapping.

Unstructured or Line Scales

0010 Scales that consist of an unbroken line that is

anchored on both ends are called unstructured,

graphic, line, or visual analog scales. Line scales can

be used to quantify both sensory dimensions and

hedonic dimensions. Several different scales were

tried in one study by Pangborn and colleagues: a 10-

cm line marked ‘dislike extremely’ and ‘like ex-

tremely’ as the two anchors on a visual analog scale

of liking, a similar line scale with the midpoint de-

fined by ‘neither like nor dislike,’ and a 100-unit scale

labeled at the ends and midpoint. Hedonic scores on

the line scale tended to be more centrally clustered

than scores of sensory intensity. Reliability of the

scores from line scaling was usually good, but with

some products subjects did not maintain the same

order of hedonic judgment. Lawless and Malone

found that line scales of sensory intensity were not

as restrictive as category scales, and could be com-

pleted quickly; however, they may be somewhat more

difficult to understand.

0011Herbert Stone and his colleagues developed the

procedure known as quantitative description analysis

(QDA), and recommended the use of line scales for

quantifying sensory dimensions, maintaining the use

of the nine-point hedonic scale for the hedonic dimen-

sion of products. In the original description of QDA,

both sensory and hedonic dimensions were measured

on 6-in (15-cm) line scales with anchors (0–5) located

1–3 cm from each end (labeled ‘dislike moderately,’

and ‘like moderately’) and a midpoint (‘neither’). It

was argued that the scales were interval scales and

empirical evidence was presented to support this.

Ratio Scales

0012In the 1950s, SS Stevens at Harvard University

developed a variety of ratio scales that permit ratio

comparisons of stimuli (twice as strong, one-third as

sweet, etc.). The most widely known (but not the

only) ratio methods to be adopted for use in food

research are magnitude estimation and its variant,

magnitude matching. The latter technique has only

been applied to basic sensory problems and will not

be further discussed.

0013Magnitude estimation was originally applied to a

large number of sensory modalities to establish

psychophysical functions relating physical intensity

to perceived (subjective) intensity. Magnitude estima-

tion was probably first applied to foods in 1971 by

Howard Moskowitz and Joel Sidel. The practical

importance of magnitude estimation was argued as

a ‘supplement’ to category scaling, and it was sug-

gested that the category scale could be used to label

the perceptual categories of acceptance, while magni-

tude estimation could be used to refine the degree of

acceptability.

0014In magnitude estimation the task for the respondent

is to rate the sensory intensity or liking of one product

in relation to another. If one product is liked twice as

much as another product, it is assigned a number

twice as large. Many studies have used magnitude

estimation to judge the perceived sensory intensity of

laboratory food models and of real foods. Far fewer

studies have used magnitude estimation to assess the

hedonic dimensions of foods, primarily because af-

fective judgments must be made by untrained con-

sumers, and magnitude estimation is a more complex

method to understanding and to use than are other

methods of scaling, e.g., the nine-point hedonic scale.

0015Several studies comparing magnitude estimation to

category and line scales for rating sensory intensity

have shown equal sensitivity, reliability and precision

among the scales. When used to scale affect, magni-

tude estimation was found by some authors to be

more sensitive than the nine-point hedonic scale,

FOOD ACCEPTABILITY/Affective Methods 2571

1. Category scales

(a) Peryam and Girardot (1952), Jones et al. (1955) − the nine-point hedonic scale

1 2 3 4 5 6 7 8 9

Dislike Dislike Dislike Dislike Neither like Like Like Like Like

extremely very much moderately slightly nor dislike slightly moderately very much extremely

(b) Lawless and Malone (1986) − nine categories

1 2 3 4 5 6 7 8 9

Little Much

2. Line scales

(a) Pangborn et al. (1989) − labeled 0−100

0 Dislike Neither like Like 100

extremely nor dislike extremely

(b) Shepherd et al. (1988), Pangborn et al. (1989) − 100 mm

Dislike Neither like Like

extremely nor dislike extremely

Dislike Like

extremely extremely

(d) Stone et al. (1974)

Dislike Like

moderately Neither moderately

3. Relative-to-ideal scales

(a) Shepherd et al. (1988) − 100 mm

Not nearly Just right Much too

strong enough strong

4. Ratio scales

Magnitude estimation:

(a) Moskowitz and Sidel (1971) − first sample served as standard; no zeros or negative numbers permitted.

(b) Pearce et al. (1986) − unipolar scale from zero to infinity (liking).

(c) Pearce et al. (1986) − bipolar scale from minus infinity (disliking) through zero (neutral) to plus infinity (liking).

(d) Giovanni and Pangborn (1983) − unipolar scale; no zeros or negative numbers permitted; fixed modulus of 10.

Labeled magnitude scales

(a) Green et al. (1993) − equally spaced numerical labels with ratio-determined verbal labels.

(b) Schutz and Cardello (2000) − equally spaced numerical labels with ratio-determined verbal labels.

Barely detectable

Weak

Moderate

Strong

Very strong

Strongest imaginable

Greatest imaginable

Like extremely

Like very much

Like moderately

Like slightly

Neither like nor dislike

Dislike slightly

Dislike moderately

Dislike very much

Dislike extremely

Greatest imaginable dislike

−100

−80

−60

−40

−20

100

(a)

(b)

2572 FOOD ACCEPTABILITY/Affective Methods

while others found the hedonic scale to be more reli-

able, although both scales were linearly related. A

collaborative study by Pearce was sponsored by the

American Society for Testing Materials (ASTM) and

compared the nine-point hedonic scale with both

unipolar and bipolar (like–dislike) magnitude estima-

tion. Eight fabric samples were rated by 20–30 panel-

ists at 23 different sites. All three scales ranked the

eight fabrics in the same order with very similar

spread of ratings; they achieved very similar degrees

of precision in differentiating samples, and similar

discrimination. Reliability of the category scale was

higher than for either form of magnitude estimation,

although higher reliability would likely be obtained

with larger sample sizes.

0016 Thus, it appears that, while magnitude estimation

provides valid and reliable data on the sensory and

hedonic attributes of food, its practical difficulty in

use has limited its broad-based application, in favor

of the simplicity of category scales. However, in a

recent effort to combine the practical simplicity of

category scales with the quantitative sophistication

of ratio scaling for rating the intensity of oral sensa-

tions, a hybrid, category-ratio scale was developed by

Barry Green. This scale places verbal labels of ex-

pressed intensity along a linear numerical scale at

specific locations that reflect the numerical ratios

among their perceived intensities, as obtained by a

magnitude estimation procedure. A critical character-

istic of this scale is the presence of a phase involving

the ‘maximal,’‘strongest imaginable’ or ‘strongest

possible’ sensation. The use of this phase is assumed

to serve as a fixed endpoint of sensation that places

judgments of different subjects on a common sensory

‘ruler.’ Subjects can quickly look at the verbal labels

and corresponding numbers and then place a slash

mark (/) through the line scale to indicate the per-

ceived strength of their sensation. The resulting data

have been shown to have ratio properties. More

recent studies of this scale have shown that the scale

is most effective for application to chemosensory

stimuli that include extremely intense (painful)

sensations.

0017Recently, the use of such labeled magnitude scales

has been extended to affective ratings by Howard

Schutz and Armand Cardello. Using Barry Green’s

approach, the semantic meaning of various verbal

phrases used to describe the liking/disliking of foods

was scaled. By locating these phrases along a visual

analog scale in accordance with their determined se-

mantic meaning, it was shown that consumers could

use these scales to rate the affective aspects of stimuli

with equal or better sensitivity, greater reliability, and

equivalent ease of use as with the nine-point hedonic

scale.

Alternative Scales: Relative-to-Ideal,

Food Frequency

0018While category, line, and ratio scales are the most

commonly used instruments for assessing food

acceptability, they are not the only ones. Another

approach to scaling affective reactions to foods is

the relative-to-ideal scale, which combines the affect-

ive or hedonic component with the sensory compon-

ent. For example, rather than separately scaling

sweetness and liking, the relative-to-ideal scale com-

bines the two dimensions to yield a scale from ‘much

too sweet’ to ‘not nearly sweet enough.’ The midpoint

of such scales is usually defined as ‘just right.’ Such a

scale implies that the respondent has an internal ideal

of a product against which to compare a stimulus.

5. Frequency scales

(a) Schutz (1965)

I would eat this every opportunity I had.

I would eat this very often.

I would frequently eat this.

I like this and would eat it now and then.

I would eat this if available but would not go out of my way.

I don't like it but would eat it on occasion.

I would hardly ever eat this.

I would eat this only if there were no other food choices.

I would eat this if I were forced to.

Never tried.

(b) Meiselman (1988) − each food can be rated for preferred frequency in terms of days per week (1−7) and

weeks per month (7−28 times) or rated 'never.' There are separate ratings for each meal.

Breakfast Mid day Evening meal

days per weeks/ days per weeks/ days per weeks/ Never

week month week month week month

Food name 1 2 3 4 5 6 7 1 2 3 4 1 2 3 4 5 6 7 1 2 3 4 1 2 3 4 5 6 7 1 2 3 4 0

fig001b Figure 1 Examples of sensory, hedonic, and frequency scales. Most of the scales can function as either sensory or hedonic scales

by changing from sensory attributes (e.g., salty) to hedonic attributes (e.g., like) or vice versa.

FOOD ACCEPTABILITY/Affective Methods 2573

Hence it is appropriate for established products, but

less appropriate or inappropriate for novel products

or for laboratory food models (e.g., sugar water).

0019 A number of studies have shown good agreement

between relative-to-ideal scales and hedonic scales.

The breakpoint in the hedonic function, at which

responses shift from increasing liking to decreasing

disliking, is highly correlated with the just-right

point, although the latter tends to be slightly higher.

Shepherd has argued that relative-to-ideal scales are

‘unfolded hedonic scales,’ i.e., scales which measure

the amount of like or dislike from the just-right point

of the scale. One advantage of the relative-to-ideal

scale is that it produces a sharp fulcrum, below which

are less-liked, weaker points, and above which are

more-liked, stronger points. Hedonic judgments

across the same range will yield a rounded break-

point, because different respondents have different

maximally preferred points.

0020 The relative-to-ideal scale appears useful for quan-

tifying the hedonic aspect of a product along a clear,

definable intensity dimension. The scale has been

used widely in the food industry because it tells prod-

uct developers in which direction to make product

changes along a sensory dimension (e.g., make it

more sweet). It is unclear how this would be applied

to complex food products with multiple physical

dimensions (many tastes, smells, textures) and to

combinations of food products. It is also unclear

what level of training or expertise is needed for opti-

mal use of this scale.

0021 Another approach for quantifying food acceptabil-

ity is through the use of food frequency scales. The

basic concept here is that the frequency with which an

individual consumes a food or expects to consume it

is a valid index of the acceptability of the food to that

person. Similarly, since the practical purpose of any

measure of food acceptability is to predict subsequent

consumer behavior toward the food, simple measures

of affect may not suffice. For example, although

cheesecake may receive a rating of eight on a nine-

point hedonic scale, while milk may receive a rating

of six, this does not mean that the individual wants to

eat cheesecake more often than he/she wants to drink

milk. Thus, hedonic judgments of the acceptability of

foods are not as useful for menu planning or for

predicting food choice as are judgments of the desired

frequency of serving.

0022 Howard Schutz developed the first form of food

frequency scale was developed as a supplement to the

nine-point hedonic scale. Known as the food action

(FACT) scale, it consists of a series of statements

concerning the frequency with which a food item is

desired to be eaten, ranging from ‘eat every opportun-

ity I get’ to ‘eat if forced.’ By combining hedonic scale

data with FACT scale data, information can be

obtained about both the affective aspects of a food

item and the relative frequency that it is desired. The

FACT scale was shown to have high group reliability

coefficients of 0.97.

0023While the FACT scale provides only relative fre-

quency estimates, other frequency techniques can be

used to provide absolute frequency measures. A fre-

quency scale was designed to assess food preferences

of military men. Respondents indicate the number of

times during a 30-day period that they would like to

have an item served. Other research from Natick

indicated high group reliability for such frequency

scales (r ¼ 0.98), moderate individual reliability coef-

ficients (r ¼ 0.58–0.60), and moderate correlation

with consumption (r ¼ 0.59–0.66).

Affective Judgments: The Role of Context

and Expectations

0024The study of food acceptability and the measurement

of its sensory and affective dimensions are compli-

cated by the fact that foods are typically eaten within

a context, i.e., the meal, and that the consumer typic-

ally brings to the dining situation a host of expect-

ations that are based on previous experience, product

and nutrition information, and packaging, all of

which can affect liking for the food. Such influences

can limit the predictive validity of laboratory-based

measures of food acceptance unless adequate research

methods are developed to understand and control

these effects. In terms of situational context, few

studies have adequately assessed the contribution of

the meal context/situation to liking. Herbert Meisel-

man and his colleagues have studied the influence of a

number of contextual variables was studied. All stud-

ies showed changes in the probability of food selection,

while some studies showed changes in acceptance and

some did not. For example, the effort required to

obtain food showed large selection rate changes but

no reliable changes in acceptance, whereas changes in

food cost changed both selection rates and acceptance.

Further studies identified a major difference in accept-

ance ratings of foods in institutional vs noninstitu-

tional food service settings, in agreement with

previous observations concerning expectations in

these two different classes of environments.

0025In an attempt to develop a measure of food

acceptability that takes into account the situational

influences of the dining situation, Howard Schutz

developed the concept of ‘situational appropriate-

ness’ as well as a method to index the degree of

appropriateness of any food in a wide range of situ-

ations. This scale has been used effectively in a variety

of survey formats, and more recently has been shown

2574 FOOD ACCEPTABILITY/Affective Methods

to be a valuable adjunct to laboratory affective testing

of foods.

0026 Concerning the expectations that consumers bring

to the dining situation, Armand Cardello and other

researchers have shown that a wide variety of extrin-

sic variables, such as product information, packaging,

brand names, and nutrition statements, can influence

the consumer’s expected liking for a product. In nu-

merous studies designed to look at the role of these

expectations on actual liking and behavior toward

the food, expectations were manipulated so as to

lead consumers to expect a better-liked or a worse-

liked product. By measuring expected liking using a

simple category scale and comparing changes in

actual liking relative to a baseline rating of the ac-

ceptance of the food in a blind (no information) con-

dition, it has been shown that product acceptance

assimilates the level of expectation. That is, if expect-

ations are higher than the baseline acceptance of the

product, rated acceptability increases. If expectations

are lower, rated acceptability decreases. These import-

ant influences of consumer beliefs and expectations

on product perception have been demonstrated for

purely sensory dimensions of food and have been

shown to have a direct influence on behavior, e.g.,

consumption.

Which Method to Use

0027 When conducting basic research or practical product

evaluations, the question arises as to which is the

best scale method for measuring food acceptability.

There are two parts to the answer. First, it is usually

far better to select a scale that has been carefully

developed and tested in a wide variety of situations,

including conditions similar to those planned.

Second, if any of the commonly used scales (Figure 1)

are employed, there is probably not a great deal of

difference among them. Category scaling and line

scaling appear to be easier for people to use than

magnitude estimation, although the use of a semantic

ratio scale will provide ratio level data with equal

ease of use to that of category and line scales. Rela-

tive-to-ideal scales combine a sensory and a hedonic

dimension, if it is thought that people have a know-

ledge of what is ‘just right’ for the product, while

frequency scales provide more meaningful data for

menu planning and predicting repeat consumption

of foods. Lastly, caution must always be used in gen-

eralizing the results of laboratory-based affective

studies to real-life eating situations, in order to

avoid the influences of both situational context and

consumer expectations. In view of the increase in

published methodological research on sensory and

affective scaling in the past decade, the strengths

and weaknesses of different scaling methods under a

variety of conditions of use will be better clarified in

the future.

See also: Sensory Evaluation: Sensory Difference

Testing; Sensory Rating and Scoring Methods;

Descriptive Analysis

Further Reading

Giovanni ME and Pangborn RM (1983) Measurement of

taste intensity and degrees of liking of beverages by

graphic scales and magnitude estimation. Journal of

Food Science 48: 1175–1182.

Green BG, Shaffer GS and Gilmore MM (1993) Derivation

and evaluation of a semantic scale of oral sensation

magnitude with apparent ratio properties. Chemical

Senses 18: 683–702.

Jones LV, Peryma R and Thurstone LL (1955) Development

of a scale for measuring soldiers’ food preferences. Food

Research 20: 512–520.

Lavenka N and Kamen J (1994) Magnitude estimation

of food acceptance. Journal of Food Science 59:

1322–1324.

Lawless HT and Heymann H (1998) Sensory Evaluation of

Food. New York: Chapman and Hall.

Lawless HT and Malone GJ (1986) The discriminative effi-

ciency of common scaling methods. Journal of Sensory

Studies 1: 85–98.

MacFie HJH and Thomson DMH (1994) Measurement

of Food Preferences. Glasgow: Blackie Academic and

Professional.

McDaniel MR and Sawyer FM (1981) Preference testing

of whiskey sour formulations: magnitude estimation

versus the 9-point hedonic. Journal of Food Science 46:

182–185.

Meiselman HL (1988) Consumer studies of food habits. In:

Piggott JR (ed.) Sensory Analysis of Foods, 2nd edn, pp.

267–334. London: Elsevier Applied Science.

Meiselman HL (2000) Dimensions of the Meal: The

Science, Culture, Business, and Art of Eating. Gaithers-

burg, MD: Aspen.

Meiselman HL and MacFie HJH (1966) Food Choice Ac-

ceptance and Consumption. London: Blackie Academic

and Professional.

Moskowitz HR (1983) Product Testing and Sensory

Evaluation of Food. Westport, CT: Food and Nutrition

Press.

Pangborn RM, Guinard JX and Meiselman HL (1989)

Evaluation of bitterness and caffeine in hot chocolate

drink by category, graphic and ratio scaling. Journal of

Sensory Studies 4: 31–53.

Peryam DR and Girardot NF (1952) Advanced taste-test

method. Food Engineering 24: 58–61.

Pearce JH, Korth B and Warren CB (1986) Evaluation of

three scaling methods for hedonics. Journal of Sensory

Studies 1: 27–46.

Schutz HG (1965) A food action rating scale for measuring

food acceptance. Journal of Food Science 30: 365.

FOOD ACCEPTABILITY/Affective Methods 2575

Schutz HG and Cardello AV (2001) Development of a

labeled affective magnitude (LAM) scale for assessing

food liking/disliking. Journal of Sensory Studies 16:

117–159.

Shepherd R, Griffiths NM and Smith K (1988) The rela-

tionship between consumer preferences and trained

panel response. Journal of Sensory Studies 3: 19–35.

Shepherd R, Smith K and Farleigh CA (1989) The relation-

ship between intensity, hedonic and relative-to-ideal

ratings. Food Quality and Preference 1: 75–80.

Stone H, Sidel J, Woolsey A and Singleton RC (1974) Sens-

ory evaluation by quantitative descriptive analysis. Food

Technology 29: 24–34.

Market Research Methods

H R Moskowitz, Moskowitz Jacobs Inc., White Plains,

NY, USA

The Market Research Role in the

Corporation

0001 Market research as a profession has the job of sup-

porting the marketing function. For the past 40 years

corporations worldwide have used market research to

identify consumer trends, measure the acceptance of

concepts, and measure the acceptance of products.

0002 Acceptance measurement takes on a variety of dif-

ferent aspects. At the one end lies small-scale taste-tests

in the corporate facility. These tests are usually exe-

cuted by the in-house sensory evaluation group. There

are slightly larger-scale research studies run with local

consumers. These are the so-called church panels.

Church panels may be executed either by a market

researcher or by the in-house sensory analyst. The

researcher contracts with local volunteer organiza-

tions to provide a source of panelists. For the larger-

scale studies, involving hundreds of consumers,

marketing researchers usually play the key role, rather

than sensory analysts. Marketing researchers run these

types of studies, often across a country to obtain geo-

graphic dispersion, and often with a relatively limited

number of products. The role of the marketing re-

searcher in these larger-scale studies is to present a

report card of the product, in terms of performance,

either in an absolute sense or relative to competition.

The Market Research Role versus the

Sensory Research Role in Acceptance

Testing

0003 In today’s corporations market researchers and

sensory analysts each have responsibility for primary

research on product acceptance. Sensory analysts

often run small-scale studies to identify which par-

ticular formulation among a set of formulations looks

promising for further development. Sensory analysts

are therefore called in to execute acceptance tests very

early in the project. Sensory analysts also measure

acceptance as part of a general profiling exercise

wherein panelists rate the different characteristics of

products. The role of the sensory analyst in accept-

ance testing is to support product development, and

not to support marketing. Therefore, the acceptance

evaluations run by sensory analysts are usually con-

sidered directional in terms of ongoing development,

rather than constituting a definitive ‘go/no’ decision

for marketing. These guidance tests, none the less,

often involve consumers.

0004Marketing researchers interact most usually with

marketers and general management rather than with

research and development (R&D). By the time a

product has reached the marketing researcher,

the product is presumed to be optimally acceptable.

The marketing researcher tests the product among

many individuals, in multiple locations, and often

against a market leader, or against the current

in-market product. The goal is not so much to

provide guidance for development (such guidance

having been assumed to be already in place and

implemented), but rather to measure the potential of

the product in terms of market share. From time

to time the acceptance testing run by market

researchers may be used to identify potentially fatal

flaws in the product. For the most part if the

sensory analysts have done their job properly such

fatal product flaws should already have been

eliminated.

The Multiple Aspects of Acceptance

Measurement

0005Market research techniques involve a number of

different aspects of acceptance measurement. Each

of these aspects is investigated regularly by the

researcher in an attempt to help the corporation

develop the best product possible. These aspects

include the following:

0006basic product acceptance without the presence of a

concept

0007product–concept fit (does the concept fit the

product, and vice versa?)

0008frequency of use

0009purchase intent coupled with price information

0010replacement of different foods by the product

0011appropriateness measurement – when would the

product be eaten (e.g., day-part)?

2576 FOOD ACCEPTABILITY/Market Research Methods

7.0012 relevance of the product as part of a line of

products

0013 loss of acceptance due to sensory-specific satiety

and/or boredom/habituation

0014 branded versus blind acceptance

Basic Product Acceptance

0015 One key corporate goal is to offer a product with a

high level of liking, packaged attractively, at an af-

fordable (and competitive) price. In order to insure

final product acceptance the market researcher uses a

variety of procedures. Two of the most common pro-

cedures are monadic tests and paired comparison

tests.

0016 Monadic tests These tests require the consumer to

rate the product on a scale of acceptability. There are

many such scales (Table 1). These tests simply ask the

consumer to act as a measuring instrument, assigning

a number or a verbal label to represent product ac-

ceptance. On the analytic side the researcher averages

the ratings to obtain a measure of acceptance. The

results are reported as a score card of the products.

Statistical tests are used in order to determine whether

or not the product being tested differs significantly

from the competition. Whatever scale is used must be

amenable to statistical analysis, such as tests of differ-

ences (viz., t-tests).

0017Paired comparison tests These tests pit one product

against another. One of the products may be the

current product, another may be a revised formula-

tion or a competitor, depending upon the business

objectives. The consumer selects which of the two

products is preferred, either overall or on specific

characteristics. Quite often commercial tests require

the consumer to indicate preference on many differ-

ent attributes (up to 20 or more). The researcher

computes the proportion of panelists who choose

one product over another, for each particular attri-

bute being investigated. The results are reported in

the form of percentages. The paired comparison tests

are used for several purposes, included measurement

of performance against the competition, as well as

claims substantiation (viz., product X is preferred to

product Y).

0018Results reported for many subgroups For both mo-

nadic and paired comparison tests the market re-

searcher looks for the data analyzed by total panel,

and by the different key subgroups in the population.

It is not unusual for a market researcher to work with

upwards of 10 or 20 different subgroups in the

population. Since the researcher deals with 300þ

panelists in many of these studies, and since there

is a classification questionnaire accompanying the

actual evaluations, the researcher can cut the data

many different ways.

Product–Concept Fit

0019Market researchers often test products in the presence

of a concept. The concept tells the consumer what to

expect. The concept may stress the benefits of the

product (positioning), or the features of the product

(product concept), or some combination of both.

The results of the product concept fit show either

the degree to which the product is acceptable based

upon expectations set up by the concept, or the

degree to which the product delivers what the concept

promises. When products are tested in the presence

of a concept the researcher gets a much better idea

of the actual performance of the product after

the consumer better understands the reason for

the product and its end uses (e.g., appropriate time

of day for consumption; pricing, etc., specific health

benefits). Occasionally the product may score poorly

without the presence of a concept, especially when

the product is a health-oriented product that by itself

is not particularly acceptable, but provides other

benefits, such as nutrition and energy. The basic

tbl0001 Table 1 Types of hedonics scales

Two-point Five-point Nine-pointhedonic scale

Like Very good Like extremely

Dislike Good Like very much

Moderate Like moderately

Tolerate Like slightly

Three-point Never tried Neither like nor dislike

Dislike slightly

Acceptable Very good Dislike moderately

Dislike Good Dislike very much

Never tried Moderate Dislike extremely

Dislike

Like a lot Tolerate

Dislike Purchase intent scale

Do not know

FA CT s c a l e Definitely would buy

Well liked Eat every opportunity Probably would buy

Indifferent Eat very often Might/might not buy

Disliked Frequently eat Probably not buy

Eat now and then Definitely not buy

Eat if available

Don’t like–eat on

occasion

Hardly every eat

Eat if no other choice

Eat if forced

Reproduced from Meiselman HL (1978) Scales for measuring food

preference. In: Petersen MS and Johnson AH (eds) Encyclopedia of Food

Science, pp. 675–678. Westport: AVI, with permission.

FOOD ACCEPTABILITY/Market Research Methods 2577