DOI: 10.1111/j.1745-459X.2009.00236.

CONSUMER ACCEPTANCE OF COW’S MILK VERSUS SOY

BEVERAGES: IMPACT OF ETHNICITY, LACTOSE TOLERANCE
AND SENSORY PREFERENCE SEGMENTATION joss_236 731..748

O.M. PALACIOS1, J. BADRAN1, M. ANNE DRAKE2, M. REISNER3 and

H.R. MOSKOWITZ3,4
1
National Dairy Council
Dairy Management Inc.
10255 West Higgins Road
Rosemont, IL 60018
2
Department of Food, Bioprocessing and Nutritional Sciences
Southeast Dairy Foods Research Center
North Carolina State University
Raleigh, NC 27695
3
Moskowitz Jacobs, Inc.
1025 Westchester Ave.
White Plains, NY 10604

Accepted for Publication January 24, 2009

ABSTRACT

In a large-scale study of milk and milk-substitute beverages appropriate

for individuals with lactose intolerance, a total of 893 consumers rated
reduced-fat (2%) cow’s milk the highest, followed by fat-free cow’s milk, and
in turn, followed by soy-based milk-substitute beverages. No interactive effect
on overall liking was found for ethnicity, age or gender, or lactose tolerance/
intolerance, with the pattern of liking following that established for the total
panel of 893 participants. Two sensory preference segments emerged – sweet
seekers (37%) and cream seekers (63%), respectively. These segments tran-
scended conventional geo-demographics and lactose tolerance. Both seg-
ments assigned higher liking ratings to white milk products than to white soy
products. However, the sweet seekers liked the reduced-fat and fat-free cow’s
milk products about equally, while the cream seekers preferred the reduced-fat
cow’s milk product.

4
Corresponding author. TEL: 914-421-7400; FAX: 914-428-8364; E-MAIL: mjihrm@[Link]

Journal of Sensory Studies 24 (2009) 731–748.

© 2009, The Author(s) 731
Journal compilation © 2009, Wiley Periodicals, Inc.
732 O.M. PALACIOS ET AL.

PRACTICAL APPLICATIONS

Cow’s milk is recommended in the Dietary Guidelines for Americans

because of its unique nutrient package. Because of beliefs about lactose
intolerance, nondairy beverages are being recommended for those with lactose
intolerance, including dairy substitutes in federal feeding programs. As there
have been limited studies regarding the acceptance of substitute beverages, we
conducted a study to determine “white” milk and milk-substitute beverage
product liking among consumers. Findings indicated that lactose-free cow’s
milk is significantly better accepted than all tested soy-based beverages.
Furthermore, the lactose-free, reduced-fat cow’s milk is better accepted than
the lactose-free, fat-free cow’s milk. Thus, the practical application of this
research is to provide new information to health professionals and other
stakeholders about taste preferences for milk, and substitute beverages to aid
in making recommendations for obtaining dairy nutrients for those with
lactose intolerance.

INTRODUCTION

The 2005 United States Department of Agriculture Dietary Guidelines

for Americans (hereafter cited as Dietary Guidelines for Americans or
USDA DGA; USDA, Center for Nutrition Policy and Promotion 2005) and
[Link] website (USDA FNS 2008b) recommend that children and
adults consume three servings of low-fat or fat-free milk or milk products
including milk, cheese or yogurt each day as part of a healthy, balanced diet
(USDA, Center for Nutrition Policy and Promotion 2005). With its unique
nutrient package, cow’s milk is an important part of federal feeding programs
such as Women, Infants and Children (hereafter cited as WIC) and the Child
and Adult Care Feeding Program (hereafter cited as CACFP; USDA FNS
2008a). Ethnic minority groups make up a higher percentage of the clientele in
some nutrition programs, such as WIC, than does the general population
(USDA FNS 2008a) Earlier tests and reports suggest that some ethnic minority
groups have a higher incidence of lactose intolerance than do Caucasians
(Rusynyk and Still 2001). More recently, several studies indicate that lactose
intolerance is often overestimated and falsely diagnosed, particularly when
the perceived condition is self-diagnosed (Savaiano 2003). In self-reported
surveys, far fewer members of minority groups, including African-Americans,
report being lactose intolerant or are avoiding dairy foods than that cited in the
widely available population-based prevalence rates (Wooten and Price 2004).
The Dietary Guidelines for Americans and the National Medical Asso-
ciation both recommend dairy foods as the first choice for individuals who are
lactose intolerant in order to help meet their nutrient needs (Wooten and Price
 COW’S MILK VERSUS SOY BEVERAGES 733

2004; U.S. Department of Health and Human Services and USDA 2005). The
Dietary Guidelines for Americans state that milk alternatives within the milk
group, such as lactose-free milk, are the most reliable and the easiest way for
individuals with lactose intolerance to acquire the health benefits associated
with milk and milk products. The consumption of milk products is associated
with overall diet quality and adequate intake of nutrients including calcium,
potassium, magnesium, zinc, iron, riboflavin, vitamin A, folate and vitamin D
(U.S. Department of Health and Human Services and USDA 2005).
Lactose intolerance can be a barrier to consuming the recommended
servings of nutrient-rich dairy foods. In light of the health benefits of milk yet
recognizing that some participants in nutrition programs have lactose intoler-
ance, nondairy alternatives are now offered in WIC (USDA FNS 2008a). WIC
allows lactose-free milk, but the new package has also added soy beverage as
a milk-substitute beverage (USDA FNS 2008a). Although various product
choices comprise an important component of food and nutrition programs,
offering nondairy options to replace milk may compromise nutritional quality
because cow’s milk (both conventional and lactose-free) and soy-based bev-
erages are not nutritionally equivalent. In order to maximize the availability of
healthful and good-tasting products to improve key dairy nutrient intake, it
may also be important to consider the taste preferences of individuals for milk
and milk-substitute beverages.

PRODUCT ACCEPTANCE

Acceptance and usage constitute an additional consideration for these

milk-substitute beverages. In addition to nutritional content, the product must
taste good or the product and the nutrients it has to offer will not be consumed.
In support of this notion, the Dietary Guidelines for Americans acknowledges
that enhancing the palatability of a product will improve nutrient intake (U.S.
Department of Health and Human Services and USDA 2005). The taste accep-
tance of current or proposed milk-substitute beverages need to be determined
by experiment. There are reports that taste, ethnicity and cultural practices can
influence food choices, and may lead to altered consumption of a food and the
nutrients it provides (Decker 2004). Recent studies looked at the acceptance
of chocolate milks and yogurts among Hispanic, Caucasian and African-
American consumers (Thompson et al. 2007a,b). These studies indicated that
ethnicity played a minor role in preferences shown for specific chocolate milk
or drinkable yogurt flavor products, tested “blind.” However, such information
about ethnic effects is not readily available in the literature regarding milk-
substitute beverages such as soy-based beverages.
A variety of studies have assessed the effects of different factors on the
sensory characterization and acceptance of cow’s milk products. These studies
734 O.M. PALACIOS ET AL.

looked at milk pasteurization temperature, fat content and feeding regime.

They did not consider lactose-free milks (Quinones et al. 1997, 1998;
Chapman and Boor 2001; Chapman et al. 2001; Clare et al. 2005; Croissant
et al. 2007; Gandy et al. 2008) Furthermore, although sensory properties of
soy-based beverages have also been documented, to our knowledge, published
studies have not compared consumer liking of lactose-free milks and soy-
based beverages (N’Kouka et al. 2004).

MULTI-ATTRIBUTE APPROACHES TO PRODUCT EVALUATION

AND THE DISCOVERY OF ACCEPTANCE DRIVERS

Despite the importance of milk and milk-substitute products to the nutri-

tion of children and adults, a literature search suggests only a relatively few
systematic explorations of the sensory attributes of these products and the
drivers of liking. Among the reasons for the paucity of such data is the fact that
many research programs focus on the evaluation of one product rather than a
number of products. With one product and many attributes, the objective is to
see how well the product performs and what is wrong with the product, if
anything, rather than looking for patterns of attributes that drive acceptance. In
addition, traditional ways of thinking about product acceptance continue to be
brand usage, and secondarily, market. There is a lack of published literature
dealing with ethnic differences in sensory and acceptance response to foods, as
well as a similar lack of studies on patterns underlying variability in sensory
preferences in actual foods (so-called sensory segmentation; Moskowitz
1994).
Our organizing principles going into this research was that one could
measure sensory and acceptance responses to a variety of milk and milk-type
beverages, that there would be clear differences in ethnicities as they respond
to the products and that there might be sensory preference segments, even for
a product that is relatively simple such as milk.

SCOPE OF THE PAPER

The objective of this study was to determine the liking of lactose-free

cow’s milk and soy-based beverages among adult (ages 18–64 years) lactose-
tolerant and lactose-intolerant Caucasian, African-American and Hispanic
adults, respectively.
This paper focuses on the acceptance response of lactose tolerant and
intolerant consumers to six different “white” lactose-free milk and soy-based
beverages. The primary objective of the paper is to provide a normative
 COW’S MILK VERSUS SOY BEVERAGES 735

database (or representative sample) for the response to these products, and to
determine whether or not lactose-intolerant individuals show an altered level
of acceptance. The secondary objective is to identify how sensory attributes of
the product drive acceptance, whether these sensory-liking patterns differ by
lactose tolerance, as well as by age, gender or other “latent” grouping of the
respondents (“sensory segments”).
Because the objective of the study is to generate a normative database or
representative sample, the study was run across the U.S.A. in four different
geographic markets among a large group of 893 lactose-intolerant and lactose-
tolerant consumers.

METHOD

Products
The products reported here comprise of six stimuli from a full set of
stimuli tested by both lactose-tolerant and lactose-intolerant respondents. The
six products were tested among a set of other products, as part of a large-scale
study of cow’s milk and soy-based beverage products, with flavors comprising
plain, vanilla and chocolate. We focus here on select white, non-chocolate
products that were appropriate for and evaluated by self-identified lactose-
intolerant as well as lactose-tolerant respondents.
The set of products whose ratings are reported here appears in Table 1.
There were other products, such as chocolate milk and soy beverages, tested
either by lactose intolerants or by lactose tolerants, whose results are not
presented here. All test products were acquired from different stores across the
U.S.A., were confirmed to be consistent with each other within the prescribed
code date, shipped to the four central location test facilities and evaluated
while still within the allowable freshness date.

TABLE 1.
PRODUCTS TESTED IN THE STUDY

Product Flavor, type

1 Cow’s milk, lactose-free, 2% (reduced-fat) White LF

2 Cow’s milk, lactose-free, fat free White LF
3 Cow’s milk, lactose-free, fat-free, fortified White LF
4 Soy beverage, plain, organic Plain soy
5 Soy beverage, vanilla Vanilla soy
6 Soy beverage, plain, fortified Plain soy

LF, low fat.

736 O.M. PALACIOS ET AL.

TABLE 2.
SCREENING REQUIREMENTS FOR THE RESPONDENTS

Aged 18–64 years old.

50% females and 50% males.
Must currently drink lactose-free milk or would be (or are) willing to try lactose-free milk.
Not on strict medical diet/does not have food allergies to flavors of milk being tested.
Security and industry screen.
No past 6 month participation in food or beverage related market research.
Lactose-intolerant respondents n = 438.
(Defined as gastrointestinal problems after drinking milk or gastrointestinal issues with dairy
products other than milk).
Lactose-tolerant respondents n = 455.
(Defined as being able to drink milk or have dairy products without any type of gastrointestinal
symptoms).

Respondents
Respondents were pre-recruited via telephone, to meet screening criteria
(age balance 18–64 years old, gender balance, ethnic balance, and equal
numbers of lactose tolerant and intolerant, etc.). The respondent quotas were
set up to allow an analysis of perception and acceptance patterns by a wide
variety of respondents, divided primarily by lactose tolerance versus intoler-
ance, but also representing different ethnicities and degree of acculturation.
Table 2 presents the screening criteria. Table 3 presents the final panel com-
position, as well as shows the consumption patterns for this large-scale study.

Evaluation Procedure
The evaluation followed the sequence of steps shown in Table 4. The use
of pre-recruited respondents who participate for several hours over a 2-day
period has become a well-accepted practice in sensory and market research.
The extended time of the interview allows the respondents to evaluate a
number of different products in a leisurely pace. The respondents are enabled
to concentrate on the task, and maintain their motivation because they are
compensated at the end of the evaluation (Moskowitz 1985).
The actual evaluation was done by recruiting the respondents, bringing
them in groups of 20–25 respondents to a well-monitored central location.
There were trained interviewers at the location. The respondents sat in front of
individual computers, with the computer directing the interview, and paced it.
Each respondent went through a different sequence of products, following the
computer prompts.
Both the computer program and a printed sequence showed the respon-
dent what product to evaluate, and thus controlled the order and pace of
 COW’S MILK VERSUS SOY BEVERAGES 737

TABLE 3.
PANEL COMPOSITION, CONSUMPTION PATTERNS AND NATURE OF LACTOSE
INTOLERANCE. WHERE APPROPRIATE. ALL NUMBERS ARE PERCENTS, BASED ON THE
TOTAL INCIDENCE FOR THE SPECIFIC COLUMN

Total Lactose Lactose

tolerant intolerant

Total incidence 891 453 438

Type of milk beverage consumed most often (%)
2% Milk (reduced-fat) 30 38 21
Lactose-free milk 17 4 31
Whole milk 17 23 11
Skim/Fat-free milk 10 14 6
1% Milk (low-fat) 10 11 8
Soy milk 9 4 14
Chocolate milk 4 4 4
When typically drink milk (%)
Breakfast 85 86 84
Nighttime snack 30 34 26
Dinner 18 20 16
Lunch 13 13 12
Morning snack 11 10 12
Afternoon snack 10 12 7
Other times 8 8 8
Other 4 5 3
Ethnic background (%)
Caucasian 36 37 36
Hispanic 32 32 32
African-American 31 30 32
Gender (%)
Female 52 51 52
Male 48 49 48
Age (%)
18–34 years old 32 30 33
35–54 years old 41 41 42
55–64 years old 27 29 25
Lactose intolerance (%)
No 68 92 43
None of the above mentioned 51 90 11
Yes 32 8 57
Gastrointestinal problems after drinking milk 28 1 55
Issues (gastrointestinal problems) with dairy products 13 1 25

Percents greater than 20 are in bold.

evaluation. The computer program presented the product code, the respondent
went to take the product from a central location and then went through the
evaluation on the computer, with the program presenting the question and
checking/accepting the typed answer. At the end of each product evaluation,
738 O.M. PALACIOS ET AL.

TABLE 4.
STEPS FOLLOWED FOR THE CONSUMER EVALUATION OF THE COW’S MILK AND
SOY BEVERAGE PRODUCTS

Step Action Rationale

1 Appearance and color evaluation The appearance and color of the milk and milk
alternative products was judged, but the products
were not tasted
2 Aroma, taste and texture evaluation Respondent sniffed the products, took 2–3 sips of
the product, and then answered questions
pertaining to taste and smell, as well as
mouthfeel, overall liking, image characteristics
and purchase intent
3 Rest Respondent waited five minutes and then evaluated
the next sample. Respondent cleared palate using
unsalted crackers and water
4 Process repeated Each day of two days the respondent evaluated a
randomized 8 of 16 test stimuli
5 Classification questionnaire Respondent answered demographic and usage
questions pertaining to the product

one of the study interviewers had to type in the appropriate “code” to permit
the respondent to go to the next product. This action allowed the interviewer to
ask the respondent a question about the rating. Such interaction maintained
interest and motivation.

Attributes and Scales

Respondents evaluated each product in a randomized sequence, evaluat-
ing the product monadically. Thus, the study is a within-subjects design. The
attributes covered appearance, aroma, taste/flavor, mouthfeel and image, as
well as overall acceptance. Respondents rated attributes in terms of “amount of
a characteristic” (sensory attribute), liking of that attribute, and directionality
(too much versus too little). We present the overall liking ratings in this paper
and where relevant to explicate results from sensory segmentation, plot overall
liking versus four different sensory attribute ratings, representing appearance,
aroma, taste/flavor, and mouthfeel, respectively. The Appendix presents the
average sensory ratings for each product from the total panel.
The attributes were evaluated on a 0–100-point scale, anchored at both
ends. This scale has been used for more than 30 years to evaluate the charac-
teristics of products (Moskowitz 1985). The anchors make the scale unam-
biguous. The respondents are oriented in evaluation, and are instructed to use
the full rating scale, but otherwise untrained. Most respondents find the scale
easy to use after a short orientation, and appear to have no problem differen-
tiating between sensory, liking and directional attributes.
 COW’S MILK VERSUS SOY BEVERAGES 739

RESULTS

How the Products Score on a Blind Basis

The key evaluative criterion in this study is overall liking, rated on an

anchored 0–100-point scale. Zero denotes “dislike completely,” while 100
denotes “like completely.” The appropriate analysis for these rating scales is
the arithmetic average. To help with understanding the “meaning of a
rating,” there are general rules of thumb about performance on overall liking.
These are rough guides, but give a sense of performance. They have been
developed across many different products. Good products typically receive
mean liking scores of 60 or higher, poor products typically receive
mean liking scores of 50 and below. These norms come from observation
made in thousands of studies run during the past 27 years at Moskowitz
Jacobs, Inc. with commercial products tested on a blind basis (see
Moskowitz 1985).
Data were analyzed using SPSS, version 11.0.1 (Copyright SPSS 1989–
2001), a statistical software package that can read and write data from ASCII
text files (including hierarchical files), other statistics packages, spreadsheets
and databases. SPSS can also read and write to external relational database
tables via ODBC and SQL. In addition, we used SYSTAT, a statistics and
statistical graphics software package (SYSTAT for Windows 2004, Version 11.
SYSTAT Software Inc., Chicago, IL). Both packages are in the class of “off the
shelf, software packages.”
Table 5 shows the overall liking rating from the total panel, key geo-
demographic groups, lactose-tolerant versus lactose-intolerant groups and
sensory-preference segments (discussed later in this paper). The results for
the total panel point to substantial differentiation among the products, with
a 34-point range in liking. This range is highly significant (P < 0.01). The
average liking data for the dairy and dairy-substitute beverage products
suggest that the lactose-free, reduced-fat (2%) cow’s milk product is extremely
acceptable, while two of the cow’s milk products, lactose-free, fat-free and
fortified lactose-free, fat-free milks are moderately acceptable. The soy prod-
ucts substantially underperformed versus other tested products, and defined
the bottom tier of liking. Keep in mind that these products were tested in
randomized order, with product type and brand hidden. Only through sensory
cues alone can respondents evaluate the products. Thus, the first conclusion is
that the lactose-free milks score substantially better than do the soy beverage
products. The second conclusion is that the reduced-fat milk (lactose-free, 2%)
scores substantially better than do the fat-free products. The third conclusion
is that the fortification tested in this panel appears to make no difference versus
the same class of product unfortified.
 TABLE 5. 740
AVERAGE OVERALL LIKING RATINGS FOR THE SIX STIMULI TESTED BY BOTH LACTOSE TOLERANT AND
LACTOSE-INTOLERANT RESPONDENTS

Product code (delete before submitting) 102 104 106 113 108 107

Cow’s milk, Cow’s milk, Cow’s milk, lactose Soy, plain, Soy, Soy, plain,
lactose free, 2% lactose free, fat free free, Fat free, fortified fortified vanilla organic

Total panel 68 52 52 39 38 34
Geo-demographics
Male 68 53 56 41 38 35
Female 68 51 48 38 38 34
18–34 66 54 54 37 38 33
35–54 68 51 50 39 37 33
55–64 70 50 53 42 39 37
African-American 66 49 51 39 37 33
Caucasian 68 55 52 38 38 35
Hispanic 70 51 53 41 40 34
Lactose tolerance
O.M. PALACIOS ET AL.

Total Lactose Tolerant 66 50 49 37 38 33

Total Lactose Intolerant 70 54 55 42 38 36
Sensory preference segments
Total Segment 1 – sweet seeker 67 62 65 36 29 26
Total Segment 2 – cream seeker 69 38 33 44 50 45
Sensory preference segment ¥ lactose tolerance
Lactose Tolerant – Segment 1 57 63 53 33 38 24
Lactose Intolerant – Segment 1 50 39 41 62 56 63
Lactose Tolerant – Segment 2 77 35 44 42 39 42
Lactose Intolerant – Segment 2 76 57 59 37 33 29

All means are on the anchored 0–100-point liking scale.

 COW’S MILK VERSUS SOY BEVERAGES 741

The average liking rating for six products, broken out by the geo-
demographic subgroups mirrors the pattern that emerged for the total panel.
There is some, albeit not a great deal, of variability around the mean when we
look at each subgroup. Milk products still score significantly and far higher
(P < 0.05) than do soy products, and reduced-fat milk scores higher than
fat-free milk.
When the liking ratings are analyzed out by lactose-tolerant versus
lactose-intolerant respondents, the picture changes a bit. The rank order of the
types of products remains the same (milk products > soy beverage products;
reduced-fat milk > fat-free milk). However, the lactose-intolerant respondents
rate the products higher, albeit not by a constant amount. This finding is not
limited to milk. It occurs as well with the soy beverage products, suggesting
that the lactose-intolerant respondents may show a slight 3–5-point response
bias to up-rate milk and milk-substitute beverages. Other research work and
observations by the authors suggest that this 3–5-point has little actual
meaning in the market, other from a small response bias.

Sensory Preference Segmentation

People differ in the types of products they like. Some individuals prefer
spicy products, others prefer bland products. Such variation in sensory pre-
ferences has been recognized for model systems of sugar-sweetened water
(Pangborn 1970), for foods (Moskowitz 1994), and has even caught the atten-
tion of the popular press (Gladwell 2004).
Sensory segments are developed by relating overall liking to sensory
attributes, creating sensory-liking curves and determining the sensory level at
which overall liking maximizes. This analysis creates a vector of sensory level
corresponding to an individual’s optimum profile. The vectors are then used as
inputs to cluster analysis, which emerges with different groups of respondents,
showing different patterns of optimal sensory levels. The patterns are devel-
oped through a scatter plot, with the abscissa coming from the ratings of the six
products on a specific, nonevaluative sensory attribute such as light-dark.
Figure 1 suggests that Segment 1 can be described as individuals who
prefer lighter, low-aroma, sweeter, smoother products. If we were to name this
segment, we might call it the “sweet seekers” segment. In contrast, Segment 2
“cream seekers” prefer stronger milk aroma and flavor, creamy and smooth.
Segment 2 does not show a pattern that we can call opposite to the pattern
shown by Segment 1, so Segment 2 is not searching for a high-impact product.
Rather, Segment 2 appears to be searching for a medium-impact product.
It is important to keep in mind that the sensory segments for milk and soy
beverage products are developed on the basis of the products actually tested in
the study, rather than from attitudes about what people want. That is, the
742 O.M. PALACIOS ET AL.

80 80

S1
Overall Liking

60 60

T
S2
40 40
T S2
S1
20 20
20 40 60 80 20 40 60 80
Appearance - Light to Dark Aroma - Strength Overall

80 80
S1

T
Overall Liking

60 60
S2
S2
40 40
S1

T
20 20
20 40 60 80 20 40 60 80
Taste - Sweetness Mouthfeel - Smooth

FIG. 1. SENSORY-LIKING CURVES FOR FOUR ATTRIBUTES, FOR TOTAL PANEL (T), AND
BOTH THE SWEET SEEKER (S1) AND THE CREAM SEEKER (S2). THE CURVES ARE
FITTED TO THE DATA BY A QUADRATIC FUNCTION TO SHOW THE IDEALIZED
NATURE OF THE SENSORY-LIKING RELATION

sensory segments reflect the patterns of what people actually like when they
are confronted by the products. The segments might both say that they want the
same sensory profile, if asked directly. However, Figure 1 suggests that the
segments will respond quite differently, especially when presented with a very
light, weak-smelling, low-sweetness product.

DISCUSSION

There is a paucity of published quantitative research on ethnic prefer-

ences, including those who are or are not lactose intolerant to dairy products.
 COW’S MILK VERSUS SOY BEVERAGES 743

Thompson et al. (2007b) reported only minor ethnic effects on liking the
charges of chocolate milks between Hispanics and Caucasian consumers.
Focus groups conducted in this project with Hispanics and Caucasians uncov-
ered different usage patterns and attitudes toward milk as well as differences in
interest for specific flavors of milks, but when commercial chocolate milks
were evaluated quantitatively, few differences resulted. Instead, distinct pref-
erence segments emerged that were independent of ethnicity and were consis-
tent with previous chocolate milk studies with only Caucasians (Thompson
et al. 2004). Similar results were documented with strawberry-flavored drink-
able yogurts (Thompson et al. 2007a). In the case of Hispanics, acculturation
may play a critical role in product liking. Hispanics in the U.S.A. have varying
levels of acculturation status depending on length of time in the country and
contact with mainstream American culture which affects consumer behavior
(Greenberg 2000; Palumbo 2005). It is possible that while ethnicity impacts
usage and attitudes toward fluid milks and dairy products (documented in
focus groups in previous studies), these differences do not impact overall
liking and are congruent with the American population as a whole.

Product Acceptance and Feeding Programs

Ethnic minorities are overrepresented in some public nutrition programs
compared with the general population. The goal of these federal feeding
programs is to ensure that Americans, particularly those of low-income house-
holds, receive adequate or improved nutrition through the Federal Food
Program (Food Research and Action Center 2008). Americans do not consume
enough nutritious foods including milk, thus it is more important than ever that
they choose these foods first as a means to meet nutrient needs. One way to
help Americans meet their nutrient needs is to offer foods high in essential
nutrients that are also well liked. Offering foods that are well liked and highly
accepted may minimize food waste and economic burden for feeding pro-
grams, while increasing diet quality. This concept is recognized by nutrition
health professionals and is acknowledged in the Dietary Guidelines for
Americans where it states that enhancing the palatability of a product will
improve nutrient intake (U.S. Department of Health and Human Services and
USDA 2005).
As a consequence of its unique nutrient package, cow’s milk is offered in
feeding programs. Nevertheless, some feeding programs now offer eligible
nondairy substitute beverages for cow’s milk including soy-based beverages.
These substitutes are available for select participants such as those with lactose
intolerance, even though the acceptance or liking of the offered beverages has
not been evaluated. In light of this, we conducted a study to determine “white”
milk or milk-substitute beverage product liking and found that lactose-free
744 O.M. PALACIOS ET AL.

cow’s milk is significantly more liked and better accepted than all tested
soy-based beverages. Furthermore, the lactose-free, reduced-fat cow’s milk is
more liked and better accepted than the lactose-free, fat-free cow’s milk.
The results of this study support choosing cow’s milk dairy foods first for
lactose-intolerant individuals, regardless of ethnicity, age or gender, as both
reduced-fat and fat-free varieties of the lactose-free cow’s milk received the far
higher “liking” rating than did the nondairy soy-based beverages. The soy-
based beverages rated significantly lower, an important consideration if
product liking, acceptance or palatability can affect the product’s consumption
and the nutrients it contains.
Ignoring liking in a product may have negative consequences that impact
the nutritional objectives. One potential consequence of decreased milk intake
is reduced intake of important nutrients such as calcium, identified as a
“nutrient of concern” by the Dietary Guidelines as it is lacking in most
American’s diet (U.S. Department of Health and Human Services and USDA
2005). Within the “space” of cow’s milk, the lactose-free, reduced-fat milk
was better liked than the lactose-free, fat-free milk. This is another important
consideration as some feeding programs have already removed reduced-fat
cow’s milk from their list of eligible foods in an attempt to limit fat intake
among participants (USDA FNS 2008a). However, if reduced-fat cow’s milk
were to be removed, and if product liking, acceptance or palatability can affect
product consumption, one potential consequence is the reduced consumption
of the nine essential nutrients that milk offers. Although increasing options and
reducing fat intake for participants of feeding programs is important to help
achieve their goal of improving Americans’ nutrient intake and overall diet
quality, product liking and acceptability must also be considered in order to
ensure that the product will actually be consumed.

Sensory Segments as a Key Factor in Preferences

We do not typically think of milk as a product that is driven by different
flavors and mouthfeel/textures in the way that we might think of pasta sauce or
fruit-flavored beverages. That being said, the strength of the sensory segmen-
tation is surprising in light of the sensory nature of milk, with low-flavor
intensity. Although fluid milk is a relatively low-impact product, based on the
intensity of its flavor, previous studies have documented distinct consumer
preferences within this category. Chapman and Boor (2001) and Gandy et al.
(2008) both documented distinct consumer preferences in degree of cooked
flavor in pasteurized milks, confirming that consumers do notice differences in
milk flavors. Previous studies have also documented that consumers prefer
higher fat content milks in blinded acceptance studies even if they do not
typically consume these products (Pangborn et al. 1985; Phillips et al. 1995).
 COW’S MILK VERSUS SOY BEVERAGES 745

A moderately cooked flavor and low but distinct milk-fat flavor as well as
moderate viscosity and opaque color are associated with conventionally pas-
teurized low and reduced fat milks. A combination of these cues appears to
signal and then drive acceptance to milk consumers (Quinones et al. 1997).
The results from this current study are consistent with these older previous
studies.

Methodological Issues and Validity of Sensory-Oriented Tests

Quite often, the demand characteristics of a consumer task generate
artifacts in the data such as exaggerated differences between products when a
respondent is required to rate the degree of difference between two products.
In those cases, the respondent is forced to focus with almost “biased,” unusual
attention to the stimuli in order to fulfill the task requirements. In some types
of studies, a responsive individual is primed, and then likely looks for differ-
ences, even when these differences may be small and irrelevant, simply
because of the desire to fulfill the instructions.
It is quite likely that these segments uncovered in this study are real,
however. The respondents evaluated the products in a sequential monadic test,
rating eight products a day in randomized order, and rating the other eight
products on the next day. This type of design reduces the respondent’s
ability to exaggerate differences because products are never compared with
each other directly, but rather, rated one at a time on a large number of
characteristics. Only when we divide respondents by the pattern of sensory
characteristics that drive acceptance do we see differences that are coherent,
meaningful and lead to hypotheses. Otherwise, with normal divisions of
respondents according to standard criteria, we find exceptional consistency in
responses by ethnicity and by one’s state of lactose tolerance (tolerant
versus intolerant).

ACKNOWLEDGMENT

The authors thank Linda Lieberman, editorial assistant at Moskowitz

Jacobs, Inc., for reviewing and preparing this manuscript for submission and
for keep the project on publication schedule.
746 O.M. PALACIOS ET AL.

APPENDIX 1: SENSORY PROFILES FOR THE SIX PRODUCTS

AND THE SENSORY RANGE. DATA OBTAINED FROM THE
TOTAL PANEL
Product 1 2 3 4 5 6 Range

Dark 43 38 44 66 70 70 33
Creamy 63 36 47 65 55 70 34
Aroma 46 39 41 54 49 48 15
Milk Aroma 56 37 44 34 35 35 21
Flavor 65 49 55 46 42 45 23
Sweet 56 48 58 40 42 42 19
Aftertaste 56 49 54 56 51 55 7
Creamy Mouthfeel 65 44 52 46 45 54 21
Smooth Mouthfeel 70 57 58 46 49 50 24

REFERENCES

CHAPMAN, K.W. and BOOR, K.J. 2001. Acceptance of 2% ultra-pasteurized

milk by consumers 6 to 11 years old. J. Dairy Sci. 84, 951–954.
CHAPMAN, K.W., LAWLESS, H.T. and BOOR, K.J. 2001. Quantitative
descriptive analysis and principal component analysis for sensory char-
acterization of ultrapasteurized milk. J. Dairy Sci. 84, 12–20.
CLARE, D.A., BANG, W., DRAKE, M.A., CARTWRIGHT, G., CORONEL,
P. and SIMUNOVIC, J. 2005. Comparison of microwave and UHT
heating methods on sensory, microbiological, and biochemical para-
meters of fluid skim milk. J. Dairy Sci. 88, 4172–4182.
CROISSANT, A.E., WASHBURN, S. and DRAKE, M.A. 2007. Evaluation of
chemical properties and consumer perception of fluid milk from conven-
tional and pasture-based production systems. J. Dairy Sci. 90, 4942–
4953.
DECKER, K.J. 2004. Crossing borders: Designing for the Hispanic demo-
graphic. Food Product Design. [Link]
articles/[Link] (accessed September 18, 2008).
FOOD RESEARCH AND ACTION CENTER. Federal Food Program. http://
[Link]/State_Of_States/2006/states/[Link] (accessed August 9,
2008).
GANDY, M., SCHILLING, W., COGGINS, P.C., WHITE, C.H., YOON, Y.
and KAMADIA, V.V. 2008. The effect of pasteurization temperature on
consumer acceptability, sensory characteristics, volatile compound com-
position, and shelf-life of fluid milk. J. Dairy Sci. 91, 1769–1777.
GLADWELL, M. 2004. The ketchup conundrum. The New Yorker, September
6, 2004, 129–135.
 COW’S MILK VERSUS SOY BEVERAGES 747

GREENBERG, K. 2000. Wakeup call for retailers. Prog. Grocer. 79(10),

114–115.
MOSKOWITZ, H.R. 1985. New Directions in Product Testing and Sensory
Analysis of Food. Food and Nutrition Press, Trumbull, CT.
MOSKOWITZ, H.R. 1994. Food Concepts and Products: Just in Time Devel-
opment. Food and Nutrition Press, Trumbull, CT.
N’KOUKA, D., KLEIN, B.P. and LEE, S.Y. 2004. Developing a lexicon for
descriptive analysis of soymilks. J. Food Sci. 69, S259–S263.
PALUMBO, F.A. 2005. Segmenting the U.S. Hispanic market based on level
of acculturation. J. Promot. Manage. 12(1), 151–173.
PANGBORN, R.M. 1970. Individual variations in affective responses to taste
stimuli. Psychon. Sci. 21, 125–128.
PANGBORN, R.M., BOS, K.E. and STERN, J.S. 1985. Dietary fat intake and
taste responses to fat in milk by under, normal and overweight women.
Appetite 6, 25–28.
PHILLIPS, L.G., MCGIFF, M.L., BARBANO, D.M. and LAWLESS, H.T.
1995. The influence of fat on the sensory properties, viscosity and color
of low-fat milks. J. Dairy Sci. 78, 1258–1266.
QUINONES, H.J., BARBANO, D.M. and PHILLIPS, L.G. 1997. Influence
of protein standardization by ultrafiltration on the viscosity, color, and
sensory properties of skim and 1% milk. J. Dairy Sci. 80, 3142–
3151.
QUINONES, H.J., BARBANO, D.M. and PHILLIPS, L.J. 1998. Influence
of protein standardization by ultrafiltration on the viscosity, color, and
sensory properties of skim and 1% milks. J. Dairy Sci. 81, 884–894.
RUSYNYK, A.R. and STILL, C.D. 2001. Lactose intolerance. J. Am. Osteo-
path. Assoc. 101, S10–S12.
SAVAIANO, D. 2003. Lactose intolerance: A self-fulfilling prophecy leading
to osteoporosis? Nutr. Rev. 61, 221–223.
THOMPSON, J.L., DRAKE, M.A., LOPETCHARAT, K. and YATES, M.D.
2004. Preference mapping of commercial chocolate milks. J. Food Sci.
69(11/12), S406–S413.
THOMPSON, J.L., LOPETCHARAT, K. and DRAKE, M.A. 2007a. Prefer-
ences for commercial strawberry drinkable yogurts among African
American, Caucasian, and Hispanic consumers in the United States. J.
Dairy Sci. 90, 974–4987.
THOMPSON, J.L., GERARD, P.D. and DRAKE, M.A. 2007b. Chocolate
milk and the Hispanic consumer. J. Food Sci. 72, S666–S675.
UNITED STATES DEPARTMENT OF AGRICULTURE, CENTER FOR
NUTRITION POLICY AND PROMOTION. 2005. Washington, DC.
[Link] (accessed August 9,
2008).
748 O.M. PALACIOS ET AL.

UNITED STATES DEPARTMENT OF AGRICULTURE, FOOD AND

NUTRITION SERVICE (USDA FNS). 2008a. Child and adult care food
program (the CACF Program), Alexandria, VA. [Link]
cnd/Care/CACF (accessed September 9, 2008).
UNITED STATES DEPARTMENT OF AGRICULTURE, FOOD AND
NUTRITION SERVICE (USDA FNS). 2008b. Special Supplemental
Nutrition Program for Women, Infants, and Children [the WIC Program],
Alexandria, VA website: [Link] (accessed Sep-
tember 9, 2008).
U.S. DEPARTMENT OF HEALTH AND HUMAN SERVICES AND U.S.
DEPARTMENT OF AGRICULTURE (USDA). 2005. Dietary Guide-
lines for Americans, 2005, 6th Ed., U.S. Government Printing Office,
Washington, DC.
WOOTEN, W.J. and PRICE, W. 2004. Consensus report of the National
Medical Association: The role of dairy and dairy nutrients in the diet of
African Americans. J. Natl. Med. Assoc. 96, 1S–31S.