Placing Food Color Experimentation Into a Valid c

29
PLACING FOOD COLOR EXPERIMENTATION INTO A VALID CONSUMER CONTEXT Lawrence L. Garber, Jr.* Assistant Professor of Marketing Marketing Department Appalachian State University Boone, NC 28608 Phone: 828 262 6219 Fax: 828 262 6192 e-mail: [email protected] Eva M. Hyatt Associate Professor of Marketing Marketing Department Appalachian State University Boone, NC 28608 Phone: 828 262 2926 e-mail:[email protected] Richard G. Starr, Jr. Senior Lecturer in Marketing University of Auckland Private Bag 92019 Auckland, NZ Phone: 011 64 21 637 915 e-mail: [email protected] * Corresponding author

Transcript of Placing Food Color Experimentation Into a Valid c

Page 1: Placing Food Color Experimentation Into a Valid c

PLACING FOOD COLOR EXPERIMENTATION INTO A VALID CONSUMER CONTEXT

Lawrence L. Garber, Jr.* Assistant Professor of Marketing

Marketing Department Appalachian State University

Boone, NC 28608 Phone: 828 262 6219 Fax: 828 262 6192

e-mail: [email protected]

Eva M. Hyatt Associate Professor of Marketing

Marketing Department Appalachian State University

Boone, NC 28608 Phone: 828 262 2926

e-mail:[email protected]

Richard G. Starr, Jr. Senior Lecturer in Marketing

University of Auckland Private Bag 92019

Auckland, NZ Phone: 011 64 21 637 915 e-mail: [email protected]

* Corresponding author

Page 2: Placing Food Color Experimentation Into a Valid c

PLACING FOOD COLOR EXPERIMENTATION

IN A VALID CONSUMER CONTEXT

ABSTRACT

Empirical research examining the effects of food color on food acceptability comes from the food science literature, but its applicability for predicting consumer purchase decisions is limited. By design or intent, that research does not consider that food choice is typically made from a consideration set filled with food brand alternatives that are close to parity, and that product and flavor information come in modes in addition to food color. We review the existing food color literature, recount its limitations for consumer behavior research, and provide a conceptual framework to guide the future empiricist in conducting externally valid food color research. Further, whereas in most prior research the food color manipulation includes a novel level merely to contrast with expected color, we consider novel food color directly, for its potential as an attention-getting and differentiating brand communications tool, and present strategies for its effective presentation despite food color’s strong link to expected flavor. Keywords: consumer research, food color, novel color, visual persuasion

Page 3: Placing Food Color Experimentation Into a Valid c

1

INTRODUCTION

In the fragmented, cluttered, global and therefore unprecedently competitive

environment that is today’s marketplace, point of purchase communications has become an

increasingly important part of the marketing communications mix (Wells, Burnett & Moriarty

2000, p. 410). This has in turn placed a greater emphasis on visual appearance as a conveyor of

brand equity (Keller 1993) in the store, as products and packages are typically first seen and

evaluated as the consumer approaches them from a distance (Cardello 1996). These

circumstances have prompted marketers of late, cognizant of the powerful role that they can play

in bringing attention and distinctiveness to their brands, to place a particular emphasis on product

and package design, and the individual visual elements that comprise them (Garber, Burke &

Jones 2000a). Within the mix of individual visual design elements that comprise a product or

package, color is known to serve as a particularly vivid, affect-loaded and memorable visual cue

(Cheskin 1957) and persuader (Garber & Hyatt 2000b). A systematic look into food color and

its selection for persuasive communications purposes in a point of purchase context is needed.

The strong link between food color and palatability has been known for a long time. For

example, dramatic and amusing (though not to the participants!) empirical demonstrations were

performed by Moir (1936) and Wheatley (1973). As reported in Moskowitz (1978, p. 163):

“Moir prepared a buffet of goods for a dinner with scientific colleagues of the Flavor Group of the Society of Chemistry and Industry in London. Many of the foods were inappropriately colored, and during the dinner several individuals complained about the off-flavor of many of the foods served. Several of the individuals reported feeling ill after eating some of the foods, despite the fact that only the color was varied. The rest of the food was perfectly wholesome, with the requisite taste, smell, and texture.”

Page 4: Placing Food Color Experimentation Into a Valid c

2

A similar result was obtained by Wheatley (1973) as reported in Cardello (1996, p. 12):

“In this study, subjects ate a meal of steak, french fries and peas under color-masking conditions. Halfway through the meal, normal lighting was restored to reveal blue steak, green french fries and red peas. The mere sight of the food was enough to induce nausea in many of the subjects. The stark novelty of the colors used in this study leads one to ask whether certain colors are innately preferred or rejected.”

The above examples underscore food color’s entrenched role as a flavor signal. This role

has been reinforced over the years by food marketers’ primary use of color as a flavor identifier,1

both in its foods and on its packages, especially for foods that come in more than one flavor and

can be seen through their packages in store displays. However, extreme competition is

prompting some food marketers to flaunt such conventions by selecting novel food colors (i.e.,

those food colors that make no reference to, or are in contradiction with, the food flavor that it

represents) for competitive branding and communications purposes (Garber, Hyatt & Starr

2000c). Examples of this include Heinz’ introduction of green ketchup (c.f., USA Today, July

10, 2000, p. 2b), and Gatorade’s Frost series of sports drinks, whose colors refer to winter

themes and chill temperatures rather than flavor.

Those marketers who have selected novel food colors have done so in spite of the

perceived risks that prevents most marketers from trying it, even though it has potential for

effective point of purchase communications impact. Such risk is exemplified by the almost

immediate demise upon its launch in the early 90’s of Pepsico’s Crystal Pepsi, a clear form of

Pepsi Cola that even regular Pepsi drinkers rejected (c.f., Triplett 1994). In this instance, it was

1 A store audit performed by the authors revealed that 97% of all food brands displayed, in all categories, used food color to indicate flavor.

Page 5: Placing Food Color Experimentation Into a Valid c

3

clear that failure resulted from marketers’ lack of realization that a strong link existed between

food color and expected flavor. Marketers cannot hope to select effective color, particularly

novel color, until they understand color’s relative nature.

In spite of color’s potency as a compelling stimulus for competitive strategy purposes and

the concomitant need for marketers to understand the nature of its effects, there has been

surprisingly little research in the marketing literature concerning how to select effective product

or package color (exceptions being Garber, et al. 2000a; Garber et al. 2000b; Garber et al.

2000c). Though food scientists have empirically examined aspects of the effects of food color

on acceptability, research on the topic remains sparse (for exceptions, see Moir 1936; Kanig

1955; Hall 1958; DuBose, Cardello & Maller 1980; Hyman 1983; Stillman 1993; Oram et al.

1995; for a review of this literature, see Cardello 1996, p. 11-19), is not theory-based, and, as a

body, is limited in terms of its applicability to the consumer purchase process (Garber et al.

2000c). Our purpose, therefore, is to extend the food science literature on color by reviewing it,

recounting its limitations as it applies to consumer behavior research, providing a conceptual

framework intended to guide the future empiricist in conducting externally valid food color

research appropriate to uncovering consumer response in a purchase decision context, and

offering a series of strategies for the effective implementation of novel food color.

BACKGROUND

Staged Models of Choice

There is much evidence that consumers go through a multistage decision process when

making a purchase (Payne 1976; Bettman 1979; Lussier & Olshavsky 1979). Following

Page 6: Placing Food Color Experimentation Into a Valid c

4

Roberts (1989), who presents, “…a model of individual-level choice as a phased process

represented by a series of nested stages where behavior at each stage is conditioned by the events

of previous stages,” Garber et al. (2000a) adapt this model to explicitly consider package color’s

main (and differential) effects on the consumer at each of several stages in the choice process. In

turn, Garber et al. (2000c) adapt this model to explicitly consider food color’s significant and

differential main effects on the consumer at these same stages in the choice process, as shown in

Figure 1. Those stages are flavor identification, flavor perception and flavor preference

formation, respectively. Therefore, in order to best examine the existing food color literature

from the marketer’s standpoint, we organize the food color literature according to which stages

of the consumer decision process are specified as the dependent variables in a given empirical

study.

Flavor Identification

Most of the research that is reviewed addresses food color’s effect on correct flavor

identification. This is, from a marketing standpoint, the least important of the consumer choice

stages discussed, and the weakest model, because it is the furthest median stage from the choice

decision itself (Huber 1975). It is generally recognized that food color aids correct flavor

identification, as affirmed by several studies reported in the food science and sensory literatures

(Moir 1936; Hall 1955; Kanig 1955; DuBose et al. 1980; Hyman 1983; Stillman 1993; Oram

et al. 1995). As Christensen (1985, p. 755) points out, “Color is virtually essential for the correct

identification of color-linked food flavors (odors and taste) such as cherry, lime and orange

[DuBose et al. 1980], particularly in those foods that assume many flavors [i.e., beverages,

Page 7: Placing Food Color Experimentation Into a Valid c

5

cakes, puddings] and have no other visual characteristics related to flavor identification.”

Specifically, these studies generally find that characteristic food color facilitates the

ability to correctly identify flavor, that an absence of food color does not facilitate correct flavor

identification, and that uncharacteristic food color degrades correct flavor identification (Moir

1936; Kanig 1955; Hall 1958; Pangborn 1960; DuBose et al. 1980; Hyman 1983; Stillman

1993). However, reports vary as to subjects’ abilities to identify uncharacteristically colored

foods, as shown in Table 1. Oram et al. (1995, p. 245) attribute this variance to “task

differences,” which we discuss in greater detail in the following sections on limitations. Garber

et al. (2000c) obtain a greater strength of flavor identification effect across color levels relative to

previous studies. They contend that is due largely to the fact that they conceal the real purpose

of the experiment from their subjects; in particular, they conceal from subjects the fact that food

colors and flavors are deliberately being mismatched for experimental purposes. Therefore, for

this reason among others, the cognitive tasks that Garber et al.’s (2000b) subjects are caused to

perform more closely resemble the choice processes that the consumer undertakes under typical

store purchase conditions.

In prior flavor identification research, color has been shown to dominate taste because

subjects exposed to atypical color often misidentify its associated flavor as being one that is

normally associated with that color, an error which Oram et al. (1995) refer to as a “color biased

identification error.” There are two possible reasons for this kind of error, according to Oram et

al. (1995, p.240): “… the color-biased identification errors suggest that the subjects are either

not aware that there is a color-flavor conflict, or, if they are aware, that they cannot ignore the

color. Consequently, it is quite possible that such color-biased identification errors may reflect

Page 8: Placing Food Color Experimentation Into a Valid c

6

color being perceptually more salient than flavor. Color may be perceptually more salient than

flavor in those contexts because color generates a stronger neural response than flavor, or

because color is typically perceived before flavor in the eating experience.”

Flavor Perception

Only a few studies have examined the effects of food color on flavor perception or

preference, and they collectively present mixed or conflicting results. With respect to

perception, experiments have been of three types, though not necessarily mutually exclusive:

those requiring subjects to make differential judgements along a single dimension (Hyman,

1983) such as sweetness or thirst-quenching-ness; those which measure the effect of different

levels of intensity or saturation of a typical food color; and those which measure the effects of

food color on simple taste sensates (sweet, sour, bitter, salty).

Several studies have examined the effect of typical food color on perceptions of

sweetness or on a sweet-sour dimension. Pangborn (1960) had panels of trained and untrained

subjects evaluate the relative sweetness of a number of typically colored fruit flavored waters,

finding that red and orange colored drinks tasted sweeter, and green drinks tasted more sour.

Johnson & Clydesdale (1987) tested the effects of typical color intensity or saturation on

perceptions of sweetness. Using forty untrained subjects, they found that level of perceived

sweetness is directly proportional to the saturation level of red color. Norton & Johnson (1987),

however, using eighteen randomly selected subjects and manipulating the intensity of four

typical colors, found no relationship between color intensity and flavor ratings on a sweet-sour

dimension, or on a distinct-indistinct flavor dimension. They further conclude that taste is a

Page 9: Placing Food Color Experimentation Into a Valid c

7

much more powerful determinant of flavor than color on these two dimensions. In two other

studies measuring the effects of food color on a single flavor dimension, Duncker (1939) finds

that four of seven subjects report that white chocolate tasted “milkier” than dark chocolate, and

another two subjects find white chocolate to have less chocolate taste or less taste in general.

And Guinard, Souchard, Picot, Rogeaux & Siefferman (1998), using twelve subjects, purport to

find that the color intensity of sixteen beers is inversely proportional to its perceived thirst-

quenching-ness (although the ten-level color manipulation was described as ranging from light to

dark, which is a range of values, not color intensity).

Using a multi-attributed approach in a crossed design, Maga (1974) examines the relative

effect of several colors (red, green, yellow and colorless) on the four taste sensates (basic taste

sensations shorn of the complexities of flavor found in whole foods) of sweet, sour, bitter and

salty, presented in water solutions. Green makes sweet drinks seem sweeter (Pangborn 1960

found the opposite), and yellow makes them seem less sweet. Yellow and green cause sour

drinks to seem less sour, and red causes bitter drinks to seem less bitter. Similarly, McCullough,

Martinson & Moinpour (1978) also manipulated basic taste sensates at two levels (sweet, sour)

and color at two levels (red, blue) to derive a perceptual space using multidimensional scaling.

Their results indicate that blue is perceived to be relatively sweet.

Again, all of these studies fall short of offering results that are useful to, or approximate,

a consumer context. Moskowitz (1978, p. 163), in reviewing perceptual food color/flavor studies

to date, concludes that, “…no definitive study had appeared that systemizes the effect of color

upon sensitivity to taste, or to pleasantness of taste.” We feel that Moskowitz’s conclusion

continues to hold true. It remains our need as managers to understand the effects of food color

Page 10: Placing Food Color Experimentation Into a Valid c

8

on full flavor profiles as defined by Wilkie & Pessemier (1973), both on their nature and on their

strength, and we propose a more comprehensive approach to the problem in the methods section.

Flavor Liking

Of those models predicting the effects of food color on flavor identification, perception

and preference respectively, of greatest concern to the marketer would be those predicting

preference (Huber 1975), yet this is the stage that has been least studied. There is no available

theory of the effects of food color on flavor preference, and the few available empirical studies

show conflicting results (Johnson et al. 1982; Compeau, Grewal & Monroe 1989; Baeyens,

Eelen & Van den Bergh 1990; Baeyens, Crombez, Hendricks & Eelen 1995; Baeyens,

Vansteenwegen, De Houwer & Crombez 1996); and, as with models of identifiability and

perception, are limited in their applicability to a consumer context, again reflecting the cross-

purposes of food scientists and marketers.

It is commonly believed that food color affects judgements of flavor and food liking,

though this belief is not unequivocally supported by the sparse literature examining this

relationship. Nonetheless, researchers are undeterred from offering broad testimonials asserting

their belief in this relationship. For example, Maga (1974) states, “Color and flavor are two

primary factors that can influence food acceptability.” Christensen (1985, p. 755) says, “Color is

recognized as an important element in consumer ratings of food palatability, although the reasons

for its importance have not been elucidated.” More poetically, Birren (1963, p. 45) avers, “Color

is forever a part of our food, a visual element to which human eyes, minds, emotions and palates

are sensitive. Perhaps through eons of time, man has come to build up strong and intuitive

Page 11: Placing Food Color Experimentation Into a Valid c

9

associations between what he sees and what he eats. A good meal, to say the least, is always a

beautiful sight to behold.” And, again, DuBose et al. (1980, p. 1393) claim that, “Color is an

extremely important attribute of most food products because it usually influences the consumer’s

first judgement of the product and also provides sensory information which may interact with

gustatory, olfactory, and textural cues to determine the overall acceptability of the product.” Yet,

oddly, neither the research these authors conduct nor the prior research they cite addresses the

relationship between color and preference. Maga (1974), for example, investigates the effect of

color on perceptual attributes such as sweet and sour, and reviews literature that examines the

effects of food color on flavor identifiability and flavor perception. Christensen (1985) examines

the effect of food color on perceptions of flavor intensity, and reviews the literature on flavor

identification. And Dubose et al. (1980) study the effects of food color on identification, and the

effects of color intensity, though not color itself, on hedonic quality, while reviewing the

literature on flavor identification. Birren’s (1963) article is descriptive in nature.

LIMITATIONS TO FLAVOR ACCEPTABILITY RESEARCH AS IT APPLIES TO COMPETITIVE BRAND STRATEGY

A premise of this paper is that the empirical research done so far examining food color’s

effects on food acceptability does not, by its nature or intention, precisely address the food

color/consumer decision making issues that concern the food marketer.

The authors found little research in marketing that addresses the role of product or

package color, much less food color. For the latter, we must turn to the food science and sensory

literatures. However, much of it does not address consumer behavior issues or the store context

Page 12: Placing Food Color Experimentation Into a Valid c

10

specifically, and its applicability to the consumer is therefore limited for methodological reasons.

For example, some studies use panels of taste experts as subjects (Moir 1936; Murphy, Bailey &

Covell 1954; Kanig 1955; Pangborn, Berg & Hansen 1963; Szczesniak & Kleyn 1963;

McCullough et al. 1978; DuBose et al. 1980; Guinard et al. 1990) or trained subjects (Hall

1958; Pangborn 1960; Johnson et al. 1982) who do not properly represent consumers. In some

cases, subjects are encumbered with: red eye-goggles (DuBose et al. 1980), blindfolds (Hyman

1983; Scanlon 1985) red lights (Hall 1958; DuBose et al. 1980) or red glass (Duncker 1939) to

mask color; funnels and jugs for spitting (Looy, Callaghan & Weingarten 1992) or otherwise

subjects were not allowed to swallow (Pangborn 1982); and sensory testing booths to isolate the

subject (Pangborn 1960; DuBose et al. 1980; Johnson et al. 1982; Baeyens et al. 1990; Baeyens

et al. 1995; Baeyens et al. 1996). These encumbrances do not lend themselves to a naturalistic

representation of point of purchase conditions.

At times, very small sample sizes (30 or less; in some instances, 10 or less) are employed,

thus limiting statistical power (Duncker 1939; Murphy et al.1954; Maga 1974; McCullough et

al. 1978; DuBose et al. 1980; Johnson et al. 1982; Rolls, Rowe & Rolls 1982; Baeyens et al.

1990; Guinard et al. 1990; Zellner, Bartoli & Eckard 1991; Looy et al. 1992; Baeyens et al.

1996; Baeyens et al. 1995).

A common element in food color research is that each study performs taste test

experiments in which food color is typically manipulated at three levels. The first, characteristic

food color, also at times referred to as “correct”, “appropriate” or “congruent” food color, is that

food color normally associated with some given color-associated food (i.e., orange color with

orange fruit or flavor). Uncharacteristic food color, also at times referred to as “incorrect”,

Page 13: Placing Food Color Experimentation Into a Valid c

11

“inappropriate”, “incongruent” or “unusual” color, is any color not normally associate with some

given color-associated food or flavor (i.e., any color other than purple with grape fruit or flavor).

And finally, a base level is included, meant to represent no color, or the absence of color. Also at

times referred to as “masked” or “ambiguous” color, it is a condition in which no flavor

information is conveyed to the subject by food color.

In some studies where subjects are exposed to an atypical food color, they are informed

of the presence of color/flavor mismatches, and explicitly assigned the task of uncovering them

(Pangborn 1960; Stillman 1993), or are able to deduce from the experimental context and the

nature of the subject task that some food colors and flavors were mismatched (Moir 1936;

DuBose et al. 1980; Oram et al. 1995), conditions which likely underestimate the effect that the

atypical and no color conditions can have on the level of flavor misidentification as it is normally

presented in the marketplace.

Some studies do not actually manipulate color at all, instead manipulating only the

typical color’s intensity or saturation (Duncker 1939; Murphy et al. 1954; Schutz 1954;

DuBose et al. 1980; Johnson et al. 1982; Rolls et al. 1982; Tuorila-Ollikainen 1982;

Christensen 1985; Norton et al. 1987); their purpose being to understand the effect of color

intensity, not the effect of color itself, on the perception of food flavor. In two cases, it is value

that is actually manipulated (presenting the subject with varying levels of light and dark), rather

than color (Weckel, Mathias, Garnatz & Lyle 1961; Guinard et al. 1998).

For experimental control purposes, some studies do not expose the consumer to the

complex flavors normally encountered in most foods and beverages, but, rather, to simple flavor

sensates (sweet, sour, bitter or salty) (Maga 1974; McCullough et al. 1978; Baeyens et al. 1990;

Page 14: Placing Food Color Experimentation Into a Valid c

12

Baeyens et al. 1995; Baeyens et al. 1996).

A number of the few perceptual studies that have been done were univariate in nature,

“…requiring subjects to make differential judgements along a single dimension,” (Hyman 1983,

p. 145). These studies explore the effects of food color on a variety of respective flavor attributes

individually, rather than on a full flavor profile, which would be able to account for all of the

dimensions of complex food flavors simultaneously. Those attributes studied individually

include thirst-quenching-ness (Guinard et al. 1990), sweetness (Pangborn 1960; Johnson et al.

1982), sweet-versus-sour (Norton et al. 1987), distinctiveness (Norton et al. 1987) and flavor

intensity (Duncker 1939; Christensen 1985). Hyman (1983) also cites Pangborn et al. (1963)

and Romeu and DeVincente (1968) as conducting univariate studies of this nature, without

specifying which perceptual dimensions they measure.

In sparse literature relating the effects of food color on liking or preference, there are

conflicting accounts as to whether food color has significant effects on various dependent

variables that may be construed to reflect some overall flavor or food assessment. For example,

Baeyens et al. (1990, 1995, 1996) find that food color, as a covariate, does not have an effect on

(dis)liking. Likewise, Compeau et al. (1989), as part of an empirical test measuring the effect of

affect on judgements of quality, found that there is not a significant interaction between food

color and flavor. Nor did Johnson et al. (1987) find that color intensity has a significant effect on

liking/disliking. On the other hand, Hall (1958) found that food color powerfully influences

consumers’ abilities to estimate the strength and quality of a food. Rolls, Rowe, Rolls, Kingston,

Megson & Gunary (1980) find that foods of a favorite color are found to be “slightly” more

pleasant. And Rolls et al. (1980) find that food presented in a variety of colors enhances food

Page 15: Placing Food Color Experimentation Into a Valid c

13

intake.

Baeyens et al. (1990, p. 434-5) assert that, “Food and liking entertain a privileged

relationship. On the one hand, few categories of objects elicit as clear-cut and immediate

reactions of liking or disliking, as food typically does in humans. On the other hand, like-dislike

reactions represent one of the major determinants of (individual differences in) food acceptance

or rejection.” This statement reveals much of the problem in seeking to adapt existing color-

flavor experiments from food science to consumer issues; namely, that the dependent variable

has valence. Food acceptability, the usual dependent variable in food science experiments, is

commonly measured on the 9-point Hedonic Scale of Food Preference (Peryam & Pilgrim 1957)

whose extreme points are “Like extremely” and “dislike extremely.” (For a discussion of this

measure and its use in the measurement of food attitudes, see Moskowitz 1978.) Stimuli are

selected to represent the full range of that scale (for examples, see Johnson et al. 1982; Looy et

al. 1992; Baeyens et al. 1995) and therefore incorporate aversive stimuli (i.e., a noxious flavor

represented by a strongly bitter sensate). In contrast, all commercial food products are carefully

designed and tested to assure that they taste generally good, not bad, to consumers, or they are

never launched. Moreover, consumers at the compensatory stage of the choice process are

presumed to be selecting from a consideration set comprised of comparable alternatives

(Bettman 1979), meaning that all food products in a given consumer’s consideration set have

been screened by the consumer to assure that they taste good to that particular consumer, who

will then select for purchase that one product that tastes the best to them, or is particularly

appealing on some other attribute, such as price.

And, finally, in no prior research were information sources other than food color or

Page 16: Placing Food Color Experimentation Into a Valid c

14

tasting available to the subjects; whereas, in a typical point-of-purchase situation, flavor

information is normally provided not just by means of food or beverage color, but also by way of

text on the label. Therefore, any experiment considering the effects of food or beverage color

must consider the effects of label information provided to the consumer, along with food color,

prior to tasting.

SELECTING PERSUASIVE FOOD COLOR FOR COMPETITIVE PURPOSES

As previously stated, marketers have a great interest in the ability to effectively

implement novel food color for persuasive purposes at the point of purchase. Garber et al.

(2000c) empirically demonstrate both the strength of effect of food color on expected flavor, a

stumbling block to the use of novel food color, and several means of unlinking the food-color-

expected-flavor flavor relationship. These are recounted in the following, once again organized

according to which stage of the consumer purchase process is being targeted for influence: flavor

identification, perception or preference.

Strategies for the Implementation of Novel Food Color

Research shows that food color affects the consumer’s ability to correctly identify flavor,

to form distinct flavor profiles and preferences, and dominates other flavor information sources,

including labeling and taste. Further, these results support the notion that food color is

inextricably linked to expected flavor in the minds of consumers, making the selection of

uncharacteristic food color problematical. In the following, we present three possible strategies

Page 17: Placing Food Color Experimentation Into a Valid c

15

for making the introduction of a novel food color viable for marketing communications purposes.

The first is to teach consumers to accept a novel color as characteristic, or emblematic, of a

particular food, as green is for peppermint of or brown is for cola. This is a strategy that is self-

defeating in some respects, but is useful under certain circumstances, as we shall discuss. The

second strategy is to celebrate the very incongruity of a novel food color, to announce to the

consumer that its novelty is there to surprise and delight, and the proper response is to have fun

and enjoy it. The third strategy for the introduction of novel food color is to sever the food color

and flavor expectations connection, making it impossible for the consumer to connect the two.

Flavor Identification. The first strategy centers around teaching consumers to associate

a given color with a given flavor, so that color becomes a flavor identification cue. However, a

drawback to rendering a novel color no longer novel is that it loses its ability to surprise the

consumer into attention, which was the prime reason for utilizing novel color in the first place.

But, we mention this strategy here because it does serve a purpose in the case of food products

that are nondescript in color or appearance. When the appearance of a food product is

nondescript, then associating it with a new, more vibrant color can enhance its noticeability, its

distinctiveness and its appeal. Such has been the case with green for peppermint or yellow for

Mountain Dew and all its me-too competition (a me-too color strategy!).

A problem with rendering a novel food color characteristic is that it will likely be a

lengthy and expensive process, requiring as it does the conditioning of consumers to accept the

new color as characteristic of a particular food product. Another obstacle is the sheer diversity

and multiplicity of food products (and their packages) on display. This makes it hard for the

Page 18: Placing Food Color Experimentation Into a Valid c

16

marketer to find an empty visual niche, when compared to the days when peppermint was made

green or cola was made brown.

Flavor Perception. A second strategy for the effective implementation of novel food

color is to make a virtue of its incongruence. This is done by featuring novel color and its very

incongruence in the shelf presentation. The consumer therefore knows that the incongruence is

intended, is meant to be perceived as amusing, and is therefore made to feel welcome to share in

the fun. The consumer’s flavor perceptions are then guided by the marketer. An example of this

is Gatorade’s Blue Raspberry drink, an uncharacteristically blue-colored beverage whose name

calls attention to the incongruence of the drink’s color and flavor. This packaging ploy assures

the consumer that the discrepancy is intentional, is for their entertainment, and cues feelings of

pleasure and enjoyment. As an example of this approach, it is interesting to note that blue foods,

once considered by food manufacturers to be unpalatable and unacceptable to the consumer

(Cheskin 1957; Hine 1996) have in recent years been utilized in this manner as a means of

creating novelty and excitement, especially in categories aimed at children.

Flavor Perception and Preference. A third approach to novel food color is to sever the

food color/ labeling connection by making it impossible for the consumer to connect the two. If

color and flavor cannot be connected, then novel food color cannot be incongruent. Given the

ingrained nature of the food color/expected flavor association, how is such an uncoupling made

possible? First, the natural tendency of the consumer to connect color and flavor must be

deliberately blocked, to permit the introduction of other color themes and associations to

Page 19: Placing Food Color Experimentation Into a Valid c

17

distinguish and contrast the brand, and lend it meaning.

The most straightforward means of unlinking food color and labeling is to mask food

color. The focus of the product can then be shifted to a more thematic association. Several drink

brands have elected this approach by packaging their drinks in opaque bottles or plastic labels

that cover the outside of the package, thus hiding the view of the actual product. Cans, of course,

have by their nature always masked their contents, but the deliberate painting of bottles, whose

virtue has always been to permit the consumer to view its contents, is a new and contrary

wrinkle.

A recent example of this approach is an iced coffee beverage line by Havana, which has

covered the bottles in plastic and placed eye-catching graphics on the package. The Havana

“Mocha Iced Cappuccino” package has a sexual appeal, with an exotic woman on a pink

background surrounded by floating coffee beans. This color/graphic combination is sure to stand

out among beverage competitors. This strategy makes novel use of color while masking the dull

brown color of the coffee-based beverage. With this motif, flavor expectations formed in the

store derive solely from label information (including package color, which, itself, may or may

not be congruent or characteristic) or, of course, from prior experience. Exposure to food color

is postponed until the time of preparation and consumption, which grants the designer and

marketer greater marketing communications latitude.

A more subtle approach to the disconnection of the food color/expected flavor

relationship is the selection or creation of food colors and flavors that are not flavor- or color-

associated. In denying the consumer the ability to readily categorize the flavor cues that food

color and labeling present, the consumer may be induced into a mode of more elaborated

Page 20: Placing Food Color Experimentation Into a Valid c

18

information processing in order to understand and evaluate the product. This opens an

opportunity for the presentation of marketing communications ideas, symbols, meanings and

associations through the medium of novel food color.

For example, Gatorade sought to extend their product line, but was limited by the fact

that there are a finite number of fruit flavors, all of which can be found in competitor product

lines. The Frost series solved this problem by replacing the usual flavor references on the

package with references to themes of winter, with colors and names that evoked images of cold,

ice and snow; images that are not inconsistent with themes normally positively associated with

fruit beverages, but which are normally evoked in advertising and not by food color and flavor

names. For example, Gatorade’s Frost series includes “Glacier Freeze,” which comes in a clear

strong blue color not unlike mouthwash, “Whitewater Splash,” which comes in a clear strong

green, and one called “Alpine Snow,” which comes in a semi-translucent white.

A key to consumer acceptance of such a thematic approach, along with its innate

vividness and appeal, is to deny the consumer ready access to any flavor cues, which, if obtained,

would likely preempt further search and evaluation. Therefore, Gatorade goes so far as to

withhold specific flavor information in its “Nutritional Facts” label, citing only “natural flavors.”

The consumer is therefore blocked from falling back on old flavor habits, and can have none of

the usual flavor expectations prior to tasting. The consumer is therefore forced to consider and

evaluate the Frost line of drinks in an entirely new context.

A related but somewhat different approach seeks an alternative appeal that is cognitive in

nature. For example, Snapple and SoBe offer beverage lines whose names and body text offer

symbolic and spiritual references not normally associated with beverages, and whose colors are

Page 21: Placing Food Color Experimentation Into a Valid c

19

designed to be consistent with those themes rather than with flavor. For example, Snapple’s

“Rain” is clear like water and contains ginseng, astragalus and agave, whose flavors most

consumers would find unfamiliar. SoBe’s “Drive” comes in a glass bottle which describes it as a

“Beach Brew,” and whose label features two green lizards. Its herbal additives include

epimedium, ginseng and muria puama, which, as with “Rain,” offer little recognizable flavor

information to the consumer. By uncoupling color from flavor, SoBe has given themselves the

marketing communications opportunity to associate color with meaning, namely “healthy.”

They added value for the health-conscious consumer by including herbal ingredients without

affecting flavor expectations.

SUGGESTIONS FOR FUTURE FOOD COLOR EXPERIMENTAL DESIGN

In summary, the following laboratory conditions have been shown to be necessary for

consumer response to a food color stimulus to be validly represented in a store context.

The Interaction of Food Color with Other Flavor Information Sources

Though research has shown that flavor expectations are formed primarily on the basis of

associated food color, food color is not the sole source of flavor information. These flavor

information sources may interact in the formation of flavor expectations and perceptions.

Therefore, these other information sources, label information in particular, must be present in a

valid consumer experiment. In particular, the intrinsic conflict that arises with the pairing of

novel color and labeling must be presented and measured (Garber et al. 2000c).

Page 22: Placing Food Color Experimentation Into a Valid c

20

Order of Presentation of Flavor Information

As mentioned previously, food color is the first piece of flavor information processed by

the consumer in the store, because it can be resolved at a distance (Cardello 1996). The primacy

of food color over other flavor information sources may explain food color’s relative strength of

effect on expected flavor (Garber et al. 2000c). However, this hypothesis remains unproven

because another possible source of food color’s relative strength of effect is color’s intrinsic

sensory qualities, including its vividness, strength of affect and memorability (Oram et al. 1995).

Though Garber et al. (2000c) empirically demonstrate food color’s relative strength of effect,

their experimental design does not allow them to disentangle these two variables, and further

research is therefore required to measure these effects separately and determine their relative

importance to food color’s effect on flavor expectations.

Demand Artifacts

To avoid demand artifacts, the subject cannot know that they are deliberately being

presented with masked, atypically colored and mismatched beverages. If they deduce the real

purpose of the experiment, they will then pursue the task of detecting the true flavor of the

beverages they sample, rather then the intended task of evaluating beverages as true product

candidates for market introduction.

FUTURE RESEARCH

As previously stated, the effects of food color, and, for that matter, the effects of all

visual elements and appearance in general, are largely under-researched, particularly from a

Page 23: Placing Food Color Experimentation Into a Valid c

21

consumer standpoint. In the following, we suggest six general areas where a greater

understanding would be beneficial to food marketers: research to disentangle primacy and

intrinsic sensory effects as sources of the effect of food color on flavor expectations; research to

understand the nature of the effects of food color on perception and preference; systematic

examination of the effects of the components of color, including hue, value and chroma; a

similar examination of all other visual elements and their interaction, including size, shape and

movement; research to uncover the cognitive underpinnings of visual persuasion; and research

examining individual differences in color perception and preference, food color perception and

preference, and flavor perception and preference.

Oram et al. (1995) demonstrated that food color is a strong determinant of an individual’s

ability to correctly identify flavor in color-associated foods. Garber et al. (2000c) extended that

research by finding that food color dominates other flavor information forms in a typical

consumer context, and is an important determinant of flavor identify, perception and preference.

Both

sets of authors speculate that food color’s strength of effect is likely due to a combination

of two conditions, the intrinsic sensory qualities of color as a vivid visual stimulus, and the fact

that food color is the first sensory information that the consumer processes in the store.

However, no prior research has sought to disentangle these effects. Further research is required

to measure these effects separately, and determine their relative importance in the formation of

flavor expectations.

Secondly, as has been stated herein, it is important that the food marketer understand the

impact of food color at each stage of the consumer purchase process. In particular, there has

Page 24: Placing Food Color Experimentation Into a Valid c

22

been very little research examining food color’s effects on food and flavor perception and

preference. Given the complex multidimensional nature of food color, there is a need to examine

these relationships further from many perspectives.

Thirdly, food color has itself been treated as a monolithic element, when, in fact, it can be

decomposed into several components. A common breakdown is that of hue, value and chroma.

Each of these components needs to be measured separately for their main and interactive effects

on identification, perception and preference to be determined. The effects of hue have been

particularly neglected. For example, the color manipulations utilized by most of the previously

discussed flavor identification experiments manipulated chroma, or different colors. These may

be characterized as large changes to color, as represented by orange vs. blue, or green vs. red.

However, the human eye can also perceive the effects of very small changes to color, as would

be represented by manipulating hue. For example, how is the consumer’s ability to correctly

identify flavor, or her or his interpretation of a given flavor, or her or his preference for it,

affected by a slight change to the shade of red or blue or some other color presented.

Fourth, all such sensory experiments examining color’s effects on identification,

perception and preference need to be replicated for all the other visual elements, including size,

shape and movement, and their interactions.

Fifth, though all the above will allow us to better predict the effects of appearance on food

purchases, they may not fully explain their nature. Further research is necessary to understand

the cognitive underpinnings of how food color and other visual elements influence flavor and

food identification, perception and purchase.

And, finally, we need also to understand individual differences. For example, all prior

Page 25: Placing Food Color Experimentation Into a Valid c

23

research into food color implicitly assumed, and therefore did not consider, that consumers do

not perceive and like all food colors equally, nor do they know, perceive and like all foods and

flavors equally. What is the range and effect of such differences? It is hoped and intended by

this research that the empirical methods proposed here for the study of the consumer will

contribute to future research that will go towards answering these questions.

Page 26: Placing Food Color Experimentation Into a Valid c

REFERENCES

Baeyens, F., Crombez, G., Hendricks, H., & Eelen, P. (1995) Parameters of Human Evaluative Flavor-Flavor conditioning. Learning and Motivation, 26, 141-160.

_______, Eelen, P., Van DEN Bergh, O., & Crombez, G. (1990) Flavor-Flavor and Color-Flavor Conditioning in Humans. Learning and Motivation, 21, 434-455.

_______, Vansteenwegen, D., De Houwer, J., & Crombez, G. (1996) Observational Conditioning of Food Valence in Humans. Appetite, 27, 235-250.

Bettman, James R. (1979) An Information Processing Theory of Consumer Choice. Chicago: Addison-Wesley.

Birren, F. (1963) Color and Human Appetite. Food Technology, (May), 45-47. Cardello, A. V. (1996) The Role of the Human Senses in Food Acceptance. In H.L.

Meiselman and H.J.H. McFie, Food Choice, Acceptance and Consumption (pp. 1-82). New York: Blackie Academic and Professional.

Cheskin, L. (1957), How to Predict What People Will Buy. New York: Liveright. Christensen, C. M. (1985) Effect of Color on Judgements of Food Aroma and Flavour

Intensity in Young and Elderly Adults. Perception, 14, 755-762. Compeau, L. D., Grewal, D., & Monroe, K. B., (1989) Role of Affect and Sensory Cues

on Consumer’s Affective and Cognitive Responses and Overall Perceptions of Quality. Journal of Business Research, 42, 295-308.

DuBose, C. N., Cardello, A. V., & Maller, O. (1980) Effects of Colorants and Flavorants on Identification Perceived Flavor Intensity, and Hedonic Quality of Fruit-Flavored Beverages and Cakes. Journal of Food Science, 45, 1393-1399, 1415.

Duncker, K. (1939) The Influence of Past Experience Upon Perceptual Properties. American Journal of Psychology, 52, 255-265.

Garber, L. L., Jr., Burke, R. R., & Jones, J. M., (2000a) The Role of Package Appearance in Consumer Purchase Consideration and Choice. Boston: Marketing Science Institute Working Paper Series.

_______, & Hyatt, E. M. (2000b) Color as a Tool for Visual Persuasion. In R. Batra, & Scott, L., Visual Persuasion, Hillsdale, NJ: Lawrence Erlbaum Associates.

_______, Hyatt, E. M., & Starr, R. G., Jr. (2000c) The Effects of Food Color on Perceived Flavor. The Journal of Marketing Theory and Practice, forthcoming.

Guinard, J.-X., Souchard, A., Picot, M., Rogeaux, M., & Siefferman, J.-M. (1998) Sensory Determinants of the Thirst-Quenching Character of Beer. Appetite, 31, 101-115.

Hall, R. L. (1958) Flavor Study Approaches at McCormick and Company, Inc. In A.D. Little, Flavor Research and Food Acceptance. New York: Reinhold.

Hine, T. (1996) The Total Package. New York: Little, Brown and Company. Huber, J. (1975) Predicting Preferences on Experimental Bundles of Attributes: A

Comparison of Models. Journal of Marketing Research, XII (August), 290-297. _______, & Holbrook, M. B. (1979) Using Attribute Ratings for Product Positioning:

Some Distinctions Among Compositional Approaches. Journal of Marketing Research, XVI (November), 507-516.

Hyman, A. (1983) The Influence of Color on the Taste Perception of Carbonated Water Preparations. Bulletin of the Psychonomic Society, 21(2), 145-148.

Page 27: Placing Food Color Experimentation Into a Valid c

Johnson, J., & Clydesdale, F. M. (1982) Perceived Sweetness and Redness in Colored Sucrose Solutions. Journal of Food Science, 47, 747-752.

Kanig, J. L. (1955) Mental Impact of Colors in Foods Studied. Food Field Reporter, 23, 57.

Keller, K. L. (1993) Conceptualizing, Measuring, and Managing Customer-Based Brand Equity. Journal of Marketing, 57 (January), 1-22.

Looy, H., Callaghan, S., & Weingarten, H.P. (1992) Hedonic Response of Sucrose Likers and Dislikers to Other Gustatory Stimuli. Physiology and Behavior, 52, 219-225.

Lussier, D. A., & Olshavsky, R. W. (1979) Task Complexity and Contingent Processing in Brand Choice. Journal of Consumer Research, 6 (September), 154-165. Maga, J.A. (1974) Influence of Color on Taste Thresholds. Chemical Senses, Flavor, 1, 115-119.

McCullough, J. M., Martinsen, C. S., & Moinpour, R. (1978) Application of Multidimensional Scaling to the Analysis of Sensory Evaluations of Stimuli with Known Attribute Structures. Journal of Applied Psychology, 65 (1), 103-109.

_______, & Tybout, A. M. (1989) Schema Incongruity as a Basis for Product Evaluation. Journal of Consumer Research, 16 (June), 39-54.

Moir, H. C. (1936) Some Observations on the Appreciation of Flavor in Food Stuffs. Chemistry and Industry, 55, 145-148.

Moskowitz, H. R. (1978) Taste and Food Technology: Acceptability, Aesthetics, and Preference. Handbook of Perception, VIA, 158-193.

Murphy, E. A., Bailey, R.M., & Covell, M.R. (1954) Observations on Methods to Determine Food Palatability and Comparative Freezing Quality of Certain New Strawberry Varieties. Food Technology, (February), 113-116.

Norton, W. E., & Johnson, F. N. (1987) The Influence of Intensity of Colour on Perceived Flavour Characteristics. Medical Science Research, 15 (5-8), 329-330.

Oram, N., Laing, D. G., Hutchinson, I., Owen, J., Rose, G., Freeman, M., & Newell, G. (1995) The Influence of Flavor and Color on Drink Identification among Children and Adults. Developmental Psychobiology, 28 (4), 234-246.

Pangborn, R. M. (1960) Influence of Color on the Discrimination of Sweetness. American Journal of Psychology, 73, 229-238.

Payne, J. W. (1976) Task Complexity and Contingent Processing in Decision Making: An Information Search and Protocol Analysis. Organizational Behavior and Human Performance, 16 (August), 366-387.

Peryam, D. R, & Pilgrim, F. J. (1957) Hedonic Scale Method of Measuring Food Preferences. Food Technology, 11, 9-14.

Roberts, J. H. (1989) A Grounded Model of Consideration Set Size and Composition. Advances in Consumer Research, 16, 749-757.

Rolls, B. J., Rowe, E. A., & Rolls, E.T. (1982) How Sensory Properties of Foods Affect Human Feeding Behavior. Physiology and Behavior, 29, 409-417.

________, Rowe, E. A., Rolls, E.T., Kingston, B., Megson, A., & Gunary, R. (1980) Variety in a Meal Enhances Food Intake in Man. Physiology and Behavior, 26, 215-221.

Romeu, J.V., & DeVincente, M.L.S. (1968) Influyen los colores en el sabor. Revista Interamericana de Psycologia, 2, 143-147.

Page 28: Placing Food Color Experimentation Into a Valid c

Scanlon, B. A. (1985) Race Differences in Selection of Cheese Color. Perceptual and Motor Skills, 61, 314.

Schutz. H. G. (1954) Color in Relation to Food Preference. In K. T. Farrell, J. R. Wagner, & M. S. Peterson, Color in Foods (pp. 16-23). Washington: National Academy of

Sciences National Research Council. Stillman, J. A. (1993) Color Influences Flavor Identification in Fruit-flavored

Beverages. Journal of Food Science, 58, 810-812. Szczesniak, A.S., & Kleyn, D.H. (1963) Consumer Awareness of Texture and Other

Food Attributes. Food Technology, (January), 74-77. Triplett, T. (1994) Consumers Show Little Taste for Clear Beverages. Marketing News,

28 (11), 2. Tuorila-Ollikainen, H., (1982) Pleasantness of Colourless and Coloured Soft Drinks and

Consumer Attitudes to Artificial Food Colours. Journal for Intake Research, 3, 369-376.

USA Today (2000) What’s Green, Easy to Squirt? Ketchup! (Monday, July 10), 2b. Weckel, K. G., Mathias, W. D., Garnatz, G. F., & Lyle, M. (1961) Effect of Added Sugar

on Consumer Acceptance of Canned Peas. Food Technology, (May), 241-242. Wells, W., Burnett, J., & Moriarty, S. (2000) Advertising: Principles and Practice, 5th

Ed. Upper Saddle River, NJ: Prentice Hall. Wheatley, J. (1973) Putting Color into Marketing. Marketing (Oct. 23-29), 67. Wilkie, W. L., & Pessemier, E. A. (1973) Issues in Marketing’s Use of

Multiattribute Attitude Models. Journal of Marketing Research, 10 (November) 428- 441. Zellner, D. A., Bartoli, A. M., & Eckard, R. (1991) Influence of Color on Odor

Identification and Liking Ratings. American Journal of Psychology, 4 (4), 547-651.

Page 29: Placing Food Color Experimentation Into a Valid c

Lawrence L. Garber, Jr. is Associate Professor of Marketing at the John A. Walker College of Business, Appalachian State University, Boone, NC. He holds a Ph.D. in business administration with a concentration in marketing from the University of North Carolina at Chapel Hill, an MBA from Yale University, an AB from Brown University, and attended the Parsons School of Design for illustration and the Cleveland Institute of Art for painting. His research interests include visual persuasion, the graphical presentation of statistical data, and nonprofit and arts marketing. He has published in the Journal of Marketing Theory and Practice, the Working Paper Series of the Marketing Science Institute, Transportation Journal, Journal of Professional Services Marketing, the Journal of Nonprofit and Public Sector Marketing, Advances in Consumer Research and other proceedings. Eva M. Hyatt is Associate Professor of Marketing at the John A. Walker College of Business, Appalachian State University, Boone, NC. She holds a Ph.D. in Business Administration from the University of South Carolina, an MBA from Louisiana State University and a BA from the University of California at Berkeley. Her research interests include the effects of cognitive processes in consumer behavior, product symbolism, demand bias issues, consumer socialization, public policy issues especially with respect to vulnerable populations, and marketing ethics. She has published in the Journal of Marketing Theory and Practice, Journal of Consumer Research, the Journal of Business Research, the Journal of Public Policy and Marketing, and Consumption, Culture and Markets, Advances in Consumer Research and other proceedings. Richard W. Starr, Jr. is Senior Lecturer in Marketing at the University of Auckland, NZ. He is ABD at the University of North Carolina at Chapel Hill, holds an MBA from Columbia University and a BA from the University of Rochester. His research interests include how firms respond to the required disclosure of nutritional information on packaged foods, the use of color in product marketing, customer perceptions of value, and the use of marketing knowledge for social marketing and environmental improvement. In terms of selected publications, he co-authored several instructors’ editions of Kotler and Armstrong’s Principles of Marketing, and his articles have appeared in Marketing Letters, Journal of Marketing Theory and Practice, and conference proceedings.