Advertising Assignment I

Shaun Webbs

AdvertisingUnit1Article2.pdf

Home >Business & Finance homework help >Marketing homework help >Advertising Assignment I

DOI: 10.1002/mar.21171

R E S E A R C H A R T I C L E

Fast food, ads, and taste in a Russian child’s mind

Tobias Effertz1 | Thorsten Teichert2 | Marina Tsoy3

1The Institute of Law & Economics, Faculty of

Business Administration/Hamburg Business

School, University of Hamburg, Hamburg,

Germany

2Chair of Marketing and Innovation, Faculty of

Economics and Social Sciences, University of

Hamburg, Hamburg, Germany

3Marketing and Service Department, Faculty

of Business, Novosibirsk State Technical

University, Novosibirsk, Russia

Correspondence

Tobias Effertz, The Institute of Law &

Economics, Faculty of Business

Administration/Hamburg Business School,

University of Hamburg, Von Melle Park 5,

Hamburg 20146, Germany.

Email: [email protected]

Abstract

Research demonstrates that brands can influence children’s food preferences and

potentially contribute to unhealthy consumption patterns. This article extends this line of

research by investigating the complex and interacting effects of food brand marketing on

experienced taste. The empirical field in a remote Russian town enabled the assessment

of branding effects when entering a newly established market. Examining the combination

of various advertisement features with emotional brand elicitation, we derive hypotheses

about the interplay of brands and advertisement components linked to the perceived

taste of a fast food meal. In Novosibirsk, Russia, 778 children and adolescents aged

10–18 years were exposed to fast food advertisements with real and imaginary brand

logos and varying advertising claims. The advertisements consisted of a warning, an

exaggerated credence claim, or both. A subsequently offered fast food meal—a portion of

French fries—was experimentally varied and prepared to be either healthier but less tasty

and unsalted, or less healthy but tastier and salted. The findings verify strong and positive

brand effects on children’s taste satisfaction. Warnings in advertisements “worked” only

for novel fast food brands by negatively affecting taste satisfaction, but increased taste

satisfaction when applied in advertisements for established brands. Single credence claims

did not improve the taste experience, but counteracted the negative effects of warnings

for novel brands. Finally, the established brand influenced taste satisfaction positively

when the fries were saltier. The findings reveal various opportunities for fast food

marketing to artificially create taste satisfaction for potentially unhealthy food. Public

health strategies that focus on advertising claim restrictions should be reconsidered in the

context of possible evasion strategies of the food industry and counter‐effects of warnings among strong fast food brands.

K E Y W O R D S

brands, emotions, fast food (FF), marketing to children, taste experience, warnings

1 | INTRODUCTION

Children’s unhealthy consumption patterns and resulting obesity are

among the most challenging health threats in many countries across

the world (World Health Organization, 2014). One influential factor

that is considered crucial for the onset and maintenance of obesity is

advertising of “unhealthy” food that contains high amounts of fat,

sugar, or salt to children and adolescents (Harris, Bargh, & Brownell,

2009; Harris, Pomeranz, Lobstein, & Brownell, 2009).

Since the 1970s, public health and marketing research has focused

on possible explanations for the links between children’s perception and

evaluation of advertised products, their subsequent eating behavior,

and obesity (Lesser, Zimmerman, & Cohen, 2013; A.C. McClure et al.,

2013). Many studies reach the conclusion that children are not aware of

the risks associated with the advertised product (Chou, Rashad, &

Grossman, 2008; Effertz, Franke, & Teichert, 2012; Halford et al., 2008).

Because of advertisements that promote unhealthy products and their

steady consumption, children become obese (Chou et al., 2008; Effertz

& Adams, 2014; Harris, Bargh et al., 2009; Harris, Pomeranz et al., 2009;

Zimmerman & Shimoga, 2014). This phenomenon has become more

widespread globally in the face of the diffusion of fast food (FF)

offerings, even in remote markets (Igumbor et al., 2012; Wang, Wang,

Xue, & Qu, 2016).

Branding is a key marketing instrument in addressing children, as

it transports the advertised product’s information, expectations, and

experiences in one image, logo, or icon. Children across nations are

already aware of hundreds of brands from the very young age of

2 years (Ji, 2002; LaTour, LaTour, & Zinkhan, 2010), and knowledge

of food brands is the main predictor of increased body‐mass‐index (BMI) (Cornwell, McAlister, & Polmear‐Swendris, 2014). The effects of advertising and branding last until adulthood (Connell, Brucks, &

Nielsen, 2014). At the same time, it is acknowledged that children are

not able to understand advertising up to the age of 8 years, and

probably even until 11 (e.g., Andronikidis & Lambrianidou, 2010). The

ability to critically evaluate the food proposals in advertisements

seems to develop only after that age.

Previous research revealed a close dependency between brands and

taste via experienced and learned emotional states that are conveyed

and embodied in the brand (e.g., Robinson, Borzekowski, Matheson, &

Kraemer, 2007), especially among children (e.g., Keller et al., 2012).

From neuroeconomic research, it is also widely known that food brands

can influence the perceived taste. Accordingly, children’s knowledge of

food brands was found to be significantly related to a higher score on

palatability scales (Cornwell & McAlister, 2011).

Contributing to previous findings, this article analyzes the joint

effects of different combinations of advertising, brand, and taste

conditions on the taste satisfaction of FF products. Insights from

schema congruity and emotional brand elicitation are joined to derive

hypotheses for various interaction effects. Research on schema

incongruity serves as a framework (Johnson & Russo, 1984; Lee &

Schumann, 2004; Mandler, 1982; Russo & Edward, 1980) to

theoretically assess varying conditions in branding and advertising.

As real‐world food advertising consists of more or less congruent information—for example, by using emotional images, warnings,

health claims, and exaggerated credence claims—it is unclear how

these components are integrated with a brand schema on which

especially children base their decisions.

Such insights are important for the design of public health

policies that aim to inform society about the harmful effects of

obesity and unhealthy eating via warnings similar to those in tobacco

advertisements. Hence, when prevention strategies aim to inform

minors about the possible health risks linked to the consumption of

unhealthy food, it is not sufficient to evaluate the proposed

instrument, for example, a warning claim, in isolation, but its effect,

when embedded in an advertisement, is also relevant.

In an experimental setting in Novosibirsk, Russia, children and

adolescents were exposed to fictitious FF advertisements from

different prototypical or novel brands with more or less congruent

advertising information. FF is still a relatively new phenomenon in

Russia; hence, it provides a unique setting to investigate children’s

early contact with FF. To assess the impact of different brand

offerings, systematically varied advertisements were shown in an

experimental setting. The children were then offered a portion of

French fries that were either salted (and more palatable) or unsalted.

They were asked about their taste satisfaction one day after

consumption in a “recall assessment.” This novel experimental setting

of combining brand effects, claims, warnings, and palatability makes a

contribution via its closeness to real‐world advertisements, branding, and food product designs.

The next section outlines previous research concerning brand

perception and the influence of brands on taste, especially when

information appears more or less congruent. Relevant theoretical

underpinnings are integrated to derive the hypotheses. Thereafter,

the experimental setting, statistical methods, and derived results are

described. Finally, possible implications for modeling the interplay of

brands and children and for policy regulations are discussed.

2 | THEORY

In marketing theory, a brand is regarded as a bandwidth of different

perceived attributes and properties related to the trade name with a

certain value for consumers (Armstrong & Kotler, 2005). Several

parts constitute a brand, such as the product itself, its attributes, a

set of characteristics, expectations, and associations consumers

perceive and project onto the product (Aaker, 1991; Keller, 1993).

A brand is built via communication processes, particularly advertis-

ing, which consist of several components, such as pictures and claims.

One explanation of how different elements interact in advertising

and are jointly evaluated by consumers is the concept of schema

congruity (Lee & Schumann, 2004; Mandler, 1982). This applies if

information, for example, in an advertisement, is compatible with

existing knowledge or an existing schema of the underlying product

and conforms to expectations.

The broader “congruity theory” of Osgood and Tannenbaum

(1955) states that a perceived incongruity between a message source

and its concept results in a dissociative relation with the product.

Schema incongruity, like other concepts of inconsistency, such as

cognitive dissonance, elicits changes in cognitions, attitudes, inten-

tions, and behavior. It is not confined to advertising and brand

information, but applies to all “violations” of expectations concerning

an object or piece of information (Bruner & Postman, 1949), for

example, also if the taste of an FF product does not match with the

schema of FF in the consumer’s mind. It is therefore important to

understand under which circumstances the perceived taste of an FF

product is influenced via advertising and brand components.

Research on schema incongruity has already investigated the

interplay of different advertising components (e.g. Heckler & Childers,

1992; Nagar, 2015; Stach, 2015; Velasco et al., 2015). Using the

concept of schema incongruity to explain attitudes and evaluations of

products allows a wider and more flexible view on combinations of

advertisement elements than other theoretical frameworks. This is

because the aggregate effect of the different interplays on the

dependent variable of interest is undetermined ex ante.

176 | EFFERTZ ET AL.

Three layers of potential (in‐)congruities that might influence a later recalled taste experience are analyzed in the following: (a) the

combination and congruity of brand with the product category, (b)

congruity between the brand and associated claims used in the

advertisement, and (c) the interplay of brand expectations and taste

experiences, such as when the product itself is manipulated to taste

more or less palatable.

Consumers’ brand perceptions are driven by their needs and

brand knowledge derived from personal experience (Keller, 1993;

Keller & Lehmann, 2006), but also by brands’ relationship to the

specific product category (Punj & Moon, 2002). Consumers form an

implicit category image based on their experiences with average

category members (Mao & Krishnan, 2006).

Vice versa, prototypicality is the degree to which a category

member is representative of a category (Nedungadi & Hutchinson,

1985; Rosch, 1975). Accordingly, Mao and Krishnan (2006) showed

that brands can be distinguished between prototype (or exemplary)

brands versus novel brands. As an example, Nike generally serves as

a prototype of the product category of quality sportswear, which is

not necessarily congruent with information about all its products

(Mao & Krishnan, 2006).

This study follows these concepts and differentiates between

prototype brands and two types of novel brands in the FF sector.

A prototype FF brand is defined as an established brand that has

achieved a high consistency with consumers’ generalized imagery

of the FF product category. “Novel fast food brands” refer to

brands that have yet to be established in the marketplace. They

can be positioned closely to the extant imagery of the product

category, or apart from this scheme. As the FF sector is

characterized by a predominance of global brands, a novel global

FF brand defines a new brand that is closely aligned to the extant

imagery of the FF product category. In contrast, a novel local

FF brand positions itself as a local market offering and

differentiates itself more from the category imagery than a novel

global FF brand.

This study empirically investigates the effect of three brand types

on taste experience when different aspects of FF advertisements are

present. To differentiate these aspects and their relationships, the

effect on taste satisfaction is disentangled in the context of potentially

conflicting information like warnings, credence claims, and brand

effects that are jointly present in advertisements. Because taste is a

primary indicator of food quality, children’s and adolescents’ taste

experience and satisfaction are used as strong predictors of product

evaluation.

2.1 | Congruity between brand and product category

Most nutritionists describe eating as an emotional act and experience

(Pudel & Ellrott, 2004). Individuals love the taste of palatable, high‐ energy food, and respond with feelings of disgust and nausea to

inedible ones (Cornwell & McAllister, 2011). Additionally induced

emotions or moods affect eating behavior (Cardi, Leppanen, &

Treasur, 2015). Food advertising cannot convey the product’s taste

directly, but has to use “proxies,” meaning visual or acoustic stimuli

and information about qualities that elicit positive emotions. These

emotions are “projected onto” the food product to pre‐assess the consequences that will stem from eating it.

In this setting, brands serve as a signal for quality and convey

information about the food product’s inherent properties. Vice versa,

brands are built either by the product itself, referring to the

consumption experience, or via its advertising. A brand might,

therefore, be seen as the representation of a learned food product’s

image, conveying promises of rewards and qualities.

Following the schema congruity theory, an advertised FF product

that is presented with a well‐known FF brand should be valued as more congruent than an unknown brand for which no previous

experience has been built up. In line with this, studies of Esch et al.

(2012) and Reimann, Castaño, Zaichkowsky, and Bechara (2012)

found that popular brands influence taste compared to unknown

brands. Furthermore, prototype brands serve as generalizations in

the consumer’s mind for an entire product category (Mao & Krishnan,

2006), such that popular FF brands like McDonald’s or Burger King

are likely to shape children’s expectations about all FF products. As a

contrary example, a well‐known brand outside the food sector, for example, Nike, would be perceived as rather incongruent with FF

products compared with an unknown brand. It can furthermore be

expected that different brand names, as well as advertisements, elicit

different emotional responses.

Previous research showed that expectations about food brands

influence the experienced taste of food (Keller et al., 2012; Koenigs &

Tranel, 2008; Kühn & Gallinat, 2013; S.M. McClure, Tomlin, Cypert,

Montague, & Montague, 2004; Robinson et al., 2007; Woods et al.,

2011). Effects were also found in neuroscientific studies: Reimann

et al. (2012) showed that novel brands pose a possible risk to

consumers and “trigger a lure for reward” compared with familiar

brands. Thomas, Hammer, Beibst, and Münte (2013) report consis-

tent results and conclusions for responses to branded and no‐name products. Bruce et al. (2014) showed that well‐known brand logos of food products activate brain regions associated with motivation.

Combined with the results of S.M. McClure et al. (2004), this suggests

that taste is at least partially determined by an emotional valuation

of the food brand, which hinges on experiences and accompanying

cognitive elements.

Linking these aspects with schema congruity, one can expect that

perceived congruity results in a more positive attitude. Given the

connection between brands and taste, this should lead to a more

positive perceived taste experience. We therefore hypothesize:

H1: A prototype FF brand exerts a stronger positive brand effect

on taste satisfaction than a novel FF brand.

Esch et al. (2012) demonstrate that strong brands are the

embodiment of experiences that form a prototype, which is retrieved

during exposure, whereas unfamiliar brands are linguistically

encoded in the brain. Brands that are considered “in between”

EFFERTZ ET AL. | 177

popular strong brands and completely unknown ones are recognized

in the relevant brain areas of both retrieval and encoding procedures.

This implies a spillover or adaptation of experience from strong

brands toward the evaluation of weak brands, so that a novel food

brand is judged at least partly by adapting consumption experiences

from prototype FF brands.

Seemingly popular but actually imaginary FF brands that

have the looks and glamour of the global prototype brand

should, therefore, be perceived as being more congruent than

novel local brands. Known brand features can be used as “gap

fillers” for the unknown qualities of an imaginary brand (Reimann

et al., 2012). This leads to the proposition that the effect of a

novel but prototype lookalike FF brand on taste should be higher

compared with a nonprototype brand. In short, it can be

postulated:

H2: The brand effect on taste satisfaction is stronger for a novel

global FF brand than for a novel local FF brand.

2.2 | Congruity between brand and advertising claims

A second layer of possible incongruity might arise between

advertised claims and the brand. Research identified various effects

of advertising claims on product evaluation (Olsen, Samuelsen, &

Gaustad, 2014), congruity issues (Florack & Scarabis, 2006), and

purchase intention (Effertz et al., 2012). Concerning FF high in fat,

sugar, or salt and mostly low in nutritional value, credence claims and

warnings are the most important types of advertising claims for real‐ world advertisements and in the context of public health regulation

(Effertz et al., 2012; Fernqvist & Ekelund, 2014; Ford, Smith, &

Swasy, 1990; Werle & Cuny, 2012).

Credence claims state beneficial aspects about the product that is

usually not verifiable by the consumer in the short term. As an

example, claims about improved health or well‐being are very common, but irrelevant from a medical point of view. Claims about

healthy food products may sometimes also reduce the product’s

attractiveness by making it less tasteful to children (Maimaran &

Fishbach, 2014).

Whether a credence claim is perceived as being congruent

enough with FF products and brands and, therefore, positively

affects taste experiences, strongly depends on the learned schema.

Adding credence claims to food is a ubiquitous phenomenon in

today’s advertising. A credence claim in general therefore high-

lights a congruity between the message and the product.

Furthermore, a learned schema of a prototype brand should be

more congruent with the credence claim than novel brands that

the consumer has not experienced. A specific brand credence claim

interaction with taste satisfaction is expected. The sign of this

interaction should depend on how the claim and schema informa-

tion is aligned. The following general hypothesis can, therefore, be

stated:

H3: The positive effect of a credence claim on taste satisfaction

varies with the prototypicality of the FF brand.

A warning, in contrast to a credence claim in advertisements, is

intended to raise awareness about possible health risks. This should

result in a product becoming less attractive, especially if it is food.

However, children might also regard a product with health warnings

as a “forbidden fruit,” which will induce the opposite of the intended

effect on their behavior (Effertz et al., 2012), especially if the warning

is emotionally framed.

Both directions of effect on taste satisfaction are possible,

depending on how the claims interact with the emotion‐eliciting brand. If the learned schema of a prototype FF brand is congruent

with the presence of warnings, a warning might even result in

increased taste satisfaction. Dahlén, Lange, Sjödin, and Törn (2005)

pointed out that incongruent ad‐brand information elicits more sophisticated processing. Maoz and Tybout (2002) elaborate on the

finding that incongruent information can positively affect product

evaluation and find that “mild” incongruity yields the highest

evaluation score when involvement is high.

In the presence of low involvement, however, evaluation becomes

less positive with increasing incongruity. A discount in recalled taste

satisfaction might result if brand and advertisement are perceived as

being too dissociative. In analogy to the credence claim, a warning

should decrease taste satisfaction, but might interact with different

brands depending on the perceived congruity between warning and

brand schema. We therefore hypothesize:

H4: The negative effect of warning claims on taste satisfaction

varies with the prototypicality of the FF brand.

A joint presentation of a warning and credence claims in

advertisements can be seen as logically conflicting information. This

may occur, for example, if the ad states that a product has positive

effects on individual well‐being while it also displays a warning claim that points to possible health impairments. The potential cognitive

dissonance of combining a warning and a credence claim on the same

product may be strengthened or reduced by combining it with a

certain brand type.

Whether a particular combination of claims, advertising, and the

brand name is perceived as congruent or not depends on some form

of “overall relation” between the components (Schmitt, Tavassoli, &

Millard, 1993), not necessarily on a logical plausibility. It seems likely

to assume that perceived incongruity increases with uncertainty

about the brand’s possible qualities (Maoz & Tybout, 2002, Ozanne,

Brucks, & Grewal, 1992), which should be the case in settings of a

novel brand. Here, the analytical reasoning might mitigate the brand

effect (Yamada et al., 2014), and incongruities might elicit cognitive

dissonance and worse evaluation, depending on involvement.

Combinations of warnings and credence claims might even be

perceived as typical for prototype FF products, and therefore as

congruent information, as consumers might expect such combina-

tions for prototype brands. A negative effect can be expected when

178 | EFFERTZ ET AL.

no emotional brand effect is at work, whereas a reduced negative

effect might occur when a popular prototype brand mitigates the

incongruity.

It is assumed that a well‐established prototype brand aligns with claims to form a coherent picture, whereas unknown brands

constitute an incongruity on their own that increases with additional

conflicting advertising components like claims, as suggested in H1

and H2. A brand combined with a single warning or credence claim

might comprise a less complex advertisement that is more easily

aligned and less incongruent with high taste satisfaction. The effect

of possible information incongruities with the brand appears to be

similar but not equal to the induced cognitive dissonance in the

“stroop test” of Keller et al. (2012). Taken together, it can be

hypothesized that:

H5: The negative effect of combining a warning and a credence

claim on taste satisfaction varies with the prototypicality of

the FF brand.

2.3 | Congruity between brand expectations and taste experience

The taste experience itself can also be a source of schema

incongruity (e.g., Hoegg & Alba, 2006). It is well‐known that children have an innate preference for sweet and salty food (Beauchamp &

Cowart, 1990; Bouhlal, Chabanet, Issanchou, & Nicklaus, 2013).

Therefore, the impact of brand effects and claims on taste might be

mitigated or enhanced if the product’s taste is more or less palatable,

assuming that it is pleasant and congruent to the information the

advertisement conveys.

Given that the average FF product contains high levels of salt,1

manipulating the salt condition by making a product saltier might

evoke positive emotions by a strong brand or congruent sound

claim information. Research on “salt” or “sodium” is well‐described in the literature: it is known that adding salt to food not only

makes it taste saltier, but also improves the perception of product

thickness and flavor intensity, enhances sweetness, masks metallic

or chemical off‐notes and rounds out overall flavor (Gillette, 1985). A saltier taste of a food product in combination with

positive brand information can, therefore, be considered as

strongly corroborating and congruent information that increases

taste satisfaction. As a saltier taste is more congruent with the

product category of FF, it can be hypothesized that taste

experience interacts with brands:

H6: The impact of objective taste manipulations on taste

satisfaction varies with the prototypicality of the FF brand.

While a salty taste is a palatable positive experience for children,

effective and cognitive congruities might lead to a more consistent

evaluation of the overall taste experience. Given the considerations

outlined before, the salty taste is supposed to result in a more

congruent perception with an offering of a prototype brand.

The positive emotionality of the salty taste experience might

reinforce positive expectations and limit the effects of warnings

and credence claims as rather negligible product characteristics.

Such an effect cannot be expected for novel brands without past

consumption experiences. The focus of a product evaluation is likely

to be based solely on the congruity between the brand and its

advertised claims, independent of the taste experience. It can,

therefore, be hypothesized that there is an interaction effect

between taste condition and advertisement features in the case of

a prototype brand:

H7: Taste satisfaction with a prototype FF brand is highest if the

advertisement claims and the objective taste manipulation

are consistent with consumers’ generalized imagery of the

underlying FF product category.

3 | METHOD

A two‐stage experiment was conducted to test the hypotheses. It started with showing only the randomly assigned brands and was

then implemented into the advertisements with different claim

variations as exposure for the children. This was done to first assess a

pure brand evaluation. Thereafter, participants filled out a ques-

tionnaire assessing several constructs about the ad and the brand, as

explained later, and were offered a portion of French fries which was

either “unsalted” (“normal” taste condition) or prepared with

additional salt (“salty” taste condition, equal to approximately

0.5–1 g of additional salt per portion).

The advertisement with different brand claim combinations

preceded the taste experience. On the basis of literature, the “salty”

taste condition should have been perceived as more palatable. The

systematic manipulation of taste was conducted similarly, by using a

salt cellar accordingly and as accurately as possible for each portion

of French fries. On the day after this procedure, participants were

asked to complete a second questionnaire that assessed their taste

satisfaction. Instructions and taste manipulation were conducted by

the same team of researchers in the same manner.

Russia was chosen as the empirical basis for our study because

most studies about marketing to children are carried out in the USA

and European countries. However, the problematic development of

overweight and obesity has also been witnessed in several remote

and emerging markets, especially in Russia and other Commonwealth

of Independent States countries. Approximately 30 million children

under the age of 17 live in Russia, comprising more than 20% of the

whole population (Russian Federation State Statistics Service, www.

gks.ru) and making them an important and critical target group. In

Novosibirsk, children comprised approximately 17.5% of the popula-

tion at the beginning of 2012 (www.novosibstat.ru). Studying brand

perception, valuation, and reactions toward FF advertising enables a

unique and purer view on the effects in a less complex environment,1See http://money.cnn.com/2015/08/18/news/salt‐fast‐food‐mcdonalds

EFFERTZ ET AL. | 179

http://www.gks.ru

http://www.novosibstat.ru

http://money.cnn.com/2015/08/18/news/salt-fast-food-mcdonalds

which is an advantage. It is likely that the results can be aligned with

findings from Western markets if additional specific market features

are accounted for.

3.1 | Sample selection and description

The city of Novosibirsk in Russia serves as sampling basis for this

study. Novosibirsk is the third largest city in the Russian Federation,

with a population of 1.5 million in 2012. The city may be regarded as

an appropriate model for studying young consumer behavior in a

market which, 25 years after the iron curtain in Europe broke down,

still had a rather novel and nonsaturated market for FF, with the only

exceptions the urban areas of Moscow and St Petersburg. Global and

local brands develop the public catering market in Novosibirsk, like

Carl’s Jr., KFC, as well as novel local brands, such as Uncle Dener and

Kroshka‐Kartoshka. Only in 2014, McDonald’s opened its first restaurant in Novosibirsk, with the prospect of expansion in the

near future. Other major market players, such as Burger King and

Pizza Hut, have also taken an interest in the Ural and West Siberian

markets.2

A research team supervised by the authors conducted a survey of

children aged 10–18 years in Novosibirsk schools in 2012/13. To

select the respondents, the idea of multilevel cluster sampling was

implemented. In the first stage, Novosibirsk schools participating in

the research were defined at random, whereas in the second stage,

single classes from Grades 5 to 11 were selected on an equal

distribution basis. Each chosen school class was then assigned to one

experimental condition such that the resulting design was nearly

balanced with regard to class cohorts. All experimental conditions

were almost equally sampled across class cohorts. A total of 780

school students of secondary general education institutions were

surveyed. Out of those, 59 were excluded due to invalid responses,

such as partly missing or incorrect completions. The sample thus

consisted of 721 subjects with 381 boys (52.8%) and 340 girls

(47.2%), whereas 43.6% of obtained responses came from 5th‐ to 6th‐grade, 31.2% from 7th‐ to 8th‐grade, 25.1% from 9th‐ to 11th‐ grade students, respectively. Parental consent for children being

surveyed was obtained at general meetings of schools’ trustee

councils.

3.2 | Experimental stimuli

Three FF brands were used as experimentally varied stimuli: a

prototype FF brand (McDonald’s), an imaginary and, therefore,

considered weak and novel global‐looking FF brand, and a fake local Russian brand. As already mentioned, McDonald’s did not have an

outlet in Novosibirsk at that time, and therefore children only knew

the brand from advertisements or experiences with the product

during visits to other cities. In addition, the advertisements were

varied by adding warnings and exaggerated credence claims, which

are the usual components of real advertisements. The subsequent FF

meal—a portion of French fries—was varied and served in a normal

and salty taste condition.

The global imaginary brand was named Fasty Ju. The brand name

was written in English letters, which made it look global and similar

to a prototype brand. The local fake brand was named TИНар (Tinar

in English) to reflect a Russian origin. The English and Russian

descriptions for the two fake brands were considered the crucial

discriminator for being perceived as global or local. The global brand

McDonald’s is considered to be a prototype brand in our setting,

whereas Fasty Ju appears to be global but also novel, hence similar

but not equal to prototype FF brands.

The presence of advertising and the two claims and brands

were also experimentally varied. The claims were composed to be

close to a real‐world advertisement. The credence claim was “A super lunch that brings you strength and energy for the whole

day!”3 and the warning stated “Excessive fat consumption leads to

obesity.”4 Figure 1 provides an example of a full advertisement

stimulus.

Given the two claims that are present or not in the experimental

settings, three brands, and two taste conditions, a 3×2×2×2 design

evolved with three brands (McDonald’s, Fasty Ju, and TИНар), two

warning conditions (yes or no), two credence claims (yes or no), and

two taste conditions (salty or not salty). Depending on the respective

combination, the fictitious product offering might be more or less

schema congruent. With regard to Keller et al. (2012), who applied a

“stroop test” to measure responses to logically conflicting stimuli, one

would assume that more congruent information combination in an

advertisement would lead to a higher taste satisfaction with more

familiar brands, representing a less incongruent picture.

3.3 | Survey measures and model

Besides the experimentally varied variable, the study design included

self‐reporting questionnaires containing 27 questions about percep- tion of the brands that were offered, and their characteristics. A 7‐ point Likert scale of importance and semantic differential (Smith &

Albaum, 2004) was used to measure behavioral and emotional

responses.

Different analysis of variance (ANOVA) analyses were conducted

to test the hypotheses. As a dependent variable, “taste satisfaction,” a

construct from Fitzsimons, Greenleaf, and Lehmann (1997), was used,

which is assessed with four 5‐point Likert‐scaled items in the recall questionnaire one day after the taste experience. Recall is often used

as a main dependent variable in research on schema congruity.

Different ANOVA models were built to account for the derived

hypotheses directly. To test hypotheses H1 and H2, a model with the

main effects of the experimental conditions and without interaction

effects was estimated (Model 1). To test H3, H4, and H5, models with

2See, for example, http://afisha.ngs.ru/blogs/more/934367

3The original Russian credence claim stated: “Супер‐обед придаст тебе сил и энергии на целый день!” (Super dinner will give you strength and energy for the whole day!).

4The original Russian warning stated: “Чрезмерное потребление жирной пищи приводит к

ожирению” (Excessive fat intake leads to obesity).

180 | EFFERTZ ET AL.

http://afisha.ngs.ru/blogs/more/934367

additional interactions between the single claims and the brand

(Model 2 and Model 3) and between the combined claims and the

brand (Model 3) were conducted. For H6, an interaction between taste

and brand was added to the main effects model (Model 4). Finally, to

test H7, a satiated model with all possible interactions was estimated

(Model 5). The model building was directly derived from the hypotheses.

Taste is not only determined by the inherent properties of the good,

but also by individual states of deprivation like hunger and accompany-

ing marketing modalities, such as price (see, e.g., Plassmann, O’Doherty,

Shiv, & Rangel, 2008). To control for such variations because of

individual states of hunger that might affect taste satisfaction, the

students were asked to indicate their level of hunger in self‐developed items that asked the children to state their level of satiety by drawing a

level line in a bowl figure. In addition, it was controlled for class (lower,

middle, and upper class in the school) and gender.

Manipulation checks for the claims and brand comprehension were

conducted to assess whether the stimuli were understood correctly.

This was done by exposing a group of 110 adolescent children aged

12–18 to the stimulus material and asking them to assign the used

claims to three categories—warning claim, neutral claim, and credence

claim. In addition, they were asked to state whether the presented

brands were “global” or “Russian” and “real” or “novel.”

To further ensure the validity of the stimulus material, especially

the differences in perceived brand qualities, a post‐hoc t‐test as in Plassmann et al. (2008) on several constructs relating to brand and

advertisement perception was conducted. The associated items were

also assessed in the questionnaire. The different constructs used for

these post‐hoc t‐tests were derived from several sources. Feltham’s (1994) persuasive disclosure inventory was used to assess the

advertisements’ trustworthiness (ethos), information (logos), and

emotional appeal (pathos) by three items each (Cronbach’s α 0.8323,

0.8405, 0.8168), Shamdasani, Stanaland & Tan’s (2001) brand

relationship scale for only the brand assessment as well as combined

with the ad (Cronbach’s α 0.8467, 0.8366), and finally the “familiarity

with the brand” scale from Simonin and Ruth (1998) (Cronbach’s α

0.7363). The claims were also checked for possible emotional

elicitation among the children.

4 | RESULTS

4.1 | Stimuli and manipulation checks

The manipulation checks yielded satisfying results. Concerning the

comprehension of the different claims, 73.6% from the group of 110

adolescents assigned the credence claim correctly, whereas 71.8%

assigned the warning correctly. Everybody recognized McDonald’s as

a real and global brand, whereas Fasty Ju and TИНар were identified

as novel brands (both 97.3%), and TИНар (100%) as a Russian (local)

brand. Claims and brands were, therefore, perceived in the intended

way. To check for the postulated properties of the advertisements

and similar to Esch et al. (2012), post‐hoc t‐tests were conducted comparing the three brands pairwise on the means of the constructs

brand awareness, familiarity with the brand, and emotional appeal of

the advertisement.

The test revealed that the popular and considered prototype FF

brand was indeed perceived significantly much more emotionally

positive compared with the other two brands, as well as displaying a

higher brand awareness and brand relation, brand credibility, and

brand information (Table 1). Regarding all perceived constructs, the

prototype global brand was therefore significantly more positively

evaluated than the other two brands. The advertisement of the Fasty

Ju brand compared with the local brand seemed to be more credible.

In addition, and most importantly, brand relationship and familiarity

were much stronger for the imaginary global brand compared with

the local brand, despite their completely imaginary nature. This

confirmed that all brand stimuli worked as intended.

FIGURE 1 Example of stimulus material

TABLE 1 Manipulation checks: Post‐hoc mean comparison tests

Brands/constructs

Prototype

brand

Novel global

brand

Novel local

brand

Mean/(SD) Mean/

(SD)

Mean/(SD)

information (logos)

0.207A −0.093B −0.113B

(1.07)(0.898) (0.996)

trustworthiness (ethos)

0.164A −0.016B −0.157B

(1.009)(0.918) (1.04)

Advertisement emo

(pathos)

0.186A −0.038B −0.149B

(1.006)(0.9435) (1.023)

Brand relationship

(brand only)

0.356A 0.015B −0.374C

(0.888)(0.3563) (0.828)

Familiarity with the brand 1.008A −0.425B −0.606C

(0.675)(0.533) (0.831)

Brand relationship

(assessed after

embedded in

advertisement)

0.240A 0.0743B −0.331C

(0.999)(0.240) (0.981)

Note. Means (standard deviations) by brands and constructs. Post‐hoc mean comparison tests using pairwise t‐test. Means with same super- scripts are not statistically different from each other (p < 0.05).

EFFERTZ ET AL. | 181

With the constructs of advertisement information and emotion, it

was checked whether an advertisement with a warning or credence

claim was perceived as being more emotional, which was not the case

(warning claim, t = 0.7507; p = 0.4531; credence claim, t = −0.8590;

p = 0.3903) or more informing, which was the case for the warning

(t = −1.8748, p < 0.05). The alternatively tested credence claim was

not perceived as contributing to the informational content of the

advertisement.

4.2 | Findings

Five nested ANOVA models were used to derive parameter

estimates at the level of main effects only, two‐way‐ and three‐way interaction effects. The experimentally varied factors of brand type,

claim, warning, and taste condition serve as independent variables to

predict taste satisfaction, controlled by covariates of class and

gender, as well as the level of hunger. Table 2 shows the means and

standard deviations of taste satisfaction in the different experimental

conditions, whereas Table 3 displays the F‐values for the variance decomposition derived from the five ANOVA models.

Brand main effects turn out to be significant in any model,

indicating the overarching impact of branding on taste satisfaction.

All models also display a significantly positive effect of the level of

hunger on taste satisfaction and a negative effect of older school

classes, meaning that older children are less likely satisfied by the

taste experience than younger children. Gender is not a significant

covariate in any of the models.

Hypothesis H1 is supported by Model 1: The main effects

ANOVA model reveals a highly significant brand effect (F (2,706) =

12.95). The main effects parameter estimate of 0.4277 (p < 0.000)

shows that the prototype global real brand compared with the novel

local brand positively affects taste satisfaction. The contrast

between the prototype and the global novel brand on taste

satisfaction is also positive and significant (0.2894, p < 0.01). H2

stated that the novel but global‐looking brand affects taste satisfaction more positively than the local Russian brand. The novel

global brand tends in the same direction, but is not significant at the

10% level (0.1383, p = 0.102). The findings, therefore, seem to

support the hypothesis that the recalled taste satisfaction one day

after the experiment increases with the prototypicality of the brand

TABLE 2 Mean taste satisfaction, standard deviations, and cell sizes

Experimental condition Taste satisfaction

Taste Warning Credence claim McDonald’s Fasty Ju Tinar

Mean/(SD)/# obs.

Unsalted No No 0.064 0.203 −0.009

(0.806) (1.026) (1.027)

35 27 53

Unsalted No Yes 0.036 −0.407 −0.114

(0.840) (0.635) (0.927)

32 14 22

Unsalted Yes No 0.026 −0.526 −0.509

(0.899) (0.794) (0.798)

31 40 35

Unsalted Yes Yes 0.518 0.304 −0.067

(1.221) (0.904) (0.896)

15 60 31

Salted No No 0.595 0.075 −0.750

(0.998) (0.933) (0.676)

34 25 10

Salted No Yes −0.364 0.056 −0.128

(0.785) (1.049) (1.088)

16 61 20

Salted Yes No 0.952 −0.643 −0.256

(1.036) (1.020) (0.906)

39 14 32

Salted Yes Yes 0.034 0.069 0.009

(0.996) (1.059) (1.124)

33 22 25

182 | EFFERTZ ET AL.

offer. The findings consistently match the ranking order already

found with the manipulation checks. Figure 2 displays the marginal

means of Model 1.

With H3, it was hypothesized that the impact of a credence

claim increases taste satisfaction, but varies with the accompany-

ing brand. H4 states a similar assumption concerning the warning,

but with the opposite sign. The ANOVA analysis of the associated

model (Table 3) corroborates H3: there is a significant interaction

between credence claim and brand (F (2,702) = 6.83), but the claim

itself does not increase taste satisfaction per se. The credence

claim significantly decreases taste satisfaction for the prototype

brand while increasing taste experience for the local Russian

brand.

Aligned with H4, the analyses above demonstrate that a warning

does also not unequivocally affect taste satisfaction for either brand

condition (Model 2). A prototype real brand combined with a single

warning but without a credence claim leads to a significantly higher

taste satisfaction if an interaction term between the claims is added in

TABLE 3 F‐values of five (nested) ANOVA models

Model 1 Model 2 Model 3 Model 4 Model 5

DOF Brand with

product